Prolactin-Induced Protein Is Required for Cell Cycle Progression in Breast Cancer

Abstract

Prolactin-induced protein (PIP) is expressed in the majority of breast cancers and is used for the diagnostic evaluation of this disease as a characteristic biomarker; however, the molecular mechanisms of PIP function in breast cancer have remained largely unknown. In this study, we carried out a comprehensive investigation of PIP function using PIP silencing in a broad group of breast cancer cell lines, analysis of expression microarray data, proteomic analysis using mass spectrometry, and biomarker studies on breast tumors. We demonstrated that PIP is required for the progression through G₁ phase, mitosis, and cytokinesis in luminal A, luminal B, and molecular apocrine breast cancer cells. In addition, PIP expression is associated with a transcriptional signature enriched with cell cycle genes and regulates key genes in this process including cyclin D1, cyclin B1, BUB1, and forkhead box M1 (FOXM1). It is notable that defects in mitotic transition and cytokinesis following PIP silencing are accompanied by an increase in aneuploidy of breast cancer cells. Importantly, we have identified novel PIP-binding partners in breast cancer and shown that PIP binds to β-tubulin and is necessary for microtubule polymerization. Furthermore, PIP interacts with actin-binding proteins including Arp2/3 and is needed for inside-out activation of integrin-β1 mediated through talin. This study suggests that PIP is required for cell cycle progression in breast cancer and provides a rationale for exploring PIP inhibition as a therapeutic approach in breast cancer that can potentially target microtubule polymerization.

Introduction

Prolactin-induced protein (PIP) is widely expressed in breast cancer and has been used as a characteristic biomarker for the diagnostic evaluation of this disease [1]. Genomic studies have revealed that PIP is highly expressed in luminal A and molecular apocrine subtypes of breast cancer [2–4]. Molecular apocrine is a subtype of estrogen receptor (ER)–negative breast cancer that is characterized by the overexpression of steroid response genes such as androgen receptor (AR) and forkhead box A1 (FOXA1) [3,5,6]. Notably, a recent study has shown that PIP is one of the best biomarkers for the immunohistochemical identification of molecular apocrine tumors [7]. It is known that PIP expression is regulated by prolactin and androgen hormones [8]. In particular, AR engages in a transcriptional cooperation with prolactin-activated Stat5 and Runx2 to regulate PIP expression [8,9]. In addition, we have demonstrated that PIP is a cAMP responsive element binding protein 1 (CREB1) target gene that is induced by a positive feedback loop between AR and extracellular signal-regulated kinase (ERK) [10].

There is limited knowledge regarding the molecular function and binding partners of PIP in breast cancer. The available data indicate that PIP is a secreted protein with aspartyl protease activity that can degrade fibronectin and has the ability to bind and modulate CD4 receptor in T lymphocytes [11,12]. In addition, early studies using actin-sepharose columns have shown possible binding of PIP to actin in seminal fluid [13]; however, this finding has not been validated using more modern proteomic methods. Furthermore, the importance of PIP in cell proliferation has been demonstrated by the fact that purified PIP promotes growth of breast cancer cells and PIP expression is necessary for the proliferation of T-47D and MDA-MB-453 cell lines [9,10,14]. Moreover, we have recently demonstrated that PIP mediates invasion of breast cancer cells in a process that partially depends on the degradation of fibronectin by this protein [10].

It is notable that the extracellular effects of PIP on fibronectin degradation are necessary for the outside-in activation of integrin-β1, which, in addition to the regulation of invasion, has a role in promoting cell proliferation [10]. Furthermore, it has been shown that coculture of PIP-silenced and naive T-47D cells does not reverse the growth inhibition induced by PIP silencing, which suggests a potential intracellular function for this protein [4]. Despite these findings, the underlying molecular mechanisms of PIP function in cell proliferation have remained largely unknown and require further studies.

In this study, we investigated PIP function in breast cancer using small interfering RNA (siRNA) silencing in a broad group of breast cancer cell lines, analysis of expression microarray data, proteomic analysis by mass spectrometry (MS), and biomarker studies on primary breast tumors. We demonstrated that PIP is required for the progression through different phases of cell cycle and identified key molecular mechanisms and binding partners for this protein in breast cancer.

Materials and Methods

Cell Culture

Breast cancer cell lines MCF-7, T-47D, BT-474, HCC-202, HCC-1954, MDA-MB-453, SK-BR-3, MFM-223, and MDA-MB-231 were obtained from American Type Culture Collection (Manassas, VA) and cultured as recommended by the provider.

RNA Interference

PIP knockdown (KD) by siRNA silencing was performed as described before [15]. The following two siRNA-duplex oligos (Sigma-Aldrich, St Louis, MO) were applied: duplex 1—sense, 5′CUCUACAAGGUGCAUUUAA and antisense, 5′UUAAAUGCACCUUGUAGAG; and duplex 2—sense, 5′CCUCUACAAGGUGCAUUUA and antisense, 5′UAAAUGCACCUUGUAGAGG. Transfections with siRNA Universal Negative Control No. 1 (Sigma-Aldrich) were used as controls. The effect of KD was assessed 72 hours after transfections. The average changes obtained for two duplexes are presented in manuscript.

Quantitative Real-Time Reverse Transcription–Polymerase Chain Reaction

Quantitative real-time reverse transcription–polymerase chain reaction (qRT-PCR) to assess the expression levels of PIP (assay ID: Hs00160082_m1), cyclin D1 (Hs00765553_m1), cyclin E1 (Hs01026536_m1), cyclin B1 (Hs01565448_g1), forkhead box M1 (FOXM1) (Hs01073586_m1), TTK (Hs01009870_m1), BUB1 (Hs01557695_m1), and cell division cycle 20 (CDC20) (Hs00426680_mH) was carried out using TaqMan Gene Expression Assays (Applied Biosystems, Grand Island, NY). Housekeeping gene ribosomal protein, large, P0 (RPLP0) was used as a control. Relative gene expression = gene expression in the KD group / average gene expression in the control group.

Western Blot Analysis

Rabbit monoclonal PIP antibody (Novus Biologicals, Littleton, CO); rabbit antibodies for ERK1/2, phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴), c-Jun, phospho–c-Jun (Ser⁶³), Stat3, phospho-Stat3 (Try⁷⁰⁵), Cdc2, phospho-Cdc2 (Tyr¹⁵), focal adhesion kinase (FAK), and Talin-1 (Cell Signaling Technology, Danvers, MA); rabbit polyclonal integrin-β1 and mouse monoclonal Arp2/3 antibodies (Millipore, Billerica, MA); rabbit monoclonal phospho-FAK (Tyr³⁹⁷) antibody (Life Technologies, Grand Island, NY); and mouse monoclonal β-tubulin antibody (Sigma-Aldrich) were applied at 1:1000 dilutions using 20 μg of each cell lysate. Rabbit α-tubulin antibody (Abcam, Cambridge, United Kingdom) was applied to assess loading. To extract protein from media, cell lines were cultured for 48 hours in serum-free media, followed by concentration using Amicon Ultra-15 (3 K) centrifugal filters (Millipore). A total of 100 μg from each concentrated sample was precipitated and used for immunoblot analysis.

Cell Proliferation Assay

Cell proliferation assays were performed using Vybrant MTT Proliferation Assay Kit (Life Technologies) in eight replicates as previously published [10].

Cell Cycle Analysis

Cell cycle analysis with propidium iodide was performed as described before [16]. Data analysis was carried using ModFit LT software (Verity Software House, Topsham, ME).

Immunohistochemistry

Three sets of breast cancer tissue microarray (TMA) slides that are constituted of duplicate cores for a total of 210 malignant breast tumors (BRC1501-3) were obtained from Pantomics (Richmond, CA). Immunohistochemistry (IHC) staining was performed as described before [17]. Staining was carried out with rabbit PIP antibody at 1:100 dilution and mouse monoclonal antibodies (Dako, Carpinteria, CA) for AR (1:75 dilution), Ki-67 (1:100 dilution), and cytokeratin 5/6 (1:100 dilution). PIP score was defined as the percentage of PIP-positive cells (0-100) multiplied by the intensity of PIP cytoplasmic staining (1-3).

Immunofluorescence

Immunofluorescence (IF) staining was performed as previously described [15,17], with mouse monoclonal β-actin and α-tubulin antibodies (Abcam) at 1:200 dilution, mitotic protein monoclonal 2 (MPM-2; Abcam) at 1:500 dilution, and rabbit polyclonal pericentrin antibody (Abcam) at 1:1000. Alexa 488 anti-mouse and Alexa 594 anti-rabbit (Life Technologies) were used as secondary antibodies. Quantification of pericentrin/nuclear ratio was performed on 100 nuclei, and experiments were carried out in duplicates. Percentage of multinucleated cells and percentage of cells stained with MPM-2 were assessed in PIP-KD and control experiments on 200 cells, and each experiment was carried out in four replicates.

Immunoprecipitation

Immunoprecipitation (IP) of integrin-β1 using CHAPS buffer was performed as previously published [10]. To perform PIP-IP, T-47D cells were grown in 10-cm dishes to 60% confluency in full media and further cultured in serum-free media containing 100 nM dihydrotestosterone, (Thermo Fisher Scientific, Waltham, MA) for 48 hours. Conditioned media were then concentrated with Amicon Ultra-15 filters, and volume was adjusted to 500 μl with a nondenaturing lysis buffer containing 20 mM Tris-HCl (pH 8), 137 mM NaCl, 10% glycerol, 2 mM EDTA, and 1% NP-40. Cell lysate from each dish was extracted using 500 μl of lysis buffer. Subsequently, the extracted medium and lysate for each sample were mixed and subjected to IP as described [10]. PIP-IP was carried out using 2 μg of rabbit monoclonal PIP antibody, and control experiments followed the same process without a PIP antibody.

Proteomics

Following PIP-IP, protein bands were separated using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie R-250 staining (Bio-Rad Laboratories, Hercules, CA). Control experiments included pulldown with Protein A-Sepharose beads (Life Technologies) without a PIP antibody. Processing of gels for MS and database search were carried out by the Proteomics Facility at the University of Iowa (Iowa City, IA). In summary, each PIP-IP and control IP lane was sliced to 16 separate bands and analyzed by MS. In-gel digestion and sample processing were carried out as described before [18]. Database searching was performed by Mascot search engine (Matrix Science, Boston, MA), Spectrum Mill MS Proteomics Workbench (Agilent Technologies, Santa Clara, CA), and X! Tandem [The Global Proteome Machine Organization (The GPM), thegpm.org; version CYCLONE (2010.12.01.1)]. Scaffold (Proteome Software Inc, Portland, OR) was used to validate MS/MS-based peptide and protein identifications. Protein identifications were accepted at greater than 99.0% probability with at least four identified peptides [19]. Two replicates of PIP-IP and control experiments were analyzed by MS, and only hits that were present in both PIP-IP replicates and absent in the controls were accepted for further analysis.

