ERCC1/NGFR affects prognosis in basal-like breast cancer

Abstract

Objectives

To study the effects of ERCC1 and NGFR on metastasis and prognosis of basal-like breast cancer, and construct a prognostic prediction model, initially to explore whether ERCC1 plays a role in BLBC metastasis through NGFR.

Methods

RNA-seq was used to identify the metastasis-related gene NGFR associated with ERCC1. ERCC1 and NGFR were transfected into BLBC cells, then western blot and transwell assays were applied to explore the effects of ERCC1 and NGFR on the migratory/invasive abilities and whether ERCC1 plays a role in BLBC cell metastasis through the NGFR. Data from the TCGA database were used to analyze the relationship between NGFR and ERCC1 expression in BLBC patients. The effects of ERCC1 and NGFR on the prognosis were analysed using K–M survival curves and multifactorial Cox regression, and a prognostic prediction model was constructed.

Results

RNA-seq results showed that NGFR was a differentially expressed gene of ERCC1, and NGFR was associated with ERCC1.After knockdown of ERCC1, the migratory and invasive abilities of cancer cells were significantly reduced; after overexpression of NGFR, the migratory and invasive abilities of cancer cells were significantly enhanced. In ERCC1 knockdown cancer cells, the expression levels of NGFR were subsequently reduced, and overexpression of NGFR resulted in an increase in migrating and invading cells and a rebound in migration and invasion capacity. NGFR was positively correlated with ERCC1 expression levels in BLBC patients (r_s = 0.24, P < 0.05). The overall survival of patients in the group with high expression of ERCC1 and NGFR was shorter. High expression of ERCC1 (HR = 2.310, 95% CI 1.008 ~ 5.293) and NGFR (HR = 2.600, 95% CI 1.126 ~ 6.002) were independent risk factor for poor outcome. The prognostic prediction model’s C-index was 0.781, and the AUC values of 3-year and 5-year survival were 0.76, 0.81, and the calibration curves for 3-year and 5-year survival were close to the reference line. The accuracy of the model was similarly validated in the METABRIC dataset.

Conclusions

A positive correlation was observed between ERCC1 and NGFR. ERCC1 may affect the migration and invasion ability of BLBC cells through NGFR; ERCC1 and NGFR were independent risk factors for the poor prognosis of BLBC patients and had a predictive effect on the 3-year and 5-year prognosis of BLBC.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-025-02807-w.

Introduction

Breast cancer (BC) has become the most commonly diagnosed cancer worldwide, according to the latest data from the World Health Organisation's International Agency for Research on Cancer (IARC) [1], which is a significant threat to women's health and lives. Although the treatment of BC has become increasingly standardized and the 5-year survival rate has improved to approximately 90% [2], recurrent metastasis remains the leading cause of death among BC patients and is closely associated with poor prognosis [3–6]. Basal-like breast cancer (BLBC) is a subtype of triple-negative breast cancer (TNBC) with poor prognosis. It accounts for nearly 15% of all BC cases [7], characterized by the expression of basal-type keratins and myoepithelial cell markers to varying degrees [8]. It shows rapid progression at both local and distant metastasis and is associated with high rates of recurrence and metastasis [9]. Clinically, the therapeutic advancement of BLBC has lagged behind that of other subtypes, with a relatively poor prognosis. Therefore, it is significant to conduct an in-depth investigation of the factors affecting BLBC metastasis and poor prognosis, along with the identification of potential tumor biomarkers and effective targets. It can provide an important theoretical basis for the clinical application in cancer detection, prognosis assessment, and the development of targeted therapies.

The DNA damage repair system plays a critical role in maintaining the stability and integrity of genetic information, and its aberrant initiation is an important feature of tumourigenesis and progression [10, 11]. ERCC1 is a key DNA damage repair gene, and many studies have shown that aberrant expression of ERCC1 is associated with tumourigenesis, progression and metastasis [12–14]. In gastric cancer, low expression of ERCC1 is associated with a favorable prognosis for overall survival, and ERCC1 expression is associated with cancer malignancy [15]. In colorectal adenocarcinoma patients, ERCC1 expression is negatively correlated with tumour metastasis [16]. In breast cancer, ERCC1 is associated with vascular infiltration and lymph node metastasis. High ERCC1 expression is considered as a risk factor for both disease-free survival and overall survival, often indicating a poor prognosis [17]. Positive ERCC1 expression is associated with shorter progression-free survival and shorter overall survival in breast cancer patients, and high ERCC1 expression is thought to be significantly associated with adverse outcomes in patients receiving neoadjuvant chemotherapies [18]. Our previous research identified ERCC1 as an independent risk factor for postoperative metastasis in breast cancer, and patients with high ERCC1 expression had a significantly higher risk of metastasis than those with low expression [19]. These studies suggest that ERCC1 is closely associated with metastasis and prognosis in breast cancer. However, whether ERCC1 is associated with the prognosis and metastasis of BLBC, and how it contributes to BLBC progression, remain unclear and require further study.

