Identification of ABCA5 among ATP-Binding Cassette Transporter Family as a New Biomarker for Colorectal Cancer

Bu, Peilong; Xiao, Yafei; Hu, Shaowen; Jiang, Xiaowei; Tan, Cong; Qiu, Mengyuan; Huang, Wanting; Li, Mengmeng; Li, Quanying; Qin, Changjiang

doi:https://doi.org/10.1155/2022/3399311

Journal of Oncology

On this page

Abstract Introduction Methods Results Discussion Conclusion Data Availability Ethical Approval Consent Disclosure Conflicts of Interest Authors’ Contributions Acknowledgments Supplementary Materials References Copyright Related Articles

Special Issue

Identification of Malignant Tumor Markers and their Mechanism

View this Special Issue

Research Article | Open Access

Volume 2022 | Article ID 3399311 | https://doi.org/10.1155/2022/3399311

Identification of ABCA5 among ATP-Binding Cassette Transporter Family as a New Biomarker for Colorectal Cancer

Peilong Bu,¹Yafei Xiao,¹Shaowen Hu,¹Xiaowei Jiang,²Cong Tan,³Mengyuan Qiu,³Wanting Huang,³Mengmeng Li,¹Quanying Li,¹and Changjiang Qin¹

Academic Editor: Mingjun Zheng

Received27 Apr 2022

Accepted25 May 2022

Published22 Jun 2022

Abstract

Background. The increasing incidence and mortality of colorectal cancer (CRC) urgently requires updated biomarkers. The ABC transporter family is a widespread family of membrane-bound proteins involved in the transportation of substrates associated with ATP hydrolysis, including metabolites, amino acids, peptides and proteins, sterols and lipids, organic and inorganic ions, sugars, metals, and drugs. They play an important role in the maintenance of homeostasis in the body. Purpose. This study aims to search for new markers in the ABC transporter gene family for diagnostic and prognostic purposes through data mining of The Cancer Genome Atlas (TCGA) and GEO (Gene Expression Omnibus) datasets. Methods. A total of 980 samples, including 684 CRC patients and 296 controls from five different datasets, were included for analysis. The construction of the PPI (protein-protein interaction) network and pathway analysis were performed in STRING database and DAVID (database for annotation, visualization, and integrated discovery), respectively. In addition, GSEA (gene set enrichment analysis) and WGCNA (weighted gene co-expression network analysis) were also used for functional analysis. Results. After several rounds of screening and validation, only the ABCB5 gene was retained among the 49 genes. Conclusions. The results demonstrated that ABCA5 expression is reduced in CRC and patients with high ABCA5 expression have better OS, which can provide guidance for better management and treatment of CRC in the future.

1. Introduction

Colorectal cancer (CRC), the third most commonly diagnosed cancer and the second leading cause of cancer death worldwide, was diagnosed in 1931,590 cases and 935,173 people died of colorectal cancer worldwide in 2021 [1]. Due to the early development of metastasis, the overall survival time of patients with CRC is generally less than 5 years [2]. Although CRC can be treated by surgery, radiotherapy, chemotherapy, immunotherapy, and other comprehensive treatment methods, the drug resistance of tumor and a series of side effects often offset the curative efficacy of the treatment schemes, resulting in a high recurrence rate (54.5%) [3] and high mortality rate (9.5%) [4]. Due to the lack of precise and effective molecular targets for CRC treatment, it remains important to explore new diagnostic, prognostic biomarkers.

The ABC transporter family is a kind of widespread membrane-bound protein, mainly distributed in the liver, intestines, blood-brain barrier, blood testosterone barrier, placenta, and kidney [5], and participates in the transportation of ATP hydrolysis-related substrates [6]. The energy obtained from ATP hydrolysis by ABC is used to overcome the concentration gradient and transport the substrates across the outer and inner membrane [7]. ABC proteins promote the transportation of heterogeneous substrates across cell membranes, including metabolites, amino acids, peptides and proteins, sterols and lipids, organic and inorganic ions, sugars, metals, and drugs [8]. The mechanism of ABC transporters plays an irreplaceable part in the formation of multidrug resistance [9].

ABC family genes are often used as a search for biomarkers due to their important biological functions. In the digestive system, Guo et al. found that ABCA8 could be used as a prognostic biomarker for gastric adenocarcinoma and was associated with immune invasion [10]. In the urinary system, our group has previously identified ABCG1 as a diagnostic and prognostic marker for clear cell renal cell carcinoma [11]. In this study, we are focusing on the expression of ABC genes in several datasets to evaluate their diagnostic and prognostic value in CRC.

2. Material and Methods

A total of 980 samples, including 684 CRC patients and 296 controls from five different datasets (GSE44076, GSE44861, GSE9348, TCGA-COADREAD, and GSE24551), were included for analysis. The further verification of mRNA level for the ABC gene was conducted on the Oncomine database (https://www.oncomine.org). Next, the prognostic analysis was performed in the TCGA Colon and Rectal Cancer (TCGA-COADREAD) dataset from UCSC Xena (https://xenabrowser.net). The construction of the PPI network and pathway analysis were performed in the STRING database and DAVID, respectively. In addition, gene set enrichment analysis (GSEA) and weighted correlation network analysis (WGCNA) were also used for functional analysis. Finally, immuno-infiltration analysis was conducted in the TIMER database (https://cistrome.shinyapps.io/timer/). The t-test is a simple, statistically based method for detecting differentially expressed genes [12]. The ROC curve is an analytical method, represented graphically, for assessing the overall diagnostic performance of a test and comparing the performance of two or more diagnostic tests [13]. Our screening conditions are progressively more lenient to more stringent, so the AUC threshold varies from 0.5 to 0.85, which naturally leads to a gradual reduction in the number of eligible genes. This serves two purposes: (1) to leave markers with high sensitivity and specificity; and (2) to reduce marker misses due to differences in the datasets. Different datasets, due to the existence of different instruments for the experiments, different proficiency of the experimenters, and even different reagents, can result in heterogeneity between different datasets (Figure S1). In addition, the validation steps are sequential, and some genes that fail to meet the conditions are eliminated, with those that meet all the conditions of the validation steps being retained in the end. A flow diagram for this study is shown in Figure 1.

