Upregulated CCNE1 Correlates with Poor Prognosis, Tumor Immune Inltration and Escape in Breast Cancer

Background Breast cancer (BC) is the most common malignant tumor in women and widely known for its poor prognosis. More and more research has discovered that cyclin E1 (CCNE1) plays an important role in progression of various types of cancer. But its specic mechanism in BC progression still needs further research to explore. Methods At rst, we determined the expression and prognostic value of CCNE1 through The Cancer Genome Atlas (TCGA) database and The Genotype-Tissue Expression (GTEx) data. Then, we predicted the upstream non-coding RNAs of CCNE1 through StarBase, GEPIA, and Kaplan-Meier plotter database. We further studied the correlation of CCNE1 expression with BC immune cell inltration, biomarkers of immune cells and immune checkpoints expression through TIMER and GEPIA databases. LINC00511 / microRNA-195-5p / could regulate through and The results suggested that the expression level of CCNE1 was signicantly positively

Introduction BC, as the most frequent female malignant tumor and the second cause of tumor-related death in women, has received more and more worldwide attention [1]. Despite the great improvement has been received in diagnosis and therapy of BC, and signi cantly improved survival rate in BC patients, there is still great challenge in BC treatment for its recurrence and metastasis [2,3]. Therefore, the current treatment of BC is still based on early diagnosis, and surgery assisted with radiotherapy, chemotherapy and targeted-therapy [4]. Therefore, we urgently need to further study the speci c molecular mechanisms of BC occurrence and development, and discover more meaningful biomarkers to improve the early diagnosis of BC and the prognosis of breast cancer patients.
Tumor immune cell in ltration has been reported to be closely related to the curative effects of chemotherapy, radiotherapy, immunotherapy in cancer treatment and prognosis of cancer patients, including BC [5][6][7]. The interaction between tumor cells and their microenvironment has got more and more attention because of its important role in tumor progression. It was widely believed that tumor could escape immune system through three processes, including immune elimination, equilibrium and escape, this ability was widely known as tumor immuno-editing [8]. The interaction between different tumor immune cells facilitates this process [9,10]. It was reported that cancers could be detected only in immune escape phase in clinic after tumor immuno-editing. And, polymorphonuclear cells, dendritic cells, macrophages, B cells, natural killer cells and T cells were all involved in tumor microenvironment, they played important roles in different cancers in varying degrees. For example, more wide immune in ltration was performed in colorectal cancer than BC [11]. This was widely believed to be because of the mutational burden degree, such as melanoma [12,13]. So, we urgently need more in-depth research on tumor immunity to better apply immunotherapy to tumor treatment. CCNE1 was widely studied for its role of binding and activating cyclin dependent kinase 2 (CDK2) to promote the transition of the cell cycle from G1 phase to S phase and start cell DNA synthesis [14]. More and more evidence has suggested that high expression level of CCNE1 was signi cantly correlated with various types of cancer, including bladder cancer [15], ovarian cancer[16] and so on, also, some opposite research had been found [17,18]. In BC, some research reported that upregulated CCNE1 was signi cantly correlated with increasing the risk of recurrence and death basal-like and triple receptor-negative BC, however, no signi cant relationship was found between CCNE1 overexpression and hormone receptorpositive and luminal BC [19]. Also, many researches had reported the role of CCNE1 played in inhibiting cell apoptosis and promoting cellular malignant phenotype in BC [20,21]. So, CCNE1 may played an important role in progression of BC. Our research aims to further discover its potential mechanism involved in the development of BC, and analyze the relationship between its expression level and the prognosis of BC, tumor immune in ltration and escape. CCNE1 expression data of the all 33 cancer types, including BLCA, BRCA, CESC, CHOL, COAD, ESCA,   HNSC, LIHC, LUAD, LUSC, PAAD, READ, STAD, UCEC, KICH, KIRC, KIRP, THCA, ACC, DLBC, GBM, LAML, LGG, MESO, OV, PCPG, PRAD, SARC, SKCM, TGCT, THYM, UCS and UVM, were collected from TCGA database (https://genome-cancer.ucsc.edu/). Then we normalized all the data and analyzed the statistical difference between CCNE1 expression of normal and tumor sample in every cancer type through taking advantage of R limmar package [22]. If p value < 0.05, we think the difference is statistically signi cant.

