Abstract
Background/Aim: The role of postoperative radiotherapy (RT) combined with chemotherapy (CT) for lymph node-positive (LN+) triple-negative breast cancer (TNBC) remains controversial. SUV39H1-mediated epigenetic regulation is associated with cancer cell migration, invasion, metastasis, and treatment resistance. This study aims to identify the role of SUV39H1 in TNBCs. Materials and Methods: Overall, 498 TNBCs with SUV39H1 RNA-seq profiles were retrieved from TCGA-BRCA and analyzed; the X-tile algorithm was used to stratify the population into low, intermediate, and high SUV39H1. Furthermore, we performed an in vitro clonogenic cell survival assay using the MDA-MB-231 cell line to assess the effects of SUV39H1 on cellular responses. Results: The results showed that SUV39H1 was significantly higher in TNBC than normal tissue and luminal subtype breast cancer. Notably, SUV39H1 is significantly expressed in the basal-like 1 (BL1) and immunomodulatory (IM) subgroups, compared to other subtypes. Compared to patients with a low or medium expression of SUV39H1, omitting RT only worsens disease-free survival (DFS) in those with high SUV39H1 expression. The experimental results showed SUV39H1 was suppressed by si-SUV39H1, and SUV39H1 knockdown in MDA-MB-231-IV2-1 cells enhanced the cellular toxicity of doxorubicin and paclitaxel. Conclusion: Targeting SUV39H1 may provide a potential guiding indication of omitting RT to avoid over-treatment and chemosensitivity for TNBC.
Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype and is characterized by the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression. Because these three receptors are lacking, TNBC is resistant to hormone therapy and HER2−targeted therapy, significantly reducing treatment options. Studies have shown that TNBCs have high local recurrence and distant metastasis rates (1, 2). TNBC also accounts for ~67% of all breast cancer deaths while present in 10-20% of cases only (3). The specific pathophysiology of TNBC remains poorly explored. Hence, studies are focused on identifying suitable targets to predict outcomes for clinical decision-making and enhance the effect of therapies on TNBC.
The standard care for TNBC is a multidisciplinary approach involving surgery, chemotherapy (CT), and radiotherapy (RT) (4). CT is generally administered to all TNBC patients with tumors >5 mm. Postoperative RT after breast-conserving surgery (BCS) reduces the local recurrence rate of TNBC patients, especially for patients with ≥4 positive axillary lymph nodes (LN) and a boost is beneficial to TNBC patients (5). Post-mastectomy radiation therapy (PMRT) is also recommended for these patients. However, the effectiveness of PMRT in patients with 1-3 positive lymph nodes remains controversial. These conflicting results show that novel biomarkers are needed to help oncologists avoid unnecessary radiation in patients unlikely to gain benefit.
Epigenetic modifications have been extensively studied in the mechanism of radioresistance of cancer cells. Emerging evidence indicates that dysregulation of DNA methylation and acetylation play pivotal roles in cell radioresistance (6). Kim et al. (7) found that aberrant methylation of the ataxia telangiectasia mutated kinase (ATM) promoter decreased the radioresistance of colorectal tumor cell in vitro. Furthermore, DNA methylation takes part in altering the mitotic checkpoint, cell cycle, DNA repair, and apoptosis, which mediate the resistance of cancer cells to ionizing radiation (7, 8). SUV39H1 is a crucial histone methyltransferase that methylates H3K9 (9). SUV39H1-mediated chromatin modification is not only associated with cell migration, invasion, and metastasis of breast cancer cells (10), but also the activation of ATM to initiate double-strand break (DSBs) repair by homologous recombination (HR) (11). Much hope is placed on SUV39H1, owing to its role in migration and invasion, as well as SUV39H1-mediated resistance to DSBs induced by cancer therapy.
Here, we conducted a comprehensive analysis of RNA sequencing data from The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) to identify the role of SUV39H1 in TNBC patients. In addition to survival, we aimed to elucidate the value of SUV39H1 in radiotherapy strategies for patients with TNBC via in vitro experiments using a TNBC cell line.
