Pan-cancer analysis of ABCC1 as a potential prognostic and immunological biomarker

Highlights

• •ABCC1 expression exhibited intricate associations with diverse immune-related genes.
• •ScRNA-seq analysis from GEO revealed a positive correlation between ABCC1 expression and macrophage infiltration.
• •Various vitro and vivo experiments substantiated the oncogenic role of ABCC1 in hepatocellular carcinoma.

Abstract

ABCC1 belongs to the ATP-binding cassette (ABC) superfamily, which encompasses a total of 48 constituent members. ABCC1 has been shown to be associated with the growth, progression, and drug resistance of various types of cancer. However, the impact of ABCC1 on cancer immune infiltration and pan-cancer prognosis has been rarely studied. Our comprehensive pan-cancer analysis unveiled elevated ABCC1 expression across various cancers. ABCC1 overexpression consistently predicted unfavorable outcomes based on TCGA data. Moreover, ABCC1 expression exhibited intricate associations with diverse immune-related genes and demonstrated a close correlation with immune scores across multiple tumor types. Analysis of scRNA-seq data from the GEO database revealed that the expression of ABCC1 in hepatocellular carcinoma (HCC) cells is significant positively correlated with macrophage infiltration. Furthermore, various in vitro and in vivo experiments substantiated the role of ABCC1 in promoting the progression of HCC, along with increased macrophage recruitment. Based on the results, we propose ABCC1 as a potentially valuable prognostic indicator and a prospective target for immune-based cancer therapies.

Introduction

Based on their sequence homology, the 48 human ABC transporters are classified into seven subfamilies (ABCA to ABCG) [1,2]. ABC transporters exhibit a distinctive four-domain structure, comprising two cytoplasmic nucleotide-binding domains (NBDs) and two transmembrane domains (TMDs) [3]. The NBDs of ABC transporters are responsible for binding and hydrolysis of ATP, while the TMDs recognize and transfer substrates, enabling the transport of various endogenous substrates and xenobiotics across cell membranes using ATP as an energy source [4,5]. There is mounting evidence that ABC transporters contribute to the onset and progression of cancer [6,7], although most research has focused on ABC transporters in the context of cancer drug resistance [8,9].

ABCC1, one of the most extensively studied ABC transporters, encodes multidrug resistance-associated protein 1 (MRP1). ABCC1 is widely expressed in different tissues and cell types, and its expression is higher in solid tumors [10,11]. ABCC1 has been demonstrated to enhance prostate cancer progression by directly exporting lysophospholipid lysophosphatidylinositol (LPI) [12,13] from cancer cells. In breast cancer, ABCC1 can export sphingosine-1-phosphate (S1P), leading to cancer cell proliferation and migration [14,15]. Furthermore, the expression of ABCC1was found to be upregulated by LINC00470, promoting glioma cell proliferation and invasion [16]. In addition, ABCC1 can also be transcriptionally activated by ATF4, enhancing chemoresistance in pancreatic ductal adenocarcinoma [17]. Despite the emerging understanding of the functions of ABCC1 in proliferation and metastasis of various cancers, the underlying mechanisms, and particularly their potential relation to antitumor immunity, remain poorly understood.

In this research, we comprehensively examined multiple public databases to evaluate the potential implications of ABCC1 expression levels for cancer diagnosis and prognosis. Our results indicate a strong connection between ABCC1 and immune cell infiltration across multiple cancer types, as well as significant co-expression trends with immune-related genes. Additionally, we verified these findings through molecular biology experiments in hepatocellular carcinoma (HCC), confirming the role of ABCC1 in cancer growth and progression using both in vitro and in vivo models. In conclusion, our pan-cancer analysis suggests that ABCC1 is a potential immunological biomarker, promoting tumor growth.

Materials and methods

Data processing and analysis of differential expression

We obtained transcriptomic and clinical data encompassing 33 cancer types from The Cancer Genome Atlas (TCGA) data portal (https://portal.gdc.cancer.gov) through the University of California Santa Cruz (UCSC) Xena platform (https://xena.ucsc.edu/) [18]. Gene mapping information for 31 normal human tissues was acquired from the Genotype-Tissue Expression (GTEx) project (https://commonfund.nih.gov/GTEx) through UCSC. ABCC1 expression in 27 tumors and corresponding normal tissues was analyzed using TCGA and GTEx datasets via the SangerBox website (http://SangerBox.com/Tool) [19]. The assessment and visualization of data was executed using R software (version 4.2.2; https://www.r-project.org/) [20].