Nerve growth factor receptor (NGFR) is widely distributed and expressed in a variety of tissues and tumours and plays an important role in biological processes, such as cell cycle and apoptosis, which in turn promotes tissue repair, metabolism, etc. [20, 21]. NGFR is abnormally expressed in a variety of tumours, such as cutaneous squamous cell carcinoma [22] and melanoma [23], and is associated with the prognosis of tumours [24]. NGFR expression has been reported to be higher in hypopharyngeal carcinoma compared to normal tissues, and this high expression has been shown to enhance cancer cell stemness, which is associated with poor prognosis [25]. Numerous studies have shown that NGFR is also associated with tumour metastasis. NGFR promotes melanoma metastasis and serves as an important biomarker for melanoma progression and early diagnosis [26]. In gastric cancer, expression of NGFR was significantly related to venous invasion, stage, and tumor infiltration pattern. Overall survival of the high expression group was significantly lower than that of the low expression group for NGFR [27]. These findings suggest that NGFR is associated with prognosis and tumour metastasis across various cancer types. A similar association is observed in breast cancer. NF-κB1 binding sites have been identified within the promoter region of NGFR, and the interdependent regulation between NGFR and NF-κB has been shown to influence breast cancer cell proliferation and migration [28]. In a mouse model of breast cancer induction, asparaginyl endopeptidase was found to exacerbate bone metastasis by increasing NGFR expression [29]. In the GOBO database, it was found that differential overexpression of NGFR in basal-like and HER2-rich tumour samples is correlated with overall survival and metastasis-free survival [30]. Recent studies indicate that NGFR expression is strongly associated with the BLBC subtype. Although it is less sensitive, its specificity is comparable or better than other basal markers, and NGFR expression may be indicative of the BLBC subtype, representing a potential co-marker for this subtype [31]. These studies suggest that NGFR is associated with breast cancer metastasis and prognosis, and is closely related to basal-like breast cancer typing and its prognosis. However, it remains unclear whether NGFR is associated with metastasis of BLBC subtypes and how NGFR is involved in metastasis and prognosis of BLBC, which needs to be further explored by researchers.

Moreover, it was found that the regulation of NGFR and ERCC1 expression is related to the cell cycle [32, 33], and these genes can maintain the tumor stem cells in a quiescent state through the cell cycle regulatory system. The quiescent cancer cells can initiate the macroscopic metastatic growth after re-entering the cell cycle, which is the underlying cause of drug resistance, recurrence, and metastasis of cancer [34].In this study, NGFR was found to be an ERCC1-related gene and associated with tumor metastasis by RNA-seq. These results suggest a potential correlation between ERCC1 and NGFR in tumors.

In summary, in this study, we examined the effects of ERCC1 and NGFR on the migration and invasion abilities of BLBC cells through in vitro cellular experiments, and investigated their roles and relevance in BLBC metastasis. In addition, we analysed the effects of ERCC1 and NGFR on BLBC metastasis and prognosis using the TCGA database, and constructed prognostic prediction models to support their potential as therapeutic targets and prognostic biomarkers for BLBC.

Materials and methods

RNA-seq screening of NGFR, a tumor metastasis-related gene associated with ERCC1

Total RNA was extracted from ERCC1 knockout and control cells using the Trizol method, and the concentration, integrity and purity of RNA were detected by Nanodrop and Agilent 5400 instruments, and the integrity and total amount of RNA were used as the main reference indexes. For the RNAs that successfully passed the quality control, sequencing was performed using Illumina instruments. P < 0.05 and |log2FC|> 1 were used as criteria of differential gene expression, and all differential genes related to ERCC1 were obtained by RNA-seq sequencing. Shanghai Genechem Co., Ltd (China) completes the gene sequencing experiments and quality control section. The specific screening criteria and quality control results are detailed in the Results section. Genes related to tumor metastasis were downloaded from the GeneCards database, and the intersection with these RNA-seq differentially expressed genes was taken as candidate genes, which are shown in the Veen diagram. Based on the KEGG database, KEGG pathway enrichment analysis of differentially expressed genes was performed using Pathview and other R packages, and the more significant KEGG pathways were selected from the enrichment results to draw scatter plots, from which we found the biological signalling pathways with significant changes related to tumor metastasis, and the genes contained in the pathways were further analysed to identify subsequent genes. The genes contained in the pathway were further analysed to determine the subsequent research gene, NGFR.

Source of information on BLBC patients

We obtained RNA sequencing data and relevant clinicopathological information on BLBC from the TCGA database, in which BLBC was defined using the PAM50 molecular typing of the TCGA whole gene expression profile [35].To ensure the reliability of the analysed results, we required that the key data of the included samples be complete. These key data cover clinical information that may be related to prognosis (e.g., age, stage, TNM, survival status, OS, etc.) and gene expression data (raw RNA-seq count data and normalised FPKM data, etc.). Patients with very short overall survival tended to be in an extremely advanced state or may have comorbid with other serious underlying conditions rather than the effects of specific genes, and to exclude acute deaths due to other factors, patients with an overall survival of less than 30 days were, therefore, excluded. Samples with missing key data were excluded. During data processing, if the difference in FPKM values was not large and there were no extreme values in the dataset, no transformation is necessary. If the difference was large, the FPKM data were transformed into log(FPKM + 1) before analysing.

Correlation of ERCC1 and NGFR with metastasis and BLBC clinical features

The SangerBox mapping module was used to understand the relationship between the ERCC1/NGFR and the age, stage, T, N, and M of BLBC patients, as well as other clinical and pathological factors. T means tumour magnitude, N means lymph node metastasis, and M means distant metastasis. To illustrate the results of the analyses, we used the “ggpubr” software package to simultaneously draw box plots.

Correlation analysis of ERCC1 and NGFR

Sangerbox 3.0 platform was used to perform correlation analyses and create charts.

Prognostic value of ERCC1 and NGFR in BLBC

In this study, we used the survival analysis panel in SangerBox to draw K–M survival curves, and then used one-way Cox regression and multifactorial Cox regression to explore the prognostic value of ERCC1 and NGFR in BLBC. ERCC1 and NGFR expression data, survival time, and survival status of 148 samples were utilised, and K–M survival curves were plotted after selecting the optimal cutoff point for ERCC1 and NGFR express. Subsequently, based on the optimal cutoff value, we explored the effects of ERCC1, NGFR and clinical characteristics on the prognosis of basal-like breast cancer using one-way Cox regression analysis. Next, multifactorial Cox regression analysis was used to assess the independent effects of ERCC1 and NGFR on prognosis using whether basal-like breast cancer is survivable or not as the dependent variable, and using univariate meaningful influences as independent variables.