Materials and Methods
Data source. The TCGA-BRCA dataset was downloaded from the cBioPortal platform (12, 13). We included 498 patients with non-metastatic invasive ductal carcinoma breast cancer with RNA-seq expression data. Baseline characteristics including age at diagnosis, subtype, tumor stage, LN stage, pathological stage, and treatment were collected from the TCGA-BRCA database. The study population was divided into TNBC and non-TNBC (Her2, luminal A, and luminal B) subgroups. In addition, LN invasion status was classified into LN+ and LN− subgroup according to the LN stage. N0 is regarded as LN− subgroup, and N1 to N3 are regarded as LN+ subgroup. Disease-free survival (DFS) is the primary endpoint for this study. The DFS interval was tracked from the date of initial diagnosis date until the disease progressed or the study ended.
SUV39H1 expression and stratification. SUV39H1 levels in RNA-seq were determined by comparing the expression z-scores of tumor sample and adjacent normal samples in the study cohort. The log2-transformed differential expression profiles were normalized using a variation of the reads per kb/million (RPKM). SUV39H1 expression from RNA-seq in TNBC and non-TNBC grouped according to LN invasion status were illustrated using a boxplot. The differences between LN invasion status were estimated using Wilcoxon rank-sum test. Furthermore, SUV39H1 expression at the transcriptome level for normal, different subtypes, and different TNBC subgroups in breast cancer were compared using UALCAN (http://ualcan.path.uab.edu, accessed on June 19, 2021). The X-tile algorithm (14) was used to divide the population into low (<0.1), intermediate (0.1 to <0.4), and high (≥0.4) SUV39H1 strata, based on the RNA-seq expression data. The X-tile algorithm was used because it provides a more comprehensive evaluation of the strength and reliability of the association between gene expression and prognosis outcome than the typical ROC method.
Cell cultures and reagents. MDA-MB-231-IV2-1 was a kind gift of Prof. Lu-Hai Wang (China Medical University, Taichung, Taiwan, ROC) (15). MDA-MB-231-derived lung metastatic sublines, including IV2-1were cultured grown in Dulbecco’s Modified Eagle’s Medium (DMEM; ThermoFisher, Rockford, IL, USA) containing 10% fetal bovine serum (FBS, Hyclone Laboratories Inc., South Logan, UT, USA) and 1% penicillin-streptomycin at 37°C in 5% CO2. IV2-1 cells were transfected with specific siRNAs, including those against SUV39H1 (sc-38463, Santa Cruz Biotechnology, Santa Cruz, CA, USA), using Lipofectamine 2000 and Opti-MEM medium according to the standard protocol provided by Invitrogen (Carlsbad, CA, USA).
Colony formation assay. IV2-1 cells transfected with siRNA were seeded at 500 cells per 6-cm dish and exposed to Doxorubicin (25 nM) and Paclitaxel (3.125 nM). Cells were incubated for 14 days at 37°C to allow colonies to form. Colonies were stained with 2% crystal violet and counted. Colonies were defined as groups of 50 or more cells.
Statistical analysis. The baseline characteristics for each SUV39H1 strata were summarized as frequency and percentage. The age at diagnosis was summarized in median and interquartile ranges. Differences between subgroups were estimated using the Kruskal-Wallis rank sum test, Pearson’s chi-squared test, and Fisher’s exact test. The distribution of TNBC and non-TNBC in each SUV39H1 strata according to LN status were presented as frequency and percentage. The differences between TNBC subgroups were estimated using Pearson’s chi-squared test. The DFS rate of the study population was estimated using the Kaplan-Meier method. The survival differences between SUV39H1 strata and treatment subgroup were tested using the log-rank test. All p-values were two-tailed, and p-values <0.05 were considered statistically significant. All analyses were performed using R 4.0.5 software (R core team, Vienna, Austria, 2021) and GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA).