Tubulin Polymerization Assay

Quantitation of polymerized and soluble tubulin was carried out as described before [20]. Immunoblot analysis was performed using mouse monoclonal β-tubulin antibody and the band intensities of polymerized and soluble β-tubulin in each PIP-KD experiment were normalized to that of control siRNA.

Bioinformatics and Statistical Analysis

Analysis of Gene Expression Data

Gene expression for 52 breast cancer cell lines was extracted from published microarray data by Neve et al. [21]. PIP transcriptional signature included genes that showed Pearson correlation coefficients (CCs) ≥ 0.5 with PIP expression (P < .001). Pearson CC analysis, proximity matrix, and clustering algorithms were performed using IBM SPSS Statistics 20 (Armonk, NY). Hierarchical clustering of the PIP signature was carried out using centroid linkage method, and intervals were measured by CC values. Functional annotation of the PIP signature based on Gene Ontology was performed using The Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatics Resources (National Institute of Allergy and Infectious Diseases, Bethesda, MD) [22,23].

Analysis of Proteomics Data

Functional classification of PIP-binding partners was carried out using DAVID Bioinformatics Resources. The following parameters were used for the analysis: κ similarity overlap = 4, similarity threshold = 0.35, and multiple linkage threshold = 0.50. Enrichment score was obtained for each functional cluster. Canonical pathways associated with PIP-binding partners were derived using Ingenuity Pathway Analysis (Ingenuity Systems, Redwood City, CA).

Statistical Analysis

Biostatistics was carried out using IBM SPSS Statistics 20. Student's t test and paired sample t test were applied for calculating the statistical significance. All error bars depict ± 2 SEM.

Results

PIP Expression Is Necessary for Cell Proliferation

We first characterized PIP expression in nine breast cancer cell lines from different molecular subtypes (Table W1). These included luminal A lines MCF-7 and T-47D, luminal B line BT-474, ER-negative luminal lines MDA-MB-453, HCC-202, MFM-223, and SK-BR-3, and ER-negative basal lines HCC-1954 and MDA-MB-231. PIP expression using qRT-PCR was high in HCC-202, T-47D, MDA-MB-453, and HCC-1954 cells, intermediate in BT-474, MFM-223, and SK-BR-3 cell lines, and very low to undetectable in MCF-7 and MDA-MB-231 cells (Table W1 and Figure W1A). Notably, PIP-KD resulted in ≥ 80% reduction in PIP transcripts across seven cell lines with measurable PIP (Figure 1, A and B). In addition, PIP-KD was validated at the protein level using cell extracts in cell lines with high PIP expression (Figure 1C). In cell lines with an intermediate PIP expression, due to low levels of cellular PIP (Figure W1B), conditioned media were used to measure secreted PIP levels (Figure 1D). Importantly, all seven cell lines with measurable PIP levels showed a marked reduction of this protein following PIP-KD by ≥ 80% (Figure 1, C and D).

Figure 1.

PIP silencing in breast cancer cell lines and the effect of PIP expression on cell proliferation. (A and B) qRT-PCR demonstrates PIP-KD efficiencies with siRNA-duplex 1 (D1) and siRNA-duplex 2 (D2). PIP expression is relative to nontargeting siRNA (CTL). (C) Immunoblot analysis shows PIP protein following PIP-KD using cell extracts or (D) conditioned media. Fold changes (RR) in band density were measured relative to the control and represent the average change for two siRNA duplexes. (E and F) MTT assay measures cell proliferation following PIP-KD. The asterisk (*) is P value for each PIP-KD versus control (CTL).

We next examined the effect of PIP silencing on the proliferation of breast cancer cells using MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay and observed a significant reduction in cell proliferation following PIP-KD in T-47D, BT-474, MFM-223, HCC-1954, HCC-202, and SK-BR-3 cell lines (P < .01-.03; Figure 1, E and F). These results are similar to the effect of PIP silencing on MDA-MB-453 cells [10] and suggest that PIP expression is necessary for cell proliferation across different molecular subtypes of breast cancer.

PIP Expression in Molecular Subtypes of Breast Cancer

We next investigated PIP expression in primary breast tumors to identify the pattern of PIP expression in various molecular subtypes of breast cancer and to study the association of biomarkers with PIP expression. In this process, we carried out IHC in a TMA cohort of 210 primary breast tumors (Table W2). To classify the cohort into established molecular subtypes, ER-positive tumors were subdivided into luminal A and luminal B groups using ErbB2 and Ki-67 expression patterns [24]. On the basis of this classification, luminal B was defined as ER-positive tumors with either ErbB2 overexpression or a Ki-67 index ≥ 14% (Table W3). In addition, ER-negative tumors were subdivided on the basis of their AR status [25], and cytokeratin 5/6 (CK5/6) staining was employed as a marker of basal-type tumors. Furthermore, we obtained IHC scores for PIP on the basis of the percentage and intensity of staining for this protein in each tumor (Figure 2, A and B).

Figure 2.

PIP expression in breast tumors. (A and B) PIP expression using IHC in breast tumors. Magnifications are at 10X. (C) PIP expression scores using IHC in molecular subtypes of breast cancer is shown. *P < .01 is for ER −/AR − versus other groups. (D) PIP-IHC scores in ER-negative tumors stratified on the basis of AR and CK5/6 status are shown. *P < .01 is for AR − tumors versus other groups.

Next, we assessed the association of PIP expression with biomarkers and molecular subtypes. We observed that AR-positive tumors have a significantly higher PIP expression compared to AR negatives (P < .01; Table W3). In addition, PIP was highly expressed in luminal A, luminal B, and ER −/AR + molecular subtypes (Figure 2C and Table W3). Furthermore, among ER-negative tumors, PIP expression was associated with AR-positive status and was unrelated to CK5/6 staining (Figure 2D). It is notable that PIP expression was not associated with either tumor size or grade in this cohort (P > .1). These findings suggest that PIP is widely expressed in luminal A, luminal B, and ER −/AR + (molecular apocrine) subtypes of breast cancer. Moreover, PIP expression is associated with AR and is present in both luminal and basal ER-negative tumors.

PIP Transcriptional Signature Is Enriched with Cell Cycle Genes

To investigate the functional role of PIP in breast cancer, we carried out a nonbiased genomic approach to study the transcriptional signature of this gene using a microarray data set of 52 breast cancer cell lines [21]. To identify PIP coregulated genes, we first calculated the Pearson CC for each gene expression in the data set with that of PIP and then obtained the list of genes that had Pearson CC values ≥ 0.5 (P < .001) with PIP expression. This PIP transcriptional signature is composed of 136 genes that had a highly coregulated expression pattern with PIP (Tables W4 and W5). We next performed hierarchical clustering analysis of PIP transcriptional signature and observed two main gene clusters in this signature on the basis of the direction of CC values with PIP expression (Figure 3A and Table W6).

Figure 3.

PIP transcriptional signature. (A) Hierarchical clustering analysis of PIP transcriptional signature was performed using centroid linkage method, and intervals were measured by Pearson CCs. Functional annotations for gene clusters are demonstrated. FE, fold enrichment. (B) Functional annotation of PIP transcriptional signature based on Gene Ontology is presented. FEs and P values are shown.

We subsequently conducted functional annotation of the PIP signature. Notably, this signature was highly enriched with cell cycle genes and in particular those related to mitotic transition and spindle checkpoint [Figure 3B, fold enrichment (FE) values = 11-56; P < .001]. These findings suggest a strong transcriptional coregulation between PIP and cell cycle–related genes that is most pronounced in relation to mitotic transition.

PIP Expression Is Required for G₁-S Progression

In view of the fact that PIP transcriptional signature is highly saturated with cell cycle genes and PIP expression is necessary for cell proliferation in breast cancer, we investigated the role of PIP in cell cycle progression using flow cytometry analysis on PIP-silenced cells. We observed that T-47D and MDA-MB-453 cells underwent a profound G₁ arrest following PIP-KD manifested by a 10% to 20% increase in G₀-G₁ cell population compared to the controls with a corresponding decrease in the percentage of cells in S phase (P < .01; Figure 4, A and B). Furthermore, this G₁ arrest was associated with a marked decrease in cyclin D1 and cyclin E1 expression following PIP-KD (P < .01; Figure 4, C and D).

Figure 4.

The effect of PIP expression on G₁ phase. (A) Cell cycle histograms following PIP-KD in T-47D are shown. (B) The percentage of cells in different phases of cell cycle following PIP-KD is shown. P < .01 is for ΔG_0 -1 between PIP-KD and CTL groups. (C) Cyclin D1 and (D) cyclin E1 expression using qRT-PCR for PIP-KD relative to control is shown. (E) Immunoblot analysis measures the ratio of phospho (Ph)-ERK to total (T) ERK, Ph–c-Jun to T–c-Jun, and Ph-Stat3 to T-Stat3 following PIP-KD. Fold changes were assessed relative to control. The average changes obtained for two duplexes are presented.

Moreover, we examined the effect of PIP-KD on the level of ERK phosphorylation that is a required step for the transcriptional activation of cyclin D1 [26]. In addition, we have previously demonstrated a reduction in ERK phosphorylation in MDA-MB-453 cells following PIP silencing [10]. Notably, phospho-/total ERK was markedly reduced by 0.14-fold following PIP-KD in T-47D cells (Figure 4E), which represents a similar pattern to that observed in MDA-MB-453 [10]. We also examined the effect of PIP silencing on phosphorylation of c-Jun and STAT3 that are also involved in G₁/S transition; however, we did not find any significant changes in the level of these proteins (Figure 4E). These data suggest that PIP silencing results in a profound G₁ arrest in T-47D and MDA-MB-453 cells associated with a reduction in the levels of cyclin D1, cyclin E1, and ERK phosphorylation.

PIP Silencing Leads to Mitotic Arrest and Aneuploidy

Cell cycle studies in MFM-223, SK-BR-3, HCC-1954, HCC-202, and BT-474 cell lines following PIP silencing revealed that these lines undergo a G₂/M arrest manifested by a significant increase in G₂/M phase and a marked increase in the percentage of aneuploidy by approximately 15% to 30% (P < .01; Figure 5, A–C). In addition, there was a moderate degree of G₁ arrest following PIP-KD in four of these cell lines (P < .01; Figure 5A). In contrast, T-47D and MDA-MB-453 cells did not have an increase in aneuploidy following PIP silencing (Figure 5, A and B). These data suggest that PIP expression is required for progression through both G₁ and G₂/M phases of the cell cycle and the occurrence of G₂/M arrest following PIP silencing is accompanied by an increase in aneuploidy.

Figure 5.

The effect of PIP expression on G₂/M and aneuploidy. (A) The percentage of cells in different phases of cell cycle following PIP-KD is shown. P values are for ΔG_0 -1 and ΔG₂/M between PIP-KD and control groups. (B) The change in percentage of aneuploidy between PIP-KD and control (CT) experiments is presented. (C) Cell cycle histograms following PIP-KD in BT-474 cell line are shown. (D) Cyclin D1 expression using qRT-PCR for PIP-KD relative to control (CTL) is shown. *P < .01 is for PIP-KD versus CTL groups. (E) T- and Ph-Cdc2 protein levels by immunoblot analysis for PIP-KD relative (RR) to control are shown. (F) Cyclin B1 expression for PIP-KD as explained in D is shown. The average changes obtained for two duplexes are presented.