Creating and validating prognostic prediction column charts for BLBC patients

Gene expression profiles, survival time, and survival status were integrated using the SangerBox tool, and prognostic modelling was performed using the Cox approach. This model can assess the prognostic significance of these characteristics in 148 samples. Tools such as the C-index, ROC curves over time, calibration curves and DCA decision curves were used to assess the predictive accuracy and stability of the model. Finally, to assess the robustness and generalisability of the model, we used the METABRIC database and downloaded the number of patients with basal-like breast cancer as an external dataset, which was cleaned and externally validated.

Cell line

We screened MDA-MB-231 and BT-549 BLBC cell lines with high ERCC1 expression, possessing high migratory and invasive abilities for experiments by RT-qPCR method [36, 37]; see Appendix Fig. 1.

Fig. 1.

Bubble map of KEGG functional enrichment analysis of 38 differential genes. The horizontal coordinate is the ratio of the number of differential genes on the KEGG pathway to the total number of differential genes; the vertical coordinate is the KEGG pathway; the size of the dots represents the number of genes annotated to the KEGG pathway; and the colour from red to violet represents the magnitude of significance of the enrichment

Acquisition and infection of ERCC1 gene-interfering lentivirus and NGFR plasmid

The construction of the ERCC1-interfering lentivirus and pre-infection experiments were completed by Shanghai Genechem Co., Ltd (China). The vector for ERCC1-interfering lentivirus and the negative control was GV493 with green fluorescence and puromycin labelling. The company designed the shRNA sequences of three target genes (see Appendix Table 1), and selected shRNA1 sequences with at least 75% knockdown efficiency and stability for the subsequent experiments. The NGFR overexpression target plasmid and the control plasmid were designed using the GV740 vector, and primer design and synthesis of the target fragment, plasmid construction, and identification were performed by Shanghai Genechem Co., Ltd (China) carried out the primer design, target fragment synthesis, plasmid construction, identification, comparison, and concentration detection.

After the cells were inoculated in 6-well plates, the cell confluence should be 20% ~ 30% when infected with the ERCC1 lentivirus and 60% ~ 70% when infected with the NGFR plasmid. After 12 h of infection, complete medium was used for fluid exchange, and the cells continued to be cultured in the cell culture incubator after fluid exchange. When the cell confluence of infected ERCC1 lentivirus reached about 70%, the fluorescence abundance was observed by an inverted fluorescence microscope, and puromycin (Beyotime, China) was added for screening. The screening concentration of puromycin was 2 µg/mL for MDA-MB-231 and 0.5 µg/mL for BT-549. After 48 h of screening, all the cells in the group without infection plus puromycin were dead, and at this time, it was considered that all the remaining cells in the group infected with the virus were positive. The above basal-like breast cancer cells successfully infected with ERCC1 knockdown lentivirus were named the ERCC1-sh group, and the corresponding cells infected with control lentivirus were named the ERCC1-ctrl group.

RT-qPCR

After infecting the cells for 24 h, extract the total cellular RNA for RT-qPCR to detect the infection efficiency. Total cellular RNA was extracted using an RNA extraction kit (Agbio, China), total RNA was reverse transcribed to cDNA using a reverse transcription kit (Vazyme, China), and the reaction was performed using an RT-qPCR kit (Vazyme, China). The housekeeping gene GAPDH was used as an internal reference and the relative expression of the target gene was calculated. Primer sequences are listed in Appendix Table 2. For the data analysis section, we used the 2^−ΔΔCt method to calculate the relative expression levels of genes in the experimental and control groups. And during the experiment, we ensured as much as possible that there should be three biological replicates for each gene in each group of samples, and each replicate needs to have three replicate wells.

Western blot

Cell proteins were extracted 48 h after cell infection. Total protein was extracted using RIPA lysate with protease inhibitor (Beyotime, China). The protein concentration was finally determined using the BCA kit (Beyotime, China). The protein samples described above were mixed with Protein Upload Buffer (Beyotime, China) at a ratio of 4:1 and then denatured by boiling. SDS–PAGE (Solarbio, China) was used to separate 20 µg of protein in each sample. The protein was then transferred to a PVDF membrane and blocked with blocking solution (Beyotime, China) for 30 min. Primary antibodies against ERCC1, NGFR, and α-Tubulin were added and then incubated three times with TBST buffer (Solabio, China) overnight at 4 °C in a shaker.

The membrane was washed three times with TBST buffer. The membrane was then incubated with horseradish peroxidase-labelled secondary antibody for 1 h at room temperature and washed three times with TBST buffer. Appendix Table 3 displays the details and dilution ratios of the primary and secondary antibodies. Adding ECL light-emitting reagent (Oriscience, China) to the membrane and then placing it into a gel imaging system to develop and capture images. Greyscale analysis was performed using Image J software, with α-Tubulin as a control and relative protein expression calculated. During the experiment, each group of samples was repeated three times and the average of the three times was taken for statistics.