Results
Clinicopathological characteristics. X-tile analysis was used to establish the optimal cut-off values of SUV39H1 expression by the DFS (Figure 1A). The optimal cut-off ranges for SUV39H1 were −1.5 to 0.1, 0.1 to 0.4, and 0.4 to 2.8 for low (n=168), medium (n=169), and high (n=161) strata, respectively (Figure 1B). The clinicopathological characteristics of 498 patients from TCGA-BRCA data set were analyzed and shown in Table I. The distribution of baseline characteristics including subtypes (p<0.001), tumor size staging (p=0.018), pathological stage (p=0.034), and treatment characteristics (p=0.015), were significantly different between three SUV39H1 strata. Furthermore, each SUV39H1 strata were further divided into lymph node-negative (LN−) and lymph node-positive (LN+) subgroups, and the proportion of TNBC in each SUV39H1 strata in both LN− and LN+ subgroups were presented in Table II. Particularly, the proportion of TNBC was gradually increased from low (10.2%), medium (25.9%) and high (50.8%) SUV39H1 strata as shown in Table I.
X-tile algorithm stratification results for SUV39H1 expression from RNA-seq data. (A) X-tile plot indicates the association between SUV39H1 expression and disease-free survival (DFS). The brightness of the grid indicates the strength of association; red color indicates poorer DFS, while green color indicates better DFS. The optimal circled area indicates the optimal cutoff values for SUV39H1 stratification. (B) Histogram revealing the optimal cutoff values and number of patients of low, intermediate, and high strata of SUV39H1.
Baseline characteristics of study population according to SUV39H1 strata (n=498).
Distribution of triple-negative breast cancer (TNBC) and non-TNBC in each SUV39H1 strata among lymph node (LN)− and LN+ subgroup.
Similar findings were found in both LN− and LN+ subgroups, TNBC showed relative higher proportion in high SUV39H1 strata compared to intermediate and low SUV39H1 strata. Notably, high levels of SUV39H1 performed in TNBC, compared to other subtypes of breast cancer. The comparison of RNA-seq expression of SUV39H1 between LN− and LN+ was further compared in both the TNBC and non-TNBC cohort, as shown in Figure 2. In the non-TNBC cohort, the LN+ subgroup showed significantly higher SUV39H1 expression compared to the LN− subgroup (p=0.047). However, no significant difference was observed between LN− and LN+ subgroups in the TNBC cohort.
Boxplot of SUV39H1 expression from RNA-seq data among the triple-negative breast cancer (TNBC) and non-TNBC subgroup according to lymph node (LN) invasion status. p-Values are estimated using Wilcoxon rank-sum test.
High levels of SUV39H1 are present in TNBC patients. We further verified SUV39H1 expression levels in different breast cancer subtypes using the UALCAN database. As shown in Figure 3A, significantly higher levels of SUV39H1 were observed in TNBC compared with other subtypes of breast cancer, except the HER2 positive subtype. SUV39H1 was significantly expressed in the BL1, basal-like 2 (BL2), IM, luminal androgen receptor (LAR), mesenchymal (M), mesenchymal stem-like (MSL), and unspecified (UNS) subgroups of TNBC compared to the luminal subtype. Compared to the HER2−positive subtype, SUV39H1 was only significantly expressed in the BL1 and IM subgroups of TNBC (Figure 3B). Evidence indicates that the BL1 subtype is characterized by the enrichment of the DNA damage response. The observation of high level SUV39H1 in BL1 subgroup suggested that SUV39H1 might be associated with DNA repair ability.
UALCAN portal analysis of breast cancer samples from the TCGA database. (A) Comparison of SUV39H1 expression in the transcriptome level between normal and different subtypes of breast cancer samples. (B) Comparison of SUV39H1 expression in the transcriptome level between different triple-negative breast cancer (TNBC) subgroups. *p<0.05. p-Values were estimated using a two-tailed Student’s t-test.
The expression of SUV39H1 does not affect the outcome of TNBC patients. To verify the impact of the levels of SUV39H1 in the outcomes of TNBC patients, Kaplan–Meier survival curves were used to compare the DFS between different SUV39H1 expression strata. As shown in Figure 4, there was no statistical difference between low, medium, and high expression strata of SUV39H1 in both LN− and LN+ subgroups in TNBC patients. TNBC patients usually receive combined therapy including surgery, CT, and RT, which may abrogate the effect of SUV39H1 expression in DFS.
