Breast Cancer Genes PSMC3IP and EPSTI1 Play a Role in Apoptosis Regulation

A key element to delineate the biology of individual tumors is the regulation of apoptosis. In this work, we functionally characterize two breast cancer associated genes, the proteasome 26S subunit ATPase 3 interacting protein (PSMC3IP) and the epithelial-stromal interaction 1 (EPSTI1), to explore their potential apoptotic role in breast cancer. We first explore the existence of direct physical interactions with annotated BC-apoptotic genes. Based on the generated interaction network, we examine several apoptotic markers to determine the effect of PSMC3IP and EPSTI1 gene expression modulation in two different human breast cancer cell lines to suggest potential molecular mechanisms to unveil their role in the disease. Our results show that PSMC3IP and EPSTI1 are able to modulate the extrinsic apoptotic pathway in estrogen receptor positive and triple negative breast cancer cell lines, highlighting them as potential therapeutic targets.


Introduction
Due to its complexity, breast cancer (BC) is often considered a broad set of diseases including multiple, distinct biological subtypes with diverse natural histories that present a varied spectrum of clinical, pathological and molecular features with different prognostic and therapeutic implications [1]. The poor prognostic outcome of breast cancer is largely due to its resistance to current cancer therapies, where the balance between cell proliferation and apoptosis plays a critical role and it is crucial in determining the overall growth or regression of the tumor in response to treatments [2]. Hence, identifying proteins involved in apoptosis resistance developed by tumorigenic cells has an essential importance in order to find new therapeutic approaches. Two major apoptosis pathways have been described: the mitochondria mediated (intrinsic) and the death receptor mediated (extrinsic), which is initiated by the binding of ligands such as TNF-α or TNF-related apoptosis inducing ligand (TRAIL) to death receptors [3]. Once the receptors are activated, they oligomerize and form complexes that recruit and activate the initiator caspase-8. Active caspase-8 subsequently cleaves effector caspases like caspase-3 and caspase-7, resulting in activation or inactivation (as well as translocation), of several substrates such as the poly ADP ribose polymerase (PARP), with the consequent induction of cell death [4,5].
In the last decade, network biology approaches have contributed to identify novel causative and susceptibility oncogenes, as well as secondary effectors that could not be highlighted by conventional analysis based on differential expression [6] Therefore, this novel approach can provide a deeper understanding of the molecular mechanisms underlying complex pathological processes, offering new biomarkers that may help to improve breast cancer diagnosis.
PSMC3IP gene is located on chromosome 17q21, proximal to BRCA1 [7] and previously linked to breast cancer predisposition [8]. It has been characterized as a nuclear receptor participating in estrogen, androgen, glucocorticoid and progesterone receptor-mediated gene regulation [9,10]. PSMC3IP is upregulated in breast cancer [7,11] and in addition, inactivating mutations [12] have also been shown to regulate DNA recombination in DNA repair [13], potentially contributing to an increased risk in familial breast and ovarian cancers.
On the other hand, EPSTI1, an interferon (IFN) response gene [14], has been identified as a stromal fibroblast-induced gene in breast cancer, being highly upregulated in invasive breast carcinomas as compared with normal breast [15,16]. More recently, it has been shown that expression levels of EPSTI1 associate with tumor initiation and migration, stem cell-like properties, epithelial-mesenchymal transition (EMT) [17] and breast cancer invasion and metastasis [18], with the highest expression observed in basal-like subtype breast cancer cells exhibiting a poor prognosis [17].
Although the relationship of PSMC3IP and EPSTI1 with BC is well established, the underlying molecular mechanisms are still unknown. In the present study, we describe novel interactions between PSMC3IP and EPSTI1 with well-established BC genes which are also related to apoptosis and cell proliferation processes. We explore the anti-apoptotic role of PSMC3IP and EPSTI1 and their contribution in breast cancer development. We have carried out a functional characterization associated to cell apoptosis by means of caspase-8 and-3 activation, PARP cleavage and DNA integrity, based on gene overexpression and silencing in two different human breast cancer derived cell lines under both basal and apoptotic-induced conditions.

Subcloning of human cDNAs into Y2H plasmids
Human ORF clones were cloned into pENTR⁄D-TOPO vector (pENTR Directional TOPO cloning kit; Life Technologies) and sequence verified. AKT1, BCAR3 and EPSTI1 clones derived from the human ORFeome v1.1 [19]; CASP8 and PSMC3IP from Life Technologies Ultimate ORF Clones [20]. All ORFs were individually transferred into yeast two-hybrid (Y2H) destination vectors by Gateway recombinational cloning (ProQuest Two-Hybrid System, Life Technologies). BC-apoptosis genes were cloned into pDEST32 to generate bait plasmids and the BC associated genes (EPSTI1 and PSMC3IP) were cloned into pDEST22 to obtain prey plasmids.