2.1. Identification of Genes That Have Not Been Reported in CRC

Search for ABC family genes associated with CRC in PubMed (https://pubmed.ncbi.nlm.nih.gov/) on June 20, 2021. Our search strategy was as follows: (“colorectal cancer” OR “colon cancer” OR “colorectal carcinoma” OR “colorectal neoplasms” OR “colonic neoplasms” OR “rectal neoplasms” OR “CRC” OR “colon tumor” OR “rectal tumor”) AND (XXX), with XXX representing the specific ABC family gene. The ABC family genes that were not found to be markers for the diagnosis or prognosis of colorectal cancer were defined as novel genes. Then, these novel genes would be screened and identified as follows.

2.2. Screening of ABC Genes in the GEO Database

ROC analysis and t-test were conducted in the GSE44076 dataset (98 patients with CRC and 98 controls) and the GSE44861 dataset (56 CRC patients and 56 controls), respectively. In the ROC analysis, genes with area under the curve (AUC) > 0.50 and t-test were deemed statistically significant and will undergo the next round of verification.

2.3. Three Rounds of Validation

The first round of verification was to perform t-test and ROC analysis in the GSE9348 dataset (70 CRC patients and 12 control groups). Genes with and AUC > 0.85 were considered to have statistical significance. The second round of validation was conducted in the TCGA-COADREAD dataset (380 patients with CRC and 51 controls). Genes with value <0.05 in t-test and AUC >0.85 in ROC analysis were considered significant. The third round of validation was to analyse the transcriptional expression level of ABC genes of CRC patients on the Oncomine database. The threshold sets were as follows: ; multiple: 2; genetic rank: top 10%; data type: mRNA. It was considered to have obvious biological significance with the ratio >2 or <0.5.

2.4. Prognostic Analysis

To explore the potential value of ABC genes in CRC, survival analysis, descriptive statistics, and univariate and multivariate Cox regression analyses were performed in the TCGA-COADREAD dataset (380 patients with CRC and 51 controls). In the survival analysis, the correlation between gene expression levels and survival status was investigated. A chi-squared test was used to observe the correlation between gene expression level and CRC patients’ age, gender, clinical stage, tumor site, and living status. In univariate Cox regression analysis, we further explored the relationship between age, sex, clinical stage, tumor site, gene expression level, and CRC survival status. Parameters that were statistically significant in the univariate Cox regression analysis were involved in the multivariate Cox regression analysis. Finally, multivariate Cox regression analysis was used to identify whether gene expression level was an independent prognostic factor in CRC patients. was considered statistically significant in survival analysis, descriptive statistics, and univariate and multivariate Cox regression analyses.

2.5. Functional Analysis

2.5.1. PPI Network Construction and Functional Analysis

The STRING (https://string-db.org/) database is used to integrate all known and predicted connections between proteins, including physical interactions and functional associations. We performed PPI (protein-protein interaction) analysis for genes in the STRING database. The basic parameters were set as follows: organism: Homo sapiens; minimum required interaction score: medium confidence 0.4; 1st shell: no more than 50 interactors; and the rest of the parameters were default. The interaction relationship for related genes was finally obtained.

To explore the potential pathways and biological processes of the related genes, we used the annotation visualization and integrated discovery database (DAVID 6.8; https://david.ncifcrf.gov/) for GO pathway analysis of interacting genes obtained from STRING database. Cytoscape is one of the most successful tools for network biology analysis and visualization, and we used Cytoscape 3.8.0 (https://cytoscape.org/) tool to construct the PPI network of related genes. In addition, we presented the results of the GO analysis in bubble charts drawn by https://www.bioinformatics.com.cn—a free online platform for analysing and visualizing the data.

2.5.2. GSEA

GSEA was performed using GSEA v4.1.0 software, which was a knowledge-based software for interpreting whole-genome expression profiles and was available for free at https://www.gsea-msigdb.org. The parameters were set as follows: gene sets database: c2.cp.kegg.v7.4.symbols.gmt; number of permutations: 100; collapse: no_collapse; and permutation type: gene_set.

2.5.3. WGCNA and GO Pathway Analysis

WGCNA was conducted using the Sangerbox tools, a free online platform for data analysis (https://www.sangerbox.com/tool). The parameters were set as follows: network type: unsigned; screening genes: no; outlier sample: filter; soft threshold: 3; minimum module size: 60; sensitivity: 2; and module merge threshold: 0.25. The identified modular genes were analysed by the GO pathway in the DAVID database.

2.6. Immuno-Infiltration Analysis

The TIMER database (https://cistrome.shinyapps.io/timer/) can be used to systematically analyse the immune infiltrates of different cancer types. Here, we analysed the relationship between the expression level of the correlated genes and the abundance of immune infiltrates in colorectal cancer, including B cells, CD4+ T cells, CD8+ T cells, neutrophils, macrophages, and dendritic cells.

2.7. Statistical Analysis

ROC analysis and ROC curves, t-test, survival analysis, and survival curves were calculated and plotted by GraphPad prism 8.0. A chi-squared test and univariate and multivariate Cox regression analyses were performed in SPSS 20.0 software. In addition, the datasets were all grouped according to median gene expression level, and were considered statistically significant.

3. Results

3.1. Identification and Screening

A PubMed search was conducted on June 20, 2021, and we found a total of 27 genes that had not been studied with CRC (Table 1). t-test and ROC analysis were performed for the 27 genes in GSE44076 and GSE44861 datasets. When for t-test and AUC > 0.5 for ROC analysis were regarded as the limiting conditions, when t-test and ROC analysis AUC > 0.5 were the constraints, only seven ABC genes—ABCA5, ABCA8, ABCA10, ABCC1, ABCC13, ABCF2, and ABCF3—could enter the validation stages (Table S1).

3.2. First Round of Validation

The seven genes screened in the previous round were validated in the GSE9348 dataset. The result showed that there were 6 genes with and AUC > 0.85, namely ABCA5, ABCA8, ABCA10, ABCC1, ABCC13, and ABCF2. Among them, ABCA5, ABCA8, ABCC1, and ABCC13 in the t-test were with , indicating that the difference between the CRC group and the control group was highly significant (Table S2).

