NRIP1 is activated by C-JUN/C-FOS and activates the expression of PGR, ESR1 and CCND1 in luminal A breast cancer

Using chip array assays, we identified differentially expressed genes via a comparison between luminal A breast cancer subtype and normal mammary ductal cells from healthy donors. In silico analysis confirmed by western blot and immunohistochemistry revealed that C-JUN and C-FOS transcription factors are activated in luminal A patients as potential upstream regulators of these differentially expressed genes. Using a chip-on-chip assay, we identified potential C-JUN and C-FOS targets. Among these genes, the NRIP1 gene was revealed to be targeted by C-JUN and C-FOS. This was confirmed after identification and validation with transfection assays specific binding of C-JUN and C-FOS at consensus binding sites. NRIP1 is not only upregulated in luminal A patients and cell lines but also regulates breast cancer-related genes, including PR, ESR1 and CCND1. These results were confirmed by NRIP1 siRNA knockdown and chip array assays, thus highlighting the putative role of NRIP1 in PGR, ESR1 and CCND1 transcriptional regulation and suggesting that NRIP1 could play an important role in breast cancer ductal cell initiation.


Gene expression comparison between luminal A ductal cells and normal ductal cells indicates that C-JUN and C-FOS are putative regulators of differential expressed genes. Using a fold
change ≥ 5 to compare gene expression between Luminal A tissues and healthy tissue, we identified 133 differentially expressed genes (DEGS) (Supplementary Table S2).
The application of an in silico analysis with MetaCore software (Clarivate Analytics, USA) indicated that C-JUN, C-FOS and C-MYC were putative candidate regulators of our differentially expressed genes (Fig. 1A).
As shown in Fig. 1B, both C-JUN and C-FOS were overexpressed in the luminal A samples. We also evaluated the expression of phosphorylated C-JUN and C-FOS proteins, and both were increased in luminal A tissues (Fig. 1B, Supplementary Fig. S1). We also performed immunohistochemistry assays using C-JUN and C-FOS phosphorylated antibodies (Fig. 2).
Together, our data indicates that activation of both the C-JUN and C-FOS proteins in luminal A tissues may be related to the increased expression of some differentially expressed genes.

NRIP1 is a target of C-JUN and C-FOS in luminal A breast cancer.
To identify the targets of C-JUN and C-FOS transcription factors, a chromatin immunoprecipitation (ChiP) assay with anti-C-JUN and anti-C-FOS antibodies was carried out, followed by a promoter chip array assay using MCF7 and HMEC cell lines. We defined the C-JUN and C-FOS targets in each cell line using a p-value of < 0.01 as a cutoff.
Using a ≥ twofold change as a cutoff (Supplementary Table S2), we found 45 DEGs in the luminal A tissues that are potential genes regulated by C-JUN and/or C-FOS transcription factors (Supplementary Table S3). The expression of these genes was confirmed in 541 luminal A breast cancer patients from TCGA-BC and 178 healthy breast tissue samples from the GTEX dataset ( Supplementary Fig. S2).
Five of these genes were tested for AP1 regulation (DNAJC10, INHABA, NRIP1, YTHDF3 and ZBTB6). For this, 2 kb of the promoters of each chosen gene were analyzed in silico, and the predicted sites were tested for AP1 direct binding.
We identified two consensus binding sites in the NRIP1 gene, five in the ZBTB6 gene, two in the DNAJC10 gene and one consensus binding site in the INHBA and in the YTHDF-3 genes.
As shown in Fig. 3, our results showed several specific bindings of C-JUN and C-FOS at consensus binding sites for these genes. However, binding of both C-JUN and C-FOS observed in the two consensus binding sites was only observed for the NRIP1 gene promoter.
These results suggest that the NRIP1 gene is a potential gene regulated by AP1. We next performed transient transfection assays using the MCF7 cell line with constructs containing two (S1) or one AP-1 (S2) consensus binding site, and luciferase activity was measured using a luciferase assay approach. As shown in Fig. 4, both promoter regions caused an upregulation of luciferase activity, mainly in the proximal region (S2). NRIP1 gene and its target genes are altered in luminal A breast cancer patients. The NRIP1 gene was overexpressed in luminal A tissues in the chip array and TGCA dataset (Supplementary Table S2, Sup-   Fig. S2, respectively). To validate this expression, we performed RT-qPCR with a larger number of patients and healthy donors (Fig. 5A). Moreover, through immunohistochemistry assays, we demonstrated an increase of the NRP1 protein in all Luminal A patients (100.0%) (Fig. 5B,C). None of the healthy donors presented NRIP1 nuclear labeling, suggesting increased nuclear protein expression (Fig. 5D). Next, an in silico analysis showed that genes related to breast cancer, including HES1, MYC, CCND1, FOS and PGR, could be regulated by NRIP1 (data not shown).
As shown in Fig. 6, the mRNA levels of HES1, MYC, CCND1, and PGR were increased in the luminal A patients compared with those in the healthy donors, indicating that NRIP1 could be important for the regulation of these genes.
Using the same TCGA and GTEX datasets, the expression profiles of the same genes were investigated and as we can observed in Fig. 7 the expression of these genes corroborating with our results.