Cell migration and invasion assay

This experiment was performed according to the manufacturer's protocol using a 24-well, 8.0 μm pore membrane (Corning, USA). When the cells were cultured to the logarithmic growth phase, single cell suspensions were made in serum-free medium for cell counting. We took 1 × 10⁵ cells per well and inoculated them in 100 μL of serum-free medium in the upper chamber, while we added 700 μL of complete medium containing 20% FBS as a chemotaxis agent in the lower chamber. After incubation at 37 ℃ for 24 h, cells remaining on the upper surface of the membrane were removed with a cotton swab, and cells on the lower surface of the membrane were migrating cells. After fixation with 4% paraformaldehyde and staining with 0.1% crystal violet solution, the upper chamber was washed twice with PBS phosphate buffer to remove residual crystal violet, and non-migrated cells were gently wiped away from the upper chamber with a cotton swab, and then the upper chamber was air-dried. Cells passing through the filter were photographed with an inverted fluorescence microscope, and the field of view with a more uniform cell distribution was selected for photographic counting. For invasion experiments, 100 μL of 1:8 DMEM-diluted matrix gel (Corning, USA) was added to each well at 37 ℃ for 6 h in advance before the cells were inoculated onto the gel-coated basement membrane. The other steps were consistent with the migration experiments. Each group of samples was repeated three times, and the average of the cell numbers of the three times was taken to count.

Data analysis

GraphPad Prism 8.0 software and SPSS 27.0 software were used to perform the analyses. Comparisons of multi-sample measurements were made using one-way ANOVA, and two-by-two comparisons were made using the LSD-t-test. P < 0.05 was considered a statistical significant difference.

Results

RNA-seq screening of the tumor metastasis-related gene NGFR associated with ERCC1

We transfected ERCC1 knockdown lentivirus in MDA-MB-231 cells in good status. Total RNA from this cell was extracted after 72 h, and real-time PCR was applied to verify the knockdown effect. Three repetitions were done, and the results of the three experiments showed that compared with the control group, the knockdown efficiency of the ERCC1 knockdown group was 75.18%, 76.33%, and 77.34%, respectively. Before formal sequencing, we accurately checked RNA concentration, purity, and integrity by Nanodrop and Agilent 2100 bioanalyzer for quality testing of the extracted RNA; see Appendix Table 4. The results showed that the integrity values of all RNA samples used were greater than 9, and the concentration, purity and integrity met the sequencing requirements. After successfully passing the RNA quality control, to ensure the quality of high-throughput sequencing data, the sequenced fragments were sequenced by the Illumina sequencing machine, and the image data measured by the high-throughput sequencer were converted into sequence data (reads) by CASAVA base recognition software. The reads were filtered through raw data, sequencing error rate check, and GC content distribution check before obtaining clean reads for use in subsequent analyses; see Appendix Table 5. The results showed that the clean reads of the samples in the ERCC1-sh group were 44,953,648, 44,374,774 and 45,454,342, respectively. And the clean reads of the samples in the ERCC1-ctrl group were 52,221,370, 45,665,732 and 46,375,112, respectively. After quality control by high-throughput sequencing, the Q20 and Q30 were both above 92%, and the G + C content was above 50%. This indicated that the data quality was high and ensured the reliability of the sequencing results.

After obtaining the clean reads data after the above quality control, the HISAT2 software was used to perform a fast and accurate comparison between the clean reads and the reference genome to obtain the positional information of the reads on the reference genome. Based on the positional information of the gene pairs on the reference genome, subread software was used for quantitative analysis, filtering out the reads with a quality value of less than 10 and the reads on the non-comparable pairs, so as to count the read count of each gene. The above read count values were normalised and corrected for sequencing depth, and then a statistical model was applied to perform hypothesis testing. The probability (p value) was calculated and finally corrected for multiple hypothesis testing. DESeq2 software was selected for gene expression difference significance analysis, and P < 0.05 and |log2FC|> 1 were used as the criteria for differential gene screening.

The results of RNA-seq showed that there were 523 differentially expressed genes in the ERCC1 knockdown group compared with the control group, of which 271 were up-regulated and 252 were down-regulated in expression. Subsequently, genes related to cancer metastasis were downloaded from the GeneCards database, and then intersections were taken with the above 523 differentially expressed genes, resulting in a total of 38 genes related to metastasis, such as NGFR; see Appendix Fig. 2 and Appendix Table 6.We performed KEGG pathway enrichment analysis on the above 38 genes related to ERCC1 and associated with tumour metastasis, and from the KEGG enrichment results, the 10 most significant KEGG pathways were selected to be plotted on a scatter plot for presentation, and the results showed that among the significantly enriched entries, cytokine–receptor binding interactions, transcriptional deregulation in cancer, the PI3K–Akt signalling pathway, and the MAPK signalling pathway were associated with tumour metastasis; see Fig. 1. Further analysis of the genes enriched in the above signalling pathways related to tumour metastasis revealed that NGFR could participate in the tumour metastasis-related signalling pathway; see Appendix Table 7, and therefore, we selected NGFR for the follow-up study.

Fig. 2.

Correlation analysis between genes and clinical characteristics of patients with BLBC. Pictures A to E refer to the differences in the expression of ERCC1 mRNA levels at different ages, M staging, N staging, T staging, and Stage staging, and the values above represent the magnitude of the P value between the two groups. Pictures F to J ditto, the difference in is refers to the difference of NGFR mRNA between different subgroups. T means tumour magnitude, N means lymph node metastasis, and M means distant metastasis. The sample size in this figure is 148 patients

Association of ERCC1 and NGFR with metastasis in patients with BLBC

A total of 148 patients with BLBC were included in this study, of whom 124 survived and 24 died. The clinical data of the included patients are detailed in Appendix Table 8. By comparing the expression levels of ERCC1 and NGFR under different clinical parameters, the findings indicated that the difference in expression between ERCC1 and NGFR in terms of metastasis in lymph nodes and distant metastasis was not statistically meaningful (P > 0.05), as shown in Fig. 2. The expression of NGFR in Stage III was higher than that in Stage II (P < 0.05), as shown in Fig. 2J.

