Epigenetic clock detected a breast cancer mitosis subtype with 1 improved immunotherapy 2

Epigenetic clock based on DNA methylation can estimate the epigenetic age of tissue and cell that can describe the process of aging. However, the exploration of diseases by 21 the epigenetic clock is still an uncharted territory. Our objective was to assess the role 22 of the epigenetic clock in breast cancer. In this study, DNA methylation data of breast tissue sample was download from TCGA 25 and GEO database. DNA methylation level of CpG sites and age of samples was 26 calculated by using pearson correlation test. Differentially expressed genes were 27 identified using the limma package and Kruskal-Wallis test was used for the difference 28 between cancer subtypes. Our model highlights that epigenetic age of breast cancer samples had an important impact on immunotherapy. We constructed a BEpiC web server (http://bio- where users submit DNA methylation data and age information to predict the epigenetic age of breast tissue and breast cancer subtypes.

of the epigenetic clock in breast cancer.

24
In this study, DNA methylation data of breast tissue sample was download from TCGA 25 and GEO database. DNA methylation level of CpG sites and age of samples was 26 calculated by using pearson correlation test. Differentially expressed genes were 27 identified using the limma package and Kruskal-Wallis test was used for the difference 28 between cancer subtypes. 29

30
We developed a workflow to construct the Breast Epigenetic Clock (BEpiC) that could 31 accurately predict the chronological age of normal breast tissue. Furthermore, the 32 BEpiC was applied to breast cancer to identify three breast cancer subtypes (including 33 development, homeostasis, and mitosis) by using the deviation between epigenetic age 34 and chronological age. Interestingly, the prognosis of the three breast cancer subtypes 35 is significantly different. In addition, the three breast cancer subtypes had distinct 36 differences in multiple immune cells and the mitosis subtype had the highest tumor 37 mutation burden that was used to estimate response to checkpoint inhibitors. human cancers(4). DNA methylation exhibits dynamic changes(5) and are more stable 54 than RNA during aging(6). Therefore, some researchers have built epigenetic clock that 55 can predict human epigenetic age based on methylation levels of multiple human tissues 56 and blood(7-11). However, DNA methylation pattern is specific in human tissues and 57 shows a negative correlation with aging in some tissues(12, 13). A more accurate 58 epigenetic clock needs to be developed for individual tissue. Epigenetic clock was also 59 used to predict the biological age based on the methylation level of multiple tissues in 60 mouse(14). Daniel et al. found that epigenetic clock could be used to evaluate 61 4 interventions that alter the rate of aging, such as calorie restricted diet(15). In addition, 62 the mitotic clock based on DNA methylation found that the epigenetic age acceleration 63 in normal tissues increases the risk of cancer (16,17). In brief, epigenetic clock could 64 measure aging of tissues that are independent of chronological age. 65 Cancer is the most common type of malignancy which is currently one of the leading 66 causes of death worldwide (18 were divided into three subtypes, which were samples with deviation between 144 epigenetic age and chronological age higher than 10.6 years, samples with deviation 145 between epigenetic age and chronological age lower than 10.6 years, and the remaining 146 samples. Up-regulated genes for each breast cancer subtype were screened in a "one vs.  subtypes based on the deviation between epigenetic age and chronological age ( Figure   243 4A). A sample whose epigenetic age was 10.6 years (twice the MAE of BEpiC in 244 normal samples) older than chronological age was defined as epigenetic age 245 acceleration. Then, a sample whose epigenetic age was 10.6 years younger than  Figure 3B). 260 We compared the overall survival and disease-free survival of breast cancer patients 261 with three subtypes ( Figure 4C). Patients of these three subtypes have significant 262 differences in overall survival (p=0.00047) and disease-free survival (Supplementary 263 Figure 3D). Patients with development subtype had the best prognosis and patients with 264 mitosis subtype had the worst prognosis. We inferred that breast cancer patients whose in T4, N3, stage iii-iv was higher than the other two subtypes (Supplementary Table 2).

295
It also demonstrated that most breast cancer patients with mitosis subtype were 296 advanced cancer patients with poor prognosis. 297 We analyzed the relationship between PAM50 signature and BEpiC and found that 298 Basal subtypes has a large overlap with mitosis subtypes ( Figure 5F). This indicates 299 that the epigenetic age of breast cancer tissues was decelerate in the Basal subtypes, so 300 the prognosis of Basal subtypes patients was worse. Immunotherapy has a better therapeutic effect for mitotic subtype 303 We focus on the treatment of patients with three subtypes, and whether the patients with 304 three subtypes are different in terms of medication. We compared the prognosis of three 305 breast cancer subtypes taking tamoxifen as a commonly used drug in the treatment of 306 breast cancer (Supplementary Figure 5). There was no significant difference in 307 prognosis of the three subtypes taking tamoxifen (p=0.57).

308
Since the patients of the three subtypes were significantly different at the molecular