Comprehensive Analysis of CCAAT/Enhancer Binding Protein Family in Ovarian Cancer

Abstract

Background:

Ovarian cancer has brought serious threats to female health. CCAAT/enhancer binding proteins (C/EBPs) are key transcription factors involved in ovarian cancer. Therefore, comprehensive profiling C/EBPs in ovarian cancer is needed.

Methods:

A comprehensive analysis concerning C/EBPs in ovarian cancer was performed. Firstly, detailed expression of C/EBP family members was integrally retrieved and then confirmed using immunohistochemistry. The regulatory effects and transcription regulatory functions of C/EBPs were studied by using regulatory network analysis and enrichment analysis. Using survival analysis, receiver operating characteristic curve analysis, and target-disease association analysis, the predictive prognostic value of C/EBPs on survival and drug responsiveness was systematically evaluated. The effects of C/EBPs on tumor immune infiltration were also assessed.

Results:

Ovarian cancer tissues expressed increased CEBPA, CEBPB, and CEBPG but decreased CEBPD when compared with normal control tissues. The overall alteration frequency of C/EBPs in ovarian cancer was approaching 30%. C/EBP family members formed a reciprocal regulatory network involving carcinogenesis and had pivotal transcription regulatory functions. C/EBPs could affect survival of ovarian cancer and correlated with poor survival outcomes (OS: HR = 1.40, P = .0053 and PFS: HR = 1.41, P = .0036). Besides, expression of CEBPA, CEBPB, CEBPD, and CEBPE could predict platinum and taxane responsiveness of ovarian cancer. C/EBPs also affected immune infiltration of ovarian cancer.

Conclusions:

C/EBPs were closely involved in ovarian cancer and exerted multiple biological functions. C/EBPs could be exploited as prognostic and predictive biomarkers in ovarian cancer.

Keywords

Biomarker CCAAT/enhancer binding protein drug responsiveness ovarian cancer survival

Introduction

With a mortality-to-incidence ratio beyond 0.6, ovarian cancer is the most lethal gynecological malignancy seriously threatening female health.¹ The lifetime risk of developing ovarian malignancies is 1.39% and the lifetime mortality risk of ovarian cancer is 1.04%.² Due to the insidious onset, approximately 70% of ovarian cancer patients are diagnosed at advanced stages, thereby exacerbating the situation.² The primary treatment for ovarian cancer constitutes cytoreductive surgery and postoperative chemotherapy, which generally incorporates platinum and taxane.³ The combination regimen of platinum and taxane has an initial response rate reaching 80% and has been recognized as the therapeutic mainstay.² However, most patients will resist, relapse, and recur.³ The 5-year survival rate of ovarian cancer at advanced stages is as low as 30% and remains unimproved for decades despite the unceasing development of medicine.⁴ The high lethality rate can be attributed to the lack of effective early screening biomarkers and predictive prognostic indicators, and ovarian cancer represents a public medical concern worldwide.^2
-4 Therefore, identification of reliable biomarkers for the diagnosis, treatment, prevention, and prediction of ovarian cancer is of great importance and brings immense benefit.

CCAAT/enhancer binding proteins (C/EBPs) comprise a family of 6 structurally and functionally homologous transcription factors, namely, CEBPA, CEBPB, CEBPD, CEBPE, CEBPG, and CHOP.⁵ C/EBP family members are related by a high degree of amino acid sequence identity to the basic leucine zipper domain, which is highly conserved among all family members and required for DNA binding and dimerization.⁶ Besides the C-terminal basic leucine zipper domain, C/EBP family members also encompass a transactivation domain and a regulatory domain in the N-terminus, which are least conserved and vary from strong activators to dominant repressors.⁷ C/EBPs can form homodimers and heterodimers with other family members.⁶ As the prerequisite for DNA binding, dimerization enables C/EBPs to precisely modulate the transcriptional activity of target genes.⁷ C/EBP family members act as key transcription factors in organisms and play pivotal roles in cell proliferation, differentiation, apoptosis, metabolism, and immunity.^6,7 Target genes of C/EBP family members are diverse and involved in a variety of cellular biological processes including hematopoiesis, adipogenesis, inflammation and immunity, metabolism, and carcinogenesis.^6,7 C/EBPs occupy a particular position at the crossroad between proliferation and differentiation and might possess both tumor promotion and suppression functions.⁸ Accumulative evidences indicate that anomalous expression of C/EBPs contributes to tumorigenesis and progression of ovarian cancer.^9
-13 However, the expression profiles and exact roles of C/EBPs in ovarian cancer have not yet been elucidated. Moreover, C/EBP family members might have similar but distinct expression patterns in ovarian cancer.

In this study, we conducted a comprehensive analysis concerning C/EBPs in ovarian cancer on the basis of various large databases, exploiting microarray technology to assist in biological and biomedical researches. We profiled the expression of C/EBPs in ovarian cancer integrally and in detail and then validated the results by using immunohistochemical staining. Besides, the regulatory effects and transcription regulatory functions of C/EBPs were explored thoroughly. Furthermore, the prognostic and predictive value of C/EBPs was also systematically evaluated.

Material and Methods

Gene expression analysis

To get a general insight concerning C/EBPs in ovarian cancer, a comprehensive analysis was performed based on several large data resources. The expression of C/EBPs was compared between tumor and normal tissues using Student’s t test in Oncomine, which has substantially facilitated genome-wide expression research, although its website has been taken offline.¹⁴ GEPIA and HPA were utilized to characterize the expression difference of C/EBP family members, which were user-friendly data mining portals.¹⁵ Using UALCAN, a comprehensive online database, the expression of C/EBP family members was compared across different tumor subgroups based on individual tumor grade and stage.¹⁶ Data concerning mutations, alteration frequency, and mRNA expression were retrieved from and analyzed using cBioPortal for Cancer Genomics, an open-source interactive exploration resource.¹⁷ By comparison of gene expression, one-way ANOVA was adopted to evaluate the statistical significance.

