Database analysis

RNA-sequencing expression (level 3) profiles and corresponding clinical information of hormone receptor-positive breast cancer were downloaded from the TCGA dataset (https://portal.gdc.com), including 615 cases of HR + breast cancer and 113 cases of normal tissues. The current-release (V8) GTEx datasets were obtained from the GTEx data portal website (https://www.gtexportal.org/home/datasets), including 459 normal tissues. A total of 2483 immune-related genes were obtained by downloading immune gene data through the ImmPort data portal (http://www.immport.org). A total of 1793 of them were unique elements.

Screening of independent prognostic differentially expressed genes

On the basis of the RBP7 expression, 615 h + BC samples from TCGA were divided into high (n = 308) or low (n = 307) groups. The limma package was used to study the differentially expressed genes (DEGs). Differential gene expression between the high and low RBP7 groups was analyzed using the DESeq R package (1.8.3). Heatmaps of differential genes were drawn using the pheatmap R package.

Kaplan-Meier (KM) survival analysis with the log-rank test was used to compare the overall survival (OS) differences between the two groups. The OS-associated genes in HR + BC were screened out by Kaplan-Meier curves using R software version v4.0.3 (The R Foundation for Statistical Computing, 2020). To filter the genes strongly associated with OS, P < 0.01 was set before calculating the crossed dataset.

The Venn diagram was drawn by the “VennDiagram package.” Three original datasets of the Venn diagram were DEGs (n = 5175), OS-related genes (P < 0.01, n = 203) and total immune-related genes (n = 1793). Univariate and multivariate analyses were performed using the Cox regression method. The forest was drawn by the ‘forestplot’ R package. Genes with P < 0.05 were considered potential independent prognostic factors. The log-rank test was used to compare differences in survival between the high and low RBP7 groups. The median survival time was calculated by R software. TimeROC (v 0.4) analysis was used to compare the predictive accuracy of RBP7 mRNA.

Functional enrichment analysis

The ClusterProfiler package (version: 3.18.0) in R was employed to analyze the GO function of potential targets and enrich the KEGG pathway. The R software ggplot2 package was used to draw boxplots. To further explore signaling pathway enrichment, gene set enrichment analysis (GSEA) was performed between the low and high RBP7 groups using GSEA Java software (https://www.gsea-msigdb.org/gsea/index.jsp). For GSEA, hallmark and C7 gene sets were utilized in this study.

Construction of a prognostic model

The clinicopathological data matching the HR + BC samples were extracted from the TCGA data resource. Statistical analysis and ggplot2 (v3.3.2) were performed using the R program, and a P value < 0.05 was indicated statistical significance.

The forest map was drawn through the “forestplot” R package to display the P value, HR and 95% confidence interval (CI) of each variable on the basis of RBP7 expression, age and T/N/M stage. The nomogram was built according to the results of multivariate Cox proportional hazards analysis, predicting the overall recurrence rate at 1, 3, and 5 years. The nomogram provided a graphical representation of the independent factors, and the recurrence risk of a single patient can be calculated through the points related to each risk factor in the “rms” R package.

Tissue samples

To detect the expression of RBP7, a total of 10 tumor samples and 10 normal samples from HR + BC patients who had not undergone therapy before surgical resection were obtained from the First Affiliated Hospital of Soochow University. Consent for participation in the research was duly obtained. Tissue microarrays (TMAs) used for immunohistochemical (IHC) staining analysis were composed of 151 samples, including 28 normal breast tissues, 69 h + tumor tissues and 54 tumor tissues of other molecular types (Bioaitech, Xi’an, China). The Ethics Committee of the First Affiliated Hospital of Soochow University authorized and supervised this study.

Cell culture

ZR-75-1, ZR-75-30, and BT474 cells were obtained from Abiowell Biotechnology Co., Ltd. (Changsha, China) and cultured in Roswell Park Memorial Institute (RPMI)-1640 medium (#61870-010, Gibco, Rockville, MD, USA) supplemented with 10% fetal bovine serum (#10270-106, Gibco, Rockville, MD, USA). T47D and MCF7 cell lines were obtained from the American Type Culture Collection (ATCC, USA). The T47D cell line was cultured in DMEM/high glucose medium (#SH30022.01B, HyClone, USA), and the MCF7 cell line was cultured in MEM (#SH30265.01B, HyClone, USA). Each medium was supplemented with 10% fetal bovine serum (#10270-106, Gibco, Rockville, MD, USA) and 1% penicillin‒streptomycin (#SH30010, HyClone, USA). Cells were cultured at 37 °C containing 5% CO2.