The immunohistochemistry (IHC) data for ABCC1 in human tumors and normal tissues were obtained from the Human Protein Atlas database (HPA: https://www.proteinatlas.org/) [21]. We obtained subcellular localization information and immunofluorescence of ABCC1 using the Cell Atlas module. Microsatellite instability (MSI) and tumor mutation burden (TMB) data were obtained from TCGA. We explored gene mutation and copy number alteration profiles of ABCC1 across various cancers using the cBioPortal for cancer genomics (http://www.cbioportal.org) [22]. To create a protein–protein interaction (PPI) network map of ABCC1, we utilized the STRING (https://string-db.org/) and GeneMANIA (http://genemania.org/) database.

Prognostic value and clinical traits associated with ABCC1

We categorized survival data from the TCGA data portal across different cancers into high- and low-expression groups of ABCC1, determined by the median ABCC1 expression value. To evaluate the prognostic significance of ABCC1, we conducted Kaplan-Meier survival assessment across various cancers. The impact of ABCC1 expression on patient prognosis was evaluated based on overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI) using the ``survival'' R package. Forest plots for Cox correlation analysis were generated using the R packages ``forestplot'' and ``survival''. The expression level of ABCC1 in patients stratified by clinical traits were analyzed using the R packages ``Limma'' and ``ggpubr''.

Character analysis of mutations

The genetic mutation rates of ABCC1 across multiple cancers were determined using the ``Gene Mutation'' section of the TIMER2.0 website (http://timer.cistrome.org/) [23]. Mutational attributes of ABCC1 in diverse cancers, were examined using the cBioPortal database (http://www.cbioportal.org/).

Immune cell infiltration and genes associated with immune response

We utilized the ``estimate'' R package to assess stromal and immune scores across diverse cancers. We explored the relationship between ABCC1 and immune cell infiltration, as well as immune checkpoint-related genes, using the R packages ``estimate'', ``reshape2'', ``ggpubr'', ``ggExtra'', and ``corrplot''. Furthermore, we investigated the association between ABCC1 expression and genes related to immunity. The results were visualized using heat maps created with the R packages ``Limma'', ``reshape2'', and ``RColorBrewer''.

Single cell sequencing analysis

This investigation utilized the GSE149614 dataset downloaded from the GEO database, from which we selected 10 HCC samples for further analysis. Single-cell RNA sequencing (scRNA-seq) analysis was performed using the ``Seurat'' R package. The percentage of mitochondrial and rRNA content was calculated using the PercentageFeatureSet function. Cells were filtered based on the criterion that the number of expressed genes per cell should be greater than 50 and less than 8000, while the mitochondrial content should be less than 5 %. After this filtration step, a total of 34,414 cells remained. Subsequently, cell types within different clusters were annotated using various HCC-specific cell markers.

Cell lines and cell culture

The MHCC97H, HCCLM3, THP-1, and HEK-293T cell lines were procured from the Hepatic Surgery Center, Tongji Hospital, Huazhong University of Science and Technology, Wuhan, China. MHCC97H, HCCLM3, THP-1, and HEK-293T cells were cultured in high-glucose Dulbecco's Modified Eagle's Medium (DMEM) (Cytiva, USA) supplemented with 10 % Fetal Bovine Serum (FBS) (Scitecher, USA). THP-1 cells were grown in RPMI 1640 (Cytiva, USA) also supplemented with 10 % FBS. All cells were incubated in a constant-temperature incubator at 37 °C with 5 % CO₂.

Short hairpin RNA (shRNA) and plasmid transfection

Lentivirus vectors pLKO.1-blast (Addgene) was used for the construction of recombinant lentiviruses. The small hairpin RNA (shRNA) target sequences for knocking down the target genes were as follows: scramble: 5′-GCCTAAGGTTAAGTCGCCCTCG-3′; sh1-ABCC1: 5′-CCTCTCAGTGTCTTACTCATT-3′; sh2-ABCC1: 5′-CCACATGAAGAGCAAAGACAA-3′; sh3-ABCC1: 5′-CCTGGGCTTATTTCGGATCAA-3′. The lentiviral plasmid packaging plasmids were combined to co-transfect HEK-293T cells using Lipofectamine® 3000 transfection reagent (Invitrogen, USA) for 48 h. The virions were isolated using 0.45-μm filters and cells were selected with blastomycin for 2∼3 weeks to establish cell lines with stable knockdown of the target gene for subsequent experiments.