Y2H matrix screens
Bait and prey plasmids were pair-wise co-transformed into MaV203 yeast strain in 96-well arrays and plated onto selective SD2 (lacking Leu and Trp amino acids) agar media and incubated for 48 hours at 30°C to detect colony growth. Co-transformant arrays were then replica plated onto different selective media for interaction screening. To assay the activation of the HIS3 reporter gene, SD3 (lacking Leu, Trp, His) agar plates were supplemented with 12 to 100 mM of 3-aminotriazole (3AT, Sigma-Aldrich), being 50 mM 3AT the optimal concentration for positive HIS3 activation colonies. Similarly, we assayed the activation of the URA3 reporter gene by plating onto SD3 (lacking Leu, Trp, Uracil). Double reporter HIS3/URA3 activation was evaluated by SD4 (lacking Leu, Trp, His, Uracil) agar plates supplemented with 20 mM of 3AT and LacZ reporter gene was tested by the β-galactosidase assay on a nylon membrane placed onto a SD2 agar plate.
In order to minimize the number of false positives (at the cost of penalizing potential false negatives as well), we subsequently scored the positive interactions based on their ability to activate at least two reporter genes or being repeatedly observed in biological replica screens, which defined our high-confidence (HC) interaction set.

Plasmids and siRNAs
Human ORF cDNA clones were acquired from Thermo Fisher Scientific. siRNAs targeting PSMC3IP and LUCIFERASE (non-targeting control siRNA) gene expression were purchased from Life Technologies. XIAP derived siRNA II was from Cell Signaling Technology and EPSTI1 derived siRNA was designed as described before [17].

Cell transfection and treatment
For expression of Myc-tagged fusion proteins, cDNA clones were subcloned into pDEST-Myctagged vector using the Gateway cloning system (Life Technologies). Cell transfection using X-tremeGENE 9 (Roche) was performed according to the manufacturer's instructions. After 24 hours of cell transfection, cells were treated for another 24 hours with recombinant human TRAIL/Apo2 ligand (PeproTech) at final concentration of 80ng/ml in MDA-MB-231 and 100ng/ml in MCF-7.
For RNA interference-mediated gene silencing, cells were seeded and exposed to 50 nM of either gene-specific siRNA or non-targeting control siRNA (siLUC), using Lipofectamine RNAiMAX transfection reagent (Life Technologies) for 48 h. For EPSTI1 silencing, IFN-α (Chemicon, Millipore) was added to a final concentration of 1000 U/ml 8 hours before harvesting the cells. TRAIL treatment was conducted in the same conditions as in overexpression assays (see above).

Caspase-8 and caspase-3 activity assays
The activities of caspase 3 and 8 were measured using APOPCYTO Caspase-3 Colorimetric Assay Kit (Medical and Biological Laboratories) and Caspase-8 Colorimetric Assay Kit (BioVision), respectively. Briefly, total cell protein was extracted using ice-cold cell lysis buffer. Then, 100-200 mg of total protein was diluted in 50 ml of lysis buffer and 50 ml of 2× reaction buffer containing 10 mM DTT. 5 ml of caspase 3 or 8 substrate were added into each well of a 96-well microplate. After incubation at 37°C for 3 hours, the absorbance was measured at 405 nm.

Propidium iodide staining and flow cytometry analysis
Apoptotic cells were quantified by flow cytometry as previously described [21]. Briefly, cells were washed with PBS, fixed in cold 70% ethanol, and then stained with propidium iodide (Sigma Aldrich) while being treated with RNase (Sigma-Aldrich). Quantitative analysis of sub-G 0 /G 1 cells was carried out in a CouLter XL cytometer using FlowJo software (Tree Star).

TUNEL assay
Apoptosis was assayed in cell culture using a commercially available kit (In Situ Cell Death Detection kit, fluorescein; Roche, Nutley, NJ) designed to detect terminal deoxynucleotidyl transferase (TdT)-mediated nick end labeling (TUNEL). Cells were fixed in 4% paraformaldehyde and permeabilized in 0.1% Triton X-100 in 0.1% sodium citrate. After washing in PBS, cells were incubated in TUNEL reaction mixture at 37°C for 60 minutes, washed, and mounted. Nuclei were counterstained with Hoechst 33342 Trihydrochloride, trihidrate (Invitrogen).

