H4K20me3, H3K4me2 and H3K9me2 mediate the effect of ER on prognosis in breast cancer

ABSTRACT

Previous studies have indicated that histone methylations act as mediators in the relationship between oestrogen receptor (ER) and breast cancer prognosis, yet the mediating role has never been assessed. Therefore, we investigated seven histone methylations (H3K4me2, H3K4me3, H3K9me1, H3K9me2, H3K9me3, H3K27me3 and H4K20me3) to determine whether they mediate the prognostic impact of ER on breast cancer. Tissue microarrays were constructed from 1045 primary invasive breast tumours, and the expressions of histone methylations were examined by immunohistochemistry. Multifactorial logistic regression was used to analyse the associations between ER and histone methylations. Cox proportional hazard model was performed to assess the relationship between histone methylations and breast cancer prognosis. The mediation effects of histone methylations were evaluated by model-based causal mediation analysis. High expressions of H3K9me1, H3K9me2, H3K4me2, H3K27me3, H4K20me3 were associated with ER positivity, while high expression of H3K9me3 was associated ER negativity. Higher H3K9me2, H3K4me2 and H4K20me3 levels were associated with better prognosis. The association between ER and breast cancer prognosis was most strongly mediated by H4K20me3 (29.07% for OS; 22.42% for PFS), followed by H3K4me2 (11.5% for OS; 10.82% for PFS) and least by H3K9me2 (9.35% for OS; 7.34% for PFS). H4K20me3, H3K4me2 and H3K9me2 mediated the relationship between ER and breast cancer prognosis, which would help to further elucidate the impact of ER on breast cancer prognosis from an epigenetic perspective and provide new ideas for breast cancer treatment.

KEYWORDS:

• Causal mediation analysis
• ER
• histone methylation
• prognosis

Introduction

Breast cancer is known as a hormone dependent malignant tumour. Currently, oestrogen receptor (ER) status is widely applied to provide prognostic information and guide treatment strategies [1]. However, the underlying molecular mechanisms of ER in relation to breast cancer prognosis has not been fully disentangled [2,3].

Oestrogen receptors are a subfamily of nuclear receptors that control cellular responses to oestrogens [4]. Upon oestrogen stimulation, ERα, the dominant form of ER expressed in breast, recruits a number of coregulators to oestrogen response elements to modulate gene activation or repression [5]. The coregulator complexes contain histone methyltransferases (HMTs) or histone demethylases (HDMs), which play a role in regulating histone methylation modifications on ERα target genes [6]. For example, lysine-specific demethylase-1 (LSD1) is co-expressed and co-localize with ER to regulate the demethylation of H3K4 and H3K9 [7]; enhancer of zeste (EZH) 2 combines with ER to catalyse the methylation of H3K27me3 [8].

Methylation modifications at different loci of histones have different biological effects, intricately modulating chromatin structure and gene expression. H3K4 and H3K36 methylation, typically occurring in transcriptionally active genomic regions, is associated with gene activation, marking genomic areas engaged in transcriptional initiation and elongation [9,10]. H3K9me3/2 and H3K27me3, are predominantly found in regions of transcriptional silence, such as transposons and inactive genes, serving a repressive role in gene expression [11]. Specifically, H3K9me3 is enriched in heterochromatin and H3K9me2 is usually involved in gene silencing in euchromatin [12]. The regulatory roles of these methylation marks extend beyond transcription to include a crucial role in regulating nucleosome movement, DNA repair processes and replication fidelity [13]. Dysregulation of these methylation patterns disrupts these critical cellular processes, contributing to the pathogenesis and progression of various malignancies, including breast cancer, by altering the expression and function of key oncogenes and tumour suppressor genes [14–16]. This intricate balance of histone methylation underscores its potential as a mediator in the complex interplay between ER signalling and breast cancer prognosis, offering insights into the epigenetic mechanisms driving cancer progression and therapeutic response.

In this context, histone methylations may act as an intermediary mechanism through which ER signalling influences the prognosis of breast cancer. However, the mediating role of histone methylations in ER to breast cancer prognosis has never been explored. In the present study, therefore, we investigated whether the impact of ER on breast cancer prognosis can be mediated by seven histone methylations (H3K4me2 [7,14], H3K4me3 [17,18], H3K9me1 [19,20], H3K9me2 [21,22], H3K9me3 [23,24], H3K27me3 [25,26] and H4K20me3 [27,28]), which were found to associated with both ER and breast cancer prognosis previously. This study aims to elucidate the mediating role of specific histone methylations in the association between ER status and breast cancer prognosis, thereby uncovering potential epigenetic mechanisms that may influence disease progression and response to therapy.