Construction of stable and transient cell lines

For downregulation of RBP7, siRNAs and shRNA targeting RBP7 were constructed by GenePharma (Suzhou, China). Three potential siRNAs were provided, including RBP7-Homo38 (5′-GCCGACCUCAGCGGUACUUTT − 3′), RBP7-Homo187 (5′- CCACACGAACAGCAGCCUATT − 3′), and RBP7-Homo309 (5′- GGCUCACCUGUAUCCAGAATT − 3′). Lipo8000™ Transfection Reagent (#C0533, Beyotime, Nanjing, China) was used for transient transfection. For RBP7 stable knockdown, MCF7 cells were transfected with shRNA-RBP7 (LV2N(U6/Puro)-RBP7-Homo-309). For upregulation of RBP7, overexpression lentiviruses for RBP7 (pLV3-CMV-RBP7) and control lentiviruses (pCDH-EnCMV-MCS-3×FLA) were purchased from MiaoLingBio (Wuhan, China). Polybrene (#H9268, Sigma, USA) was used for transfection. Follow-up experiments were carried out.

Quantitative reverse transcriptase polymerase chain reaction (qRT‒PCR)

The RNA-Quick Purification Kit (#RN001, Yishan Biotechnology, Shanghai, China) and PrimeScriptTM RT Master Mix Kit (#RR036A; Takara, Dalian, China) were used for total RNA extraction and cDNA synthesis, respectively. An AceQ qPCR SYBR Green Master Mix (without ROX) kit (#Q121-02-AA; Vazyme, Nanjing, China) was used to perform real-time quantitative PCR. The specific scheme was conducted in accordance with the instructions of the reagent manufacturer. The relative gene expression levels were normalized to GAPDH and calculated by the 2−∆∆Ct method. The sequences of primers used for qRT‒PCR are as follows:

GAPDH: Forward, 5’-AATCCCATCACCATCTTCCA-3’

GAPDH: Reverse, 5’-TGGACTCCACGACGTACTCA-3’

RBP7: Forward, 5’-CTCAGCGGTACTTGGACCC-3’

RBP7: Reverse, 5’-CGAGTGGCAAAGTCAATACCT-3’

Cell counting Kit-8 (CCK-8) assay

Cell viability was determined by Cell Counting Kit-8 (CCK-8) (#C6005, NCM Biotech, Suzhou, China) assay. A total of 2000 breast cancer cells were seeded in 96-well plates and incubated for 24 h, 48 h, 72 and 96 h. After that, 10 µL CCK-8/well was added to the culture medium and incubated for another 2 h. The OD450 was determined on an enzyme plate analyzer.

Western blotting

Tumor cells were lysed with RIPA lysis buffer (#P0013B, Beyotime, Shanghai, China) supplemented with protease and phosphatase inhibitor cocktail (#P1045, Beyotime, Shanghai, China) and then sonicated. After centrifuging the lysate, the supernatant was collected and the concentration was determined with the bicinchoninic acid method (BCA Protein Assay Kit, #P0012S, Beyotime, Shanghai, China). Total protein (20 µg) was separated by 12.5% SDS/PAGE gel (#P2013, NCM, Suzhou, China) and then transferred to polyvinylidene fluoride (PVDF) membranes (#10,600,023, GE, USA), blocked with 5% skim milk (#P0216, Beyotime, Shanghai, China), and incubated with primary antibodies, such as mAbs anti-RBP7 Ig (1:1000, #14541-1-AP, Proteintech, Wuhan, China), anti-GAPDH Ig (1:50000, #60004-1-Ig, Proteintech, Wuhan, China), anti-AKT Ig (1:1000, #4691; Cell Signaling Technology, Shanghai, China), anti-p-AKT Ig (1:2000, #4060; Cell Signaling Technology, Shanghai, China) and anti-SREBP1 Ig (1:1000, #14088-1-AP, Proteintech, Wuhan, China), overnight at 4 °C. The membrane was then incubated with the secondary antibody (1:5000, #SA00001-1, #SA00001-2, Proteintech, Wuhan, China) at room temperature for 1 h, and ECL Western blot reagent (#36222ES, YEASEN, Shanghai, China) was added after washing. A Chemi DocTM MP Imaging System was used for detection.