Prediction of potential upstream miRNA
To predicting the potential upstream miRNAs that could bind to CCNE1, we took advantage of serval prediction programs to nish this work, including TargetScan, PITA, miRmap, miRanda, microT, PicTar and RNA22. We downloaded and normalized all the data, chose the miRNAs that appeared in at least two databases at the same time for the further research. And, these chosen as miRNAs of CCNE1 would be subjected to further correlation analysis.

Kaplan-Meier plotter database
As a professional bioinformatics web database, Kaplan-Meier plotter (http://kmplot.com/analysis/) is widely used to analyze the correlation between genes or miRNAs expression and the survival of various cancer types, including BC. We took advantage of this database to analyze the survival of miR-195-5p in BC. If log rank p value < 0.05, we considered the difference is statistically signi cant.

GEPIA database
As a professional online bioinformatics tool, GEPIA database (http://gepia.cancer-pku.cn/) is widely used in expression and interactive analysis of normal genes based on TCGA and GTEx data [24]. We took advantage of GEPIA to determine the expression level of CCNE1 and LIINC00511 in various cancer types.
If p value < 0.05, we think the difference is statistically signi cant. Overall survival (OS) and disease-free survival (DFS) analysis for CCNE1 in 14 cancer types were conducted by taking advantage of GEPIA database. GEPIA was also used to analyze the prognostic value of LINC00511 in BC. If log rank p value < 0.05, we considered the difference is statistically signi cant. It was identi ed to be statistically signi cant, if |R| > 0.1 and p value < 0.05 at same time.
TIMER database TIMER (https://cistrome.shinyapps.io/timer/) is a professional online bioinformatics database and widely known for its role in analysis of tumor immune in ltration cells [25]. We took advantage of this tool to conduct the analysis of correlation between the expression level of CCNE1 and the in ltration level of tumor immune cell or the expression level of tumor immune checkpoints in BC. It was identi ed to be statistically signi cant, if p value < 0.05.

Statistical analysis
The statistical analysis involved in our research was all automatically calculated through taking advantage of the corresponding online tool listed above. It was identi ed to be statistically signi cant, if p value < 0.05 or log rank p value < 0.05.