Immunohistochemistry analysis

Tissues were formalin-fixed, dehydrated in ascending gradient of ethanol, paraffin-embedded, and then sliced in series. After deparaffinization in xylene and rehydration in ethanol, heat-induced antigen retrieval was performed. The sections were then blocked with bovine serum following inhibition of endogenous peroxides using hydrogen peroxide. After incubation with the corresponding primary antibody (anti-CEBPA, ab140479, Abcam; anti-CEBPB, 23431-1-AP, Proteintech; anti-CEBPD, ab245214, Abcam; anti-CEBPE, ab246861, Abcam; anti-CEBPG, 12997-1-AP, Proteintech; anti-CHOP, 15204-1-AP, Proteintech) overnight, the slides were exposed to horseradish peroxides linked secondary antibody. The staining was detected using diaminobenzidine followed by counterstaining using hematoxylin. The staining intensity was graded into 0 to 3 (0 indicates absent staining, 1 means weak staining, 2 denotes moderate staining, and 3 represents strong staining). Protein expression was calculated and quantified as a sum of the percentage of positive cells multiplied by the respective staining intensity. To compare C/EBPs expression between tumor and normal tissues, Mann-Whitney test was adopted.

Samples (tumor, N = 20; normal, N = 20) were collected at the First People’s Hospital of Yunnan Province from 2021 to 2023 with written informed consent. Tumor tissues, comprising mainly of serous subtype, were collected at initial debulking surgery prior to chemotherapy, radiation, hormone therapy, or targeted therapy. Normal tissues were obtained at oophorectomy owing to benign conditions such as ovarian chocolate cyst and leiomyoma. Tumor and normal ovarian tissues were fixed in formaldehyde, embedded in paraffin and then analyzed using immunohistochemistry. There was no statistically significant difference in age between the 2 groups (median age 58.5 and 55.5, respectively). The study was conducted in accordance with the World Medical Association Declaration of Helsinki, and approved by the Institutional Review Board of the First People’s Hospital of Yunnan Province (No. 164).

Regulatory network analysis

The protein-protein interactions between C/EBP family proteins and other closely related proteins were explored by using STRING, an online web portal for interaction network analysis.¹⁸ The edge confidence of the interaction network was then scored. GO and KEGG pathway enrichment analysis was also conducted by using STRING. Correlation analysis between C/EBPs was performed using Pearson’s correlation test. Pathway and process enrichment analysis for C/EPP family members was performed in Metascape.¹⁹ All statistically enriched terms were identified and accumulative P values and enrichment factors were calculated to cast the proteins into term clusters. Terms with a P value < .01, a minimum count of 3, and an enrichment factor > 1.5 were included and clusters with a similarity score > 0.3 were linked.

Survival analysis

Using gene expression data and survival information from GEO, EGA, and TCGA databases, KM plotter represents a tool for meta-analysis-based discovery and validation of survival biomarkers for cancer research. Information concerning 1816 ovarian cancer patients was integrated in this web-based survival analysis tool.²⁰ Briefly, ovarian cancer patients were subdivided according to their C/EBPs expression, then the probability of overall survival (OS) and progression-free survival (PFS) was calculated and compared. Hazard ratio (HR) and a 95% confidence interval was determined and the log-rank test was employed to assess the statistical significance.

Receiver operating characteristic curve analysis

ROC plotter is the first online transcriptome-level validation tool for predictive biomarkers, which links gene expression and response to therapy and helps predict benefits from a specific treatment.²¹ Currently, the database has been established with 70 632 gene symbols that have been identified in breast cancer, ovarian cancer, colorectal cancer, and glioblastoma.²¹ Ovarian cancer patients were classified based on their treatment response to platin and taxane and the Mann-Whitney test was used to compare the expression of C/EBPs between responders and non-responders. A receiver operating characteristic curve was drawn and the predictive power of C/EBP family members on the treatment response of ovarian cancer patients was evaluated.

Target-disease association analysis

Open Targets has aggregated and merged genetic associations from the literature, newly-derived loci from functional genomics data including chromatin conformation and interactions, and quantitative trait loci like eQTLS, serving as an exploratory tool that enables detailed biological insight and causal gene prioritization based on trait-associated loci and identification of potential drug targets.²² Using the Open Targets platform, the target-disease associations of C/EBP family members were assessed to inform target decision-making and summarized in a heatmap.

Immune infiltrate analysis

To evaluate the association between immune infiltrates and gene expression in various cancer types, TIMER2.0 provides a platform to estimate immune infiltration abundance using multiple immune deconvolution methods.²³ The purity-adjusted Spearman’s Rho and Spearman’s correlation P value were calculated. The effects of C/EBP family members’ expression on immune infiltration of ovarian cancer were also estimated using the quanTIseq algorithm, wherein the abundance of tumor-infiltrating immune cells was assessed using cell fractions and a Spearman’s correlation coefficient was calculated.

Transcription regulatory function analysis

JASPAR stores manually curated transcription factors binding profiles and represents a regularly maintained open-access database.²⁴ We have searched the 6 C/EBP family members in JASPAR to summarize and depict their matrix profiles, which were stored as position frequency matrices and derived from occurrences of each nucleotide at each position in a set of observed transcription factor-DNA interactions. ChIPBase is an online data resource, which has identified approximately 151 187 000 regulatory relationships between 171 600 genes and 3000 regulators through analyzing ChIP sequencing dataset.²⁵ The ChIP-Function tool helps predict the function of DNA-binding proteins on their transcriptional targets using a hypergeometric method and FDR correction. By analyzing ChIP-sequencing data, the Protein module helps identify genes that could be upstream or downstream occupied by DNA-binding proteins.

Statistical analysis

Data analysis was performed as described above in detail. Statistical significance was restricted to the level of P < .05. Asterisks represented the level of statistical significance (*P < .05, **P < .0.1, ***P < .001, and **** P < .0001) and ns represented not significant.