Kaplan-Meier plot for disease-free survival (DFS) of the (A) lymph node (LN)- subgroup and (B) LN+ subgroup in triple-negative breast cancer (TNBC) according to high-, medium-, and low-RNA-seq expression strata of SUV39H1. p-Values were estimated using the log-rank test.
Omitting RT only leads to worse disease-free survival in TNBC patients with high expression of SUV39H1. In order to verify the role of SUV39H1 in the TNBC cohort characterized as LN+, the survival analysis results of the study cohort were analyzed according to the three SUV39H1 strata and two treatment combinations (CT only: chemotherapy alone, CT+RT: chemotherapy combined with radiotherapy). As shown in Figure 5A, the CT-only subgroup showed a significantly worse DFS compared with the CT+RT subgroup in high RNA-seq expression strata of SUV39H1 (p=0.027). Notably, although the treatment subgroups in each medium or low strata of SUV39H1 showed similar expression with high strata of SUV29H1, no significant survival differences were observed between the treatment subgroups (Figure 5B, C). Accordingly, high SUV39H1 strata could exhibit significant poor DFS compared with low or medium expression strata of SUV39H1 in CT only patients (Figure 5D), while different expression strata of SUV39H1 displayed similar DFS outcome in CT+RT patients (Figure 5E). This data suggest that omitting RT does not affect the survival rate of TNBC patients with node-positive with medium or low levels of SUV39H1 expression. In short, these data indicate that the expression level of SUV39H1 may be used as a surrogate prognosis guide for whether RT can be omitted for LN+ TNBC patients after surgery and systemic treatment.
Kaplan–Meier plot for disease-free survival (DFS) of lymph node (LN)+ subgroup in triple-negative breast cancer (TNBC) among (A) high-RNA-seq expression strata, (B) medium-RNA-seq expression strata, and (C) low-RNA-seq expression strata of SUV39H1, according to treatment subgroup. DFS of LN+ subgroup in TNBC among (D) Chemotherapy (CT) only and (E) CT+radiotherapy (RT) treatment subgroup according to high-, medium-, and low-RNA-seq expression strata of SUV39H1. p-Values were estimated using the log-rank test.
Expression of SUV39H1 impacts survival after CT in TNBC cells in vitro. We hypothesize that high expression of SUV39H1 is able to overcome CT for TNBC, leading to failure of adjuvant therapy without RT. To verify this hypothesis, we performed an in vitro clonogenic cell survival assay to assess the effects of SUV39H1 on the cellular response. As shown in Figure 6A, the expression of SUV39H1 was suppressed by si-SUV39H1 in MDA-MB-231-IV2-1 cells. SUV39H1 knockdown in MDA-MB-231-IV2-1 cells enhanced the cellular toxicity of doxorubicin and paclitaxel (Figure 6B).
SUV39H1 depletion renders cells hypersensitive to doxorubicin and paclitaxel. (A) MDA-MB-231-IV2-1 cells were transfected with control or SUV39H1 siRNA and (B) exposed to doxorubicin and paclitaxel. Fourteen days later, cells were stained with crystal violet. Colonies containing more than 50 cells were counted. Each value represents the mean±standard deviation of three independent experiments. p-Values were estimated using a two-tailed Student’s t-test.