Material and methods

Study population

A total of 1063 patients with pathologically diagnosed primary invasive breast cancer and tumour sizes exceeding 1 cm in diameter were recruited between January 2008 and December 2015 from the Cancer Center of Sun Yat-sen University in Guangzhou, China. Most of the included patients (N = 1045) were successfully followed up until 31 December 2022. The study was approved by the Ethics Committee of the School of Public Health, Sun Yat-sen University. Written informed consent was obtained from all participants.

Baseline data collection

Demographic characteristics, including age and menopausal status, were collected at diagnosis by trained investigators using a structured questionnaire. Body mass index (BMI) and clinicopathological characteristics including clinical stage, histological grade, ER, progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2) status and proliferation index factor Ki-67 (Ki-67) were collected from medical records. Detailed definitions of ER, PR, and HER2 have been described previously [29].

Construction of tissue microarray (TMA)

Formalin-fixed and paraffin-embedded tissues were obtained from enrolled patients. Hematoxylin and eosin (HE)-stained sections of tissue samples were reviewed by two experienced pathologists, followed by re-sectioning and re-staining with HE. Representative tumour tissue regions and adjacent normal tissue regions (if available) were marked on the restained HE sections. From the marked regions, two tumour tissue cylinders and one adjacent normal tissue cylinder (if not available, it would be replaced by the tumour tissue) with a diameter of 1 mm were punched from the corresponding paraffin block as donor blocks and placed in the TMA paraffin block using an automated tissue arrayer (MiniCore®, Mitogen, UK). The layout of the cores was advanced in TMA Designer 2 software. Sections of 4 μm were cut from the TMA blocks, pasted onto the coded glass slides and then placed in an oven at 65°C for 30 min. The tissue surface was sealed with paraffin.

Immunohistochemistry (IHC)

The TMAs were subjected to a series of preparatory procedures, which included baking at 60°C for 2 hours, dewaxing with xylene, and rehydration in graded ethanol followed by distilled water immersion for 10 minutes. Antigen retrieval was carried out using EDTA buffer (pH 9.0) in a superheated pressure cooker. Endogenous peroxidase activity was blocked using 3% H₂O₂. Subsequently, histone methylation monoclonal antibodies (dilution concentrations in the Supplementary Table S1) were added dropwise to the slides separately and incubated at a constant temperature.

Detection was achieved using the EnVision Detection System (Peroxidase/DAB, Rabbit/Mouse) (Dako K5007). Visualization was facilitated by diaminobenzidine (DAB) staining and counterstaining with haematoxylin. Finally, the slides were dehydrated and mounted for analysis.

IHC stained sections were digitally imaged using the Pannoramic scanner and CaseViewer software. An experienced pathologist, who was blinded to the clinical data, assessed the IHC staining. Staining intensity was graded on a scale of 0 to 3 (0 = no staining, 1 = weak, 2 = moderate, 3 = strong), and the percentage of tumour cell nuclear staining was quantified (ranging from 0% to 100%). The H score was computed by multiplying the staining intensity by the percentage of tumour cell nuclear staining, resulting in a range from 0 to 300. To maintain consistency, the average H score from duplicate cores was determined for each case.

Follow‑up and outcomes

Patients were followed up by telephone calls or out-patient visits every 3 months in the first year, every 6 months in the second and third year, and annually thereafter. Overall survival (OS) and progression-free survival (PFS) were regarded as endpoints of the study. OS was defined as the time from diagnosis to death from any cause. PFS was determined as the duration from diagnosis to either disease progression, including recurrence, metastasis, or death from any cause. Survival status was censored at the last follow-up date or 31 December 2022.