NRIP1 alters the expression of important genes in the transformation of luminal A ductal cells.
To verify the impact of the NRIP1 gene on luminal A breast cancer gene expression, we performed a functional analysis of NRIP1 depletion in the MCF7 cell line using an siRNA approach. We verified the depletion 24, 48 and 72 h after transfection, and at 24 h, the mRNA levels were depleted by 80% compared with the scrambled control.
Using mRNA levels from 24 h and considering a ≥ 1.5-fold change as the cutoff to define overexpression or downregulation in an expression chip array assay, we identified 762 differentially expressed genes related to NRIP1 silencing (Supplementary Table S4), including the PGR, ESR and CCDN1 genes.
The in silico analysis of the signaling pathways that could be associated with the differentially expressed genes was performed using MetaCore software (Clarivate Analytics, USA), and it revealed that the main pathway related to NRIP1 silencing was "PR action in breast cancer-stimulation of cell growth and proliferation" (Fig. 8). The PGR, ESR1 and CCND1 genes were downregulated when NRIP1 was silenced, indicating the possible role of NRIP1 in breast cancer development.
To corroborate with these results, we also applied siRNA approach to deplete NRIP1 using another Luminal A cell line (T47D). We also verified the depletion 24, 48 and 72 h after transfection, and 24 h was the depletion chosen time. To verify if PGR, ESR1 and CCND1 genes expression were also downregulated in this cell line when NRIP1 was silenced, we performed RT-qPCR ( Supplementary Fig. S3). This result is in accordance with that obtained with MCF7 cell line, further reinforcing the possible role of NRIP1 in breast cancer development.