Discussion
SUV39H1 is a mammalian lysine methyltransferase, catalyzes the methylation of histone 3 lysine 9 (16). SUV39H1 forms the SUV39H1/H3K9me3/HP1 structure, which maintains chromatin compaction and inhibits transcription in the heterochromatin region containing inactive oncogenes. Loss of this structure during breast cancer carcinogenesis leads to oncogene transcription (17). Previous studies have also demonstrated that the inhibition of SPRY4 expression, a factor known to suppress migration and stem cell-related properties in breast carcinoma cells (18), can be achieved through the recruitment of SUV39H1 by lncRNA interaction. This interaction suggests that targeting PRC2 and SUV39H1 inhibition could potentially serve as therapeutic strategies in breast cancer treatment (19). As of now, the prognostic significance of SUV39H1 in TNBC largely remains uncertain. In the clinical setting, TNBC patients usually receive high-intensity combined therapy including surgery, CT, and RT. By analyzing a TNBC cohort in the present study, we identified SUV39H1 as a potential biomarker and its association with the outcome of TNBC patients receiving postoperative therapy (16). Data reported that patients with TNBC are more likely to have high expression levels of SUV39H1 than those without TNBC. Despite observing similar survival rates between patients with high expression levels of SUV39H1 and those with moderate-to-low expression levels in TNBC, there exists a noteworthy correlation between the SUV39H1 levels and the treatment received, predominantly either chemotherapy alone or high-intensity combined therapy involving chemotherapy and radiation. In this study, high expression level of SUV39H1 indicates that concurrent RT and CT have better results than CT alone. Compared with CT alone, concurrent RT and CT can improve the survival rate of patients with positive lymph nodes and high levels of SUV39H1 expression while omitting RT has similar survival outcomes.
Through the development of gene sequencing, TNBC could be classified into different subtypes by gene expression profiling: BL1, BL2, M, MSL, IM, and LAR (20). The clinicopathological features, treatment response, and prognosis vary with subtypes. Therefore, for the complex heterogeneity of TNBC, it should not be determined solely by TNM staging. It is reasonable to formulate treatment strategies through gene expression.
TNBC is likely to be more radioresistant than other breast cancer subtypes. Higher locoregional recurrence implies its radioresistance (21). Currently, the mechanisms underlying radioresistance in TNBC are not well understood. It is hypothesized that the radioresistance observed in TNBC may be attributed to the alteration of multiple genes. Published evidence has elucidated that gene alteration in TNBC can activate the PI3K/Akt pathway, resulting in resistance to ionizing radiation-induced DNA damage and inhibition of subsequent apoptosis (22). In this regard, maternal embryonic leucine zipper kinase (MELK) is a member of the snf1/AMPK family of serine/threonine kinases which has been shown to be enriched in TNBC compared with normal tissue and non-TNBC (23, 24). Inhibition of MELK RNA and protein expression significantly induces the radiation sensitivity in vitro and in vivo (24). miRNA also participates in regulation of radioresistance of TNBC. Studies showed that MiR-27a is significantly overexpressed TNBC cells, and downregulates the expression of CDC27, which causes increased radioresistance (25, 26).
Interestingly, integrated profiling data indicated that high levels of SUV39H1 detected in the BL1 subgroup suggest that SUV39H1 might associated with DNA repair ability. These findings are consistent with previous studies of SUV39H1 in DDR. ATM plays a vital role in the opening of the heterochromatin region within the HR pathway, thereby facilitating the recruitment of HR proteins and regulating end resection through the phosphorylation and activation of nuclease enzymes (27-30). ATM and ATM-dependent phosphorylation of DSB repair proteins is activated by Tip60 histone acetyltransferase (31-33). Activation of the Tip60 acetyltransferase requires histone H3K9me3 (11, 31, 34). SUV39H1 is rapidly recruited to DSBs and increases H3K9me3 at DSBs, formation of transient repressive chromatin to stabilize the damaged chromatin, and activation of Tip60’s acetyltransferase activity, promoting the subsequent acetylation of ATM (11). Cells lacking SUV39H1 show defective activation of both Tip60 and ATM, decreased HR repair efficiency, and increased radiosensitivity (11). Furthermore, Mo et al. revealed inhibition of SUV39H1 reduces HR repair and overexpression of SUV39H1 rescues HR repair suppressed by mTOR inhibitors in vitro (35). These results demonstrated that the crucial role of SUV39H1 plays in HR repair makes it a potential target for developing new radiosensitizers that inhibit the repair of DNA, increase the degree of radiation-induced DNA damage, and can overcome the inherent and acquired radioresistance of TNBC. HR repair is also essential for cancer cells to repair the DNA damage induced by widely used chemotherapy, including anthracycline and alkylating agents, for TNBC. Dysfunction of HR repair was associated with more pathologic complete response, longer failure-free interval, and better DFS in TNBC patients (36-38). It is rational that high expression of SUV39H1 restores HR repair and increases the resistance to those chemotherapy agents mostly prescribed in TNBC patients.