Results

Integrated expression of C/EBP family members in ovarian cancer

Six C/EBP family members have been detected in various human cancers. By using the Oncomine data portal, we compared the expression of C/EBPs in cancer tissues with that in normal tissues (Figure 1). The expression of C/EBPs in ovarian cancer and normal ovary was compared and listed in detail (Table 1). In 4 datasets, ovarian cancer tissues expressed more CEBPA than normal controls. The maximal fold change is 1.481 (P < .001). There was divergence in CEBPB expression difference. Four datasets exhibited downregulated CEBPB expression in ovarian cancer tissues (maximal fold change = −2.459, P < .0001), while one had overexpressed CEBPB (fold change = 2.149, P < .0001). CEBPD was found to be decreased in ovarian cancer tissues in 11 datasets with the greatest fold change of −3.975 (P < .0001), whereas upregulation of CEBPD was detected in 2 datasets (maximal fold change = 1.769, P = .002). As for CEBPE, one dataset had increased expression in malignant tissues (fold change = 1.074, P < .0001), while the other one had decreased expression (fold change = −2.915, P < .0001). Twelve datasets showed overexpressed CEBPG in ovarian cancer tissues and the maximal fold change reached 22.989 (P < .0001). Concerning CHOP, most datasets had lower expression in ovarian cancer tissues (maximal fold change = −1.302, P = .026), and only one dataset exhibited higher CHOP expression in ovarian cancer tissues (fold change = 1.927, P < .001).

Figure 1.

Expression of C/EBPs in different cancers. The expression of C/EBP family members was compared between tumor tissues and normal controls. Datasets with statistically significant expression difference were summarized.

Table 1.

Comparison of C/EBPs expression in ovarian cancer and normal tissues.

Gene	Type of ovarian cancer vs Normal	Fold change	t Test	P value	PMID
CEBPA	Ovarian serous adenocarcinoma vs Normal	1.338	3.675	.004	14 760 385
	Ovarian serous surface papillary carcinoma vs Normal	1.074	5.971	1.01E-06	11 158 614
	Ovarian mucinous adenocarcinoma vs Normal	1.171	2.583	.012	15 161 682
	Ovarian carcinoma vs Normal	1.481	5.155	1.65E-04	18 593 951
CEBPB	Ovarian carcinoma vs Normal	2.149	7.825	8.51E-07	18 593 951
	Ovarian endometrioid adenocarcinoma vs Normal	−1.605	−2.542	.013	15 161 682
	Ovarian mucinous adenocarcinoma vs Normal	−1.238	−4.329	3.01E-04	16 452 189
	Ovarian endometrioid adenocarcinoma vs Normal	−1.096	−2.701	.01	16 452 189
	Ovarian serous adenocarcinoma vs Normal	−2.459	−4.395	8.04E-05	19 486 012
CEBPD	Ovarian carcinoma vs Normal	1.769	3.617	.002	18 593 951
	Ovarian serous cystadenocarcinoma vs Normal	1.1	1.928	4.60E-02	NA
	Ovarian serous surface papillary carcinoma vs Normal	−2.732	−8.461	9.77E-10	11 158 614
	Ovarian serous adenocarcinoma vs Normal	−1.241	−4.574	7.96E-05	15 161 682
	Ovarian endometrioid adenocarcinoma vs Normal	−1.175	−3.965	9.56E-04	15 161 682
	Ovarian mucinous adenocarcinoma vs Normal	−1.229	−3.547	2.00E-03	15 161 682
	Ovarian clear cell adenocarcinoma vs Normal	−1.622	−2.705	1.30E-02	15 161 682
	Ovarian serous adenocarcinoma vs Normal	−1.226	−7.364	2.02E-09	16 452 189
	Ovarian endometrioid adenocarcinoma vs Normal	−1.257	−6.954	1.25E-08	16 452 189
	Ovarian mucinous adenocarcinoma vs Normal	−1.277	−5.643	4.13E-05	16 452 189
	Ovarian clear cell adenocarcinoma vs Normal	−1.051	−2.544	1.50E-02	16 452 189
	Ovarian serous adenocarcinoma vs Normal	−2.21	−3.452	1.10E-02	14 760 385
	Ovarian serous adenocarcinoma vs Normal	−3.975	−5.889	1.59E-05	19 486 012
CEBPE	Ovarian serous surface papillary carcinoma vs Normal	1.074	5.971	1.01E-06	11 158 614
CEBPE	Ovarian serous adenocarcinoma vs Normal	−2.915	−5.365	3.04E-05	19 486 012
CEBPG	Ovarian serous surface papillary carcinoma vs Normal	22.989	8.418	2.48E-09	11 158 614
	Ovarian mucinous adenocarcinoma vs Normal	1.761	4.427	4.49E-04	15 161 682
	Ovarian serous adenocarcinoma vs Normal	2.19	5.645	1.04E-05	15 161 682
	Ovarian endometrioid adenocarcinoma vs Normal	1.602	3.612	2.00E-03	15 161 682
	Ovarian clear cell adenocarcinoma vs Normal	1.436	2.78	1.00E-02	15 161 682
	Ovarian carcinoma vs Normal	2.017	11.817	2.37E-12	18 593 951
	Ovarian serous adenocarcinoma vs Normal	2.181	2.652	2.00E-02	14 760 385
	Ovarian serous cystadenocarcinoma vs Normal	1.936	7.835	1.93E-05	NA
	Ovarian serous adenocarcinoma vs Normal	1.152	4.82	7.65E-05	16 452 189
	Ovarian mucinous adenocarcinoma vs Normal	1.18	3.385	.002	16 452 189
	Ovarian endometrioid adenocarcinoma vs Normal	1.082	2.649	.008	16 452 189
	Ovarian clear cell adenocarcinoma vs Normal	1.114	2.59	.014	16 452 189
CHOP	Ovarian serous cystadenocarcinoma vs Normal	1.927	6.39	1.34E-04	NA
	Ovarian mucinous adenocarcinoma vs Normal	−1.179	−4.665	5.64E-04	16 452 189
	Ovarian endometrioid adenocarcinoma vs Normal	−1.164	−4.642	8.57E-04	16 452 189
	Ovarian clear cell adenocarcinoma vs Normal	−1.175	−3.404	.003	16 452 189
	Ovarian serous adenocarcinoma vs Normal	−1.134	−4.058	.003	16 452 189
	Ovarian carcinoma vs Normal	−1.169	−2.453	.016	18 593 951
	Ovarian serous adenocarcinoma vs Normal	−1.302	−2.04	.026	19 486 012