Statistical analysis

The H scores of histone methylations, initially quantified on a continuous scale, were transformed into categorical variables to facilitate analysis and interpretation. To determine the optimal cut-off values for categorizing these histone methylations, we used the X-tile 3.6.1 software from Yale University, New Haven, CT, USA [30]. This software employs a robust algorithm to analyse biomarker data, identifying the cut-off point that maximizes the statistical difference in outcomes between groups. The process began with the software examining the entire range of H scores for each histone methylation and tentatively dividing the patient cohort based on every possible cut-off value. For each division, the software performed log-rank tests to compare PFS between the two groups defined by the cut-off. The cut-off value that resulted in the minimum P-value from these log-rank chi-square statistics was selected as the optimal threshold. This methodological approach ensures that the chosen cut-offs are not arbitrary but are statistically validated to best differentiate patient outcomes based on the levels of histone methylations. Univariate survival analyses were carried out according to Kaplan-Meier method. Associations of ER with histone methylations were evaluated using multifactorial logistic regression. Cox proportional hazard model was performed to investigate the associations between histone methylations and breast cancer prognosis, adjusting for age at diagnosis, clinical stage, ER status, HER2 and histological grade.

The model-based causal mediation analysis was performed using a hypothetical model (Figure 1). The average direct effect (ADE) represented the effect of ER on prognosis that was independent of histone methylations. The average causal mediation effect (ACME) was defined as the effect of ER on prognosis mediated by histone methylations. To quantify the magnitude of mediation, the proportion of the association mediated by histone methylations was estimated by ACME/[ADE + ACME]. The model-based causal mediation analysis was conducted by R package: mediation [31]. All analyses were performed using R 4.2.1, with a two‐sided significance level of P below 0.05.

Figure 1. Model of histone methylations mediating the association between ER and breast cancer prognosis.

ACME: average causal mediation effect; ADE: average direct effect.

Results

Characteristics at baseline and their associations with breast cancer prognosis

As shown in Table 1, the median age of the study subjects at diagnosis was 48 (interquartile range: 41–56) years old. The majority of the women were premenopausal (59.0%), with an educational level below secondary school (56.0%) and 51.8% had a BMI below 23.0. Most women were diagnosed with early clinical stage (stage I/II: 70.0%), ER-positive (73.6%), PR-positive (72.7%), HER2-negative (66.9%) or low histological grade (grade I/II: 73.3%). During a median follow-up of 96.4 months, a total of 265 patients experienced breast cancer progression and 181 died. Significantly, the result of Kaplan-Meier test showed that ER status, clinical stage, tumour size and nodal status were associated with both overall mortality and disease progression.

Table 1. Characteristics at baseline and the associations with prognosis of 1045 breast cancer patients.

		Overall mortality		Disease progression
Characteristics	Total^a (%)	Events	HR (95%CI)	Events		HR (95%CI)
Age at diagnosis	Continuous (Median, IQR) 48 (41–56)
≤40	244 (23.4)	43	1.15 (0.81–1.64)	74		1.44 (1.09–1.90)
41–60	677 (65.0)	110	1.00	158		1.00
≥61	121 (11.6)	28	1.38 (0.91–2.09)	33		1.14 (0.78–1.66)
Menopausal status
Pre-menopausal	585 (59.0)	88	1.00	135		1.00
Post-menopausal	406 (41.0)	84	1.35 (1.00–1.82)	116		1.22 (0.96–1.57)
Educational level
Below junior school	290 (56.0)	41	1.00	64		1.00
Senior high school	113 (21.8)	23	1.24 (0.74–2.06)	35		1.35 (0.89–2.03)
College or above	115 (22.2)	14	0.78 (0.42–1.43)	30		1.14 (0.74–1.76)
Parity
0	38 (3.8)	8	1.00	13		1.00
≥1	964 (96.2)	163	0.73 (0.36–1.48)	241		0.65 (0.37–1.13)
BMI (kg/m^2)
<23	511 (51.8)	96	1.00	137		1.00
23–24.9	229 (23.2)	32	0.74 (0.49–1.10)	52		0.82 (0.59–1.13)
≥25	246 (24.6)	43	0.94 (0.65–1.34)	64		0.94 (0.70–1.27)
Clinical stage
I/II	706 (70.0)	86	1.00	137		1.00
III/IV	303 (30.0)	89	2.99 (2.22–4.02)	118		2.50 (1.95–3.20)
ER status
Negative	263 (26.4)	56	1.00	78		1.00
Positive	733 (73.6)	115	0.67 (0.48–0.92)	175		0.73 (0.56–0.95)
PR status
Negative	271 (27.3)	54	1.00	73		1.00
Positive	723 (72.7)	117	0.72 (0.52–1.00)	179		0.83 (0.63–1.09)
HER2 status
Negative	697 (66.9)	118	1.00	180		1.00
Equivocal	90 (8.6)	20	1.27 (0.79–2.04)	26		1.13 (0.75–1.70)
Positive	255 (24.5)	43	1.05 (0.74–1.49)	59		0.93 (0.69–1.25)
Ki-67
≤14%	232 (29.6)	28	1.00	47		1.00
>14%	553 (70.4)	96	1.48 (0.97–2.26)	148		1.38 (0.99–1.92)
Histological grade
I/II	696 (73.3)	116	1.00	172		1.00
III	253 (26.7)	52	1.37 (0.99–1.91)	67		1.17 (0.88–1.56)
Tumor size (cm)
≤2	299 (29.9)	30	1.00	52		1.00
>2	702 (70.1)	142	2.14 (1.44–3.17)	198		1.72 (1.27–2.34)
Nodal status
Negative	442 (44.2)	46	1.00	72		1.00
Positive	559 (44.8)	125	2.48 (1.77–3.48)	177		2.30 (1.75–3.02)