3.3. Second Round of Validation

The second round of validation for these 6 genes was performed in the TCGA-COADREAD dataset. The result showed that they could go to the next round of validation with and AUC >0.85 (Table 2), namely ABCA5, ABCA8, ABCA10, ABCC13, ABCC1, and ABCF2. Among them, five genes in the ROC analysis were with AUC > 0.90, indicating that their sensitivity and specificity are relatively high (Figure 2).

3.4. Third Round of Validation

The transcriptional expression level of these six genes was validated in the Oncomine database. The results of ABCA5, ABCA8, and ABCC1 were shown in Figure 2S. The figure showed that they were differentially expressed in 20 different types of tumors. ABCA5 and ABCA8 were lowly expressed in cancer tissues, and ABCC1 was highly expressed in colorectal cancer tissues. There were 12 datasets for ABCA5, 17 datasets for ABCA8, and 7 datasets for ABCC1 (Tables 3S–S5). No dataset was found about ABCA10, ABCC13, and ABCF2. These datasets could prove that the transcriptional expression levels of ABCA5, ABCA8, and ABCC1 are significantly different between the CRC group and the control group, and these three genes would enter the next round of verification.

3.5. Prognostic Analysis

The survival analysis of ABCA5, ABCA8, and ABCC1 genes was performed in the TCGA colon and rectal cancer (TCGA-COADREAD) and GSE24551-GPL5175 datasets. It was found that only ABCA5 had prognostic significance. In colorectal cancer patients, patients with high expression of ABCA5 had a better prognosis than those with low expression of ABCA5 (Figure 3). A chi-squared test was conducted on ABCA5 to determine whether the expression of ABCA5 was related to age, gender, and other clinical indicators. The results showed a significant correlation between ABCA5 gene expression level and the survival status of CRC patients (Table 3). To find out the relationship between patient survival status and clinical characteristics, we performed a univariate Cox regression analysis, including age, gender, clinical stage, T stage, M stage, N stage, tumor tissue site, and transcript expression level of ABCA5 gene. The result indicated that age, clinical stage, M stage, and ABCA5 expression level were significantly related to the prognosis (Table 4).

(a)

(b)

Multivariate Cox regression analysis was performed on age, clinical stage, M stage, and ABCA5 expression level to determine whether they were independent prognostic factors. The result showed that they were all independent prognostic factors (Table 4). The hazard ratio (HR) for age was 2.866, indicating that the prognosis of CRC patients older than 60 years was worse than that of younger. The hazard ratio for the clinical stage was 2.407, which illustrated that patients with stage III/IV showed a worse prognosis compared to stage I/II. The hazard ratio for the M stage was 2.841, which suggested that patients with distant metastases have a worse prognosis than those without metastases. The hazard ratio for ABCA5 expression was 0.491, revealing that patients with high expression had a longer survival period than patients with low expression.

3.6. Functional Analysis

3.6.1. PPI Network Construction and Functional Analysis for ABCA5

According to the settings, we obtained 31 genes in the STRING database that have direct interaction with ABCA5 (Figure 4(a)). The thickness of the line represents the strength of the relationship, and the direction of the arrow represents the role of the relationship. The figure shows that there are 10 genes as the source genes of ABCA5 and 21 genes as the target genes of ABCA5. Among them, ELP6 and CCDC137, as the target genes of ABCA5, are more closely related to ABCA5 (Figure 4(a)).

(a)

(b)

To further understand the biological processes in which ABCA5 may be involved, we performed GO enrichment analysis on these 31 genes. The result of the biological process (BP) indicated that ABCA5 is involved in lipid transport and cholesterol efflux, etc (Figure 4(b), Table S6).

3.6.2. GSEA for ABCA5

To understand how ABCA5 plays a role in the pathway, we did GSEA with TCGA-COADREAD as the background dataset. A total of 176 pathways were enriched, 53 pathways with and FDR< 0.25, including 33 pathways associated with ABCA5 low expression activation and 20 pathways associated with ABCA5 high expression activation. We focused on three pathways for analysis: unsaturated fatty acid synthesis pathway, ABC transport pathway, and drug metabolism pathway. In the unsaturated fatty acid synthesis pathway, the peak was in the left low expression group, indicating that the fatty acid synthesis pathway was activated when ABCA5 was expressed at a low level (normalized enrichment score, NES = 1.89) (Figure 5(a)). In the ABC transport pathway, the peak was in the high expression group on the right, indicating that it was activated when ABCA5 was expressed at a high level (NES = −1.72) (Figure 5(b)). In the drug metabolism pathway, the peak was in the high expression group on the right, indicating that it was activated when ABCA5 was expressed at a high level (NES = −1.70) (Figure 5(c)).

(a)

(b)

(c)

3.6.3. WGCNA and GO Analysis for ABCA5

To discover more potential pathways, we performed WGCNA on ABCA5. Gene modules associated with ABCA5 expression in CRC were identified, and then key modules were selected for GO analysis. Specifically, the network topology under different soft threshold powers was first analysed so that the WGCNA had relatively balanced scale independence and average connectivity. A β = 3 power was chosen as a soft threshold in order to ensure that we developed a scale-free network (Figure 6(a)). The average linkage method and Pearson correlation analysis were used to cluster the sample of 383 CRC patients (Figure 6(b)). Next, 24 different gene modules were identified after setting the sensitivity to 2 and combining the modules with less than 25% difference (Figures 6(c) and 6(d)). Finally, a yellow module was identified based on the highest correlation of the module with ABCA5 expression (Cor = 0.36) (Figures 6(d) and 6(e)).

(a)

(b)

(c)

(d)

(e)

(f)

Figure 6

Identification of modules associated with the ABCA5 expression in the TCGA-COADREAD dataset. (a) Determination of the optimal soft threshold. (b) The cluster dendrogram of co-expression network modules was ordered by a hierarchical clustering of genes based on the 1-TOM matrix. Each module was assigned to different colors. (c) Heat map of module feature vector clustering. (d) Module-trait relationships. Each row corresponds to a color module and column corresponds to a clinical trait. Each cell contains the corresponding correlation and value. (e) Scatterplot of GS vs. MM correlation for yellow module (the corresponding correlation and value). (f) GO analysis for the hub genes of the yellow module.

Subsequently, we extracted 274 hub genes from the yellow module by setting the MM (module membership) threshold to 0.60 and the GS (gene significance) threshold to 0.1. GO enrichment analysis of these genes revealed that they were enriched to 14 biological process (BP) pathways, including transcriptional regulation and transport of fatty acids, leukotriene signalling and macrophage regulation, among other pathways (Figure 6(f)).