Detailed expression profiles of C/EBP family members in ovarian cancer

We then dissected the expression profiles of C/EBPs in ovarian cancer. Expression of C/EBPs were detected using immunohistochemistry in tumor and normal tissues, which were collected at initial debulking surgery of ovarian cancer or oophorectomy owing to benign conditions correspondingly (N = 20, each separately) (Figure 2A-F). Consistently, ovarian cancer had overexpressed CEBPA and CEBPG and decreased CEBPD. Similarly, there was no statistically significance in CEBPE and CHOP expression. However, we found significantly elevated CEBPB expression in ovarian cancer. These results were further confirmed by analyzing GEPIA and HAP (Supplemental Figure s1A-s1F). The relationship between C/EBPs expression and the clinicopathological parameters of ovarian cancer patients was also explored. When compared based on individual tumor grade, the expression of CEBPA and CEBPG varied significantly, while the other 4 C/EBP family members did not significantly differ (Supplemental Figure s2A-s2F). As for different tumor stages, only CHOP had significantly varied expression (Supplemental Figure s3A-s3F). The mutation profiles of C/EBP family members were further analyzed using the cBioPortal for Cancer Genomics. Their genetic mutation rates were calculated and depicted as oncoprint (Supplemental Figure s4A). Among all 6 C/EBP family members, CEBPG had the highest mutation rate in ovarian cancer (12%) and CEBPD took the last place with a mutation rate of 4%. Alteration frequencies of C/EBPs were also assessed and the integrated alteration frequency was approaching 30% (Supplemental Figure s4B).

Figure 2.

Verification of C/EBPs in ovarian cancer. Expression of C/EBPs were detected using immunohistochemistry in ovarian tumor and normal tissues: (A) CEBPA, (B) CEBPB, (C) CEBPD, (D) CEBPE, (E) CEBPG, and (F) CHOP. P value was denoted as *P < .05, **P < .0.1, ***P < .001, and ****P < .0001 and ns represented not significant.

Regulatory network involving C/EBPs in ovarian cancer

To get a general insight concerning C/EBPs in ovarian cancer, we then interrogated their expression and processed an incorporated overview in heatmap (Figure 3A). The correlations of C/EBPs with each other were subsequently analyzed and summarized as a matrix heatmap (Figure 3B, Supplemental Figure s5A). The results showed a significant positive correlation between CEBPA and CEBPB (R = .29, P < .0001), CEBPA and CEBPD (R = .29, P < .001), CEBPA and CEBPE (R = .28, P < .0001), CEBPA and CEBPG (R = .27, P < .0001), CEBPB and CEBPD (R = .33, P < .0001), CEBPB and CEBPE (R = .27, P < .0001), and CEBPG and CHOP (R = .21, P < .01). However, the correlation between the other pair of C/EBPs was not statistically significant.

Figure 3.

Regulatory network analysis of C/EBP family members. (A) Expression of C/EBPs in ovarian cancer was incorporated as a heatmap. (B) The results of correlation analysis between C/EBP family members were summarized as a matrix heatmap. (C) Protein-protein interaction network of C/EBPs. Regulatory function analysis was performed. (D) C/EBP family members were classified into 4 clusters and shown in an enrichment network. (E) Representative clusters of C/EBPs were exported as a box plot. (F) Terms significantly affected by C/EBPs. (G) DisGeNET ontology affected by C/EBPs.

Protein-protein interaction network of C/EBPs was then constructed (Figure 3C). The edge confidence for each interaction was calculated and a confidence score was assigned (Supplemental Figure s5B). Besides, proteins closely related to C/EBPs were identified, such as CDK2, ESR1, and E2F4 (Supplemental Figure s5C). GO enrichment analysis was further performed to predict the functional roles of C/EBPs. The C/EBP family members were prominently enriched in cellular components including CHOP-C/EBP complex, RNA polymerase II transcription factor complex, and nucleoplasm (Supplemental Table s1). The molecular function of C/EBPs was mainly transcription and protein dimerization, such as protein homodimerization activity, protein heterodimerization, transcription factor binding, and transcription coregulator activity (Supplemental Table s2). C/EBP family members also significantly affected biological processes of differentiation, cellular response, and transcription, wherein GO pathways like granulocyte differentiation, response to external stimulus, and transcription by RNA polymerase II were significantly regulated by C/EBPs (Supplemental Table s3). Additionally, pathways associated with C/EBP family members were further delineated through KEGG enrichment analysis (Supplemental Table s4). Among the enriched KEGG pathways, transcription misregulation in cancer was the most significantly affected pathway by C/EBPs.

To further understand the regulatory function of C/EBPs, Metascape was exploited. Consistently, C/EBPs mainly affected differentiation, cellular response, and transcription and the significantly enriched clusters were transcriptional cascade regulating adipogenesis, response of EIF2AK1 (HRI) to heme deficiency, differentiation of white and brown adipocyte, and transcriptional misregulation in cancer (Figure 3D and E, Supplemental Figure s5D). In process analysis, terms such as immune system process, response to stimulus, and positive regulation of biological process were significantly affected by C/EBPs (Figure 3F). The enrichment analysis results regarding DisGeNET ontology were also summarized and C/EBPs were found to be significantly correlated with disease conditions including urothelial carcinoma, carcinoma transitional cell, immunosuppression, and acute promyelocytic leukemia (Figure 3G). Taken together, C/EBPs formed a reciprocal regulatory network and regulated differentiation, cellular response, and transcription, thereby involving carcinogenesis.