^aInconsistency in the total number of cases is due to missing.

IQR: interquartile range. Bold indicates statistically significant values.

Associations between ER and histone methylations

After adjusting for potential confounders, the results of multivariate logistic regression model showed that ER status was significantly positively associated with the expressions of H3K9me1, H3K9me2, H3K4me2, H3K27me3, and H4K20me3, whereas ER status was negatively associated with the expression of H3K9me3. The relationship between ER and H3K4me3 was not statistically significant (Table 2).

Table 2. The associations between ER and histone methylations.

Histone methylations	OR (95%CI)	P
H3K9me1	2.41 (1.68–3.46)	<0.001
H3K9me2	1.82 (1.15–2.81)	0.011
H3K9me3	0.70 (0.49–1.00)	0.048
H3K4me2	1.57 (1.13–2.20)	0.008
H3K4me3	1.04 (0.74–1.45)	0.819
H3K27me3	2.81 (1.99–3.98)	<0.001
H4K20me3	2.43 (1.72–3.43)	<0.001

ER negative was the reference group. The logistic regression model adjusted age at diagnosis, ER status, HER2 status, clinical stage, and histologic grade. Bold indicates statistically significant values.

Associations between histone methylations and breast cancer prognosis

The univariate analysis showed that high expressions of almost all histone methylations were significantly associated with a favourable prognosis for both OS and PFS. After adjusting for confounders, the significance remained for H3K9me2, H3K4me2 and H4K20me3 (Table 3).

Table 3. Associations between histone methylations and breast cancer prognosis.

		Overall mortality			Disease progression
Histone methylations	Total (%)	Events	HR (95%CI)	HR (95%CI)^a	Events	HR (95%CI)	HR (95%CI)^a
H3K9me1
≤3.3	278 (26.8)	58	1.00	1.00	87	1.00	1.00
>3.3	761 (73.2)	122	0.70 (0.51–0.96)	0.96 (0.67–1.36)	176	0.68 (0.53–0.88)	0.87 (0.65–1.16)
H3K9me2
≤11.7	143 (14.4)	37	1.00	1.00	51	1.00	1.00
>11.7	850 (85.6)	131	0.54 (0.37–0.77)	0.54 (0.36–0.80)	198	0.58 (0.45–0.74)	0.63 (0.45–0.89)
H3K9me3
≤240	753 (72.9)	143	1.00	1.00	213	1.00	1.00
>240	280 (27.1)	37	0.75 (0.52–1.08)	0.90 (0.61–1.33)	51	0.66 (0.48–0.89)	0.76 (0.54–1.06)
H3K4me2
≤118	482 (46.4)	110	1.00	1.00	155	1.00	1.00
>118	556 (53.6)	70	0.65 (0.48–0.87)	0.63 (0.46–0.88)	109	0.58 (0.45–0.74)	0.66 (0.50–0.86)
H3K4me3
≤275	518 (49.8)	103	1.00	1.00	149	1.00	1.00
>275	522 (50.2)	77	0.65 (0.48–0.87)	0.75 (0.54–1.04)	115	0.68 (0.53–0.87)	0.78 (0.59–1.02)
H3K27me3
≤183.3	329 (31.7)	74	1.00	1.00	104	1.00	1.00
>183.3	710 (68.3)	106	0.63 (0.47–0.84)	0.81 (0.58–1.14)	160	0.66 (0.51–0.84)	0.81(0.61–1.08)
H4K20me3
≤75	494 (47.5)	116	1.00	1.00	157	1.00	1.00
>75	546 (52.5)	64	0.47 (0.34–0.64)	0.58 (0.41–0.81)	107	0.68 (0.53–0.87)	0.66 (0.50–0.88)

^aAdjusting for age at diagnosis, ER status, HER2 status, clinical stage, histologic grade.