Western blot analysis

Proteins were extracted using radioimmunoprecipitation assay (RIPA) buffer, supplemented with protease and phosphatase inhibitors. Protein concentrations were determined using a bicinchoninic acid (BCA) quantification kit. The proteins were then separated on 10 % acrylamide sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels and subsequently transferred to a polyvinylidene difluoride (PVDF) membrane. Following blocking with 5 % skim milk, the membrane was incubated overnight at 4 °C with primary antibodies. Then, the membrane underwent three 10-min washes with TBST (TBS + Tween), followed by incubation with secondary antibodies at room temperature for 1 h. Immunoreactive signals were detected using an enhanced chemiluminescence (ECL) development kit (Merck Millipore, Billerica, MA) and analyzed using Image Lab 6.1 software (Bio-Rad Laboratories, USA). The following antibodies were used: GAPDH (60004-1-Ig, 1:50000, Proteintech); ABCC1 (67228-1-Ig, 1:5000, Proteintech).

CCK-8 viability assay

A total of 1000 cells per well were initially seeded into 96-well plates. Measurements were started after 24 h and extended for five consecutive days. Cell Counting Kit-8 (CCK8) was combined with serum-free medium at a 1:10 ratio and introduced into the 96-well plates at, 100 μl per well, followed by a 1.5-h incubation at 37 °C. Subsequently, the microplate reader was used to quantify absorbance at 450 nm.

Wound healing assay

HCC cells were grown until 90 % confluence in six-well plates. A horizontal scratch was made using autoclaved 200 µl pipette tips, followed by gentle washing with phosphate-buffered saline (PBS) to remove any detached cells. Micrographs of wound closure were recorded using an inverted microscope at 0, 24, and 36 h.

Transwell assays

For the migration assay, cells were placed in the upper chamber of Transwell inserts with serum-free DMEM medium. The lower chambers were filled with DMEM medium containing 10 % FBS. Cells on the upper surface were gently removed after 24 h of incubation. Subsequently, migrated cells were fixed in 4 % paraformaldehyde, stained with 1 % crystal violet, and counted. For the invasion assays, the upper Transwell chamber was coated with a matrigel-DMEM mixture (1:4) before following the same steps as the migration assay. Cell counts were performed in four randomly chosen microscopic fields.

Macrophage infiltration assay

THP-1 cells were induced to differentiate into macrophages using 150 nM phorbol 12-myristate 13-acetate (PMA; Sigma, USA) for 24 h. Macrophage infiltration assays were conducted by placing 1.0 × 10⁵ macrophages (200 μl) without serum in the upper chamber of a Transwell plate for 48 h. In the bottom plate, MHCC97H and HCCLM3 cells (1.0 × 10⁵) were cultured with 10 % FBS in DMEM (800 ml). After 48 h of incubation, cells in the upper chamber were fixed with 4 % paraformaldehyde and stained with 0.1 % crystal violet. Infiltrating macrophages were counted in three randomly chosen microscopic fields.

Subcutaneous xenograft experiment

We conducted all mouse experiments according to protocols approved by the Ethics Committee of Tongji Hospital affiliated with Huazhong University of Science and Technology Animal Care and Use Committee, following guidelines for animal welfare. Five-week-old BALB/c nude mice were procured from Gempharmatech Co., LTD. Subsequently, 1 × 10⁶ tumor cells suspended in 100 μl of PBS were subcutaneously injected into the flanks of nude mice. Euthanasia was performed 3 weeks post-injection, after which point the subcutaneous tumors were excised, measured, weighed, and imaged.

Statistical analysis

Two-group comparisons were conducted using Student's t-test. Spearman and Pearson coefficients were employed for correlation analysis. For data involving three or more groups, one-way ANOVA was utilized, while the Wilcoxon test was applied for non-parametric data. Survival analyses were carried out using the

Kaplan-Meier method, paired with the log-rank test. Statistical computations were executed using R software, with a significance threshold of p < 0.05.