To identify a molecular basis for the observed cell cycle findings, we assessed the effect of PIP silencing on the expression of key genes involved in G₁, G₂/M checkpoint, and mitotic transition. We observed that cyclin D1 expression, a key regulator of G₁, was significantly reduced in MFM-223, HCC1954, HCC202, and BT-474 cells following PIP-KD (P < .01; Figure 5D). We next examined the effect of PIP silencing on Cdc2 (Cdk1) and cyclin B1 as main regulators of G₂/M checkpoint [27]. Notably, there was a proportionate reduction in total and phospho-Cdc2 protein levels following PIP-KD in HCC-1954 and HCC-202 cell lines that suggests a decrease in the expression of this protein following PIP silencing (Figure 5E). Furthermore, cyclin B1 expression was markedly reduced by approximately 50% to 80% in all five lines (P < .01; Figure 5F).

In view of the fact that functional annotation showed a strong correlation between PIP and mitotic transition (Figure 3), we also studied the effect of PIP silencing on some of the key mitotic genes in PIP signature. In this respect, we examined FOXM1, TTK, BUB1, and CDC20, which have a critical role in mitotic transition [28–30]. Expression levels of these genes were assessed following PIP-KD in BT-474, HCC-1954, MMF-223, SK-BR-3, HCC-202, and MDA-MB-453 cell lines. Importantly, there was a significant reduction in FOXM1, TTK, BUB1, and CDC20 expression following PIP-KD in these cells that supports a functional role for PIP in the regulation of mitotic transition (P < .01; Figure 6, A–D).

Figure 6.

The effect of PIP expression on mitosis. (A–D) The effects of PIP silencing on mitotic transition genes are presented. Expression of FOXM1, TTK, BUB1, and CDC20 are measured using qRT-PCR for PIP-KD relative to control. P values are for PIP-KD versus CTL groups in each cell line. The asterisk (*) denotes P < .01 in BUB1 experiments. The average changes obtained for two duplexes are presented. (E and F) IF with the mitotic marker MPM-2. IF staining for MPM-2/Alexa 488 was carried for PIP-KD and control siRNA experiments in MFM-223, BT-474, and HCC-1954 cell lines. Percentage of MPM-2 staining was calculated in 200 nuclei for each experiment, and the average changes obtained for two duplexes are presented. 4',6-diamidino-2-phenylindole (DAPI) DAPI staining was used to assess the nuclei. *P value is for PIP-KD versus control groups.

We next investigated whether the observed G₂/M accumulation following PIP silencing is a result of an arrest in G₂ or M phase of the cell cycle. This was studied using IF staining with the mitotic marker MPM-2 that stains mitotic cells after G₂ phase [31]. IF for MPM-2 was carried out in MFM-223, BT-474, and HCC-1954 cell lines, and the percentage of nuclei stained with MPM-2 was assessed for PIP-KD and control experiments (Figure 6, E and F). We observed a significant increase in MPM-2 staining by two- to three-fold in PIP-KD cells compared to the control, suggesting an arrest in mitotic transition following PIP silencing (P < .01; Figure 6, E and F). Overall, these findings indicate that PIP expression is required for mitotic transition in breast cancer and the effect of PIP silencing on cell cycle corresponds to the transcriptional changes in key cell cycle genes.

PIP Silencing Results in a Cytokinesis Defect

We subsequently studied the effect of PIP expression on cytokinesis. It is established that microtubules and actin organization are essential for this process [32]. Therefore, we first examined the effect of PIP silencing on actin microfilaments using IF in BT-474, HCC-1954, and MFM-223 cell lines. We observed that actin organization was markedly disrupted after PIP-KD, resulting in abnormally shaped and large filopodia protrusions and irregular lamellipodia that were formed in multinucleated cells (Figures 7A and W2, A and B). In addition, as opposed to the control cells that demonstrated polarity in actin organization as evidenced by the orientation of filopodia and retraction fibers, there was a loss of polarity in actin filaments following PIP silencing (Figure 7A). Furthermore, formation of cleavage furrow, an essential step in cytokinesis, was disrupted in some dividing PIP-KD cells (Figure W2A).

Figure 7.

The effect of PIP on cytokinesis and integrin signaling. (A) IF staining for β-actin/Alexa488 following PIP-KD in BT-474 cells is shown. Control and PIP-KD cells are shown during cell division (top panels), and a multinucleated cell following PIP-KD is shown in bottom panel. White arrow, filopodia (fil); yellow arrow, lamellipodia (lam); orange arrow, retraction fibers (rf); and magenta arrow, direction of actin polarity. (B) IF staining for α-tubulin (Tub) and pericentrin following PIP-KD in MFM-223 cells is shown. (C) Pericentrin to nuclear ratios following PIP-KD are presented. DAPI staining was used to assess the nuclei. *P value is for PIP-KD versus control groups. (D) Change in percentage of multinucleated cells between PIP-KD and control cell lines is shown. IF following β-actin and DAPI staining was used to assess multinucleated cells. *P value is for PIP-KD versus control groups. (E) Immunoblot analysis measures the ratio of Ph-FAK (Tyr³⁹⁷) to T-FAK following PIP-KD in cell lines. Fold changes were assessed relative to control. Experiments were carried out in four replicates using two PIP-siRNA duplexes or control siRNA, and mean changes (± SEM) were shown. (F) IP assesses integrin-β1 (ITG-β1) binding to talin-1 following PIP-KD. IP and immunoblot analysis were carried out with ITG-β1 and talin-1 antibodies, respectively. Membrane was stripped, and immunoblot analysis for ITG-β1 was used to assess loading. ITG-β1 immunoblot for input control is shown. Fold changes were assessed relative to control. Experiments were carried out in four replicates using two PIP-siRNA duplexes or control siRNA, and mean change (± SEM) is shown for each cell line.

We next assessed the formation of microtubules and pericentrin to nuclear ratio following PIP silencing in these cells, because supernumerary percentrins during cell division are associated with cytokinesis defect and multinucleation [33]. Notably, pericentrin/nuclear ratios significantly increased following PIP-KD in all three cell lines by 1.4- to 2.2-fold compared to controls, suggesting that there are supernumerary percentrins (P < .03; Figures 7, B and C, and W2C). In addition, α-tubulin staining revealed that PIP silencing leads to multipolar spindle formation during cell division and an absence of distinct microtubules in multinuclear cells (Figures 7B and W2C).

Moreover, we observed that there is a 20% to 60% increase in the number of multinucleated cells following PIP silencing in MFM-223, SK-BR-3, BT-474, HCC-1954, and HCC-202 cell lines (P < .01; Figure 7D). However, T-47D and MDA-MB-453 cells did not demonstrate a significant increase in the number of multinucleated cells after PIP-KD. Importantly, these findings are in agreement with the occurrence of aneuploidy among these cell lines. Overall, our data suggest that PIP silencing disrupts the organization of actin microfilaments and microtubules and leads to an increase in the number of pericentrins and multinucleated cells that are all indicators of a cytokinesis defect.

PIP Is Required for Inside-Out Activation of Integrin-β1 Signaling

It is known that dysregulation of integrin-β1 signaling results in cell cycle defects in G₁ progression and cytokinesis [34]. Therefore, we investigated the effect of PIP silencing on FAK phosphorylation (Tyr³⁹⁷), which is a key downstream mediator of integrin-β1 signaling and integrin-β1 binding to talin-1 that is a required step for inside-out activation of integrins [35,36].

We observed that the level of phospho-FAK was generally low in breast cancer cell lines and was only detectable in HCC-1954 and HCC-202 cells. In addition, PIP-KD resulted in the reduction of phospho-/total FAK ratios by approximately 50-70% in these two lines, which was partly related to a relative increase in the total FAK levels following PIP-KD (Figure 7E). Furthermore, we found a baseline interaction between integrin-β1 and talin-1 in T-47D and MFM-223 cell lines using IP with integrin-β1 and immunoblot analysis with talin-1 (Figure 7F). Importantly, integrin-β1 binding to talin-1 was abrogated in T-47D cells and reduced by 0.5-fold in MFM-233 following PIP silencing (Figure 7E). These data suggest that PIP expression is necessary for inside-out activation and signaling effects of integrin-β1.

Identification of PIP-Binding Partners

To identify protein-binding partners for endogenous PIP, we carried out MS. These experiments were performed in T-47D cell line, which has a high level of endogenous PIP expression that was further induced using dihydrotestosterone. Because PIP is a secreted protein, IP was performed using a combination of cell extracts and conditioned media. The result of PIP pulldown was first validated using immunoblot analysis with PIP antibody (Figure 8A). PIP-IP bands were then separated using SDS-PAGE and Coomassie staining, followed by MS analysis (Figure 8B).

Figure 8.

Identification of PIP-binding partners. (A) IP and immunoblot analysis (IB) with PIP antibody. The low molecular weight band may represent a PIP fragment product. (B) Coomassie staining of SDS-PAGE for PIP-IP and control pulldowns is shown. (C) Functional classification of PIP-binding partners is shown. Enrichment score and P value are shown. (D) IP with PIP antibody and IB with β-tubulin, PIP, and Arp2/3 antibodies in T-47D cells are shown. A nonspecific rabbit IgG was used for control IP. Membrane was stripped, and IB for PIP was used to assess loading. PIP immunoblot for input control is shown. (E) Microtubule polymerization assay measures polymerized and soluble tubulin fractions following PIP-KD. The amount of each fraction following PIP-KD was normalized to that of control, and the relative ratio of Pol/Sol fractions was obtained for each cell line. The average changes obtained for two duplexes are presented.

We identified a total of 156 protein-binding partners for PIP that were reproducible between two replicate experiments (Table W7). Functional classification of these binding partners using bioinformatics revealed six highly significant clusters (Figure 8C and Table W8). These functional clusters included translational elongation, ribonuclear protein, nucleosome and chromatin assembly, regulation of actin and cytoskeleton, clathrin-coated pit, and small guanosine diphosphate (GDP)-binding protein. In addition, we studied the functional association of PIP-binding partners with the canonical pathways using Ingenuity Pathway Analysis. The top identified pathways included eukaryotic Initiation Factor 2 (eIF2) and eIF4 signaling, followed by remodeling of epithelial adherens junctions, clathrin-mediated endocytosis, regulation of actin-based motility by Rho, and integrin signaling (Table W9).

We further validated two of the identified PIP-binding partners that have key functions in cell cycle (Table W7). Notably, we identified β-tubulin as one of the top PIP-binding partners, which has a well-established role in mitosis [37]. In addition, Arp2/3 protein is another PIP-binding partner with a significant role in cytokinesis and promoting talin binding to integrins [38,39]. To validate these interactions, we carried out PIP-IP and performed immunoblot analysis with β-tubulin and Arp2/3 antibodies on pulldowns. Immunoblot analysis with PIP antibody was used to confirm a successful PIP pulldown (Figure 8D). Notably, we observed a strong interaction between PIP and β-tubulin and detected a specific interaction between PIP and Arp2/3 protein in PIP pulldown (Figure 8D). Taken together, our proteomic studies identified novel PIP-binding proteins.