Bold indicates statistically significant values.

Mediation effects of histone methylations on the relationship between ER and breast cancer prognosis

The results of mediation analysis showed the effects in different pathways, including ADE, ACME, and total effects. The ACMEs of H3K9me2, H3K4me2, and H4K20me3 were found to be statistically significant for both OS and PFS, suggesting that these histone methylations mediated the association between ER and breast cancer prognosis. The proportions of mediation of H3K9me2, H3K4me2, and H4K20me3 were 9.35%, 11.5%, and 29.07% for OS, and 7.34%, 10.82%, and 22.42% for PFS, respectively (Table 4).

Table 4. Mediation analysis^a of histone methylations between ER and breast cancer prognosis.

Histone methylations	ACME^b				ADE^c		Total effect^d			Percentage^e (%)
	β	95%CI	P	β	95%CI	P	β	95%CI	P
OS
H3K9me1	−0.63	−24.48–20.26	0.992	144.61	12.55–287.41	0.032	143.98	17.99–285.48	0.032	0.43
H3K9me2	15.00	0.74–35.89	0.030	145.48	0.47–310.50	0.048	160.48	14.34–333.00	0.028	9.35
H3K9me3	−2.45	−18.85–8.68	0.700	137.13	12.45–279.88	0.032	134.68	10.14–276.02	0.032	1.76
H3K4me2	16.80	0.51–42.90	0.036	129.71	−5.62–277.94	0.072	146.51	8.59–298.92	0.038	11.5
H3K4me3	0.85	−8.71–12.71	0.868	136.57	8.37–278.92	0.032	137.41	9.77–278.61	0.034	0.62
H3K27me3	14.76	−11.40–42.44	0.238	123.17	−3.07–267.87	0.058	137.93	9.58–278.56	0.038	10.70
H4K20me3	39.86	10.39–80.71	<0.001	97.25	−47.72–248.21	0.176	137.11	−12.12–288.17	0.062	29.07
PFS
H3K9me1	10.17	−12.58–34.70	0.348	161.19	20.43–304.77	0.022	171.36	38.36–315.25	0.010	5.93
H3K9me2	13.43	0.48–32.48	0.038	169.55	26.38–316.27	0.014	182.98	39.95–338.66	0.010	7.34
H3K9me3	−8.29	−28.76–3.12	0.230	171.85	31.19–318.73	0.014	163.56	22.44–313.57	0.028	4.60
H3K4me2	19.01	0.87–16.14	0.038	156.71	262.02–300.13	0.020	175.72	42.86–328.43	0.008	10.82
H3K4me3	0.94	−10.27–13.68	0.842	169.94	41.55–315.81	0.014	170.88	43.43–319.61	0.012	0.55
H3K27me3	17.67	−9.09–48.17	0.182	145.05	5.18–281.77	0.044	162.72	21.72–296.53	0.016	10.86
H4K20me3	36.35	7.54–77.92	0.010	125.77	2.34–259.15	0.048	162.12	42.21–305.70	0.014	22.42

^aThe CIs in brackets were 95% CIs from 1000 nonparametric bootstrap replicates. The analysis was adjusted for age at diagnosis, HER2, clinical stage, histologic grade.

^bACME, average causal mediation effect.

^cADE, average direct effect.

^dTotal effect, represents the sum of ACME and ADE.

^ePercentage describes the proportion of the effect of ER on the prognosis mediated by histone methylation; Percentage=ACME/[ADE + ACME].

OS: overall survival; PFS: progression-free survival.

Bold indicates statistically significant values.

Discussion

In this investigation of 1045 breast cancer patients, our analysis revealed the mediation effects of histone methylations between ER and breast cancer prognosis. Notably, the association between ER and breast cancer prognosis was most strongly mediated by H4K20me3 (29.07% for OS; 22.42% for PFS), followed by H3K4me2 (11.5% for OS; 10.82% for PFS) and least by H3K9me2 (9.35% for OS; 7.34% for PFS).