Discussion
Breast cancer is a heterogeneous and complex disease, and the treatment decisions and prognosis are determined according to the different subtypes, which are based on histopathological, biological, and molecular characteristics 9,10 . The values were compared to pGL3 (Mock), which was used as a negative control. pGL3-S1 contains proximal and distal sites, and pGL3-S2 contains only proximal sites. After 24 h, the S2 region showed twice the activity compared to the S1 region, which spans distal and proximal regions. After 48 h and 72 h, similar behavior was observed but with less intensity. Each bar represents the mean ± SD. *p < 0.05, **p < 0.01. www.nature.com/scientificreports/ Although breast cancer is a highly studied tumor, most studies are related to the classification, treatment and progression of the disease to metastasis. Studies investigating events related to its initiation and initial progression to an invasive tumor are scarce. The transition of DCIS to invasive breast cancer has recently been studied at the molecular level and showed a gene expression signature that could predict DCIS progression 11 . However, the molecular mechanisms responsible for breast tumor initiation and the overall mechanisms that lead healthy cells to transform into luminal A tumors remain poorly understood.
Results of DEGS analysis indicates C-JUN and C-FOS as potential regulators. These genes have shown before important roles in the regulation of several cellular processes implicated in cancer acting in dimers called AP1. The Jun and Fos subfamilies are the most important AP-1 proteins that bind TPA responsive elements (TREs) 12 .
Some studies evaluated the expression of AP-1 family components from primary breast tumors from patients with invasive ductal and lobular carcinomas with adjacent nontumor tissue 13 . In our study, while the mRNA levels of C-JUN and C-FOS did not differ between the tumor and normal tissues, the C-JUN and C-FOS proteins were not only increased in expression but also activated in the luminal A patients, indicating a potential regulatory role. To test this hypothesis, we combined several experimental approaches to identify NRIP1 as one of the regulated genes.
Nuclear receptor interacting protein 1 (NRIP1) is a coregulator of several nuclear receptors and transcription factors that can act as a coactivator or a corepressor. NRIP1 is essential for normal mammary gland development and functions as a component in estrogen signaling [14][15][16][17] . www.nature.com/scientificreports/ In the mammary gland, NRIP1 modulates the expression of several ER target genes, such as AREG, PGR, CCND1 and STAT5a, thus acting as a coregulator with ER. When NRIP1 expression is ablated, the expression of AREG and PGR is lost, which leads to developmental defects 17,18 .
NRIP1 has shown to be related to human cancers 16,18 . In BC, it was associated with estrogen receptors (ERs) and suggested to regulate proliferation and invasion 15 . Moreover, it was associated with the risk of breast cancer 16 .
NRIP1 expression was found to be elevated in ductal carcinomas in situ 14 and overexpressed in human breast cancer tissue, and its expression in BC cell lines was elevated compared with that in MCF10A 16 , suggesting that NRIP1 is indeed altered in breast cancer cells.
As NRIP1 is an overexpressed transcriptional coregulator in breast cancer, we focused on evaluating the genes that could be regulated by NRIP1. To evaluate the importance of NRIP1, we depleted this gene in the MCF7 and T47D cell lines using an siRNA approach and performed a chip array or RT-PCR assays, respectively. The results from chip array assay using MCF7 cell line, revealed 762 differentially expressed genes related to NRIP1 silencing, including the PGR, ESR and CCDN1 genes, which we have shown to be upregulated in luminal A patients of our cohort and that of TCGA. When we depleted NRIP1 using T47D cell line, the results reinforced our findings.
Aziz and coworkers and Yuan and coworkers showed that the inhibition of NRIP1 expression in siRNA assay are related to apoptosis and cell growth inhibition 16,18,19 . Another group, using the same approach experiments, revealed that NRIP1 is needed to the regulatory complex required to stimulate breast cancer proliferation. Moreover, the genes that changed as a result of NRIP1 knockdown in MCF7 cells were used to stratify patients with breast cancer who received adjuvant tamoxifen treatment 18 . Corroborating these studies, our in silico analysis of NRIP1 silencing showed that signaling pathways related to breast cancer were altered.
Taken together, our results indicate that in the development of breast cancer, NRIP1 is targeted by C-JUN and C-FOS in luminal A cells and plays a role in ductal cell transformation by regulating genes related to the disease. Thus, NRIP1 is a target for future breast cancer therapy.

Methods
Study design. All patients included in this study were female (mean age: 56; age range 30-80 years) and were diagnosed with molecular subtype luminal A breast cancer, and samples were collected before the first chemotherapeutic regimen. Clinicopathological data were obtained from medical records (Table 1). These patients were stratified into four cohorts Microarray cohort (n = 4), Western Blot cohort (n = 4), RT-qPCR cohort (n = 28) and Immunohistochemistry cohort (n = 25) (  20 . Partek software was used for data analysis 21 and MetaCore software was used to evaluate pathway analysis and related processes. (Clarivate Analytics, USA). All data have been deposited in the NCBI's Gene Expression Omnibus (GEO) (GSE58102 and GSE180835). Chip-on-chip assay. To identify the promoter targets of C-JUN and C-FOS, we used SimpleChIP Enzymatic Chromatin IP kit (Magnetic Beads) to perform chromatin immunoprecipitation (ChIP) assays according to the manufacturer's instructions (Cell Signaling Technology, MA, USA), followed by a promoter chip array assay (Affymetrix, CA, USA). Briefly, chromatin from the MCF7 and HMEC cell lines, which was previously prepared and digested with micrococcal nuclease, was incubated with 2 μg of C-JUN antibody, C-FOS antibody (Santa Cruz Biotechnology, TX, USA) or the negative immunoprecipitation control normal anti-IgG rabbit antibody (#2729, Cell Signaling Technology, MA, USA). Then, the immunoprecipitated DNA was purified according to the manufacturer's instructions (Affymetrix CA, USA) and subsequently hybridized to a Human Promoter 1.0R Array (Affymetrix, CA, USA). After that, the arrays were washed, stained and scanned according to the manufacturer's protocols.
Promoter array analysis. The binding sites of C-FOS and C-JUN were identified using Tiling Analysis Software (TAS) version 1.1. Quantile normalization of the probe intensity data was applied to the enriched samples and the controls of both cell types. The probe analysis was conducted with a bandwidth of 500 bp, and the interval analysis selected promoter regions of at least 200 bp in length and a maximum gap of 50 bp between significantly enriched probes (pvalue < 0.01). These regions were subsequently matched with the promoter library of NCBI v.36 Human genome using BEDTools intersect program, which revealed the genes associated with the promoter regions.