Transcription regulatory functions of C/EBP family members

Since C/EBP family members represent classical transcription factors, we turned to study their biological functions as DNA-binding proteins. By analyzing JASPAR, the binding motifs of C/EBPs were retrieved and there was similarity conserved among C/EBP family members (Figure 4A-F). Additionally, ChIPBase data resource was explored to further clarify their transcriptional function. Using ChIP-function tool, the transcription function of C/EBPs was predicted (Supplemental Table s5). CEBPA could regulate 27 terms such as response to endoplasmic reticulum stress and protein sumoylation. Thirty-four terms were under the control of CEBPB like cell division and autophagy. CEBPD has affected 253 terms including DNA repair and G2/M transition of mitotic cell cycle. Concerning CEBPE, we could recall no records. A total of 15 terms were controlled by CEBPG like translation and base-excision repair. CHOP was found to regulate 3 terms, namely, RNA splicing, negative regulation of catalytic activity, and male gonad development. Genes occupied and targeted by C/EBPs were also identified (Supplemental Table s6). CEBPA had 2272 targets like BRCA1 and MSH2. 2330 genes were predicted to be targets of CEBPB such as BRAF, BRCA1, and BRD4. CEBPD has targeted 10 266 genes including ATM and ATR. CEBPE failed in target prediction. 1300 genes were upstream or downstream occupied by CEBPG like TP53 and TP53BP1. CHOP could target 27 genes including ANKS4B and LAG3. In summary, C/EBPs had pivotal roles in transcription regulation.

Figure 4.

Binding motif sequences of C/EBPs. The binding motifs of C/EBPs were retrieved and depicted: (A) CEBPA, (B) CEBPB, (C) CEBPD, (D) CEBPE, (E) CEBPG, and (F) CHOP.

Prognostic and predictive value of C/EBP family members in ovarian cancer

Using the KM plotter, the prognostic significance of C/EBPs in ovarian cancer was assessed. Combined C/EBPs denoted unfavorable survival of ovarian cancer patients [OS: HR = 1.40 (1.10-1.77), P = .0053 and PFS: HR = 1.41 (1.12-1.78), P = .0036] (Figure 5A and B). More specifically, the prognostic significance of each C/EBP family member was further evaluated. There was no significant association between CEBPA expression and patient survival in ovarian cancer (Supplemental Figure s6A). High CEBPB expression was significantly correlated with unfavorable survival of ovarian cancer patients (Supplemental Figure s6B). Elevated CEBPD expression predicted poor OS and PFS of ovarian cancer patients (Supplemental Figure s6C). Increased CEBPE expression was beneficial to PFS of ovarian cancer patients, while the correlation between CEBPE expression and OS did not have statistical significance (Supplemental Figure s6D). Overexpressed CEBPG indicated favorable OS in ovarian cancer patients, whereas its prognostic effect on PFS was not statistically significant (Supplemental Figure s6E). Expression of CHOP tended to not significantly affect survival of ovarian cancer patients (Supplemental Figure s6F).

Figure 5.

Prognostic and predictive effects of C/EBPs in ovarian cancer. C/EBP family members were integrated for survival analysis and Kaplan-Meier survival curves were drawn. (A) OS. (B) PFS. The effects of C/EBPs expression on platinum and taxane responsiveness of ovarian cancer were assessed and receiver operating characteristic curve analysis was performed. (C) CEBPA. (D) CEBPB. (E) CEBPD. (F) CEBPE. (G) CEBPG. (H). CHOP.

Expression of C/EBPs could also predict drug responsiveness of ovarian cancer. To verify the effects of C/EBPs on treatment benefits of ovarian cancer, ROC plotter was explored. Decreased CEBPA expression could predict platinum and taxane responsiveness and the area under the curve (AUC) value was 0.644 and 0.623, respectively (Figure 5C and Supplemental Figure s7A). CEBPB also harnessed significant potential of drug responsiveness prediction in ovarian cancer and reduced CEBPB expression indicated treatment benefit (platinum AUC 0.605, taxane AUC 0.584) (Figure 5D, Supplemental Figure s7B). Patients with declined CEBPD expression harbored pronounced responsiveness toward platinum (AUC 0.587) and taxane (AUC 0.584) (Figure 5E, Supplemental Figure s7C). A low expression level of CEBPE endowed ovarian cancer with prominent platinum responsiveness (AUC 0.612) and taxane responsiveness (AUC 0.605) (Figure 5F, Supplemental Figure s7D). However, CEBPG and CHOP could not recognize platinum and taxane responsiveness and lacked significant predictive power (Figure 5G and H, Supplemental Figure s7E and s7F).

The Open Targets platform was also searched to retrieve the target-disease associations of C/EBPs, thereby informing their potential as targets in certain disease conditions (Supplemental Figure s8). CEBPA was closely related to the blood system and involved in diseases such as acute myeloid leukemia. CEBPB had roles in various diseases like diabetes mellitus, ulcer disease, and breast carcinoma. CEBPD and CEBPE mainly affected the blood system and participated in infection. CEBPG was primarily negatively associated with metabolism. CHOP was correlated with a plethora of malignancies including liposarcoma, esophageal squamous cell carcinoma, and bladder carcinoma. To summarize, C/EBPs had prognostic and predictive value in ovarian cancer.

Relationship between C/EBPs and tumor immune infiltrates

Tumor-infiltrating immune cells are key components in tumor microenvironment and affect tumorigenesis, progression, and oncotherapy.^2,23 We therefore explored TIMER2.0 database to estimate the effects that C/EBPs exerted on immune infiltration abundance in ovarian cancer. Increase in CEBPA expression was significantly correlated with enhanced B cell, macrophage, monocyte, T cell CD4+, and T cell CD8+ infiltration in ovarian cancer (Figure 6A). Overexpressed CEBPB expression remarkably promoted infiltration of B cell, macrophage, monocyte, T cell CD4+, and T cell CD8+ in ovarian cancer (Figure 6B). Results concerning CEBPD, CEBPE, and CEBPG had diverse inclinations. Elevated CEBPD was negatively related to B cell and T cell CD8+ infiltration, while it was positively associated with macrophage, monocyte, and T cell CD4+ infiltration (Figure 6C). Increased CEBPE enhanced B cell, macrophage, monocyte, and T cell CD8+ infiltration but reduced T cell CD4+ abundance (Figure 6D). Elevated CEBPG potentiated extensive macrophage, T cell CD4+, and T cell CD8+ infiltrate, whereas suppressed B cell and monocyte infiltration (Figure 6E). However, CHOP could hardly affect immune cell infiltration in ovarian cancer (Figure 6F).

Figure 6.