Cell viability assay
Cell viability was determined by adding a final concentration of 1.1 mg/mL MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma Aldrich) into the cell culture. Cells were incubated for 2 hours at 37°C, then the medium was removed, MTT was dissolved in DMSO (Sigma Aldrich) and its absorbance was determined at 570nm.

Correlation in gene expression profiles
We used the microarray data from [22], a compendium of gene expression profiles from 73 normal tissue and cell types, and the protein data from the Human Protein Atlas [23]. We applied a mixture model in order to obtain correlation coefficients that are robust under the presence of noise. We fit the model using the Expectation-Maximization (EM) algorithm [24]. We defined two genes as co-expressed if their EM correlation coefficient was greater than 0.5 and the probability of noise less than 0.5.

Statistical analysis
Results are expressed as the mean AE standard deviation (SD) of three independent experiments. Statistically significant differences were determined by one-way ANOVA followed by Tukey-Kramer post-test to identify pair wise differences. Differences were considered significant at P < 0.05Ã. Statistical analyses were carried out using GraphPad Prism (GraphPad Software Inc V4.03, San Diego, CA, USA) software.

Results and Discussion
PSMC3IP and EPSTI1 interact with key apoptotic proteins PSMC3IP and EPSTI1 present relevant transcriptomics/genomics aberrations in breast cancer [7,11,[15][16][17][18], although the molecular mechanisms are still unknown. Since cell proliferation and apoptosis are key for the development of breast tumorigenesis, we sought to explore the eventual implication of these genes in the regulation of these processes. To this end, we exploited the high interconnectivity observed among disease genes [25] to reveal novel direct relationships between well-established BC genes and PSMC3IP and EPSTI1, which could provide a molecular rationale for the implication of these genes in the disease.
An important factor when performing interaction discovery strategies is the selection of the core genes. We extensively mined the literature and the OMIM database [26] to select the most relevant BC genes that are involved in apoptosis or cell proliferation, ending up with 9 BC-apoptosis genes to be included in our interaction screen: AKT1, BAG4, BCAR3, CASP8, CDKN2A, CKN2C, CHEK2, IGF1R and PARP1 (See S1 Table).
We then performed pairwise yeast two-hybrid (Y2H) assays to identify novel interactions between BC-apoptosis genes and PSMC3IP and EPSTI1. Among the 7 high-confidence interactions we identified (see Methods and S1 Table), those of PSMC3IP and EPSTI1 with central proteins in the apoptosis extrinsic pathway, i.e. caspase 8, appeared as the most promising ones. Interestingly, PSMC3IP and EPSTI1 both show a strong co-expression profile with CASP8 in breast normal and cancer tissues, also reflected in higher protein levels (see Methods and S1 Table for further details). In addition, PSMC3IP and EPSTI1 also interact with breast cancer anti-estrogen resistance 3 (BCAR3), which is known to regulate proliferation and induce antiestrogen resistance in ZR75-1 and MCF-7 breast cancer cells [27]. Finally, we also found EPSTI1 to interact with AKT1 (unpublished data), an antiapoptotic protein that in response to specific stimuli phosphorylates and inactivates certain components of the apoptotic machinery, such as the Bcl2 antagonist of cell death (BAD) and caspase-9 [28,29]. In breast cancer, the PI3K-Akt pathway is a critical downstream effector of growth factor receptors such as HER2/ErbB2, insulin-like growth factor receptor (IGFR) and epidermal growth factor receptor (EGFR) [30][31][32][33]. Taken together, these findings suggest that PSMC3IP and EPSTI1 could play a role in the regulation of the apoptotic response.