Results

ABCC1 expression is abnormally upregulated in multiple cancers

We assessed ABCC1 expression across various healthy individuals using GTEx data. ABCC1 exhibited widespread expression across organs, with notably high values in the thyroid gland and esophagus (Fig. 1A). Subsequently, we examined ABCC1 mRNA levels in tumor tissues across 33 different cancer types using TCGA data. We found that ABCC1 mRNA levels were increased in invasive breast carcinoma (BRCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), head and neck cancer (HNSC), kidney clear cell carcinoma (KIRC), kidney papillary cell carcinoma (KIRP), hepatocellular carcinoma (LIHC), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD), recta adenocarcinoma (READ), stomach adenocarcinoma (STAD), and uterine corpus endometrial carcinoma (UCEC), but decreased in kidney chromophobe (KICH) and thyroid carcinoma (THCA) (Fig. 1B). Furthermore, when we combined the data from TCGA and GTEx, ABCC1 expression was notably elevated in 26 out of the 34 cancers, and downregulated in the same cancer types as observed in the TCGA database alone (Fig. 1C). Moreover, we investigated the protein expression patterns of ABCC1 protein using the HPA database, which showed that ABCC1 protein was localized in the plasma membrane and cell junctions, indicating its possible function as a transporter or connector between the cytoplasm and extracellular or intercellular spaces (Figs. 1D and 2A). We also examined the IHC results of ABCC1 in the HPA database, and similarly, ABCC1 was found to be highly expressed in LIHC, COAD, BRCA, PRAD, LUSC, KIRP, brain lower grade glioma (LGG), and UCEC (Fig. 2B). Thus, these results highlight the abnormal upregulation of ABCC1 expression across various cancer types, underscoring its crucial role in cancer progression.

Fig. 1.

Pan-cancer analysis of ABCC1 expression and its subcellular location. (A) ABCC1 expression in normal samples from the GTEx database. (B) Pan-cancer comparison of ABCC1 expression between tumor and normal samples using TCGA data. (C) Pan-cancer comparison of ABCC1 expression after combining the data from TCGA and GTEx. (D) Subcellular localization profile of ABCC1 in human cells from the HPA database. (E) Chromosomal location of the ABCC1 gene based on the UCSC database. (*, p < 0.05; **, p < 0.01; ***, p < 0.001).

Fig. 2.

Expression of ABCC1 in cancer tissues and cells. (A) Immunofluorescence staining showing the subcellular localization of ABCC1. (B) ABCC1 expression in LIHC, COAD, BRCA, PRAD, LUSC, KIRP, LGG and UCEC.

ABCC1 expression is significantly correlated with the prognosis of diverse tumors

To evaluate the prognostic importance of ABCC1 in cancers, we investigated the link between ABCC1 levels and survival outcomes. including overall survival (OS, Fig. 3A, B), disease-specific survival (DSS, Fig. 3C, Supp. Fig. 1A), disease-free interval (DFI, Supp. Fig. 1B, D), and progression-free interval (PFI, Supp. Fig. 1C, E). The Cox proportional hazards model revealed a notable connection between ABCC1 expression levels and OS in various cancers, including, bladder urothelial carcinoma (BLCA, p = 0.017), KIRP (p < 0.001), LGG (p < 0.001), LIHC (p < 0.001), lung adenocarcinoma (LUAD, p = 0.048), mesothelioma (MESO, p = 0.020), READ (p = 0.027), pancreatic adenocarcinoma (PAAD, p = 0.041), skin cutaneous melanoma (SKCM, p = 0.018), and uveal melanoma (UVM, p = 0.002) (Fig. 3B). Kaplan-Meier survival analysis indicated reduced overall survival in high ABCC1 expression groups of BLCA (p = 0.049), KIRP (p = 0.008), LGG (p < 0.001), LIHC (p < 0.001), MESO (p = 0.010), PAAD (p = 0.018), THCA (p = 0.044), and UVM (p < 0.001). However, high ABCC1 expression was correlated with improved OS in THYM (p = 0.035) (Fig. 3A).

Fig. 3.