Effects of C/EBPs on immune infiltration abundance in ovarian cancer. The effects of C/EBPs on ovarian cancer immune infiltrate were investigated. Immune cells including B cell, macrophage, monocyte, T cell CD4+, and T cell CD8+ were analyzed. (A) CEBPA. (B) CEBPB. (C) CEBPD. (D) CEBPE. (E) CEBPG. (F) CHOP.

Effects of C/EBPs on tumor immune infiltration were further estimated using the quanTIseq algorithm and summarized as a heatmap (Supplemental Figure s9). Although there was some inconsistency when compared with the results analyzed using TIMER2.0, Tregs, neutropil, NK cell, and myeloid dendritic cell were included. Expression of CEBPA, CEBPB, CEBPD, and CEBPE was negatively related to Tregs infiltration abundance. CEBPG was inversely correlated with neutropil, NK cell, and myeloid dendritic cell infiltration abundance. CHOP was also found to be negatively associated with NK cell infiltration abundance. Together, C/EBPs could affect immune cell infiltration in ovarian cancer.

Discussion

C/EBPs have been reported to play multifaced roles in ovarian cancer and have been described as both tumor promoters and tumor suppressors.⁵ In the present study, we systematically studied the expression profiles, regulatory effects, transcription regulatory functions, prognostic significance, and predictive value of C/EBPs in ovarian cancer. Aberrant expression and significant associations have been noted between C/EBPs and ovarian cancer. C/EBP family members regulated a panel of genes and pathways and formed a complex regulatory network. C/EBPs could affect survival, drug responsiveness, and immune infiltration of ovarian cancer. Our results could add to the growing evidence regarding the relevance and complexity of C/EBP family members and offer rationales for exploration of C/EBPs-based prognosis prediction and development of C/EBPs-mediated targeted therapy.

Deregulation of CEBPA can predispose to different malignancies including hematological neoplasms, breast cancer, and lung cancer, and has been reported to correlate with tumor size and progression as well as poor prognosis.⁹ CEBPA could activate leptin transcription in ovarian cancer, thereby mediating reduced chemo-sensitivity to paclitaxel/docetaxel.²⁶ Moreover, polymorphism of CEBPA gene together with its mRNA expression upregulation has deteriorated ovarian cancer outcome and was correlated with poor response to platinum/cyclophosphamide therapy.²⁷ Consistently, in the current study, we found that ovarian cancer had overexpressed CEBPA. Besides, CEBPA could promote immune cell infiltration in ovarian cancer and was involved in diseases such as acute myeloid leukemia. Although CEBPA did not significantly affect survival prognosis, its expression could predict platinum and taxane responsiveness of ovarian cancer patients. CEBPA could also regulate several transcription terms and target genes, which were implicated in ovarian cancer. Therefore, CEBPA could play a tumor-promoting role in ovarian cancer.

There are similarities regarding biological properties of CEBPA and CEBPB.⁷ CEBPB is widely implicated in cell proliferation, differentiation, apoptosis, senescence, and inflammation.¹⁰ Cell-penetrating CEBPB leucine zipper decoys could function as broadly acting anti-cancer agents in breast cancer, glioblastoma, melanoma, colon cancer, and lung cancer.²⁸ Targeting CEBPB could compromise cancer cell survival in chronic myeloid leukemia, glioblastoma, prostate cancer, and breast cancer.¹⁰ Expression of CEBPB was found to be increased in ovarian cancer and correlated with tumor progression.²⁹ CEBPB was reported to promote ascites formation and inhibition of CEBPB could elicit growth inhibitory effect in ovarian cancer.³⁰ In this study, ovarian cancer had increased CEBPB expression and overexpressed CEBPB indicated unfavorable survival prognosis and poor response to platinum and taxane. Besides, CEBPB was involved in various disease conditions and could enhance immune infiltrates in ovarian cancer. Transcription terms like autophagy and target genes including BRCA1 and BRD4 were under the control of CEBPB. These results suggested that CEBPB could exert pro-oncogenic effects in ovarian cancer.

CEBPD was found to be a tumor suppressor in prostate cancer, breast cancer, and myeloid cancers but a tumor promoter in glioblastoma, pancreatic cancer, and urothelial cancers.¹¹ CEBPD could inhibit cell growth and promote mesenchymal to epithelial transition while enhance the migration potential of fallopian epithelial cells, showing a dichotomous role in initiation and promotion of ovarian carcinoma.⁸ In our report, increased CEBPD was associated with poor survival outcomes and reduced platinum and taxane responsiveness. CEBPD affected transcription terms such as DNA repair and G2/M transition of mitotic cell cycle and targeted genes like ATM and ATR. However, CEBPD was downregulated in ovarian cancer and its effect on immune infiltrate was divergent. Accordingly, CEBPD could be exploited as a predictive prognostic biomarker in ovarian cancer, but its exact role warrants further investigation.

CEBPE principally regulates myeloid differentiation and can reprogram myeloid lineage commitment.³¹ Currently, there have been no data concerning a specific tumor suppressor or tumor promoter role of CEBPE.³¹ In the present study, we found that high CEBPE expression predicted longer PFS and decreased platinum and taxane responsiveness. CEBPE participated principally in infection and leukemia and promoted immune infiltration. CEBPE thus had prognostic effect and predictive value in ovarian cancer. However, the expression pattern and the regulatory function of CEBPE need further exploration.

As the truncated isoform of the C/EBP family transcription factors, CEBPG is neither an activator nor a repressor.⁵ CEBPG regulates transcription by heterodimerizing with other C/EBPs and acts as a trans-dominant regulator.¹² Here, CEBPG was found to be significantly overexpressed in ovarian cancer and varied in different tumor grades. Additionally, CEBPG could regulate transcription terms like base-excision repair and target genes including E2F2 and HMGB1, which were all involved in ovarian cancer.^32,33 However, its effects on survival and immune infiltration were divergent. Expression of CEBPG was primarily negatively related to metabolism but failed to affect drug sensitivity. Consequently, CEBPG might serve as a regulator of other C/EBPs in ovarian cancer.