PIP Is Necessary for Microtubule Polymerization

To assess whether PIP binding to β-tubulin has a regulatory effect on microtubule polymerization, we measured the effect of PIP silencing on the polymerized and soluble fractions of tubulin in T-47D and HCC-1954 cell lines. Immunoblot analysis was carried out using a β-tubulin antibody to measure polymerized and soluble tubulin fractions. The amount of each fraction following PIP silencing was normalized to that of siRNA control, and the relative ratio of polymerized to soluble (Pol/Sol) fractions was obtained for each cell line. We observed that Pol/Sol tubulin ratio was markedly reduced following PIP-KD to 0.31- and 0.06-fold of controls in T-47D and HCC-1954 cell lines, respectively (Figure 8E). These findings suggest that PIP is required for tubulin polymerization in breast cancer cells.

Discussion

PIP is widely expressed in breast cancer and is used as a characteristic biomarker in this disease; however, the molecular functions of PIP have remained largely unknown. Therefore, we carried out a comprehensive study to identify the underlying mechanisms for PIP function in breast cancer. In this process, we employed seven breast cancer cell lines that encompass luminal A, luminal B, and molecular apocrine subtypes. It is notable that these molecular subtypes also correspond to the pattern of PIP expression among primary breast tumors. Previous studies have shown a positive regulatory role for PIP in cell proliferation [9,10]. In this study, we demonstrated that PIP expression is necessary for the proliferation of all breast cancer cell lines that have a detectable level of PIP. Although PIP expression varied among different cell lines, the impact of PIP on cell proliferation was similar across these cells. Moreover, the only two cell lines with undetectable levels of PIP were MCF-7 and MDA-MB-231, which do not have AR expression. This association between AR and PIP expression is also present among primary breast tumors and can be explained by the fact that AR is a transcriptional regulator of PIP [8–10]. Overall, the pattern of PIP association with molecular subtypes and biomarkers are similar between breast cancer cell lines and breast tumors.

The effect of PIP on proliferation is explained by the fact that PIP expression is required for cell cycle progression in breast cancer cells and PIP silencing leads to defects in G₁, mitosis, and cytokinesis. In addition, we show that PIP interacts with β-tubulin and is necessary for tubulin polymerization. This interaction is particularly significant because microtubules are known to have a critical role in mitotic transition, spindle assembly, and cytokinesis [37,40]. Moreover, proteomic data provide further evidence for the importance of PIP interaction with β-tubulin in the functional characterization of PIP-binding partners. In fact, a major portion of PIP interactions that contribute to the observed functional classification are known “Tubulin-Binding Proteins” (Figure 8C and Table W7). These include RNA-binding proteins, proteins involved in translation such as eIF4, heat shock proteins, clathrin, and 14-3-3 protein family [41–43]. It is notable that 14-3-3 protein is a key regulator of cell cycle and clathrin is required for mitotic spindle function and endocytosis [42,43]. In addition, mRNA localization to microtubules contributes to the translation of genes involved in cell division such as cyclin B1 [44]. Therefore, PIP regulation of microtubular polymerization and its interaction with other tubulin-binding proteins would have a profound effect on cell cycle (Figure 9).

Figure 9.

A schematic model for PIP regulation of cell cycle. The proposed mechanisms by which extracellular and intracellular PIP can regulate cell cycle are depicted. Fn, fibronectin; Fn-f, fibronectin fragments; ITG, integrin-β1. Red arrows indicate positive regulation. Cell membrane has been depicted by a circular line. Arp2/3, β-tubulin, and histones interact with PIP based on our study.

Another major group of PIP interactions involves actin-binding proteins. Among these, Arp2/3 has a critical role in actin polymerization, and along with vinculin and talin, it provides a physical link between actin cytoskeleton and the integrin scaffold, which is needed to transform talin binding to integrin-β1 from a low- to high-affinity state [38]. Therefore, PIP interaction with Arp2/3 can explain the marked reduction in talin binding to integrin-β1, which is a required step for inside-out activation of integrins, and a decrease in FAK phosphorylation following PIP silencing. It is notable that outside-in signaling of integrin-β1 binding to integrin-linked kinase 1 (ILK1) is also regulated by PIP in a process that partially depends on fibronectin fragmentation (Figure 9), [10]. In addition to Arp2/3, some of the other PIP-binding partners including gelsolin, cofilin 1, F-actin–capping protein, α-actinin, and small GTP-binding proteins have key roles in actin organization, formation of focal adhesions, and cytokinesis [39,45]. In fact, our study provides functional evidence for the importance of these interactions as shown by a defect in cytokinesis following PIP silencing that is associated with abnormal actin organization (Figures 7 and W2). Importantly, some of the PIP-binding partners such as Rho-GTPase are known to coordinate cytokinesis and cell polarity [46]. Overall, our findings indicate that a key feature of PIP function is the regulation of cytoskeleton (Figure 9).

Our data suggest that PIP expression is necessary for both G₁/S and mitotic progression in breast cancer cells. However, the impact of PIP on each phase of cell cycle varies among breast cancer lines. In this respect, we observed two main patterns for the effect of PIP silencing on cell cycle. In one group constituted of T-47D and MDA-MB-453 cell lines, there was a severe degree of G₁ arrest that was not associated with mitotic arrest or aneuploidy. Importantly, this pattern was associated with a profound reduction in cyclin D1 expression following PIP silencing. In comparison, the remaining cell lines developed a moderate degree of G₁ arrest accompanied by mitotic arrest and a marked increase in aneuploidy (Figure 5, A and B). Furthermore, our genomics data on PIP transcriptional signature and a marked reduction in key mitotic transition genes such as cyclin B1, BUB1, and CDC20 following PIP down-regulation suggest the importance of PIP expression in mitosis. It is notable that the emergence of aneuploidy can be a consequence of defects in both mitotic checkpoint and cytokinesis [47]. In particular, dysregulation of mitotic and spindle checkpoint genes such as BUB1 and FOXM1 have been associated with aneuploidy [48,49]. Importantly, this increase in aneuploidy can further contribute to a reduction in cell proliferation [50].

We observed that PIP expression is necessary for the transcription of multiple genes involved in cell cycle progression. Some of these effects can be explained by PIP regulation of the upstream signaling pathways. For example, cyclin D1 is a target of integrin-β1 mediated through the activation of ERK [34]. Therefore, the effect of PIP on cyclin D1 can be explained by the fact that G₁ arrest in T-47D cells is accompanied by a marked reduction in ERK phosphorylation associated with an abrogation of talin binding to integrin-β1. In addition, PIP transcriptional signature shows a robust pattern of coregulation between PIP and mitotic transition genes. This finding along with a profound decrease in the expression of mitotic genes observed following PIP silencing suggest that PIP is required to maintain a balance in the expression of key genes involved in mitotic transition. This is especially important because both overexpression and down-regulation of some of the mitotic genes such as BUB1 can lead to abnormal mitosis and aneuploidy [48,51]. Notably, FOXM1 is known to be a transcription factor for multiple genes involved in cell cycle progression including cyclin B1, BUB1, CDC20, and polo-like kinase (PLK) [28]. Therefore, PIP regulation of FOXM1 expression can explain many of the transcriptional changes observed following PIP silencing. In addition, soluble tubulin has been shown to interact with histones and regulate transcription [52]. In view of the fact that PIP interacts with histones (Table W7) and also regulates soluble tubulin levels, the possibility of a transcriptional role for PIP deserves further studies (Figure 9).

In summary, we propose that PIP has a versatile function in breast cancer resulting from a diverse range of both intracellular and extracellular binding partners (Figure 9). As a consequence of these functional interactions, PIP can regulate key cellular processes including outside-in and inside-out activation of integrin-β1, transcription of key cell cycle genes such as FOXM1, and cytoskeletal organization including microtubule polymerization. The net effect of these molecular functions is the fact that PIP can profoundly influence cell cycle progression in breast cancer cells (Figure 9). Importantly, our findings provide a rationale for the tantalizing possibility of exploring PIP as a therapeutic target in breast cancer.

Supplementary Materials.

Figure W1.

PIP expression in breast cancer cell lines. (A) Relative PIP expression in breast cancer cell lines to that of HCC202 cells presented in a logarithmic scale (base 2). (B) PIP protein expression in BT-474 and MFM-223 cell lines following PIP-knockdown (KD) using cell lysate samples. CTL: cells transfected with the control-siRNA. RR: relative ratio to control.

Figure W2.

PIP effect on cytokinesis. (A-B) Immunofluorescence (IF) staining for β-Actin in HCC-1954 and MFM-223 cell lines. White arrow: filopodia (fil), blue arrow: cleavage furrow (cf). (A) Cleavage furrow formation is present in dividing control cells (CT) and is absent in dividing cells with PIP-knockdown (KD). Multinucleated cells are shown following PIP-KD in HCC-1954 (A) and MFM-223 cells (B). IF was carried out with β-Actin antibody and Alexa488 was used as secondary antibody. (C) IF staining for α-Tub/Pericentrin. IF was carried out with, α-Tubulin and Pericentrin antibodies in BT-474 cell line. Alexa488 and Alexa 594 antibodies were used as secondaries. There is an absence of distinct microtubules and presence of supernumerary pericentrins in PIP-KD cells. Magnifications are shown for each panel.

Table W1.

Characteristics of Breast Cancer Cell Lines Used in the Functional Study of PIP.

Cell Line	ER	ErbB2	AR	Subtype	PIP − ΔCT
MCF-7	POS	NEG	NEG	Luminal A	− 17.5 (± 0.08)
T-47D	POS	NEG	POS	Luminal A	− 4.9 (± 0.04)
BT-474	POS	POS	POS	Luminal B	− 12.2 (± 0.08)
HCC-202	NEG	POS	POS	Luminal	− 1.1 (± 0.13)
HCC-1954	NEG	POS	POS	Basal A	− 7.3 (± 0.02)
MDA-MB-453	NEG	POS	POS	Luminal	− 6.2 (± 0.06)
SK-BR-3	NEG	POS	NEG	Luminal	− 11.1 (± 0.08)
MFM-223	NEG	NEG	POS	Luminal	− 11.8 (± 0.03)
MDA-MB-231	NEG	NEG	NEG	Basal B	− 20 (± 0)⁎

Cell subtype and AR status were obtained from Lehmann BD et al. (2011) [53], Magklara A et al.(2002) [54], Heiser LM et al. (2012) [55], Naderi A et al. (2010) [56], Naderi A et al. (2008) [57], Doane AS et al. (2006) [58], and Neve RM et al. (2006) [59]. ErbB2 status was obtained from Subik K et al. (2010) [60], Ginestier C et al. (2007) [61], Lombardi DP et al. (2004) [62], and Agarwal R et al. (2009) [63]. POS, positive; NEG, negative. PIP − ΔCT is − Δ cycle threshold value (± SEM) for PIP expression using qRT-PCR.