Correlation analysis of ERCC1 and NGFR

NGFR was found to be positively correlated with ERCC1 in BLBC by Spearman correlation analysis (r_s = 0.24, P < 0.05), as shown in Fig. 3.

Fig. 3.

Correlation analysis of ERCC1 with NGFR. This figure refers to ERCC1 mRNA correlation with NGFR mRNA in the above 148 BLBC patients

Effect of ERCC1 and NGFR on the prognosis of BLBC

The optimal cutoff values for ERCC1 and NGFR expression were obtained using the RiskScore algorithm in the maxstat package of the R language software, which were 11.2841 and 4.7863, respectively. Based on the optimal cutoff values, the patients were divided into the ERCC1 high-expression/low-expression group and the NGFR high-expression/low-expression group, respectively, to perform the survival analysis of BLBC. The results of K–M survival analysis indicated that the BLBC patients in the ERCC1 highly expressed group had a much shorter overall survival than that in the low-expression group (P < 0.05) (Fig. 4A), as did the NGFR (see Fig. 4B).

Fig. 4.

K–M survival curves for overall survival in patients with BLBC. A refers to the results of the survival analysis of 148 BLBC patients divided into ERCC1 high and low expression groups according to the optimal cutoff value of ERCC1, with the ERCC1 high expression group in red and the low expression group in blue, which indicates that the survival of the low expression group is higher than that of the high expression group at the same risk. B ditto; the difference lies in the results of the analysis referring to the NGFR

Analysis of the effects of ERCC1 and NGFR on the prognosis of BLBC

Based on the optimal cutoff value, patients were divided into the ERCC1 high/low expression group and the NGFR high/low expression group, respectively, and one-way regression analysis was used to explore the effects of ERCC1, NGFR, and clinical characteristics on the prognosis of BLBC. The results of one-way Cox regression analysis showed that ERCC1, NGFR, N stage, and M stage were associated with the prognosis of BLBC patients (P < 0.05), as shown in Appendix Table 9. Multifactorial Cox regression analysis was performed with whether BLBC was survivable or not as the dependent variable and ERCC1, NGFR, N stage, and M stage, which were significant in one way, as the independent variables, with the assignment of values shown in Appendix Table 10. The results showed that N stage (HR = 4.170, 95% CI: 1.709 ~ 10.17, P < 0.05), ERCC1 (HR = 2.310, 95% CI: 1.008 ~ 5.293, P < 0.05), and NGFR (HR = 2.600, 95% CI: 1.126 ~ 6.002, P < 0.05) were basal-like independent risk factors for adverse outcomes in BLBC patients, as detailed in Appendix Table 11.

Establishment and validation of prognostic prediction models for BLBC patients based on ERCC1 and NGFR

To specifically quantify the survival probability of BLBC patients, a predictive model for predicting the survival rate of BLBC patients at 1, 3, and 5 years was constructed using the Cox method by integrating data on survival time, survival status, and 3 characteristics (ERCC1, NGFR, and N staging) using the R software (histograms), as shown in Fig. 5A. However, as the 1-year survival model was empirically proven to be not statistically significant (P > 0.05), it is, therefore, not shown in the figure. The modelled C index was 0.781 (95% CI 0.677–0.885, P < 0.001).Model calibration curves at 3 and 5 years are close to the standard line, indicating that the expected survival values in the model are in good agreement with the observed values (see Fig. 5B). Time-dependent ROC analyses showed AUC values of 0.76 and 0.81 for 3-year and 5-year survival, respectively, for this model (see Fig. 5C). The predictive model of the columnogram with the DCA decision curve demonstrated the clinical utility of this survival column line plot, showing greater net benefit over a range of mortality risks (see Fig. 5D). The C index, ROC curve, calibration curve and DCA judgement curve collectively demonstrated the accuracy and consistency of this model in forecasting the prognosis of BLBC patients. Next, we similarly validated this result with base-like breast cancer data from the METABRIC dataset; see Fig. 6. This demonstrates the extrapolatability of the prognostic prediction model we developed.

Fig. 5.

Prognostic prediction model for patients with basal-like BC alignment diagram. A Column line graph: each factor has a corresponding individual score on the corresponding axis in the graph, summing the scores of each factor gives the total score, the larger the total score, the lower the survival rate of patients at 3 and 5 years; B Calibration curve of the column line graph; C ROC curve of time-dependence of the column line graph; D Decision curve of DCA of the column line graph (*P < 0.05, **P < 0.01)

Fig. 6.

K–M survival curves for overall survival in patients with BLBC (METABRIC dataset). This figure refers to the results of the survival analysis of 83 BLBC patients screened in the METABRIC dataset divided into ERCC1/NGFR high and low expression groups, with the high expression group in red and the low expression group in blue, which suggests that the survival rate of the low expression group is higher than that of the high expression group at the same risk

Knockdown of ERCC1 inhibits migration and invasion of BLBC cells

To test the effect of ERCC1 on the migration and invasion ability of BLBC cells, ERCC1 knockdown lentiviral breast cancer stably transfected cell lines were first constructed and the knockdown effect was verified using RT-qPCR and WB.

The results indicated that the ERCC1-sh group showed decreased mRNA and protein expression levels compared with the blank and negative control groups (see Fig. 7A, B). It indicates that ERCC1 was successfully knocked down and can be used for subsequent experiments.

Fig. 7.