3.7. Immune Infiltration Analysis for ABCA5

The results of immune infiltration analysis showed the correlation between ABCA5 expression and immune infiltration of colorectal cancer. A positive association was found between the expression level of ABCA5 and the infiltration levels of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells in colon cancer tissues (). In rectal cancer, the expression level of ABCA5 was positively correlated with the infiltration level of B cells and CD8+ T cells, while it was not significantly different from the infiltration level of CD4+ T cells, macrophages, and neutrophils (Figure 7).

In addition, we have done a pan-cancer analysis of ABCA5 with immune cells. It can be seen from the figure that ABCA5 is involved in the immune response to a variety of cancers, including BLCA (bladder cancer), COAD (colon adenocarcinoma), BRCA (breast cancer), KICH (kidney chromophobe), KIRC (kidney renal clear cell carcinoma), THCA (thyroid carcinoma), THYM (thymoma), and UCEC (uterine corpus endometrial carcinoma), with strong relevance to immune cells, including KICH, TGCT (tenosynovial giant cell tumor), and THYM (Figure S3).

4. Discussion

Many studies have demonstrated that ABC family genes have an important position in the development of diseases [14, 15], cholesterol metabolism [5, 16], and drug resistance [17]. Forty-nine human ABC transporter proteins, about half of which are considered to be involved in the transport of lipids and lipid-related compounds [15]. Jamie et al. proposed that the role of ABC transporter proteins in tumor biology is linked, such as conferring resistance to drug or cancer-related substrates (i.e., phospholipids and cholesterol) [18]. The ABCA family consists of 12 isoforms and is the second largest ABC gene family after the ABCC family [19]. Evolutionary analysis revealed that the cluster of genes encoding ABCA5, ABCA6, ABCA8, ABCA9, and ABCA10, referred to as ABCA5-like transporter genes, clustered on chromosome 17q24 [19, 20]. Comparative genomic analysis showed that the ABCA5 gene is relatively evolutionarily conserved, indicating the importance and stability of ABCA5 in organisms [21, 22]. ABCA5 gene has been found to act as a biomarker for a variety of diseases. For example, it can be used as a diagnostic marker for prostate cancer and melanoma [23, 24], a key prognostic molecular marker for ovarian cancer, osteosarcoma, and childhood acute myeloid leukemia (AML) [24–26]. After screening, identification, and multiple rounds of validation, ABCA5 was screened out from 49 ABC transporter protein genes. It was the first time we demonstrated the diagnostic and prognostic value of ABCA5 in patients with colorectal cancer.

The results of functional analysis from different perspectives suggested that ABCA5 may be involved in the synthesis of unsaturated fatty acids, cholesterol efflux, and drug metabolism. In recent years, the ABCA5 gene is reported to be involved in the development and treatment of many diseases, mainly in the aspects of drug resistance and cholesterol transport. ABCA5 is engaged in resistance processes in the treatment of many diseases, including resistance in melanoma [27], resistance to the immunosuppressant tacrolimus [28], resistance to 5-FU in laryngeal squamous cell carcinoma [29], resistance to arabinoside (Ara-C) and erythromycin (Dnr) in acute myeloid leukemia [30], resistance to doxorubicin (Dox) in malignant mesothelioma (MM) [31], and resistance to cisplatin in lung cancer [32]. Diseases associated with ABCA5 cholesterol transport function include Alzheimer’s disease [33], Parkinson’s disease [34], ST-segment elevation myocardial infarction [35], atherosclerosis [36, 37], and excessive hair growth [22]. Notably, the overexpression of ABCA5 is a protective response in these diseases, and upregulation protects cells from the accumulation of intracellular cholesterol and other sterols [22]. The increased expression level of ABCA5 promotes cholesterol outflow and tends to maintain the balance of cholesterol transport [35, 38].

The accumulation of cholesterol is a well-known feature of cancer [39]. To meet their continuous growth and proliferation needs, cancer cells would increase the uptake of exogenous lipids or upregulate the synthesis of endogenous cholesterol [40]. Lipid droplets are dynamic organelles that store triglycerides (TG) and cholesterol esters (CE) and are highly accumulated in colon cancer cells [41]. There is a consensus that high cholesterol diet and high level of total serum cholesterol increase the risk of colorectal cancer [42–45]. Cholesterol is metabolized to active derivatives, including cholesterol oxidation products (COP), called oxysterols, which have been shown to alter cell proliferation [46, 47]. For example, a high concentration of 27-hydroxycholesterol (27HC), a kind of oxysterol, significantly increases the release of pro-inflammatory interleukins 6 and 8 and exerts tumor-promoting properties in colorectal cancer [48, 49]. The ABCA5 gene assists in the transport of cellular cholesterol and is involved in the process of lipid metabolism [50]. The expression of ABCA5 is strongly induced in the presence of cholesterol, which promotes intracellular cholesterol efflux [38, 50], and the effect is evident in macrophages [36, 50, 51]. Macrophages are involved in the formation of tumor microenvironment [52, 53], growth and metastasis [54–57], and apoptosis [58] and play a critical role in tumor immunity [59]. Cholesterol can modulate the expression of inflammatory cytokines, chemokines, and lymphocyte proliferation in macrophages [60]. This indicates that ABCA5 may have an important effect in the development of colorectal cancer through cholesterol and macrophages. It is consistent with the results of our previous functional analysis and immune infiltration analysis.