^⁎PIP expression was not detectable after 40 cycles of qRT-PCR.

Table W2.

Characteristics of the TMA Cohort.

Feature	Status	Percentage
Histology	IDC⁎	85%
	Others⁎⁎	15%
ER	Negative	51%
	Positive	49%
ErbB2	0-1	52%
	2-3	48%
p53	0-1	68%
	> 2	32%

Percentage is calculated in a total of 210 primary breast tumors.

^⁎IDC, invasive ductal carcinoma.

^⁎⁎Others: Ductal carcinoma in situ, Paget disease, invasive lobular carcinoma, invasive micropapillary carcinoma, invasive tubulo-lobular carcinoma, invasive papillary carcinoma, lobular carcinoma in situ, invasive carcinoma with apocrine features, invasive tubular carcinoma, intraductal carcinoma, and tubular mixed carcinoma.

Table W3.

Association of PIP Expression with Molecular Features in Breast Cancer Cohort.

Biomarkers
Marker	ErbB2		ER		AR		p53
Status	< 2	2-3	Neg	Pos	Neg	Pos	Neg	Pos
PIP (SEM)	112 (8)	126 (9)	111 (9)	126 (9)	80 (11)	132 (7)	119 (7)	118 (11)
P value	> .1		> .1		< .01⁎		> .1

Molecular Subtypes
Subtype	Luminal A	Luminal B	ER −/AR +	ER −/AR −
Frequency	28%	21%	31%	20% (Basal: 10%)
PIP (SEM)	116 (12)	136 (12)	141 (11)	62 (12)
P value	P < .01⁎⁎	P < .01⁎⁎	P < .01⁎⁎

Mean scores for PIP expression are demonstrated in 210 primary breast tumors. Luminal B, ER + with Ki-67 index ≥ 14% or ErbB2 staining score of 3; AR +, ≥ 10% nuclear staining; Basal, CK5/6 +; SEM, standard error of mean.

^⁎P value for AR-negative (Neg) versus AR-positive (Pos) tumors.

^⁎⁎P values versus ER −/AR − group.

Table W4.

Proximity Matrix for PIP Coregulated Genes. Each Gene has Pearson CC ≥ 0.5 with PIP Expression at a Significant Level of P < .001.

Table W5.

Transcriptional Signature of PIP Coregulated Genes.

Gene	CC	P <	Function⁎
ACOX1	0.518	.001	Fatty acid pathway
ACY1	0.504	.001	Hydrolysis
AGA	0.525	.001	Lysosomal
AGR2	0.592	.001	Cell migration
ALDH6A1	0.515	.001	Mitochondrial
AR	0.59	.001	Steroid receptor
ARHGEF16	0.5	.001	Cell migration
ARHH	0.572	.001	Survival and migration
ARRB1	0.547	.001	G-protein receptor
ATP2A3	0.554	.001	Ca2 + function
BIK	0.545	.001	Apoptosis
ANKRA2	0.527	.001	Cytoskeletal
CACFD1	0.522	.001
CPD	0.606	.001	Peptidase
CRAT	0.51	.001	Mitochondrial
SGSM3	0.702	.001	neurofibromatosis 2 (NF-2) signaling
LRP10	0.603	.001	Lipid metabolism
DUSP4	0.51	.001	mitogen-activated protein kinase (MAPK) signaling
EGFL3	0.506	.001
ERRB3	0.501	.001	epidermal growth factor receptor (EGFR) signaling
FKSG28	0.583	.001	Proliferation
SRD5A3	0.572	.001	Androgen metabolism
TMEM132A	0.563	.001	Cell death
TCTN1	0.523	.001	Hedgehog signaling
RHBDF1	0.523	.001	EGFR signaling
FMO4	0.502	.001	Metabolism
FRAT1	0.503	.001	Wnt signaling
FZD4	0.569	.001	Wnt signaling
GABRA3	0.513	.001	Neurotransmitter
HMIC	0.504	.001	Metabolism
HNF3A	0.518	.001	Steroid response
HOXC10	0.501	.001	Transcription factor
HPIP	0.564	.001	estrogen receptor-alpha (ESR) signaling
ICA1	0.517	.001	Secretory function
ITPR1	0.545	.001	Ca2 + signaling
KHNYN	0.562	.001
TRIL	0.587	.001	Cytokine secretion
KIF13B	0.632	.001	Cytoskeletal
KMO	0.52	.001	Metabolism
LDB1	0.532	.001	Transcription
LDB3	0.598	.001	Cytoskeletal
LFG	0.501	.001	Apoptosis
NFATC4	0.513	.001	Transcription
P24B	0.561	.001	Protein trafficking
P2RX4	0.547	.001	Ion channel
P2RY6	0.504	.001	G-protein receptor
PAPSS2	0.521	.001	Metabolism
PCK2	0.501	.001	Metabolism
PDEF	0.521	.001	Transcription
PISD	0.565	.001	Mitochondrial
PRKAG1	0.514	.001	Metabolism
PRKCH	0.641	.001	Protein kinase
PRSS8	0.535	.001	Sserine protease
PXMP4	0.518	.001
RAB5B	0.552	.001	Protein transport
SEMA3F	0.511	.001	Cell motility
SERF2	0.513	.001
SERHL	0.518	.001	Serine hydrolase
SH3GLB2	0.512	.001
NHE2	0.517	.001	Channel protein
SPRY1	0.521	.001	Fibroblast growth factor signaling
SPTLC2	0.645	.001	Metabolism
SSBP2	0.502	.001	Genome stability
SUOX	0.595	.001	Mitochondrial
TJP3	0.52	.001	Cytoskeletal mitosis
TM7SF1	0.505	.001
TM7SF2	0.646	.001	Sterol metabolism
TM9SF1	0.503	.001	Autophagy
TMEM8	0.506	.001	Adhesion
TSPAN1	0.552	.001	Growth and motility
TST	0.506	.001	Mitochondrial
ULK1	0.527	.001	Autophagy
VIPR1	0.5	.001	G-protein receptor
WSB1	0.551	.001	Proteasome
XBP1	0.505	.001	Transcription
ZDHHC3	0.538	.001	Cell surface stability
PATZ1	0.516	.001	Transcription
BUB1	− 0.607	.001	Mitosis
CCNA2	− 0.584	.001	Cell cycle
CCNB2	− 0.574	.001	Mitosis
CDC20	− 0.588	.001	Mitosis
CDC5L	− 0.531	.001	Cell cycle
CENPE	− 0.565	.001	Mitosis
CTPS	− 0.531	.001	Cell growth
DDX18	− 0.54	.001	Cell growth and division
HJURP	− 0.662	.001	Centrosome function
DNMT1	− 0.563	.001	Epigenetics
E2_EPF	− 0.52	.001
E2F3	− 0.504	.001	Cell cycle
FBXO5	− 0.566	.001	Mitosis
FAM64A	− 0.521	.001	Mitosis
PRR11	− 0.599	.001
QTRTD1	− 0.503	.001	RNA synthesis
NCAPG2	− 0.589	.001	Mitosis
SPDL1	− 0.536	.001	Mitosis
IARS	− 0.504	.001	RNA synthesis
FASTKD1	− 0.512	.001
FOXM1	− 0.54	.001	Cell cycle
GLS_C	− 0.52	.001	Metabolism
HCAP_G	− 0.557	.001	Mitosis
HNRPD_E	− 0.562	.001	RNA function
HNRPH1	− 0.613	.001	RNA function
DLGAP5	− 0.548	.001	Mitosis
NUP205	− 0.54	.001	Nuclear transport
EHBP1	− 0.542	.001	Cytoskeletal actin
KIF4A	− 0.515	.001	Cytokinesis
MAD2L1	− 0.561	.001	Mitosis
METAP1	− 0.515	.001	Cell cycle
MSH2	− 0.507	.001	Genome stability
NCL	− 0.559	.001	Transcription
NUP54	− 0.578	.001	Nuclear transport
OSBPL11	− 0.559	.001	Lipid metabolism
PDCD5	− 0.64	.001	Apoptosis
PLK	− 0.535	.001	Cell cycle
PLS3	− 0.616	.001	Actin binding
PMS1	− 0.514	.001	Genome stability
PMSCL1	− 0.578	.001	RNA function
PSMD1	− 0.522	.001	Proteasome
RAB6KIFL	− 0.522	.001	Cytokinesis
RHEB2	− 0.518	.001	Ras-GTPase
RUVBL2	− 0.521	.001	DNA repair
SFRS10	− 0.533	.001	Splicing factor
SFRS7	− 0.539	.001	Splicing factor
SIL	− 0.59	.001	Mitosis
ORNT1	− 0.53	.001	Mitochondrial
SMC4L1	− 0.578	.001	Mitosis
SNRPD1	− 0.584	.001	RNA function
STK12 (Aurora B)	− 0.5	.001	Mitosis and cytokinesis
TSN	− 0.512	.001	Chromosomal function
TTK	− 0.639	.001	Mitosis
UBA2	− 0.581	.001	Protein modification
UBE2C	− 0.541	.001	Mitosis
UPF3B	− 0.539	.001	RNA function
USP1	− 0.555	.001	DNA repair
XPO1	− 0.563	.001	Protein export
ZRF1	− 0.562	.001	Transcription

List of genes that have Pearson CCs ≥ 0.5 with PIP expression at a significance level of P < .001.

Raw data for gene expression values were extracted from Affymetrix microarray data set published by Neve RM et al. 2006.

^⁎Proposed gene Function is derived from GeneCards (www.genecards.org).

Table W6.

Dendrogram Using Centroid Linkage.

Table W7.

Identified PIP-Binding Proteins (Protein Threshold at 99% and Peptide Threshold at 0.1% false discovery rate [FDR]).