Knockdown of ERCC1 inhibits migration and invasion of BLBC cells. A Relative expression of ERCC1 mRNA in MDA-MB-231 and BT-549 cells infected with ERCC1 knockdown lentivirus; B Gel plots of ERCC1 protein expression and relative expression in MDA-MB-231 and BT-549 cells infected with ERCC1 knockdown lentivirus; C Representative graphs of each group of MDA-MB-231 cell migration assay and invasion assay taken at 40 ×; D Representative graphs of each group of BT-549 cell migration assay and invasion assay taken at 40 ×; E Statistically significant bar graphs of the difference in the number of migrating and invading BLBC cells in each group of MDA-MB-231 cells; F Statistically significant bar graphs of the difference in the number of migrating and invading BLBC cells in each group of BT-549 cells. ERCC1-sh is the ERCC1 knockdown group, and ERCC1-sh is the ERCC1 knockdown group. Low group, ERCC1-Ctrl is the ERCC1 negative control group, and Blank Control is the ERCC1 blank control group (*P < 0.05, ***P < 0.001)

Transwell assay demonstrated that compared with other groups, the number of cells migrating and invading in the ERCC1-sh group was significantly reduced (P < 0.05), indicating that knockdown of ERCC1 decreased the migratory and invasive abilities of both MDA-MB-231 and BT-549 cells (see Fig. 7C, D, E, F).

Overexpression of NGFR enhances migration and invasion of BLBC cells

To detect the effect of NGFR on the migration and invasion ability of BLBC cells, we constructed an NGFR overexpression plasmid and verified the overexpression effect using RT-qPCR and WB. As the results displayed, both mRNA levels and protein levels were higher in the NGFR-OE group compared with the other groups (P < 0.05); see Fig. 8A, B. The results indicated that NGFR could be successfully overexpressed and could be used for subsequent experiments.

Fig. 8.

Overexpression of NGFR enhances migration and invasion of BLBC cells. A Relative expression of NGFR mRNA in MDA-MB-231 and BT-549 cells infected with NGFR oe plasmid; B Gel plots of NGFR protein expression and relative expression in MDA-MB-231 and BT-549 cells infected with NGFR oe plasmid; C Representative graphs of each group of MDA-MB-231 cell migration assay and invasion assay taken at 40 ×; D Representative graphs of each group of BT-549 cell migration assay and invasion assay taken at 40 ×; E Statistically significant bar graphs of the difference in the number of migrating and invading BLBC cells in each group of MDA-MB-231 cells; F Statistically significant bar graphs of the difference in the number of migrating and invading BLBC cells in each group of BT-549 cells. NGFR-OE is the NGFR overexpression group. NGFR-Ctrl is the NGFR negative control group, and Blank Control is the NGFR blank control group (*P < 0.05, **P < 0.01, ***P < 0.001)

The transwell assay indicated a significant increase in the number of cells migrating and invading in the NGFR-OE group compared to the other groups (P < 0.05) (see Fig. 8C–F).The results indicated that overexpression of NGFR enhanced the migration and invasion ability of BLBC cells.

ERCC1 affects NGFR expression

To investigate whether ERCC1 affects the expression of NGFR, we detected changes in NGFR expression in MDA-MB-231 and BT-549 cells knocked down by RT-qPCR and WB assay. The results showed that when ERCC1 was knocked down, the mRNA and protein levels of NGFR were subsequently decreased (P < 0.05); see Fig. 9.

Fig. 9.

NGFR expression in BLBC cells with knockdown of ERCC1. A relative expression of NGFR mRNA in MDA-MB-231 and BT-549 cells with knockdown of ERCC1 in each group; B gel plots and relative expression of NGFR protein expression in MDA-MB-231 and BT-549 cells with knockdown of ERCC1 in each group. (*P < 0.05, ***P < 0.001)

ERCC1 affects the migration and invasion ability of BLBC cells through NGFR

In previous experiments, we clarified that ERCC1 and NGFR can affect BLBC cells'migration and invasion ability, and ERCC1 can affect NGFR expression. However, whether ERCC1 plays a role in the migration and invasion of BLBC cells through NGFR needs to be further verified. To this end, we transfected the NGFR overexpression plasmid in the constructed ERCC1 stable knockdown cell line, detected the protein expression of NGFR by Western Blot, determined that the transfection was successful (P < 0.05), see Fig. 10, and then detected the changes in the migratory and invasive ability of BLBC cells by a Transwell assay. Transwell results displayed that the number of cell migration and invasion was reduced in the ERCC1-sh + NGFR-Ctrl group compared with the ERCC1-sh + NGFR-Ctrl control group (P < 0.05). The quantity of cell migration and invasion increased in the ERCC1-sh + NGFR-OE group compared with the ERCC1-sh + NGFR-Ctrl control group, and the migration and invasion ability was enhanced (P < 0.05).

Fig. 10.

NGFR overexpression plasmid was transfected in the constructed ERCC1 stable knockout cell line. A MDA-MB-231 cell line gel plots and relative expression of NGFR protein in each group; B BT-549 cell line gel plots and relative expression of NGFR protein in each group (**P < 0.01, ***P < 0.001)

These results proved that knocking down ERCC1 could greatly reduce the migratory and invasive abilities of BLBC cells, and overexpression of NGFR on this basis restored their migratory and invasive abilities, thus demonstrating that ERCC1 could affect BLBC cells through NGFR (see Figs. 11 and 12).

Fig. 11.

ERCC1 affects the migration and invasion ability of MDA-MB-231 through NGFR. A Representative images of each group under the 40 × lens shown, verified by salvage assay that in MDA-MB-231 cells, ERCC1 can affect the migration ability through NGFR; B Representative images of each group under the 40 × lens shown, verified by salvage assay that in MDA-MB-231 cells, ERCC1 can affect the invasion ability through NGFR; C Histogram of the number of migrated MDA-MB-231 cells in each experimental group, with statistically significant differences between groups; D Histogram of the number of invaded MDA-MB-231 cells in each experimental group, with statistically significant differences between groups (**P < 0.01, ***P < 0.001)

Fig. 12.