Our study showed that ABCA5 was down-expressed in colorectal cancer patient tissues compared to normal controls. This is consistent with the expression pattern of ABCA5 in breast cancer [61], suggesting that ABCA5 plays an oncogene role in colorectal and breast cancers. In addition, in colorectal cancer patients, low expression level of ABCA5 was associated with shorter overall survival, similar to the expression pattern of ABCA5 in plasmacytic ovarian cancer [25]. High expression was associated with a better prognosis, suggesting that ABCA5 overexpression was a protective response in colorectal cancer. This may be due to the fact that ABCA5 is strongly induced by cholesterol accumulation and begins to express and act as an oncogene. Perhaps it can be understood that the high expression of ABCA5 is to maintain the balance of cholesterol transport, which would reduce the cholesterol content in cancer cells. Wu et al. suggested that promoting cholesterol efflux inhibits cancer cell growth and affects cancer cell survival [62]. Lipid metabolism is extensively mediated by the LXR (liver X receptor) and PPAR (proliferator-activated receptor) transcription factor families [63]. Ray et al. found that ABCA5 is a PPAR family regulatory protein with high specificity for the γ isoform. The PPARγ agonists, troglitazone and rosiglitazone, resulted in a robust increase of ABCA5 expression level in macrophages. In addition, they found that the lipid-lowering drugs, fenofibrate and atorvastatin, also increased the mRNA level of ABCA5 in macrophages [50]. Some studies have found promising results for statins in the prevention and treatment of colorectal cancer [64–66]. Statins increase the expression of ABCA5 and are associated with longer patient survival [64, 67]. The topics that are worth exploring in the future include what are the mechanisms involved?; and can other drugs that increase the ABCA5 expression level also prolong the survival of CRC patients?

In addition, 66 potentially active drugs were identified from 119 substances interacting with ABCA5 in the Comparative Toxicogenomics Database (CTD) (https://ctdbase.org). The expression of ABCA5 could be upregulated at the mRNA level by 34 of these drugs, whereas 32 drugs could decrease the expression of ABCA5 (Table S7). Moreover, five ABCA5 source genes were identified among the seven ABCA5-interacting genes: ABCA13 [68], ABCA6 [68], ABCA8A [68], ABCA8B [68], and PPARG [69] (Table S8). We hope that these drugs and genes will serve as a reference for future researchers.

5. Conclusion

After one round of identification, two rounds of screening, and three rounds of validation, we identified a new diagnostic and prognostic biomarker for colorectal cancer at the mRNA expression level, namely ABCA5. Based on the low expression of ABCA5 in CRC and the correlation with poor prognosis, it may contribute to the diagnosis and provide new ideas for the development of molecularly targeted drugs in CRC.

Data Availability

The data used to support the findings of this study are included within the article.

Ethical Approval

The authors declare that they have no ethical approval.

The authors assure that the material is original, and it has not been published elsewhere yet.

Disclosure

Peilong Bu, Yafei Xiao, and Shaowen Hu are the co-first authors.

Conflicts of Interest

The authors declare that they have no competing interest.

Authors’ Contributions

CQ contributed to the conception and design of this article. PB, YX, and SH downloaded and organized related papers. PB, YX, SH, XJ, and CT contributed to all tables and figures and the main manuscript. MQ, WH, ML, and QL revised and supplemented the manuscript. All of the authors read and approved the final manuscript.

Acknowledgments

The authors thank Professor Longxiang Xie, Institute of Biomedical Informatics, School of Basic Medical Sciences, Henan University, for his informatics and statistical advice and support for this article. This study was supported by the National Natural Science Foundation of China (NSFC–U1504818), Science and Technology Foundation of Henan Province (172102310152, SBGJ202002097, 192102310099), Natural Science Foundation of Henan Province (182300410359), Henan Provincial Education Fund (19A320020), and Henan University Graduate Program of Excellence (SYL20060192).

Supplementary Materials

Tables 1S–8S and Figures 1S–3S have been added to Supplementary Materials, which are freely available in Supplementary materials. Table S1: t-test and ROC analysis of ABC transporter family members according to the GSE44861 database. Table S2: t-test and ROC analysis of ABC transporter family members according to the GSE9348 database. Table S3: ABCA5 expression analysis according to the Oncomine database. Table S4: ABCA8 expression analysis according to the Oncomine database. Table S5: ABCC1 expression analysis according to the Oncomine database. Table S6: Results of GO enrichment analysis in biological process (BP). Table S7: Chemical interactions of ABCA5 according to CTD database. Table S8: Target genes of ABCA5 according to CTD database. Figure S1: Distribution of the ABCA5 gene in different datasets. Figure S2: The mRNA expression of ABCA5, ABCA8, and ABCC1 in different types of cancer by Oncomine analysis. (The number in the colored cell represents the number of analyses meeting thresholds. The cell color is determined by the gene rank. The more intense red (overexpression) or blue (underexpression) indicates a more highly significant overexpressed or underexpressed gene.) Figure S3: Pan-cancer analysis of ABCA5 and immune cells. (Immune checkpoint: The heat map of immune checkpoint-related gene expression. The abscissa represents different immune checkpoint-related gene, and the ordinate represents different tumor tissues. Each box in the figure represents the correlation analysis between the expression of the selected gene and the immune checkpoint in corresponding tumors. , , , asterisks stand for significance levels. Different colors represent the changes of correlation coefficients. Immune correlations: The heat map of ABCA5 immune score and ABCA5 gene expression in multiple tumor tissues. The abscissa represents different tumor tissues, and the ordinate represents different immune score. Different colors represent the correlation coefficients. Negative values indicate negative correlations and positive values indicate positive correlations; and the deeper the color, the stronger the correlation. , , , asterisks stand for significance levels. The statistical difference between two groups was compared through the Wilcoxon test). (Supplementary Materials)