No		Accession No.
1	Cluster of tubulin β chain; Organism Species (OS) = Homo sapiens; Gene Name (GN) = TUBB; PE = 1; Splice Variant (SV) = 2 (TBB5_HUMAN)	TBB5_HUMAN [2]
1.1	Tubulin β chain; OS = H sapiens; GN = TUBB; PE = 1; SV = 2	TBB5_HUMAN
1.2	Tubulin β-4B chain; OS = H sapiens; GN = TUBB4B; Protein Existence (PE) = 1; SV = 1	TBB4B_HUMAN
2	Heat shock 70-kDa protein 1A/1B; OS = H sapiens; GN = HSPA1A; PE = 1; SV = 5	HSP71_HUMAN
3	Unconventional myosin-Id; OS = H sapiens; GN = MYO1D; PE = 1; SV = 2	MYO1D_HUMAN
4	Cluster of heat shock protein HSP 90-β; OS = H sapiens; GN = HSP90AB1; PE = 1; SV = 4 (HS90B_HUMAN)	HS90B_HUMAN
4.1	Heat shock protein HSP 90-β; OS = H sapiens; GN = HSP90AB1; PE = 1; SV = 4	HS90B_HUMAN
5	Heterogeneous nuclear ribonucleoproteins A2/B1; OS = H sapiens; GN = HNRNPA2B1; PE = 1; SV = 2	ROA2_HUMAN
6	Neuroblast differentiation-associated protein AHNAK; OS = H sapiens; GN = AHNAK; PE = 1; SV = 2	AHNK_HUMAN
7	Cluster of prelamin-A/C; OS = H sapiens; GN = LMNA; PE = 1; SV = 1 (LMNA_HUMAN)	LMNA_HUMAN
7.1	Prelamin-A/C; OS = H sapiens; GN = LMNA; PE = 1; SV = 1	LMNA_HUMAN
8	Cluster of 40S ribosomal protein S3a; OS = H sapiens; GN = RPS3A; PE = 1; SV = 2 (RS3A_HUMAN)	RS3A_HUMAN
8.1	40S ribosomal protein S3a; OS = H sapiens; GN = RPS3A; PE = 1; SV = 2	RS3A_HUMAN
9	40S ribosomal protein S3; OS = H sapiens; GN = RPS3; PE = 1; SV = 2	RS3_HUMAN
10	Cluster of α-enolase; OS = H sapiens; GN = ENO1; PE = 1; SV = 2 (ENOA_HUMAN)	ENOA_HUMAN
10.1	α-enolase; OS = H sapiens; GN = ENO1; PE = 1; SV = 2	ENOA_HUMAN
11	60S ribosomal protein L10a; OS = H sapiens; GN = RPL10A; PE = 1; SV = 2	RL10A_HUMAN
12	Cluster of serine/arginine-rich splicing factor 6; OS = H sapiens; GN = SRSF6; PE = 1; SV = 2 (SRSF6_HUMAN)	SRSF6_HUMAN [4]
12.1	Serine/arginine-rich splicing factor 6; OS = H sapiens; GN = SRSF6; PE = 1; SV = 2	SRSF6_HUMAN
12.2	Serine/arginine-rich splicing factor 4; OS = H sapiens; GN = SRSF4; PE = 1; SV = 2	SRSF4_HUMAN (+ 2)
13	60S ribosomal protein L8; OS = H sapiens; GN = RPL8; PE = 1; SV = 2	RL8_HUMAN
14	40S ribosomal protein S6; OS = H sapiens; GN = RPS6; PE = 1; SV = 1	RS6_HUMAN
15	Pyruvate kinase PKM; OS = H sapiens; GN = PKM; PE = 1; SV = 4	KPYM_HUMAN
16	40S ribosomal protein S18; OS = H sapiens; GN = RPS18; PE = 1; SV = 3	RS18_HUMAN
17	Ras GTPase-activating protein-binding protein 1; OS = H sapiens; GN = G3BP1; PE = 1; SV = 1	G3BP1_HUMAN
18	UPF0568 protein C14orf166; OS = H sapiens; GN = C14orf166; PE = 1; SV = 1	CN166_HUMAN
19	Heterogeneous nuclear ribonucleoprotein A1; OS = H sapiens; GN = HNRNPA1; PE = 1; SV = 5	ROA1_HUMAN
20	Heat shock protein β-1; OS = H sapiens; GN = HSPB1; PE = 1; SV = 2	HSPB1_HUMAN
21	60S ribosomal protein L10; OS = H sapiens; GN = RPL10; PE = 1; SV = 4	RL10_HUMAN
22	Cluster of histone H1.3; OS = H sapiens; GN = HIST1H1D; PE = 1; SV = 2 (H13_HUMAN)	H13_HUMAN [2]
22.1	Histone H1.3; OS = H sapiens; GN = HIST1H1D; PE = 1; SV = 2	H13_HUMAN
22.2	Histone H1.2; OS = H sapiens; GN = HIST1H1C; PE = 1; SV = 2	H12_HUMAN
23	60S acidic ribosomal protein P0; OS = H sapiens; GN = RPLP0; PE = 1; SV = 1	RLA0_HUMAN (+ 1)
24	tRNA-splicing ligase RtcB homolog; OS = H sapiens; GN = C22orf28; PE = 1; SV = 1	RTCB_HUMAN
25	Serine/arginine-rich splicing factor 9; OS = H sapiens; GN = SRSF9; PE = 1; SV = 1	SRSF9_HUMAN
26	40S ribosomal protein S11; OS = H sapiens; GN = RPS11; PE = 1; SV = 3	RS11_HUMAN (+ 1)
27	Clathrin heavy chain 1; OS = H sapiens; GN = CLTC; PE = 1; SV = 5	CLH1_HUMAN
28	40S ribosomal protein S16; OS = H sapiens; GN = RPS16; PE = 1; SV = 2	RS16_HUMAN (+ 1)
29	Cluster of annexin A2; OS = H sapiens; GN = ANXA2; PE = 1; SV = 2 (ANXA2_HUMAN)	ANXA2_HUMAN [4]
29.1	Annexin A2; OS = H sapiens; GN = ANXA2; PE = 1; SV = 2	ANXA2_HUMAN (+ 3)
30	Cluster of Q5VU59_HUMAN	Q5VU59_HUMAN [2]
30.1	Q5VU59_HUMAN	Q5VU59_HUMAN
30.2	Tropomyosin α-3 chain; OS = H sapiens; GN = TPM3; PE = 1; SV = 2	TPM3_HUMAN
31	ADP/ATP translocase 2; OS = H sapiens; GN = SLC25A5; PE = 1; SV = 7	ADT2_HUMAN
32	Cluster of eukaryotic translation initiation factor 4 γ 1; OS = H sapiens; GN = EIF4G1; PE = 1; SV = 4 (IF4G1_HUMAN)	IF4G1_HUMAN
32.1	Eukaryotic translation initiation factor 4 γ 1; OS = H sapiens; GN = EIF4G1; PE = 1; SV = 4	IF4G1_HUMAN
33	60S ribosomal protein L19; OS = H sapiens; GN = RPL19; PE = 1; SV = 1	RL19_HUMAN
34	Cluster of dystonin; OS = H sapiens; GN = DST; PE = 1; SV = 4 (DYST_HUMAN)	DYST_HUMAN
34	Cluster of dystonin; OS = H sapiens; GN = DST; PE = 1; SV = 4 (DYST_HUMAN)	DYST_HUMAN
35	EH domain-containing protein 1; OS = H sapiens; GN = EHD1; PE = 1; SV = 2	EHD1_HUMAN
36	Peptidyl-prolyl cis-trans isomerase A; OS = H sapiens; GN = PPIA; PE = 1; SV = 2	PPIA_HUMAN (+ 4)
37	GTP-binding nuclear protein Ran; OS = H sapiens; GN = RAN; PE = 1; SV = 3	RAN_HUMAN
38	60S ribosomal protein L23a; OS = H sapiens; GN = RPL23A; PE = 1; SV = 1	RL23A_HUMAN
39	Cluster of AP-2 complex subunit α-1; OS = H sapiens; GN = AP2A1; PE = 1; SV = 3 (AP2A1_HUMAN)	AP2A1_HUMAN
39.1	AP-2 complex subunit α-1; OS = H sapiens; GN = AP2A1; PE = 1; SV = 3	AP2A1_HUMAN
40	Elongation factor 2; OS = H sapiens; GN = EEF2; PE = 1; SV = 4	EF2_HUMAN
41	60S ribosomal protein L13a; OS = H sapiens; GN = RPL13A; PE = 1; SV = 2	RL13A_HUMAN
42	40S ribosomal protein S25; OS = H sapiens; GN = RPS25; PE = 1; SV = 1	RS25_HUMAN
43	EF-hand domain-containing protein D1; OS = H sapiens; GN = EFHD1; PE = 1; SV = 1	EFHD1_HUMAN
44	60S ribosomal protein L26; OS = H sapiens; GN = RPL26; PE = 1; SV = 1	RL26_HUMAN
45	AP-2 complex subunit μ; OS = H sapiens; GN = AP2M1; PE = 1; SV = 2	AP2M1_HUMAN (+ 2)
46	Ataxin-2-like protein; OS = H sapiens; GN = ATXN2L; PE = 1; SV = 2	ATX2L_HUMAN
47	60S ribosomal protein L15; OS = H sapiens; GN = RPL15; PE = 1; SV = 2	RL15_HUMAN (+ 2)
48	60S ribosomal protein L24; OS = H sapiens; GN = RPL24; PE = 1; SV = 1	RL24_HUMAN (+ 2)
49	Cluster of thyroid hormone receptor–associated protein 3; OS = H sapiens; GN = THRAP3; PE = 1; SV = 2 (TR150_HUMAN)	TR150_HUMAN [5]
49.1	Thyroid hormone receptor–associated protein 3; OS = H sapiens; GN = THRAP3; PE = 1; SV = 2	TR150_HUMAN
49.2	THRAP3 protein (fragment); OS = H sapiens; GN = THRAP3; PE = 2; SV = 1	Q05D20_HUMAN (+ 3)
50	Gelsolin; OS = H sapiens; GN = GSN; PE = 1; SV = 1	GELS_HUMAN
51	Nucleophosmin; OS = H sapiens; GN = NPM1; PE = 1; SV = 2	NPM_HUMAN
52	Histone H4; OS = H sapiens; GN = HIST1H4A; PE = 1; SV = 2	H4_HUMAN
53	Cofilin 1; OS = H sapiens; GN = CFL1; PE = 1; SV = 3	COF1_HUMAN
54	Glutathione S-transferase μ 3; OS = H sapiens; GN = GSTM3; PE = 1; SV = 3	GSTM3_HUMAN
55	Prolactin-inducible protein; OS = H sapiens; GN = PIP; PE = 1; SV = 1	PIP_HUMAN
56	Cluster of F-actin–capping protein subunit β; OS = H sapiens; GN = CAPZB; PE = 1; SV = 4 (CAPZB_HUMAN)	CAPZB_HUMAN
56.1	F-actin–capping protein subunit β; OS = H sapiens; GN = CAPZB; PE = 1; SV = 4	CAPZB_HUMAN
57	Cluster of peroxiredoxin-1; OS = H sapiens; GN = PRDX1; PE = 1; SV = 1 (PRDX1_HUMAN)	PRDX1_HUMAN [2]
57.1	Peroxiredoxin-1; OS = H sapiens; GN = PRDX1; PE = 1; SV = 1	PRDX1_HUMAN
57.2	Peroxiredoxin-2; OS = H sapiens; GN = PRDX2; PE = 1; SV = 5	PRDX2_HUMAN
58	Fructose-bisphosphate aldolase A; OS = H sapiens; GN = ALDOA; PE = 1; SV = 2	ALDOA_HUMAN
59	40S ribosomal protein S15a; OS = H sapiens; GN = RPS15A; PE = 1; SV = 2	RS15A_HUMAN
60	KH domain-containing, RNA-binding, signal transduction-associated protein 1; OS = H sapiens; GN = KHDRBS1; PE = 1; SV = 1	KHDR1_HUMAN
61	Leucine-rich repeat-containing protein 59; OS = H sapiens; GN = LRRC59; PE = 1; SV = 1	LRC59_HUMAN
62	60S ribosomal protein L12; OS = H sapiens; GN = RPL12; PE = 1; SV = 1	RL12_HUMAN
63	Tropomodulin-3; OS = H sapiens; GN = TMOD3; PE = 1; SV = 1	TMOD3_HUMAN
64	Heterogeneous nuclear ribonucleoprotein H; OS = H sapiens; GN = HNRNPH1; PE = 1; SV = 4	HNRH1_HUMAN (+ 2)
65	60S ribosomal protein L21; OS = H sapiens; GN = RPL21; PE = 1; SV = 2	RL21_HUMAN (+ 1)
66	Cluster of ADP-ribosylation factor 1; OS = H sapiens; GN = ARF1; PE = 1; SV = 2 (ARF1_HUMAN)	ARF1_HUMAN [4]
66.1	ADP-ribosylation factor 1; OS = H sapiens; GN = ARF1; PE = 1; SV = 2	ARF1_HUMAN (+ 1)
66.2	ADP-ribosylation factor 4; OS = H sapiens; GN = ARF4; PE = 1; SV = 3	ARF4_HUMAN (+ 1)
67	Cluster of eukaryotic initiation factor 4A-I; OS = H sapiens; GN = EIF4A1; PE = 1; SV = 1 (IF4A1_HUMAN)	IF4A1_HUMAN