Modulation of PSMC3IP and EPSTI1 expression in breast cancer cells
We first investigated whether PSMC3IP and EPSTI1 endogenous protein levels are indeed upregulated in carcinoma cells such as MCF-7 and MDA-MB-231 (Fig. 1), which represent common breast cancer subtypes differently graded upon hormone dependency and aggressiveness. MCF-7 is a weakly invasive luminal cell line, representative of estrogen receptor (ER)-positive tumors [34]. On the other hand, MDA-MB-231 is a highly invasive basal cell line, and it is often used as model for ER-negative tumors [35]. PSMC3IP is highly overexpressed in both MDA-MB-231 and MCF-7, respectively, compared to the non-tumorigenic breast cancer cell line (MCF-10A) (Fig. 1A). Conversely, EPSTI1 shows a more moderate increase, particularly in MCF-7 cells (1.3-fold, Fig. 1B), as previously observed [17], providing evidence about the heteroclonal nature of MCF-7 sublines [36].
Furthermore, to enhance the phenotypic response that PSMC3IP and EPSTI1 might have on caspase-8, we induced the extrinsic apoptotic pathway using TRAIL, which typically generated about 25 and 35% cell viability decrease in non-transfected MDA-MB-231 and MCF-7 cells, respectively (Fig. 2), in agreement with the data reported in the literature [37][38][39].
We subsequently performed a variety of functional assays to measure the apoptotic activity of the two candidate genes in BC cells under basal or apoptotic induced conditions.
Collectively, these results indicate that PSMC3IP and EPSTI1 do modulate caspase-8 activity, suggesting their involvement in the extrinsic apoptotic pathway in breast cancer cells.

PSMC3IP and EPSTI1 expression modulates caspase-3 activity and PARP cleavage
It is well known that the activation of initiator caspases, like caspase-8, leads to the activation of the executioner caspases, such as caspase-3 in MDA-MB-231 cells [42]. Therefore, we investigated whether PSMC3IP or EPSTI1 expression affects caspase-3 activity under basal or apoptotic conditions in MDA-MB-231 cells. We observed that overexpression of either gene does not alter caspase-3 activity levels (Fig. 5A). Yet, EPSTI1 silencing results in an increased caspase-3 activity in both basal conditions (2.6-fold, P<0.001) and upon TRAIL treatment (2.4fold, P<0.001), giving similar results as the silencing of the anti-apoptotic gene XIAP. On other hand, PSMC3IP silencing is only able to increase caspase-3 activity under TRAIL treatment (1.5-fold, P<0.05) (Fig. 5B). These results indicate that indeed EPSTI1 and PSMC3IP modulate caspase-3 activity in MDA-MB.231 cells, albeit at varying degrees of apoptotic stimulation.   Caspase-7 and caspase-3 coordinate the last phase of apoptosis by cleaving protein substrates such as PARP [43,44], which has an essential role in repairing single-strand breaks (SSBs). PARP is inactivated by caspase cleavage, causing SSB repair inhibition that can result in lethal DNA damage [45]. Interestingly, overexpressing PSMC3IP or EPSTI1 show a significant decrease in PARP cleavage in MDA-MB-231 and MCF-7 cells (1.3-fold, P<0.01; 1.8-fold, P<0.01) (1.2-fold, P<0.05; 2.9-fold, P<0.001) (Fig. 5C-D). Furthermore, in agreement with our caspase-3 activity results, there is an increase of cleaved PARP levels after PSMC3IP or EPSTI1 gene depletion in MDA-MB-231 cells upon TRAIL induction (PSMC3IP, 3.2-fold, P<0.001; EPSTI1, 1.5-fold, P<0.05) (Fig. 5E). This effect is more pronounced in MCF-7 cells, where PARP cleavage is increased even in basal conditions (EPSTI1, 4.2-fold, P<0.001; PSMC3IP, 3.6-fold, P<0.05) (Fig. 5F). Taken together, although preclinical studies have shown that ER-negative breast cancer cell lines are more sensitive to PARP inhibitors compared to luminal cells [46], our results indicate that PARP cleavage is similarly affected by the expression of putative extrinsic regulators such as PSMC3IP and EPSTI1 in both ER-negative (i.e. MDA-MB-231) and in luminal breast cancer cell line (i.e. MCF-7). In addition, since EPSTI1 also interacts with AKT1, it could play an alternative role in modulating apoptosis through PI3K pathway, in line with preclinical data that demonstrate synergistic activity when PARP inhibitors are combined with PI3K inhibitors [47].
Increased DNA fragmentation and reduced cell viability are associated with PSMC3IP and EPSTI1 down-regulation DNA fragmentation, resulting from apoptotic signalling cascades, is a hallmark of late-stage apoptosis [48]. Hence, we wanted to examine whether EPSTI1 and PSMC3IP are able to alter the final apoptotic response beyond the modification of caspase activity. To this end, we quantified the number of apoptotic cells by flow cytometry (i.e. measurement of the sub-G 0 /G 1 peak in the fluorescence histograms) and we further examined DNA fragmentation by TUNEL assays (Fig. 6).
In MDA-MB-231 cells, we detect an increased number of apoptotic cells upon PSMC3IP silencing under TRAIL conditions (1.6-fold, P<0.01), while EPSTI1 depletion has already a similar effect under basal conditions (1.7-fold, P<0.05) (Fig. 6A). Yet, TUNEL-based fluorescent microscopy images do not show conclusive results (Fig. 6B). Interestingly, MCF-7 cells show a clearer phenotype, where PSMC3IP depletion is able to increase the number of apoptotic cells in basal conditions (4.3-fold, P<0.01) (Fig. 6C) and EPSTI1 silencing is now able to induce a high increase of apoptotic cells under both conditions (17.7-fold, P<0.001; 3.5-fold, P<0.001) (Fig. 6C). These findings are in accordance with apoptotic positive cells observed in TUNEL assays (Fig. 6D).
Lastly, we sought to determine breast cancer cells viability after PSMC3IP or EPSTI1 expression modulation. As observed in Fig. 7, up-regulation of either gene is not able to recapitulate the viability of neither TRAIL-treated MDA-MB-231 (Fig. 7A) nor MCF-7 cells (Fig. 7B). Conversely, EPSTI1 down-regulation does decrease the viability of MDA-MB-231 cells treated with using MYC-tag transfection vector (Vector) as a negative control, reveals an attenuation of downstream apoptotic cascades. (D) This effect is much more pronounced in TRAIL-treated MCF-7 cells. (E) Analysis of cleaved PARP protein levels after gene silencing in MDA-MB-231 cells. siLUC was used as a negative control. (F) The same analysis in MCF-7 cells shows a more pronounced effect, even under basal conditions. EPSTI1-depleted cells were previously treated with IFN-α at 1000 U/ml for 8h. In apoptosis-induced conditions, MDA-MB-231 and MCF-7 cells were treated with TRAIL for 24h at 80 or 100ng/mL, respectively. XIAP was used as an anti-apoptotic reference in all experiments. Each bar represents the mean AESD of three experiments performed in duplicate (*P <0.05, **P <0.01, ***P <0.001 vs MYC-tag vector in overexpression assays and vs siLUCIFERASE in silencing).