References

R. L. Siegel, K. D. Miller, H. E. Fuchs, and A. Jemal, “Cancer statistics, 2021,” CA: A Cancer Journal for Clinicians, vol. 71, no. 1, pp. 7–33, 2021.
View at: Publisher Site | Google Scholar
P. Landreau, A. Drouillard, G. Launoy et al., “Incidence and survival in late liver metastases of colorectal cancer,” Journal of Gastroenterology and Hepatology, vol. 30, no. 1, pp. 82–85, 2015.
View at: Publisher Site | Google Scholar
H. Asano, K. Kojima, N. Ogino, H. Fukano, Y. Ohara, and N. Shinozuka, “Postoperative recurrence and risk factors of colorectal cancer perforation,” International Journal of Colorectal Disease, vol. 32, no. 3, pp. 419–424, 2017.
View at: Publisher Site | Google Scholar
F. Bray, J. Ferlay, I. Soerjomataram, R. L. Siegel, L. A. Torre, and A. Jemal, “Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA: A Cancer Journal for Clinicians, vol. 68, no. 6, pp. 394–424, 2018.
View at: Publisher Site | Google Scholar
V. Vasiliou, K. Vasiliou, and D. W. Nebert, “Human ATP-binding cassette (ABC) transporter family,” Human Genomics, vol. 3, no. 3, p. 281, 2008.
View at: Publisher Site | Google Scholar
S. Ohtsuki, M. Kamoi, Y. Watanabe, H. Suzuki, S. Hori, and A. Terasaki, “Correlation of induction of ATP binding cassette transporter A5 (ABCA5) and ABCB1 mRNAs with differentiation state of human colon tumor,” Biological and Pharmaceutical Bulletin, vol. 30, no. 6, pp. 1144–1146, 2007.
View at: Publisher Site | Google Scholar
C. F. Higgins, “ABC transporters: from microorganisms to man,” Annual Review of Cell Biology, vol. 8, pp. 67–113, 1992.
View at: Publisher Site | Google Scholar
J. L. Mendoza, E. Urcelay, R. Lana et al., “MDR1 polymorphisms and response to azathioprine therapy in patients with Crohnʼs disease,” Inflammatory Bowel Diseases, vol. 13, no. 5, pp. 585–590, 2007.
View at: Publisher Site | Google Scholar
S. Cole, G. Bhardwaj, J. Gerlach et al., “Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line,” Science, vol. 258, no. 5088, pp. 1650–1654, 1992.
View at: Publisher Site | Google Scholar
Y. Guo, Z. W. Wang, W. H. Su, J. Chen, and Y. L. Wang, “Prognostic value and immune infiltrates of ABCA8 and FABP4 in stomach adenocarcinoma,” BioMed Research International, vol. 2020, Article ID 4145164, 12 pages, 2020.
View at: Publisher Site | Google Scholar
F. Meng, Y. Xiao, L. Xie, and Q. K. Liu, “Diagnostic and prognostic value of ABC transporter family member ABCG1 gene in clear cell renal cell carcinoma,” Channels, vol. 15, no. 1, pp. 375–385, 2021.
View at: Publisher Site | Google Scholar
X. Cui and G. A. Churchill, “Statistical tests for differential expression in cDNA microarray experiments,” Genome Biology, vol. 4, no. 4, p. 210, 2003.
View at: Publisher Site | Google Scholar
F. S. Nahm, “Receiver operating characteristic curve: overview and practical use for clinicians,” Korean Journal of Anesthesiology, vol. 75, no. 1, pp. 25–36, 2022.
View at: Publisher Site | Google Scholar
T. Schumacher and R. A. Benndorf, “ABC transport proteins in cardiovascular disease-a brief summary,” Molecules, vol. 22, no. 4, p. 589, 2017.
View at: Publisher Site | Google Scholar
E. J. Tarling, T. Q. d A. Vallim, and P. A. Edwards, “Role of ABC transporters in lipid transport and human disease,” Trends in Endocrinology and Metabolism, vol. 24, no. 7, pp. 342–350, 2013.
View at: Publisher Site | Google Scholar
M. Dean, Y. Hamon, and G. Chimini, “The human ATP-binding cassette (ABC) transporter superfamily,” Genome Research, vol. 11, no. 7, pp. 1156–1166, 2001.
View at: Publisher Site | Google Scholar
O. Briz, L. Perez-Silva, R. Al-Abdulla et al., “What “the cancer genome atlas” database tells us about the role of ATP-binding cassette (ABC) proteins in chemoresistance to anticancer drugs,” Expert Opinion on Drug Metabolism and Toxicology, vol. 15, no. 7, pp. 577–593, 2019.
View at: Publisher Site | Google Scholar
J. I. Fletcher, M. Haber, M. J. Henderson, and M. D. Norris, “ABC transporters in cancer: more than just drug efflux pumps,” Nature Reviews Cancer, vol. 10, no. 2, pp. 147–156, 2010.
View at: Publisher Site | Google Scholar
T. Annilo, Z.-Q. Chen, S. Shulenin, and M. Dean, “Evolutionary analysis of a cluster of ATP-binding cassette (ABC) genes,” Mammalian Genome, vol. 14, no. 1, pp. 7–20, 2003.
View at: Publisher Site | Google Scholar
F. Peelman, C. Labeur, B. Vanloo et al., “Characterization of the ABCA transporter subfamily: identification of prokaryotic and eukaryotic members, phylogeny and topology,” Journal of Molecular Biology, vol. 325, no. 2, pp. 259–274, 2003.
View at: Publisher Site | Google Scholar
G. Li, P. Shi, and Y. Wang, “Evolutionary dynamics of the ABCA chromosome 17q24 cluster genes in vertebrates,” Genomics, vol. 89, no. 3, pp. 385–391, 2007.
View at: Publisher Site | Google Scholar
G. M. DeStefano, M. Kurban, K. Anyane-Yeboa et al., “Mutations in the cholesterol transporter gene ABCA5 are associated with excessive hair overgrowth,” PLoS Genetics, vol. 10, no. 5, Article ID e1004333, 2014.
View at: Publisher Site | Google Scholar
O. F. Karatas, E. Guzel, M. B. Duz, and M. M. Ittmann, “The role of ATP-binding cassette transporter genes in the progression of prostate cancer,” The Prostate, vol. 76, no. 5, pp. 434–444, 2016.
View at: Publisher Site | Google Scholar
I. Vásquez-Moctezuma, M. A. Meraz-Ríos, C. G. Villanueva-López et al., “ATP-binding cassette transporter ABCB5 gene is expressed with variability in malignant melanoma,” Actas Dermo-Sifiliográficas, vol. 101, no. 4, pp. 341–348, 2010.
View at: Publisher Site | Google Scholar