67.1	Eukaryotic initiation factor 4A-I; OS = H sapiens; GN = EIF4A1; PE = 1; SV = 1	IF4A1_HUMAN
68	60S ribosomal protein L14; OS = H sapiens; GN = RPL14; PE = 1; SV = 4	RL14_HUMAN
69	ATP synthase subunit α, mitochondrial; OS = H sapiens; GN = ATP5A1; PE = 1; SV = 1	ATPA_HUMAN
70	40S ribosomal protein S7; OS = H sapiens; GN = RPS7; PE = 1; SV = 1	RS7_HUMAN
71	Histone H2A type 1-B/E; OS = H sapiens; GN = HIST1H2AB; PE = 1; SV = 2	H2A1B_HUMAN (+ 14)
72	AP-2 complex subunit β; OS = H sapiens; GN = AP2B1; PE = 1; SV = 1	AP2B1_HUMAN
73	ELAV-like protein 1; OS = H sapiens; GN = ELAVL1; PE = 1; SV = 2	ELAV1_HUMAN
74	Peroxiredoxin-6; OS = H sapiens; GN = PRDX6; PE = 1; SV = 3	PRDX6_HUMAN
75	60S ribosomal protein L11; OS = H sapiens; GN = RPL11; PE = 1; SV = 2	RL11_HUMAN (+ 2)
76	78-kDa glucose-regulated protein; OS = H sapiens; GN = HSPA5; PE = 1; SV = 2	GRP78_HUMAN
77	40S ribosomal protein S17-like; OS = H sapiens; GN = RPS17L; PE = 1; SV = 1	RS17L_HUMAN (+ 1)
78	Cluster of F-actin–capping protein subunit α-1; OS = H sapiens; GN = CAPZA1; PE = 1; SV = 3 (CAZA1_HUMAN)	CAZA1_HUMAN
78.1	F-actin–capping protein subunit α-1; OS = H sapiens; GN = CAPZA1; PE = 1; SV = 3	CAZA1_HUMAN
79	Phosphoglycerate mutase 1; OS = H sapiens; GN = PGAM1; PE = 1; SV = 2	PGAM1_HUMAN
80	Phosphoglycerate kinase 1; OS = H sapiens; GN = PGK1; PE = 1; SV = 3	PGK1_HUMAN
81	Cluster of stress-70 protein, mitochondrial; OS = H sapiens; GN = HSPA9; PE = 1; SV = 2 (GRP75_HUMAN)	GRP75_HUMAN
81.1	Stress-70 protein, mitochondrial; OS = H sapiens; GN = HSPA9; PE = 1; SV = 2	GRP75_HUMAN
82	Cluster of transformer-2 protein homolog β; OS = H sapiens; GN = TRA2B; PE = 1; SV = 1 (TRA2B_HUMAN)	TRA2B_HUMAN [2]
82.1	Transformer-2 protein homolog β; OS = H sapiens; GN = TRA2B; PE = 1; SV = 1	TRA2B_HUMAN
82.2	Transformer-2 protein homolog α; OS = H sapiens; GN = TRA2A; PE = 1; SV = 1	TRA2A_HUMAN
83	Heterogeneous nuclear ribonucleoprotein M; OS = H sapiens; GN = HNRNPM; PE = 1; SV = 3	HNRPM_HUMAN
84	Heterogeneous nuclear ribonucleoprotein A3; OS = H sapiens; GN = HNRNPA3; PE = 1; SV = 2	ROA3_HUMAN
85	40S ribosomal protein S10; OS = H sapiens; GN = RPS10; PE = 1; SV = 1	RS10_HUMAN
86	60S ribosomal protein L18a; OS = H sapiens; GN = RPL18A; PE = 1; SV = 2	RL18A_HUMAN
87	Malate dehydrogenase, mitochondrial; OS = H sapiens; GN = MDH2; PE = 1; SV = 3	MDHM_HUMAN (+ 5)
88	Cluster of 14–3–3 protein β/α; OS = H sapiens; GN = YWHAB; PE = 1; SV = 3 (1433B_HUMAN)	1433B_HUMAN [3]
88.1	14-3-3 protein β/α; OS = H sapiens; GN = YWHAB; PE = 1; SV = 3	1433B_HUMAN (+ 1)
88.2	14-3-3 protein ζ/δ; OS = H sapiens; GN = YWHAZ; PE = 1; SV = 1	1433Z_HUMAN
89	Ras GTPase-activating protein-binding protein 2; OS = H sapiens; GN = G3BP2; PE = 1; SV = 2	G3BP2_HUMAN (+ 1)
90	Ras-related protein Rab-1B; OS = H sapiens; GN = RAB1B; PE = 1; SV = 1	RAB1B_HUMAN (+ 1)
91	Catenin α-1; OS = H sapiens; GN = CTNNA1; PE = 1; SV = 1	CTNA1_HUMAN (+ 1)
92	RNA-binding protein EWS; OS = H sapiens; GN = EWSR1; PE = 1; SV = 1	EWS_HUMAN (+ 3)
93	60S ribosomal protein L3; OS = H sapiens; GN = RPL3; PE = 1; SV = 2	RL3_HUMAN (+ 6)
94	Cathepsin D; OS = H sapiens; GN = CTSD; PE = 1; SV = 1	CATD_HUMAN
95	Chloride intracellular channel protein 1; OS = H sapiens; GN = CLIC1; PE = 1; SV = 4	CLIC1_HUMAN
96	Fatty acid synthase; OS = H sapiens; GN = FASN; PE = 1; SV = 3	FAS_HUMAN
97	Guanine nucleotide-binding protein subunit β-2–like 1; OS = H sapiens; GN =; GNB2L1; PE = 1; SV = 3	GBLP_HUMAN
98	Α-actinin-4; OS = H sapiens; GN = ACTN4; PE = 1; SV = 2	ACTN4_HUMAN (+ 2)
99	Splicing factor, proline- and glutamine-rich; OS = H sapiens; GN = SFPQ; PE = 1; SV = 2	SFPQ_HUMAN
100	ARF6 protein; OS = H sapiens; GN = ARF6; PE = 2; SV = 1	Q6FH17_HUMAN
101	40S ribosomal protein S5; OS = H sapiens; GN = RPS5; PE = 1; SV = 4	RS5_HUMAN (+ 4)
102	Barrier-to-autointegration factor; OS = H sapiens; GN = BANF1; PE = 1; SV = 1	BAF_HUMAN
103	60S ribosomal protein L28; OS = H sapiens; GN = RPL28; PE = 1; SV = 3	RL28_HUMAN (+ 1)
104	Ribosomal protein S19 (Fragment); OS = H sapiens; PE = 2; SV = 1	Q8WVX7_HUMAN
105	Histone H1.5; OS = H sapiens; GN = HIST1H1B; PE = 1; SV = 3	H15_HUMAN
106	Single-stranded DNA-binding protein, mitochondrial; OS = H sapiens; GN = SSBP1; PE = 1; SV = 1	SSBP_HUMAN
107	Transferrin receptor protein 1; OS = H sapiens; GN = TFRC; PE = 1; SV = 2	TFR1_HUMAN (+ 1)
108	40S ribosomal protein S14; OS = H sapiens; GN = RPS14; PE = 1; SV = 3	RS14_HUMAN
109	60S ribosomal protein L22; OS = H sapiens; GN = RPL22; PE = 1; SV = 2	RL22_HUMAN (+ 3)
110	60S ribosomal protein L31; OS = H sapiens; GN = RPL31; PE = 1; SV = 1	RL31_HUMAN
111	60S ribosomal protein L35; OS = H sapiens; GN = RPL35; PE = 1; SV = 2	RL35_HUMAN
112	THO complex subunit 4; OS = H sapiens; GN = ALYREF; PE = 1; SV = 3	THOC4_HUMAN (+ 1)
113	Serine/arginine repetitive matrix protein 2; OS = H sapiens; GN = SRRM2; PE = 1; SV = 2	SRRM2_HUMAN
114	Peptidyl-prolyl cis-trans isomerase FKBP4; OS = H sapiens; GN = FKBP4; PE = 1; SV = 3	FKBP4_HUMAN
115	Actin-related protein 2/3 complex subunit 3; OS = H sapiens; GN = ARPC3; PE = 1; SV = 3	ARPC3_HUMAN (+ 2)
116	40S ribosomal protein S26; OS = H sapiens; GN = RPS26; PE = 1; SV = 3	RS26_HUMAN
117	Transcription factor A, mitochondrial; OS = H sapiens; GN = TFAM; PE = 1; SV = 1	TFAM_HUMAN
118	Polypyrimidine tract-binding protein 1; OS = H sapiens; GN = PTBP1; PE = 1; SV = 1	PTBP1_HUMAN
119	Adenine phosphoribosyltransferase; OS = H sapiens; GN = APRT; PE = 1; SV = 2	APT_HUMAN
120	Protein FAM98A; OS = H sapiens; GN = FAM98A; PE = 1; SV = 1	FA98A_HUMAN (+ 4)
121	Galectin-7; OS = H sapiens; GN = LGALS7; PE = 1; SV = 2	LEG7_HUMAN
122	Triosephosphate isomerase; OS = H sapiens; GN = TPI1; PE = 1; SV = 3	TPIS_HUMAN
123	Histone H2B type 1-A; OS = H sapiens; GN = HIST1H2BA; PE = 1; SV = 3	H2B1A_HUMAN (+ 19)
124	DNA repair protein XRCC1; OS = H sapiens; GN = XRCC1; PE = 1; SV = 2	XRCC1_HUMAN (+ 3)
125	l-lactate dehydrogenase A chain; OS = H sapiens; GN = LDHA; PE = 1; SV = 2	LDHA_HUMAN (+ 1)
126	Cluster of serine/threonine-protein phosphatase PP1-α catalytic subunit; OS = H sapiens; GN = PPP1CA; PE = 1; SV = 1 (PP1A_HU	PP1A_HUMAN
126	Cluster of serine/threonine-protein phosphatase PP1-α catalytic subunit; OS = H sapiens; GN = PPP1CA; PE = 1; SV = 1 (PP1A_HU	PP1A_HUMAN
127	Cluster of chloride intracellular channel protein 3; OS = H sapiens; GN = CLIC3; PE = 1; SV = 2 (CLIC3_HUMAN)	CLIC3_HUMAN
127.1	Chloride intracellular channel protein 3; OS = H sapiens; GN = CLIC3; PE = 1; SV = 2	CLIC3_HUMAN
128	14-3-3 protein ε; OS = H sapiens; GN = YWHAE; PE = 1; SV = 1	1433E_HUMAN
129	MUC1 isoform J14; OS = H sapiens; GN = MUC1; PE = 2; SV = 1	B6ECA3_HUMAN
130	cDNA FLJ59433, highly similar to elongation factor 1-γ; OS = H sapiens; PE = 2; SV = 1	B4DUP0_HUMAN
131	Non-POU domain-containing octamer-binding protein; OS = H sapiens; GN = NONO; PE = 1; SV = 4	NONO_HUMAN (+ 1)
132	Actin-related protein 2/3 complex subunit 4; OS = H sapiens; GN = ARPC4; PE = 1; SV = 3	ARPC4_HUMAN (+ 1)
133	60-kDa heat shock protein, mitochondrial; OS = H sapiens; GN = HSPD1; PE = 1; SV = 2	CH60_HUMAN
134	DNA ligase 3; OS = H sapiens; GN = LIG3; PE = 1; SV = 2	DNLI3_HUMAN
135	40S ribosomal protein S15; OS = H sapiens; GN = RPS15; PE = 1; SV = 2	RS15_HUMAN (+ 4)
136	Matrin-3; OS = H sapiens; GN = MATR3; PE = 1; SV = 2	MATR3_HUMAN (+ 3)
137	Serine/arginine-rich splicing factor 10; OS = H sapiens; GN = SRSF10; PE = 1; SV = 1	SRS10_HUMAN
138	Cluster of Ras-related protein Rab-10; OS = H sapiens; GN = RAB10; PE = 1; SV = 1 (RAB10_HUMAN)	RAB10_HUMAN [3]
138.1	Ras-related protein Rab-10; OS = H sapiens; GN = RAB10; PE = 1; SV = 1	RAB10_HUMAN
138.2	Ras-related protein Rab-13; OS = H sapiens; GN = RAB13; PE = 1; SV = 1	RAB13_HUMAN (+ 1)
139	Actin-related protein 3; OS = H sapiens; GN = ACTR3; PE = 1; SV = 3	ARP3_HUMAN (+ 2)
140	U1 small nuclear ribonucleoprotein A; OS = H sapiens; GN = SNRPA; PE = 1; SV = 3	SNRPA_HUMAN (+ 6)
141	Fragile X mental retardation syndrome–related protein 2; OS = H sapiens; GN = FXR2; PE = 1; SV = 2	FXR2_HUMAN
142	Anterior gradient protein 2 homolog; OS = H sapiens; GN = AGR2; PE = 1; SV = 1	AGR2_HUMAN
143	Clathrin interactor 1; OS = H sapiens; GN = CLINT1; PE = 1; SV = 1	EPN4_HUMAN
144	Cell division control protein 42 homolog; OS = H sapiens; GN = CDC42; PE = 1; SV = 2	CDC42_HUMAN