Previous studies have found that these histone methylations were associated with both ER status and breast cancer prognosis, but they have not further explored the role of histone methylations in this association [27,32]. Except for mediating factors, histone methylations may also be confounder factors in the relationship between ER and breast cancer prognosis [33]. However, numerous studies have demonstrated histone methylation modifications in the pathway of ER affecting breast cancer prognosis, indicating that histone methylations are more likely to act as mediators. For example, U-H Park et al. [19] found that ER combined with additional sex comb-like 2 (ASXL2) erased H3K9me2 and H3K27me3 markers on target genes, which promoted the expressions of oncogenes such as NCOA3, BMP7, CA4, and RSP6KB1; Mark et al. [34] found that KDM3A regulated the expressions of a number of ER-target genes involved in breast cancer progression and proliferation by controlling demethylation of H3K9me1/me2. In addition, adjusting confounders in our analysis aims to isolate the effect of ER status on breast cancer prognosis by controlling for other variables that might influence the outcome. Conversely, when we adjusted for mediating variables, such as specific histone methylation patterns, we aimed to assess how much of the ER effect on prognosis was channelled through these epigenetic modifications [35]. Our findings indicated a nuanced interaction where the protective effect of positive ER status against adverse breast cancer outcomes was modulated to varying degrees by the presence of certain histone methylations (Supplement Table 2). Hence, histone methylations are more likely to be mediating factors between ER status and breast cancer prognosis.

A better prognosis for ER-positive breast cancer can be achieved through two mechanisms. One is the effects of endocrine therapy in ER-positive breast cancer [36], and the other is ER’s modulation of the expression of target genes to inhibit breast cancer cell proliferation and promote apoptosis [37]. Correspondingly, the mediation effects of histone methylations also occurred through two pathways. The first pathway is that oestrogen signalling mediates histone methylation alterations, thereby affecting the proliferation and migration of breast cancer cells. For example, oestrogen signalling regulates the demethylation of H3K9me2, leading to the activation of genes involved in cell survival, such as bcl-2 and pS2 [38]. The second pathway is that oestrogen signalling mediates histone methylation alterations, thereby influencing the responsiveness of breast cancer cells to endocrine therapy. For example, oestrogen signalling facilitates the demethylation of H3K9me3/me2 to control the expression of the FOXA1 gene, which is crucial for oestrogen-ER activity and endocrine response in breast cancer cells [39,40]. ERα participates in catalysing the demethylation of H3K4, which is important for maintaining the oestrogen-dependent growth of breast cancer cells [41,42]. In addition, the specific mechanism by which H4K20me3 mediates the prognostic impact of ER in breast cancer remains largely unexplored and requires further investigation.

While our study did not directly investigate treatment outcomes, the independent association of specific histone modifications with breast cancer prognosis suggests a potential link to therapy responsiveness. Histone methylation patterns have been implicated in the modulation of gene expression related to drug metabolism and efflux [43,44], providing a plausible biological basis for their influence on treatment efficacy. Further research is needed to explore this connection and its implications for overcoming drug resistance in ER-positive breast cancer.

Our study had several limitations. Firstly, we only measured global levels of histone methylations rather than gene-specific enrichment, which would make the relationship more precise. Nevertheless, it’s significant that our study provided feasible ideas and directions for further research. Secondly, only patients with tumours >1 cm were included, which may lead to selective bias. However, the expressions of histone methylations in this study were independent of tumour size (Supplement Table 3), causing non-differential bias on the associations between histone methylations and prognosis. Finally, our study did not include specific treatment information, which is relevant to patient outcomes. However, we have attempted to mitigate this limitation by adjusting for clinicopathological characteristics in our analysis, as these factors largely guide therapy decisions according to clinical guidelines, thus potentially controlling for some confounding effects of treatment. Future studies incorporating comprehensive treatment information are essential to fully understand the implications of our findings.

Conclusion

In conclusion, this study firstly found that the association between ER and breast cancer was mediated by H4K20me3, H3K4me2 and H3K9me2, which may help to further elucidate the impact of ER on breast cancer prognosis from an epigenetic perspective and provide new ideas for breast cancer treatment.

Acknowledgments

We sincerely thank the patients who participated in this study, the staff who conducted the baseline and the follow up data collection, and the medical staff in the breast departments of the Third Affiliated Hospital, and the Cancer Center of Sun Yat-Sen University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15592294.2024.2343593

Additional information

Funding

This study was funded by the National Natural Science Foundation of China [81973115]. The founders have no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.