Concluding Remarks
As shown in the summary Tables 1 and 2, collectively, our findings reveal that PSMC3IP and EPSTI1 have a strong anti-apoptotic role in breast cancer cells, particularly in estrogen receptor positive and triple negative, by means of physical interaction with the apoptosis initiator caspase-8 (Fig. 8). Interestingly, PSCM3IP or EPSTI1 depletion in breast cancer cells show increased DNA fragmentation and reduced cell viability even in the absence of apoptotic stimuli, indicating that they might also modulate the apoptotic pathway trough alternative mechanisms, such as by BCAR3 or AKT1 interaction, particularly in case of EPSTI1. Although under basal conditions (left). (D) TUNEL positive nuclei displaying green fluorescence are observed under basal conditions (right), although a higher number is clearly observed under TRAIL-induced conditions (left). siLUC was used as a negative control, XIAP was used as an anti-apoptotic reference in all experiments. EPSTI1-depleted cells were previously treated with IFN-α at 1000 U/ml for 8h. In apoptosis-induced conditions, cells were treated with TRAIL for 24h, at 80 or 100ng/mL respectively. Each bar represents the mean AESD of three experiments performed in duplicate (*P <0.05, **P <0.01, ***P <0.001 vs siLUC).  further studies are required to gain deeper insight into the molecular mechanisms underlying the anti-apoptotic role of PSMC3IP and EPSTI1 in breast cancer, our findings highlight them beforehand as very interesting therapeutic targets, preeminently for their ability to apoptosis sensitization. The extrinsic apoptosis pathway is initiated by the ligand binding to the death receptor, forming the DISC complex and leading to the activation of the caspase signaling cascade. The overexpression of EGFR and HER2 or activation of components of their downstream signaling pathways (i.e. PI3K pathway) induce an anti-apoptotic signaling through PI3K-Akt pathway in endocrine resistant breast cancer cells. Based on our findings, we suggest that PSMC3IP and EPSTI1 may regulate the apoptotic pathway via the physical interaction mainly with the apoptosis initiator CASP8, but also with AKT1 and BCAR3. Pro-apoptotic proteins are displayed in red and anti-apoptotic in green. Candidate proteins are displayed in purple. doi:10.1371/journal.pone.0115352.g008 Supporting Information S1 Table. PSMC3IP and EPSTI1 interactions with BC-apoptosis genes. List of selected BC-apoptosis genes and detected interactions with PSMC3IP and EPSTI1 observed by matrix Y2H screens. Candidate gene expression in healthy and carcinoma breast tissues and their co-expression with core BC-apoptotic genes. (XLSX)