Analysis of CCNE1 expression level in pan-cancer types
First, we explore CCNE1 expression in a total of 33 cancer types through TCGA database. The results suggested that CCNE1 was signi cantly upregulated in 18 cancer types (BLCA, BRCA, CESC, CHOL, COAD OS and DFS were included to evaluated its prognostic value. The results suggested that high expression level of CCNE1 in BRCA, LIHC and LUAD had more unfavorable prognosis than downregulation of CCNE1, but no signi cant difference was discovered in the rest types of cancer ( Figure 2). For DFS, we nd that upregulation of CCNE1 suggested poor prognosis in BRCA, LIHC and UCEC, also no signi cant difference was observed in the other types of cancer ( Figure 3). Taken together, CCNE1 may be a potential biomarker for predicting the poor prognosis of BC.
3. MiR-195-5p was predicted to be the upstream miRNA of CCNE1 in BC MiRNAs have been widely studied for its role of inhibiting its target gene expression through binding to the 3'UTR of mRNA. We speculated that CCND1 can also be regulated by miRNAs, so we predicted the potential upstream miRNAs that can bind to it through StarBase database. And, we found out 26 potential miRNAs that may can bind to CCNE1 ( Figure 4A). Next, according to the speci c mechanism of action between miRNAs and target gene, their expression levels should be negatively correlated. Based on this principle, we nally selected the following 5 miRNAs, including miR-26a-5p, miR-26b-5p, miR-101-3p, miR-195-5p and miR-497-5p ( Figure 4B). No signi cant expression level correlation was found between the rest of miRNAs and CCNE1. Among the 5 selected miRNAs, we found that miR-195-5p has the most signi cant relevance and expression relationship with CCNE1. So, we chose miR-195-5p as the rst choice for the following research. Furthermore, we determined miR-195-5p expression and its prognostic value in BC. The results suggested that miR-195-5p was signi cantly downregulated in BC compared to normal samples ( Figure 4C), and, its expression level was signi cantly negatively correlated with CCNE1 expression level ( Figure 4D). Also, high expression level of miR-195-5p was positively correlated with favorable prognosis in BC patients ( Figure 4E). Summarize the above results, miR-195-5p might be the most potential miRNA that can regulated CCNE1 expression in BC. 4. LINC00511 was predicted to be the upstream lncRNA of miR-195-5p in BC For further research, we screened the potential lncRNA upstream of miR-195-5p in BC through the StarBase database. A total of 97 lncRNAs that may bind to miR-195-5p were screened out ( Figure 5A). It was widely known that lncRNA may play a role of the competing endogenous RNA (ceRNA) in the progression of various cancers. Based on this mechanism, the expression of lncRNA should be negatively correlated with the miRNA expression level and positively correlated with the mRNA expression level. So, we veri ed the expression levels of these 97 lncRNAs in BC through GEPIA database. And, the results suggested that only the 4 lncRNAs (HOXC-AS3, LINC00511, LINC00665, TRPM2-AS) were signi cantly upregulated in BC ( Figure 5B-E). Next, we analyzed the correlation between the 4 lncRNAs and miR-195-5p or CCNE1. We nd out that only LINC00511 expression was negatively correlated with miR-195-5p expression and positively correlated with CCNE1 expression at the same time ( Figure 5F). So, we chose LINC00511 as the upstream lncRNA for the next research. Furthermore, we veri ed this result and explored the potential prognostic value of LINC00511. And, LINC00511 expression level was signi cantly negatively correlated with miR-195-5p expression level ( Figure 5G) and positively correlated with CCNE1 expression ( Figure 5H). Also, high expression level of LINC00511 predicted the poor prognosis in BC patients ( Figure 5I). Taken together, LINC00511 might upregulated CCNE1 expression through competitively binding to miR-195-5p and inhibiting its expression in BC.

Analysis of the correlation between CCNE1 and immune cell in ltration in BC
CCNE1, as a member of cyclin, was believed to be closely related to protein synthesis and DNA replication. Therefore, we boldly assumed that it is also inseparable from tumor cell immunity. So, we analyzed the correlation between CCNE1 and immune cell in ltration in BC through TIMER database to investigate whether CCNE1 could be a potential biomarker for immunotherapy in BC. We compared the tumor immune cell in ltration level in BC with various somatic copy number alterations (SCNA) of CCNE1. The results suggested that in addition to CD8+ T cell, normal copy number or deletions or amplications of CCNE1 was observed to be correlated with increased immune cell in ltration in BC, including B cell, CD4+ T cell, macrophage, neutrophil, and dendritic cell ( Figure 6A). Next, we determined the relationship between CCNE1 expression level and immune cell in ltration level in BC. We found out that high expression level of CCNE1 was signi cantly positively with immune cell in ltration level in BC, including B cell, macrophage, neutrophil, and dendritic cell, and, no signi cant difference was observed in CD4+ T cell and CD8+ T cell ( Figure 6B-G). So, these ndings suggested that CCNE1 may play a very signi cant role in immune cell in ltration in BC.

Page 7/23
Furthermore, we analyzed the relationship between CCNE1 and biomarkers of immune cell in BC through GEPIA database, to better verify the correlation of CCNE1 with tumor immunity in BC. As shown in Table  1, CCNE1 was signi cantly positively correlated with the biomarkers of B cell (CD19), CD8+ T cell (CD8B), M1 macrophage (IRF5), M2 macrophage (CD163 and MS4A4A), Neutrophil (CEACAM8 and CCR7). And, CCNE1 was negatively correlated with the biomarker of Dendritic cell (CD1C), no signi cant difference was observed between CCNE1 and CD4+ T cell biomarker. In summary, CCNE1 was still signi cantly positively related to immune cell in ltration in BC.  Figure 7A-C). And, the positive correlation between CCNE1 and PD1 was also observed through GEPIA database ( Figure 7D-F). Taken together, we considered that CCNE1 may promote the progression of BC through mediating tumor immune escape. CCNE1 might be the potential target for BC tumor immunotherapy.