E. L. Hedditch, B. Gao, A. J. Russell et al., “ABCA transporter gene expression and poor outcome in epithelial ovarian cancer,” Journal of the National Cancer Institute: Journal of the National Cancer Institute, vol. 106, no. 7, 2014.
View at: Publisher Site | Google Scholar
P. Niu, B. Yao, L. Wei, H. Zhu, C. Fang, and Y. Zhao, “Construction of prognostic risk prediction model based on high-throughput sequencing expression profile data in childhood acute myeloid leukemia,” Blood Cells, Molecules, and Diseases, vol. 77, pp. 43–50, 2019.
View at: Publisher Site | Google Scholar
V. Alla, B. S. Kowtharapu, D. Engelmann et al., “E2F1 confers anticancer drug resistance by targeting ABC transporter family members and Bcl-2 via the p73/DNp73-miR-205 circuitry,” Cell Cycle, vol. 11, no. 16, pp. 3067–3078, 2012.
View at: Publisher Site | Google Scholar
C. A. Quezada, W. X. Garrido, M. A. González-Oyarzún et al., “Effect of tacrolimus on activity and expression of P-glycoprotein and ATP-binding cassette transporter A5 (ABCA5) proteins in hematoencephalic barrier cells,” Biological and Pharmaceutical Bulletin, vol. 31, no. 10, pp. 1911–1916, 2008.
View at: Publisher Site | Google Scholar
M. B. Duz and O. F. Karatas, “Expression profile of stem cell markers and ABC transporters in 5-fluorouracil resistant Hep-2 cells,” Molecular Biology Reports, vol. 47, no. 7, pp. 5431–5438, 2020.
View at: Publisher Site | Google Scholar
S. Varatharajan, A. Abraham, S. Karathedath et al., “ATP-binding cassette transporter expression in acute myeloid leukemia: association with in vitro cytotoxicity and prognostic markers,” Pharmacogenomics, vol. 18, no. 3, pp. 235–244, 2017.
View at: Publisher Site | Google Scholar
A. Shukla, J. M. Hillegass, M. B. MacPherson et al., “Blocking of ERK1 and ERK2 sensitizes human mesothelioma cells to doxorubicin,” Molecular Cancer, vol. 9, no. 1, p. 314, 2010.
View at: Publisher Site | Google Scholar
Y. Gao, W. Li, X. Liu, F. X. Gao, and X. Zhao, “Reversing effect and mechanism of soluble resistance-related calcium-binding protein on multidrug resistance in human lung cancer A549/DDP cells,” Molecular Medicine Reports, vol. 11, no. 3, pp. 2118–2124, 2015.
View at: Publisher Site | Google Scholar
Y. Fu, J. H. T. Hsiao, G. Paxinos, G. M. Halliday, and W. S. Kim, “ABCA5 regulates amyloid-β peptide production and is associated with Alzheimer’s disease neuropathology,” Journal of Alzheimer’s Disease, vol. 43, no. 3, pp. 857–869, 2014.
View at: Publisher Site | Google Scholar
W. S. Kim and G. M. Halliday, “Changes in sphingomyelin level affect alpha-synuclein and ABCA5 expression,” Journal of Parkinson’s Disease, vol. 2, no. 1, pp. 41–46, 2012.
View at: Publisher Site | Google Scholar
A. A. Das, K. R. Choudhury, M. G. Jagadeeshaprasad, M. J. Kulkarni, P. C. Mondal, and A. Bandyopadhyay, “Proteomic analysis detects deregulated reverse cholesterol transport in human subjects with ST-segment elevation myocardial infarction,” Journal of Proteomics, vol. 222, Article ID 103796, 2020.
View at: Publisher Site | Google Scholar
D. Ye, I. Meurs, M. Ohigashi et al., “Macrophage ABCA5 deficiency influences cellular cholesterol efflux and increases susceptibility to atherosclerosis in female LDLr knockout mice,” Biochemical and Biophysical Research Communications, vol. 395, no. 3, pp. 387–394, 2010.
View at: Publisher Site | Google Scholar
T. Shimizu, S. i Miura, H. Tanigawa et al., “Rosuvastatin activates ATP-binding cassette transporter A1-dependent efflux ex vivo and promotes reverse cholesterol transport in macrophage cells in mice fed a high-fat diet,” Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 34, no. 10, pp. 2246–2253, 2014.
View at: Publisher Site | Google Scholar
G. Szakacs and R. Abele, “An inventory of lysosomal ABC transporters,” FEBS Letters, vol. 594, no. 23, pp. 3965–3985, 2020.
View at: Publisher Site | Google Scholar
B. Sharma, V. Gupta, D. Dahiya, H. Kumar, K. Vaiphei, and N. Agnihotri, “Clinical relevance of cholesterol homeostasis genes in colorectal cancer,” Biochimica et Biophysica Acta (BBA)—Molecular and Cell Biology of Lipids, vol. 1864, no. 10, pp. 1314–1327, 2019.
View at: Publisher Site | Google Scholar
E. Piccinin, M. Cariello, and A. Moschetta, “Lipid metabolism in colon cancer: role of liver X receptor (LXR) and stearoyl-CoA desaturase 1 (SCD1),” Molecular Aspects of Medicine, vol. 78, Article ID 100933, 2021.
View at: Publisher Site | Google Scholar
R. Munir, J. Lisec, C. Jaeger, and N. Zaidi, “Abundance, fatty acid composition and saturation index of neutral lipids in colorectal cancer cell lines,” Acta Biochimica Polonica, vol. 68, no. 1, pp. 115–118, 2021.
View at: Publisher Site | Google Scholar
P. J. T. López, J. S. Albero, and J. A. Rodríguez-Montes, “Primary and secondary prevention of colorectal cancer,” Clinical Medicine Insights: Gastroenterology, vol. 7, 2014.
View at: Publisher Site | Google Scholar
R. Tayyem, H. Bawadi, I. Shehadah et al., “Macro-and micronutrients consumption and the risk for colorectal cancer among Jordanians,” Nutrients, vol. 7, no. 3, pp. 1769–1786, 2015.
View at: Publisher Site | Google Scholar
H. Rodriguez‐Broadbent, P. J. Law, A. Sud et al., “Mendelian randomisation implicates hyperlipidaemia as a risk factor for colorectal cancer,” International Journal of Cancer, vol. 140, no. 12, pp. 2701–2708, 2017.
View at: Publisher Site | Google Scholar
V. Ulaganathan, M. Kandiah, and Z. Shariff, “A case-control study on the association of abdominal obesity and hypercholesterolemia with the risk of colorectal cancer,” Journal of Carcinogenesis, vol. 17, no. 1, p. 4, 2018.
View at: Publisher Site | Google Scholar