CHOP functions as a cellular stress sensor and can be activated by cellular stress conditions such as endoplasmic reticulum stress, DNA damage, growth arrest, hypoxia, nutrient deprivation, and genotoxic agents.¹³ The results herein showed that CHOP had similar expression in tumor and normal tissues but altered significantly in different tumor stages. Besides, CHOP was implicated in several malignancies and targeted genes like ANKS4B and LAG3, which were reported to be potential candidate therapeutic targets in ovarian cancer.^34,35 Nevertheless, expression of CHOP did not have prognostic effect and predictive power in ovarian cancer and could hardly affect immune cell infiltration. Our results implied that CHOP might act as a passenger driver gene in ovarian cancer whose expression altered along with tumor stage.

C/EBP family members had complex reciprocal relationships with each other and formed a multi-functional regulatory network. In addition to expression correlations, C/EBP family proteins have constructed a protein-protein interaction network, wherein some proteins involved in ovarian cancer prognosis like CDK2, ESR1, and E2F4 were also seized. Their cross-talk could regulate pathways concerning differentiation, cellular response, and transcription in cancer. Besides, in regulatory function analysis, C/EBPs were classified into the cluster “transcriptional misregulation in cancer” and enriched in DisGeNET ontology such as urothelial carcinoma and acute promyelocytic leukemia. These results indicated that C/EBPs were involved in carcinogenesis and ovarian carcinoma. Although the type of dimerization could affect the actual function of C/EBPs, the balance was tipped toward pro-oncogenic action. Integrated analysis of all 6 C/EBP family members showed that a combination of C/EBPs could predict poor survival outcomes of ovarian cancer. Elevated C/EBPs indicated unfavorable OS and PFS of ovarian cancer patients. Besides, decreased CEBPA, CEBPB, CEBPD, and CEBPE indicated pronounced responsiveness of ovarian cancer patients toward platinum and taxane. Therefore, C/EBPs could be exploited as potential prognostic and predictive biomarkers for ovarian cancer.

Ovarian cancer is a heterogeneous disease comprising tumors with different histology. Epithelial ovarian cancer makes up over 95% of ovarian malignancies, among which more than 80% pertain to the serous subtype.^2,3 In our research, we surveyed and analyzed databases including GEPIA, HPA, cBioPortal for Cancer Genomics, UALCAN, KM plotter, ROC plotter, and TIMER2.0. All these aforementioned databases are primarily derived from the TCGA dataset, which consists mainly of serous cancer. Expression of C/EBPs were further confirmed using tumor tissues comprising mainly of serous subtype. Inclusion of the other tumor subtypes might bring confounding effects to the results. Another limitation was that, by predictive value analysis, we focused on platinum and taxane. Despite that platinum and taxane comprise standard drug regimens for ovarian cancer, immune therapy and PARP inhibitors are getting increasing clinical applications.^2,36 C/EBPs were found to affect tumor immune infiltration and DNA damage repair. The predictive power of C/EBPs on immune therapy and PARP inhibitors deserves further exploration.

Conclusions

In summary, we have provided a thorough understanding concerning C/EBP family in ovarian cancer, thereby a deeper comprehension of tumorigenesis of ovarian cancer. Ovarian cancer had pronounced CEBPA, CEBPB, and CEBPG but decreased CEBPD expression. Nearly 30% of ovarian cancer had alterations in C/EBPs. C/EBP family members formed a reciprocal regulatory network involving carcinogenesis and had pivotal transcription regulatory functions. A combination of C/EBP family members could predict poor survival outcomes in ovarian cancer. Besides, CEBPA, CEBPB, CEBPD, and CEBPE could predict platinum and taxane responsiveness of ovarian cancer. C/EBPs also exerted effects on tumor immune infiltration. Our results indicated that C/EBPs could be exploited as prognostic and predictive biomarkers in ovarian cancer, which warrants deeper exploration and further validation.

Supplemental Material

sj-docx-1-cix-10.1177_11769351241275877 – Supplemental material for Comprehensive Analysis of CCAAT/Enhancer Binding Protein Family in Ovarian Cancer

Supplemental material, sj-docx-1-cix-10.1177_11769351241275877 for Comprehensive Analysis of CCAAT/Enhancer Binding Protein Family in Ovarian Cancer by Jiahong Tan, Daoqi Wang, Wei Dong, Lei Nian, Fen Zhang, Han Zhao, Jie Zhang and Yun Feng in Cancer Informatics

Footnotes

Acknowledgements

We sincerely thank all participants included in this study and those excellent and dedicated researchers, programmers, funders, and policymakers who have made contributions to the online databases.

Author Contributions

This study was designed by Jiahong Tan and Daoqi Wang. Jiahong Tan, Daoqi Wang, Wei Dong, Lei Nian, Fen Zhang, Han Zhao, Jie Zhang, and Yun Feng performed the experiments, analyzed and interpreted the data. Jiahong Tan, Daoqi Wang, and Yun Feng were major contributors in writing the manuscript. All authors have read and approved the manuscript. All authors have read and approved the manuscript for publication.

Declaration Of Conflicting Interests:

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding:

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by National Natural Science Foundation of China (82203145), National Key Clinical Specialty of Gynecology.First People’s Hospital of Yunnan Province (2022FKZDZK-16), Yunnan Provincial Clinical Medical Center for Reproductive, Gynecological and Obstetrical Diseases (2022LCZXKF-SZ08), Doctoral Research Fund Program of the First People’s Hospital of Yunnan Province (KHBS-2022-026), Basic Research Project of Yunnan Province (202301AT070041), and High-level Personnel Introduction Fund of the First People’s Hospital of Yunnan Province (2022-KHRCBZ-B05). The funding resources had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Availability of Data and Materials

The data presented in the study are openly available and included in the article and supplementary materials.

Ethics Approval

The study was conducted according to the guidelines of the World Medical Association Declaration of Helsinki, and approved by the Institutional Review Board of the First People’s Hospital of Yunnan Province (No. 164).

Consent to Participate

Written informed consent was obtained from all subjects involved in the study.

Consent for Publication

Not applicable.

ORCID iDs

Jiahong Tan

Daoqi Wang

Supplemental Material

Supplemental material for this article is available online.