Transwell assays

24-well Transwell plates (8 μm pore size) were used for invasion and migration assays. For migration assays, a total of 5 × 104 cells resuspended in serum-free medium were seeded in the inserts, while 800 µL of medium containing 10% FBS was added into the lower chamber.

In the invasion assay, Matrigel (#C0371, Beyotime, Shanghai, China) was added to the Transwell chamber 2 h in advance, and then the cell suspension was added. The remaining steps were the same as above.After 24–48 h, invaded cells were fixed with 4% paraformaldehyde (#P0099, Beyotime, Shanghai, China) and stained with 1% crystal violet (#C0121, Beyotime, Shanghai, China). Six random fields per well were photographed by microscopy and counted.

Colony formation

After transfection, 1000 cells were seeded in 6-well plates and incubated in a cell incubator at 37 °C with 5% CO2 and saturated humidity for 14 days. Cells were fixed using 4% paraformaldehyde and stained with 1% crystal violet for colony enumeration.

Wound healing assays

To generate a confluent monolayer, 5 × 105 cells were seeded in six-well plates and cultured. Wound areas were scraped using 200-µl pipette tips, followed by three washes with PBS to eliminate debris. Subsequently, culture media without FBS was added. The closure of the wounds was observed and captured using an inverted microscope at 0 and 48 h.

Cell cycle analysis

To analyze the cell cycle, the collected cells were fixed with ethanol at -20 °C for 24 h. After two washes with 1×PBS, the cells were incubated in the dark at room temperature with PI/RNase staining solution (#C1052, Beyotime, Shanghai, China) for 15 min. The DNA content of each cell phase was determined using a flow cytometer.

Xenograft mouse models

Animal experiments were carried out in accordance with the guidelines and approval of the Institutional Animal Care and Use Committee of Soochow University (Suzhou, China). For the experiment, five-week-old SPF NCG female mice (Gempharmatech Co., Ltd, Nanjing, China) were randomly allocated into three groups: kd-RBP7 group, WT (wild-type) group and oe-RBP7 group, with 5 mice in each group. Subcutaneous injections of kd-RBP7, wild-type and oe-RBP7 MCF7 cells were administered to the mice on the same side. After 36 days, all mice were humanely sacrificed through cervical dislocation, and the tumors were collected for further measurement.

Nile red staining and oil red O staining

Oil red O staining was used for lipid droplet quantification. For Oil Red O staining, frozen cells were fixed with 4% paraformaldehyde for 10 min and then incubated with Oil Red O solution (six parts Oil Red O stock solution and four Parts H2O; Oil Red O stock solution was 0.5% Oil Red O in 100% isopropanol) (#C0157S, Beyotime, Shanghai, China) for 15 min. Nile red is a fluorescent stain for intracellular lipid droplets. For Nile red staining, the fixed cells were incubated with 0.1 µg/mL Nile red (#7385-67-3, Angene, Nanjing, China) for 15 min. DAPI (#C1005, Beyotime, Shanghai, China) was used to stain the nucleus.

Free fatty acid measurement

The free fatty acid determination was carried out using a free fatty acid content assay kit (#BL869B, Biosharp, Hefei, China). The extraction solution was added according to the number of cells or the quality of the tumor samples according to the instructions. The cells were lysed with an ultrasound instrument and centrifuged to obtain the supernatant. A testing group and a control group were established according to the instructions of the reagent kit. After sufficient shaking, the supernatant was taken, and the absorbance was measured at 715 nm using an enzyme-linked immunosorbent assay. The absorbance value was substituted into the standard curve equation to obtain the free fatty acid content.

Statistical analysis

Each experiment in the present work was performed in triplicate. The statistical analysis was conducted using SPSS software (version 13.0). Student’s t test was used to compare the two groups, and ANOVA was used to compare multiple groups. A significance level of P < 0.05 was used to determine statistical significance.