ERCC1 affects the migration and invasion ability of BT-549 through NGFR. A Representative images of each group under the 40 × lens shown, verified by salvage assay that in BT-549 cells, ERCC1 can affect the migration ability through NGFR; B Representative images of each group under the 40 × lens shown, verified by salvage assay that in BT-549 cells, ERCC1 can affect the invasion ability through NGFR; C Histogram of the number of migrated BT-549 cells in each experimental group, with statistically significant differences between groups; D Histogram of the number of invaded BT-549 cells in each experimental group, with statistically significant differences between groups (**P < 0.01, ***P < 0.001)

Discussion

BC is a highly heterogeneous malignancy, with heterogeneity originating from different molecular typings, the refinement of which can be better applied to precise clinical treatment [38, 39]. BLBC, initially identified by gene expression profiles, is a typical subtype of BC [40]. Currently, clinical diagnosis and staging of BC often rely on immunohistochemistry, but the accuracy of immunohistochemistry test results can vary depending on technicians. With the increasing availability and lower cost of second-generation sequencing technology, target-guided, individualised genetic diagnosis and precision treatment of BC are expected to gradually enter clinical applications. Moreover, BLBC is known for its high resistance to drugs, elevated recurrence rates, invasive and metastatic tendencies, and greater malignancy compared to other subtypes [41], and there is a lack of precision target therapy in clinical settings. Using cellular experiments and population studies, this study investigated the metastatic and prognostic effects of ERCC1 and NGFR on BLBC. It has great significance for enhancing the accuracy of prognostic prediction and precise targeted therapy to obtain better efficacy and prognosis for patients.

ERCC1 is aberrantly expressed in various tumors and is also related to the progression of cancer. Studies have shown that enhanced ERCC1 expression can influence cancer cell proliferation and migration through the DNA repair pathway [42]. In oesophageal cancer, assessment of ERCC1 in oesophageal cancer cells can predict the invasiveness of oesophageal cancer [43]. Similarly, proteomic analysis in primary and metastatic foci of pancreatic tumors revealed that ERCC1 was elevated in metastatic lesions [44]. In esophageal squamous cell carcinoma, ERCC1 overexpression is associated with the late tumour stage and lymph-node metastasis, which are strongly correlated with cancer progression and outcome [45]. In BLBC, no direct relationship between ERCC1 and metastasis been previously established. However, the present study continued the survival analysis of BLBC patients and indicated that the BLBC patients in the ERCC1 high expression group had a shorter overall survival, and the multifactorial Cox even indicated that ERCC1 could be an independent risk factor for the poor prognosis of BLBC patients (HR = 2.310, 95% CI:1.008 ~ 5.293). In an in vitro cellular assay, it is identified that ERCC1 was expressed in MDA-MB-231 and BT-549 cells at a higher level than in normal mammary epithelial cells. Transwell migration/invasion assays indicated that knockdown of ERCC1 dramatically reduced the migration and invasion ability of both cancer cells. These results suggest that ERCC1 is closely associated with metastasis and the prognosis of BLBC, and may serve as a key driver gene contributing to its metastasis and the poor prognosis of BLBC.

NGFR is considered to be a receptor for all neurotrophins and has an extremely wide range of biological effects, including apoptosis, migration, differentiation, neurotrophic factor binding, and interactive transmembrane co-receptor expression [46, 47]. Moreover, studies have shown that NGFR can trigger diverse biological responses across various cancer types by activating distinct signaling pathways [48]. The direct relationship between NGFR and metastasis could not be analysed in the population study, but the expression of NGFR in Stage III was higher than that in Stage II, which indicated that NGFR might be related to the progression of BLBC. This might be attributed to NGFR expression activates cancer cell-related signalling pathways, which affects the invasive infiltration of the cancer cells, leading to the further deterioration of BLBC and the increased staging. In this study, we further conducted survival analysis of BLBC patients in the NGFR high-expression and low-expression groups. We found that the overall survival of BLBC patients in the NGFR high-expression group was shorter than that in the low-expression group; the results of multifactorial Cox regression analysis indicated that NGFR (HR = 2.600, 95% CI. 1.126 ~ 6.002) overexpression was an independent risk factor for poor outcome. In this study, we also found that when NGFR was overexpressed in MDA-MB-231 and BT-549 BLBC cells in an in vitro cellular assay, the number of migrations and invasions of both cancer cells was significantly increased compared with the control group. In ex vivo and in vivo studies of Kaposi's sarcoma, NGFR was found to enhance the autophagic process and influence tumor progression through the PI3K signalling pathway [49]. In gliomas, NGFR can also restrict signal transduction and affect the invasive and proliferative capacity of tumor cells [50]. It has been found that NGFR expression is upregulated in BC and can promote malignant progression and metastasis by activating the NF-κB signalling pathway, the PI3K/AKT signalling pathway, and its downstream targets [51–53]. In BLBC, NGFR may also be able to influence the migration and invasive ability of cancer cells through these pathways, thus giving patients a poor prognosis. These results suggest that NGFR enhances the migration and invasion ability of BLBC cells, serving as a risk factor for metastasis and poor prognosis of BLBC, which may have predictive value for poor prognosis and become a key target for prediction and treatment. Studies have shown that NGFR is a potential therapeutic target for many tumors [54], especially for immunotherapy [55]. In melanoma, drugs that inhibit NGFR restore tumor sensitivity to T-cell attack in vitro and in melanoma xenografts [56].NGFR also leads to down-regulation of NK cell-activating ligands, orchestrating immune control of antagonistic pathways to protect melanoma cells from NK cell clearance [57]. In head and neck squamous cell carcinoma, NGFR modulates macrophage polarization, and targeting NGFR may improve immunosuppression and enhance treatment efficacy and prognosis [58]. Similarly, in colorectal cancer, NGFR is associated with tumor immunity and drug sensitivity [59, 60] and is an immune-related gene. In uveal melanoma, an ultrasound-responsive nanoparticle can be combined with paclitaxel to inhibit cancer metastasis and recurrence by targeting NGFR via iron death-apoptosis [61].Although little research exists on the therapeutic role of targeting NGFR in BLBC, further studies are needed to determine the clinical utility of this disease. However, the various studies mentioned above have demonstrated that NGFR is a potential therapeutic target for many tumors, and we are informed about the future of NGFR-targeted therapy for BLBC.