References

Sung

Ferlay

Siegel

, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209-249.

Lheureux

Braunstein

Oza

AM.

Epithelial ovarian cancer: evolution of management in the era of precision medicine. CA Cancer J Clin. 2019;69:280-304.

Lheureux

Gourley

Vergote

Oza

AM.

Epithelial ovarian cancer. Lancet. 2019;393:1240-1253.

Siegel

Miller

Fuchs

Jemal

Cancer statistics, 2022. CA Cancer J Clin. 2022;72:7-33.

Tolomeo

Grimaudo

The “Janus” role of C/EBPs family members in cancer progression. Int J Mol Sci. 2020;21:12. doi:10.3390/ijms21124308

Wang

Xia

Mao

Wang

The CCAAT/enhancer-binding protein family: its roles in MDSC expansion and function. Front Immunol. 2019;10:10.

Martin

Nguyen

HT.

Basic leucine zipper transcription factors as important regulators of Leydig cells’ functions. Int J Mol Sci.2022;23:21. doi:10.3390/ijms232112887

Sowamber

Chehade

Bitar

, et al. CCAAT/enhancer binding protein delta (C/EBPδ) demonstrates a dichotomous role in tumour initiation and promotion of epithelial carcinoma. EBioMedicine. 2019;44:261-274.

Koschmieder

Halmos

Levantini

Tenen

DG.

Dysregulation of the C/EBPα differentiation pathway in human cancer. J Clin Oncol. 2009;27:619-628.

10.

Sun

Jefferson

Zhou

Angelastro

Greene

LA.

Dominant-negative ATF5 compromises cancer cell survival by targeting CEBPB and CEBPD. Mol Cancer Res. 2020;18:216-228.

11.

Balamurugan

Sterneck

The many faces of C/EBPδ and their relevance for inflammation and cancer. Int J Biol Sci. 2013;9:917-933.

12.

Luo

Gui

Circulating exosomal circFoxp1 confers cisplatin resistance in epithelial ovarian cancer cells. J Gynecol Oncol. 2020;31:e75.

13.

Jauhiainen

Thomsen

Strömbom

, et al. Distinct cytoplasmic and nuclear functions of the stress induced protein DDIT3/CHOP/GADD153. PLoS One. 2012;7:4.

14.

Rhodes

Shanker

, et al. ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia. 2004;6:1-6.

15.

Tang

Kang

, et al. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 2017;45:W98-W102.

16.

Chandrashekar

Bashel

Balasubramanya

SAH

, et al. UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses. Neoplasia. 2017;19:649-658.

17.

Cerami

Gao

Dogrusoz

, et al. The cbio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401-404.

18.

Szklarczyk

Gable

Lyon

, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47:D607-D613.

19.

Zhou

Pache

, et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun. 2019;10:1523.

20.

Nagy

Lánczky

Menyhárt

Győrffy

. Validation of miRNA prognostic power in hepatocellular carcinoma using expression data of independent datasets. Sci Rep. 2018;8:9227.

21.

Fekete

Ősz

Pete

, et al. Predictive biomarkers of platinum and taxane resistance using the transcriptomic data of 1816 ovarian cancer patients. Gynecol Oncol. 2020;156:654-661.

22.

Carvalho-Silva

Pierleoni

Pignatelli

, et al. Open targets platform: new developments and updates two years on. Nucleic Acids Res. 2019;47:D1056-D1065.

23.

Zeng

, et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 2020;48:W509-W514.

24.

Fornes

Castro-Mondragon

Khan

, et al. JASPAR 2020: update of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 2020;48:D87-D92.

25.

Huang

Zheng

Zhang

, et al.ChIPBase v3.0: the encyclopedia of transcriptional regulations of non-coding RNAs and protein-coding genes. Nucleic Acids Res. 2022;51:D46-D56.

26.

Zhang

Yao

, et al. Leptin contributes to the taxol chemoresistance in epithelial ovarian cancer. Oncol Lett. 2019;18:561-570.

27.

Konopka

Szafron

Kwiatkowska

, et al. The significance of c.690G>T polymorphism (rs34529039) and expression of the CEBPA gene in ovarian cancer outcome. Oncotarget. 2016;7:67412-67424.

28.

Zhou

Sun

Pasquier

, et al. Cell-penetrating CEBPB and CEBPD leucine zipper decoys as broadly acting anti-cancer agents. Cancers. 2021;13:10. doi:10.3390/cancers13102504

29.

Shi

Chen

Han

, et al. WISP2 promotes cell proliferation via targeting ERK and YAP in ovarian cancer cells. J Ovarian Res. 2020;13:85.

30.

Liu

Lin

Shen

Pan

WW.

DAXX promotes ovarian cancer ascites cell proliferation and migration by activating the ERK signaling pathway. J Ovarian Res. 2018;11:90.

31.

Wei

D-Q

Zhang

CEBPE expression is an independent prognostic factor for acute myeloid leukemia. J Transl Med. 2019;17:188.

32.

Cámara-Quílez

Barreiro-Alonso

Vizoso-Vázquez

, et al. The HMGB1-2 ovarian cancer interactome. The role of HMGB proteins and their interacting partners MIEN1 and NOP53 in ovary cancer and drug-response. Cancers. 2020;12:9. doi:10.3390/cancers12092435

33.

Zhou

Zhang

Zuo

MZ.

E2F2/5/8 serve as potential prognostic biomarkers and targets for human ovarian cancer. Front Oncol. 2019;9:161.

34.

Wang

Lei

Chi

Liu

Yang

BP.

A comprehensive understanding of ovarian carcinoma survival prognosis by novel biomarkers. Eur Rev Med Pharmacol Sci. 2019;23:8257-8264.

35.

Zheng

Guo

Zhu

Zhou

Tong

Comprehensive analyses of glycolysis-related lncRNAs for ovarian cancer patients. J Ovarian Res. 2021;14:124.

36.

Marchetti

De Felice

Romito

, et al. Chemotherapy resistance in epithelial ovarian cancer: mechanisms and emerging treatments. Semin Cancer Biol. 2021;77:144-166.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

7.48 MB