Radiotherapy is used for all TNBC patients after breast-conserving surgery and for high-risk patients after mastectomy. However, previous studies showed contradictory results in the survival benefits according to the status of pathological lymph nodes of TNBC after mastectomy. Some studies showed that there was no significant survival benefit for TNBC patients with positive lymph nodes (21, 39, 40), while some other studies showed that radiotherapy could significantly reduce recurrence and mortality of TNBC patients with positive lymph nodes (41-43). The controversial results suggest that the present indication of radiotherapy after mastectomy based on the pathological N stage might not be sufficient. Furthermore, a higher risk of cardiac morbidity and mortality was observed in breast cancer patients after RT (44). Therefore, identifying a prognostic marker in patients with RT after mastectomy has clinical benefits in personalized therapy.
As shown in Figure 5, medium to low expression of SUV39H1 in TNBC patients with positive lymph nodes did not show any significant difference between CT alone and CT coupled with RT for DFS. Our observations suggest that TNBC patients with HR repair restored by high expression of SUV39H1 are able to overcome the current CT regimen. Consistent with this observation, SUV39H1 silencing in MDA-MB-231-IV2-1 cells enhanced the chemosensitivity of doxorubicin and paclitaxel. Anthracycline and taxane-based chemotherapy is still the standard approach in TNBC treatment. Therefore, although there is some degree of radioresistance in TNBC, RT is not omittable in the adjuvant treatment of TNBC patients with high expression of SUV39H1.
We acknowledge the nature of this retrospective study that might limit the correction for all relevant prognostic variables. In addition, the low number of cases limits statistical power. The results of the current study should be interpreted carefully. However, the evidence of epigenetic regulation in radioresistance mainly came from in vitro and in vivo experiments, and the impact in the clinical practice is limited. Despite the aforementioned limitations, this study proposes a potential biomarker that could offer a chance to avoid unnecessary treatment and to ensure proper allocation of the healthcare resources. Whether to omit RT still needs further clinical trials.
Conclusion
With the progress in understanding TNBC, standard markers are not enough to account for the heterogeneity of TNBC. The expression of SUV39H1 functions as a potential biomarker that guides RT therapy omission in LN+ TNBC patients to avoid unnecessary treatment, and the in vitro experiments also showed that knockdown of SUV39H1 could enhance the cellular toxicity of doxorubicin and paclitaxel. Hence, the above evidence suggests that targeting SUV39H1 may provide a potential guiding indication of omitting RT to avoid overtreatment and chemosensitivity for TNBC.
Acknowledgements
We acknowledge the support from the following grants: 1) 111-2320-B-037-030, 111-2628-B-037-008, 111-2321-B-037-002, 112-2320-B-037-002, 112-2628-B-037-002, 112-2314-B-037-045, and 112-2314-B-037-044 from the Ministry of Science and Technology, Taiwan, ROC; 2) KMU-KI110001, KMU-TC109A03, and KMU-TC109B05 from the Kaohsiung Medical University Research Center; 3) 21020347-00111001 from the Kaohsiung Veterans General Hospital, Taiwan, ROC.
Footnotes
Conflicts of Interest
The Authors declare no conflicts of interest.
Authors’ Contributions
Wei-Lun Huang: Investigation, Formal analysis, Writing – original draft. Chi-Wen Luo: Investigation, Formal analysis. Huei-Shan Lin: Investigation. Chao-Ming Hung: Investigation. Fang-Ming Chen: Investigation. Sin-Hua Moi: Investigation, Formal analysis, Writing – review & editing. Mei-Ren Pan: Conceptualization, Supervision, Writing – review & editing.
- Received June 1, 2023.
- Revision received July 20, 2023.
- Accepted July 24, 2023.
- Copyright © 2023, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