Prediction of C-JUN and C-FOS binding sites.
To screen for putative C-JUN and C-FOS consensus binding sites, 2 kb upstream of the transcription start site from the selected genes was obtained from the NCBI database. The consensus binding sites were obtained using online prediction tools. The online software programs used were TRANSFAC (http:// www. gene-regul ation. com) and Genomatix (www. genom atix. de).

Real-time polymerase chain reaction (RT-qPCR) analysis.
To evaluate the putative C-JUN and C-FOS binding sites in the selected genes, purified DNA from the MCF7 cell lines after the ChIP assays with C-JUN, C-FOS or normal anti-IgG was used in RT-qPCR assays with specific primers (Supplementary Table S1) for each putative binding site. Rotor-Gene 6000 thermocycler (Qiagen, NWR, Germany) was used to perform the reactions using the following program: 95 °C for 10 min; 40 cycles at 95 °C for 20 s and 60 °C for 30 s; and a diagnosed from 2012 to 2018 and histologically confirmed by hematoxylin and eosin (H&E) staining were randomly chosen. Formalin-fixed paraffin-embedded normal breast tissue from mammoplasty (n = 3) was used as a procedure control. The samples were incubated with primary antibodies for 18 h at 4 °C. For the positive control, tissues determined by antibody manufacturer datasheets were used. The reaction was revealed with diaminobenzidine (DAB), followed by hematoxylin counterstaining. The negative controls were prepared with an antibody dilution solution without the primary antibody. The level of the proteins available in the control tissue and IBC biopsies was compared and analyzed with an unpaired t-test, and a Pearson test was performed to identify direct or indirect correlations.
RNAi knockdown (siRNA). 1 × 10 5 cell/mL of MCF-7 and T47D cells were plated in a 24-well plate and incubated overnight with RPMI-1640 media without antibiotics. NRIP1 siRNA (100 nM) (S15702, Thermo Fisher Scientific, MA, USA) and 2 μL of Lipofectamine 3000 (Thermo Fisher Scientific, MA, USA) were incubated separately in a final volume of 50 μL of RPMI-1640 media for 5 min. Subsequently, the siRNA and Lipofectamine were mixed, incubated for 30 min and then applied dropwise to the cell cultures. Scrambled siRNA (100 nM) (SC-37007, Santa Cruz) was used as a siRNA negative control. FITC-conjugated siRNA was used to evaluate the transfection efficiency by FACS. siRNA transfections were conducted for up to 72 h, and RT-qPCR analysis of the NRIP1 gene was used to evaluate the inhibition rates after 24 h, 48 and 72 h. The best inhibition rate was used to perform the ChIP array assay as described above or RT-qPCR using MCF7 or T47D cell lines, respectively. A pathway analysis and related processes were performed using MetaCore software (Clarivate Analytics, USA). mRNA expression data from the genotype-tissue expression project and the Cancer Genome Atlas. Healthy breast tissue samples (n = 178) from the public resource Genotype-Tissue Expression (GTEx) project and luminal A (n = 541) breast cancer (BC) patients were sampled using the UCSC Xena online tool (https:// xena. ucsc. edu). The mRNA expression data were obtained by the RNA-seq strategy RSEM expected_ count (DESeq standardized) dataset with DESeq normalization. This process is recommended for comparisons of tumor tissue with normal tissue. The TCGA-BC dataset was categorized according to the BC molecular subtypes by the cBioPortal online tool (https:// www. cbiop ortal. org/) and TCGAbiolinks (R/Bioconductor package) 25 . Patients with DCIS (ductal in situ), neoadjuvant therapy, prior treatment/malignancy and inconclusive classification were further excluded from all analyses.

Statistical analysis.
All experiments were carried out in triplicate, and the data are expressed as the mean ± standard error of the mean. Unpaired Mann-Whitney test was used for all comparisons and a p-value < 0.05 was considered statistically significant. The analysis was performed using GraphPad Prism software (GraphPad Software Inc., CA, USA).