Table W8.

Functional Classification of PIP Proteomics Data.

Cluster 1	Enrichment Score = 33.10409316169969	P Value
UNIPROT_ID	Protein Name	9.3E-76
823648	Ribosomal protein L23a pseudogene 63
779090	Ribosomal protein L15 pseudogene 22
801125	Ribosomal protein L8; ribosomal protein L8 pseudogene 2
781935	Ribosomal protein S25 pseudogene 8; ribosomal protein S25
783263	Ribosomal protein L12 pseudogene 2; ribosomal protein L12
793235	Ribosomal protein L21 pseudogene 134; ribosomal protein L21
772335	Ribosomal protein S26 pseudogene 38; ribosomal protein S26
782492	Ribosomal protein L22 pseudogene 11; ribosomal protein L22
778688	Ribosomal protein L18a pseudogene 6; ribosomal protein L18a
797877	Ribosomal protein S5
808990	Ribosomal protein S10; ribosomal protein S10 pseudogene 4
796446	Ribosomal protein L31 pseudogene 49; ribosomal protein L31
810885	Ribosomal protein L11
808508	Ribosomal protein S6 pseudogene 25; ribosomal protein S6
803377	Ribosomal protein L35; ribosomal protein L35 pseudogene 1
802424	Ribosomal protein L3; similar to 60S ribosomal protein L3 (L4)
796817	Ribosomal protein L10a pseudogene 6; ribosomal protein L10a
791016	Ribosomal protein L13a pseudogene 7; ribosomal protein L13a
816191	Ribosomal protein S3 pseudogene 3; ribosomal protein S3
789968	Ribosomal protein S3A pseudogene 5; ribosomal protein S3a
799749	Ribosomal protein S19 pseudogene 3; ribosomal protein S19
800698	Ribosomal protein L24; ribosomal protein L24 pseudogene 6
784832	Ribosomal protein S15 pseudogene 5; ribosomal protein S15
821689	Fragile X mental retardation, autosomal homolog 2
796227	Ribosomal protein, large, P0 pseudogene 2; ribosomal protein
817032	Ribosomal protein S14
785872	Ribosomal protein L26 pseudogene 33; ribosomal protein L26
777984	Ribosomal protein L14
773042	Ribosomal protein S7; ribosomal protein S7 pseudogene 11
786531	Ribosomal protein L19; ribosomal protein L19 pseudogene 12
788428	Ribosomal protein S15a pseudogene 17; ribosomal protein S15a
790706	Ribosomal protein S16 pseudogene 1
786137	Ribosomal protein S11 pseudogene 5; ribosomal protein S11
799716	Ribosomal protein L10; ribosomal protein L10 pseudogene 15
814248	Ribosomal protein S18 pseudogene 12
820040	Ribosomal protein L28

Cluster 2	Enrichment Score: 13.296661105241336	P Value
UNIPROT_ID	Protein Name	5.14E-32
814227	Heterogeneous nuclear ribonucleoprotein A2/B1
824519	Heterogeneous nuclear ribonucleoprotein A1-like 3
816714	FUS-interacting protein (serine/arginine-rich) 1
806558	THO complex 4
786723	Matrin 3
826437	Heterogeneous nuclear ribonucleoprotein H1 (H)
810069	Heterogeneous nuclear ribonucleoprotein A3
797048	Transformer 2 β homolog (Drosophila)
825011	Splicing factor, arginine/serine-rich 4
826656	Splicing factor proline/glutamine-rich
825901	Polypyrimidine tract binding protein 1
800520	Serine/arginine repetitive matrix 2
776456	Non-POU domain containing, octamer-binding
798354	Small nuclear ribonucleoprotein polypeptide A
776781	Transformer 2 α homolog (Drosophila)
806156	Splicing factor, arginine/serine-rich 6
791398	Heterogeneous nuclear ribonucleoprotein M
802485	ELAV (embryonic lethal, abnormal vision, Drosophila)-like 1
816065	Splicing factor, arginine/serine-rich 9

Cluster 3	Enrichment Score: 6.102262729332349	P Value
UNIPROT_ID	Protein Name	2.47E-12
775719	Histone cluster 1, H2ae; histone cluster 1, H2ab
784692	Histone cluster 1, H4l; histone cluster 1
781820	Histone cluster 1, H1d
784865	Histone cluster 1, H1c
799393	Histone cluster 1, H1b
783255	Histone cluster 1, H2ba

Cluster 4	Enrichment Score: 5.218508620408835	P Value
UNIPROT_ID	Protein Name	6.01E-08
775114	Gelsolin (amyloidosis, Finnish type)
819298	Capping protein (actin filament) muscle Z-line, β
799685	Capping protein (actin filament) muscle Z-line, α 1
773758	Similar to actin related protein 2/3 complex subunit 3

Cluster 5	Enrichment Score: 3.1707699720411564	P Value
UNIPROT_ID	Protein Name	4.18E-09
777280	Adaptor-related protein complex 2, β 1 subunit
797086	Adaptor-related protein complex 2, α 1 subunit
789199	Adaptor-related protein complex 2, mu 1 subunit
783500	Clathrin, heavy chain (Hc)

Cluster 6	Enrichment Score: 2.0551591175667023	P Value
UNIPROT_ID	Protein Name	9.54E-11
794683	ADP-ribosylation factor 4
824702	RAB13, member RAS oncogene family; similar to hCG24991
823808	RAB10, member RAS oncogene family
819977	RAB1B, member RAS oncogene family
825257	ADP-ribosylation factor 6
803034	ADP-ribosylation factor 1

Table W9.

Canonical Pathways Associated with PIP-Binding Partners.

Canonical Pathways	− Log (P Value)	Ratio to Total	Molecules
EIF2 Signaling	4.37E+01	1.94E-01	RPL11, RPL24, RPL22, RPS18, RPL14, RPL26, EIF4G1
			RPS17/RPS17L, RPS11, RPS7, RPL35, RPS3A, RPL18A, RPL19
			RPL12, RPL8, PPP1CA, RPS5, RPS3, RPS10, RPL31, RPL3, RPS19
			RPL21, RPL23A, RPLP0, RPL10A, RPS6, RPL15, RPS16, RPS26
			RPL28, RPL10, EIF4A1, RPS15, RPS15A, RPS25, RPL13A, RPS14
Regulation of eIF4 and p70S6K signaling	1.55E+01	1.03E-01	RPS18, RPS19, RPS17/RPS17L, EIF4G1, RPS11, RPS7, RPS6
			RPS6, RPS3A, RPS16, RPS26, EIF4A1, RPS15, RPS15A, RPS25
			RPS5, RPS3, RPS10, RPS14
Remodeling of epithelial adherens junctions	7.03E+00	1.14E-01	ARF6, ACTR3, TUBB4B, CTNNA1, ARPC3, ACTN4, TUBB, ARPC4
Clathrin-mediated endocytosis signaling	5.46E+00	5.05E-02	AP2B1, AP2M1, AP2A1, ARF6, ACTR3, CLTC, TFRC, ARPC3, CDC42 ARPC4
Regulation of actin-based motility by Rho	4.20E+00	6.59E-02	ACTR3, CFL1, ARPC3, CDC42, GSN, ARPC4
Integrin signaling	3.60E+00	3.85E-02	ARF1, ARF6, ACTR3, ARF4, ARPC3, ACTN4, CDC42, ARPC4