J. Warns, G. Marwarha, N. Freking, and O. Ghribi, “27-hydroxycholesterol decreases cell proliferation in colon cancer cell lines,” Biochimie, vol. 153, pp. 171–180, 2018.
View at: Publisher Site | Google Scholar
D. Rossin, S. Calfapietra, B. Sottero, and G. F. Poli, “HNE and cholesterol oxidation products in colorectal inflammation and carcinogenesis,” Free Radical Biology and Medicine, vol. 111, pp. 186–195, 2017.
View at: Publisher Site | Google Scholar
D. Rossin, I. H. K. Dias, M. Solej et al., “Increased production of 27-hydroxycholesterol in human colorectal cancer advanced stage: possible contribution to cancer cell survival and infiltration,” Free Radical Biology and Medicine, vol. 136, pp. 35–44, 2019.
View at: Publisher Site | Google Scholar
L. Ma, W. Cho, and E. R. Nelson, “Our evolving understanding of how 27-hydroxycholesterol influences cancer,” Biochemical Pharmacology, vol. 196, Article ID 114621, 2022.
View at: Publisher Site | Google Scholar
A. G. Ray, K. R. Choudhury, S. Chakraborty et al., “Novel mechanism of cholesterol transport by ABCA5 in macrophages and its role in dyslipidemia,” Journal of Molecular Biology, vol. 432, no. 17, pp. 4922–4941, 2020.
View at: Publisher Site | Google Scholar
D. Ye, M. Hoekstra, R. Out et al., “Hepatic cell-specific ATP-binding cassette (ABC) transporter profiling identifies putative novel candidates for lipid homeostasis in mice,” Atherosclerosis, vol. 196, no. 2, pp. 650–658, 2008.
View at: Publisher Site | Google Scholar
C. E. Lewis and J. W. Pollard, “Distinct role of macrophages in different tumor microenvironments,” Cancer Research, vol. 66, no. 2, pp. 605–612, 2006.
View at: Publisher Site | Google Scholar
K. Movahedi, D. Laoui, C. Gysemans et al., “Different tumor microenvironments contain functionally distinct subsets of macrophages derived from Ly6C (high) monocytes,” Cancer Research, vol. 70, no. 14, pp. 5728–5739, 2010.
View at: Publisher Site | Google Scholar
J. W. Pollard, “Tumour-educated macrophages promote tumour progression and metastasis,” Nature Reviews Cancer, vol. 4, no. 1, pp. 71–78, 2004.
View at: Publisher Site | Google Scholar
E. Y. Lin, J.-F. Li, L. Gnatovskiy et al., “Macrophages regulate the angiogenic switch in a mouse model of breast cancer,” Cancer Research, vol. 66, no. 23, pp. 11238–11246, 2006.
View at: Publisher Site | Google Scholar
A. Sica, P. Larghi, A. Mancino et al., “Macrophage polarization in tumour progression,” Seminars in Cancer Biology, vol. 18, no. 5, pp. 349–355, 2008.
View at: Publisher Site | Google Scholar
K. Hiraoka, M. Zenmyo, K. Watari et al., “Inhibition of bone and muscle metastases of lung cancer cells by a decrease in the number of monocytes/macrophages,” Cancer Science, vol. 99, no. 8, pp. 1595–1602, 2008.
View at: Publisher Site | Google Scholar
Y. Li, R. F. Schwabe, T. DeVries-Seimon et al., “Free cholesterol-loaded macrophages are an abundant source of tumor necrosis factor-α and interleukin-6,” Journal of Biological Chemistry, vol. 280, no. 23, pp. 21763–21772, 2005.
View at: Publisher Site | Google Scholar
G. Solinas, G. Germano, A. Mantovani, and P. Allavena, “Tumor-associated macrophages (TAM) as major players of the cancer-related inflammation,” Journal of Leukocyte Biology, vol. 86, no. 5, pp. 1065–1073, 2009.
View at: Publisher Site | Google Scholar
L. Yvan-Charvet, N. Wang, and A. R. Tall, “Role of HDL, ABCA1, and ABCG1 transporters in cholesterol efflux and immune responses,” Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 30, no. 2, pp. 139–143, 2010.
View at: Publisher Site | Google Scholar
V. Hlaváč, V. Brynychová, R. Václavíková et al., “The expression profile of ATP-binding cassette transporter genes in breast carcinoma,” Pharmacogenomics, vol. 14, no. 5, pp. 515–529, 2013.
View at: Publisher Site | Google Scholar
G. Wu, Q. Wang, Y. Xu et al., “Targeting the transcription factor receptor LXR to treat clear cell renal cell carcinoma: agonist or inverse agonist?” Cell Death and Disease, vol. 10, no. 6, p. 416, 2019.
View at: Publisher Site | Google Scholar
A. Remmerie and C. L. Scott, “Macrophages and lipid metabolism,” Cellular Immunology, vol. 330, pp. 27–42, 2018.
View at: Publisher Site | Google Scholar
B. A. Al-Husein, B. Dawah, S. Bani-Hani, S. M. Al Bashir, K. M. Al-Sawalmeh, and N. M. Ayoub, “Immunomodulatory effect of statins on regulatory T lymphocytes in human colorectal cancer is determined by the stage of disease,” Oncotarget, vol. 9, no. 87, pp. 35752–35761, 2018.
View at: Publisher Site | Google Scholar
C. R. Cardwell, B. M. Hicks, C. Hughes, and L. J. Murray, “Statin use after colorectal cancer diagnosis and survival: a population-based cohort study,” Journal of Clinical Oncology, vol. 32, no. 28, pp. 3177–3183, 2014.
View at: Publisher Site | Google Scholar
Y. K. Chae, M. Yousaf, M. K. Malecek et al., “Statins as anti-cancer therapy; can we translate preclinical and epidemiologic data into clinical benefit?” Discovery Medicine, vol. 20, no. 112, pp. 413–427, 2015.
View at: Google Scholar
Y. Ling, L. Yang, H. Huang et al., “Prognostic significance of statin use in colorectal cancer,” Medicine, vol. 94, no. 25, 2015.
View at: Publisher Site | Google Scholar
R. Pourhaghighi, P. E. A. Ash, S. Phanse et al., “BraInMap elucidates the macromolecular connectivity landscape of mammalian brain,” Cell Systems, vol. 10, no. 4, pp. 333–350, 2020.
View at: Publisher Site | Google Scholar
K. W. Lee, S. H. Kwak, Y. D. Koo et al., “F-box only protein 9 is an E3 ubiquitin ligase of PPARγ,” Experimental and Molecular Medicine, vol. 48, no. 5, 2016.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Peilong Bu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

563

Downloads

681

Citations