Discussion
BC is widely known as one of the most common women malignant tumor all over the world, and its mortality rate ranks second in female cancers due to the annual growth [26,27]. In the recent years, great achievements have been got in the diagnosis and therapy of BC, its mortality rate has decreased[28]. BC is also widely known as a highly heterogeneous tumor and has various types of etiology and pathological performance in different patients [29]. More and more evidence suggested that the prognosis of BC was signi cantly correlated with tumor immunity [30]. And, some research reported that lots of in ammatory cells could in ltrate in the tumor microenvironment of BC [31]. A lot of evidence suggested that CD8 + T cell was signi cantly related to immune escape in BC, and, CD8 + T cell and CD4 + T cell in ltration was greatly correlated with the prognosis of BC patients [32]. Macrophage also played a very important role in tumor immune in ltration and was responsible for cleaning up cell debris and antigen response in BC [33]. Therefore, more in-depth research on tumor immunity will provide more guidance for the treatment and prognosis of BC.
lncRNAs is a group of RNA that molecules with transcripts longer than 200 nt [34]. They are not involved in encoding proteins, but played a role of regulating the expression level of target genes at different stages in progression of tumors, including epigenetic, transcriptional, post-transcriptional regulation and so on [34]. More and more research found that abnormally expressed lncRNAs was signi cantly correlated with poor prognosis and cell proliferation, invasion, apoptosis of BC [35]. Such as, lncRNA-Hh was reported that could promote the generation of tumor steam cell in twist-positive BC through activating the hedgehog signaling pathway [36]. Upregulated LncRNA-HOXA11-AS was reported to be involved in promoting cell invasion and migration in BC [37]. lncRNAs is receiving more and more attention for the role it played in tumor immunity [38], and, more and more evidence suggested that the types and number in ltrating immune cells were signi cantly correlated with the prognosis of BC [39,40].
In this study, we rst analyzed the expression level of CCNE1 in pan-cancer types, and we found that CCNE1 was signi cantly upregulated in in the 14 cancer types, including BLCA, BRCA, CESC, CHOL, COAD, ESCA, HNSC, LIHC, LUAD, LUSC, PAAD, READ, STAD and UCEC. Next, the analysis of prognosis value of CCNE1 in the 14 cancer types was conducted by GEPIA database. The results suggested that upregulation of CCNE1 was correlated with poor prognosis in BRCA, LIHC, LUAD and UCEC. For further research, we conducted the prediction of upstream noncoding RNAs of CCNE1. And, miR-195-5p was chosen as the upstream miRNA of CCNE1, LINC00511 was predicted as the upstream lncRNA of miR-195-5p. The expression and correlation analysis suggested that miR-195-5p was downregulated in BC and negatively correlated with CCNE1 expression, its downregulation was also related to poor prognosis of BC patients. LINC00511 was found signi cantly upregulated in BC and negatively correlated with miR-195-5p expression. At the same time, the expression level of LINC00511 was positively correlated with CCNE1 expression and its upregulation was signi cantly related to poor prognosis of BC patients. This nding was fully in line with the mechanism of ceRNA. So, we think that LINC00511 / miR-195-5p axis is the most potential that could regulate CCNE1 expression in BC. Furthermore, our research in the correlation between CCNE1 and tumor immune in ltration suggested that CCNE1 was signi cantly positively with immune cell in ltration level in BC, including B cell, macrophage, neutrophil, and dendritic cell. CCNE1 was also positively correlated with some biomarkers of these immune cells, including B cell (CD19), CD8+