In recent years, many complex cancer signalling pathways have been elucidated, among which those related to tumour metastasis are more likely to provide a basis for drug targets. It has been found that both NGFR and ERCC1 can participate in signalling pathways related to tumour metastasis, such as PI3K/AKT, MAPK and NF-κB. The NF-κB signalling pathway is a classical inflammatory pathway that can promote tumor metastasis by promoting the production of inflammatory factors such as TNF-α. In gliomas, NGFR can mediate NF-κB activation and influence tumor progression through the NGFR/NF-κB axis [62, 63]. These findings suggest that ERCC1 and NGFR can participate in the same signalling pathway, and there may be a certain regulatory mechanism between them. However, before exploring the mechanism in depth, it has to be clarified whether ERCC1 is related to NGFR. In this study, RNA-seq analysis indicated a potential association between ERCC1 and NGFR. In cellular experiments, we found that knockdown of ERCC1 in BLBC cell lines resulted in a decrease in mRNA and protein expression of NGFR. In a population study, Spearman correlation analysis also revealed a positive correlation between the expression of ERCC1 and the expression of NGFR in BLBCs. expression in BLBC. To further explore the association between ERCC1 and NGFR, transwell migration/invasion assays were performed next. The NGFR overexpression plasmid was added to the knockdown ERCC1 BLBC cells for rescue. Both cells exhibited a certain degree of migration and invasion ability compared with the control group. These results support that ERCC1 may influence the migration and invasion abilities of BLBC cells through NGFR.

This study demonstrated that ERCC1 and NGFR can serve as prognostic indicators in BLBC, and are associated with the migratory and invasive capacities of BLBC cells. Further studies also found a positive correlation between ERCC1 and NGFR, and that ERCC1 may have an effect on the migration and invasive ability of BLBC cells through NGFR. However, the specific mechanism by which ERCC1 and NGFR influence BLBC cell migration and invasion remains unclear. To facilitate the next step in delving into the specific mechanisms, we reviewed the relevant literature and found some insights. Some studies have shown that both ERCC1 and NGFR can affect cancer metastasis by regulating epithelial–mesenchymal transition (EMT)-related transcription factors [64, 65]. Many studies have found that EMT is inextricably linked to the MAPK signalling pathway. Consistently, our KEGG results show that NGFR is also rich in MAPK signalling. The MAPK signalling pathway is extremely common in tumours, and its aberrant activation is closely related to tumour proliferation, apoptosis and metastasis [65]. Given these findings, is it possible that ERCCI affects BLBC development through NGFR activation of the downstream EMT/MAPK signalling pathway? More interestingly, NGFR lacks intrinsic catalytic activity and tends to activate signalling cascades by forming oligomeric protein complexes with other proteins via dead structural domains. This revealed protein–protein interactions and how they coordinate the hierarchical activation of pathways downstream of non-catalytic receptors [66].Therefore, can ERCCI activate the EMT/MAPK signalling pathway by binding to the death domain of NGFR? This awaits more in-depth exploration and research in the future. Moreover, this study did not establish an animal model for in vitro validation. In the follow-up, this study will collect more representative population samples to demonstrate again, and further validate the conclusions of the initial exploration of this study with in vivo animal experiments.

Conclusion

ERCC1 and NGFR can influence the ability of BLBC cells to migrate and invade, and there is a positive correlation between these two genes. ERCC1 may have an effect on the migration and invasion abilities of BLBC cells through NGFR. ERCC1 and NGFR are independent risk factors for the prognosis of patients with BLBC, and they can serve as predictive markers for the 3-year and 5-year prognosis of BLBC.

Abbreviations

BC

Breast cancer

ERCC1

Excision repair cross-complementation group 1

NGFR

Nerve growth factor receptor

BLBC

Basal-like breast cancer

RNA-seq

RNA sequencing

K–M survival analysis

Kaplan–Meier survival analysis

RT-qPCR

Real-time quantitative reverse transcription PCR

EMT

Epithelial–mesenchymal transition

Author contributions

Yang Tang designed, guided the study; Fancai Zeng guided the study; Yuxi Lei, Xiabin Li and Jungfang He performed the experiments; Yuxi Lei, Fan Fan, Ruiwen Feng and Jing Zhong analyzed the data; Yuxi Lei wrote the paper. The authors read and approved the final manuscript.

Funding

This work was supported by Sichuan Science and Technology Program (2022NSFSC0741), Luzhou Science and Technology Program (2022-RCM-177) and the School-level Research Program of Southwest Medical University (2021ZKZD001).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

TCGA is part of a public database. The patients covered in the databases have received ethical approval. Users can download the data free of charge for research and publication. Our research is based on open-source data, so there are no ethical issues or other conflicts of interest.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.