Association of ABCC1 expression with the prognosis of various cancers. (A) Kaplan−Meier analysis of the association between ABCC1 expression and overall survival (OS) (BLCA: p = 0.049, KIRP: p = 0.008, LGG: p < 0.001, LIHC: p < 0.001, THYM: p = 0.035, MESO: p = 0.010, PAAD: p = 0.018, THCA: p = 0.044, UVM: p < 0.001). (B) Cox proportional hazards model for the correlation of ABCC1 expression with OS and (C)DSS.

Subsequent analysis utilizing the Cox proportional hazards model unveiled a notable link between ABCC1 expression and patient outcomes. For instance, in BLCA (p = 0.040), KIRP (p < 0.001), LGG (p < 0.001), LIHC (p = 0.045), LUSC (p = 0.005), MESO (p = 0.014), SKCM (p = 0.023), THCA (p = 0.015), and UVM (p < 0.001), ABCC1 expression was significantly correlated with DSS. Similarly, in KIRP (p = 0.009), PRAD (p = 0.026), TGCT (p = 0.006), and UCS (p = 0.007), ABCC1 was linked with DFI, while in KIRP (p < 0.001), LGG (p < 0.001), LUSC (p = 0.021), PAAD (p = 0.019), PRAD (p = 0.049), TGCT (p = 0.008), THYM (p = 0.024), UCS (p = 0.002), and UVM (p < 0.001), it was related to PFI. Kaplan-Meier survival analysis also indicated that elevated ABCC1 expression was linked to reduced DFS in KIRP (p = 0.019), LGG (p < 0.001), MESO (p = 0.001), THCA (p = 0.016), and UVM (p < 0.001); shorter DFI in KIRP (p = 0.027), PRAD (p = 0.012), and UCS (p = 0.002); as well as reduced PFI in BLCA (p = 0.043), LGG (p < 0.001), PAAD (p = 0.028), PRAD (p = 0.042), UCS (p = 0.003), and UVM (p = 0.001). Conversely, higher ABCC1 levels were associated with a favorable DSS in LUSC (p = 0.009), positive DFI in TGCT (p = 0.021), and improved PFI in LUSC (p = 0.030) and TGCT (p = 0.025). In conclusion, the aforementioned analysis underscores the potential prognostic significance of ABCC1 expression in various cancer types.

ABCC1 mutations in the TCGA pan-cancer cohort

We analyzed the variant landscape of ABCC1 within TCGA cohorts, encompassing mutations, structural variations, deep deletions, and amplifications. Using the cBioPortal tool, we uncovered ABCC1 alterations in 27 cancer types, with uterine corpus endometrial carcinoma exhibiting the highest prevalence of 10.21 % (Fig. 4A). Further analysis using the TIMER2.0 database revealed that UCEC (51/531), SKCM (36/468), and COAD (25/406) were the top three cancers with the most frequent ABCC1 gene mutations (Fig. 4B). The cBioPortal tool was also employed to visualize the specific sites, types, and quantities of ABCC1 alterations, revealing missense mutations as the predominant mutation type in the relevant cancers (Fig. 4C). Additionally, the positioning of ABCC1 on human chromosome 16p13.11 was depicted using the UCSC database (Fig. 4D).

Fig. 4.

Mutational characteristics of ABCC1 in different tumors. (A) Copy number variations of ABCC1 in different tumors from the cBioPortal database. Green represents mutation, purple represents structural variant, red represents amplification, blue represents deep deletion, gray represents multiple alterations. (B) Genetic alterations of ABCC1 in The TIMER2.0 database. (C) Mutations of ABCC1 in the pan−cancer profile, most of which are missense mutations.

Clinical landscape of ABCC1 expression in cancers

We selected LIHC, LGG, and UVM as examples to investigate the connection between ABCC1 expression and diverse clinical characteristics. This approach was selected in orderto delve into the broader influence of ABCC1 on the prognosis across diversity cancers. In LGG, high ABCC1 expression was related to more severe clinical traits (Fig. 5A). Patients with higher ABCC1 expression had higher histological grades (p < 0.001). ABCC1 expression was additionally linked to histological type (p < 0.001). For example, patients with astrocytoma had higher ABCC1 expression than those with oligoastrocytoma or oligodendroglioma. Moreover, patients who received radiation therapy (p < 0.001) or survived with a remnant tumor (p = 0.014) had higher ABCC1 expression, indicating that ABCC1 is related to cancer status and therapy. Surprisingly, patients with Karnofsky performance scores of less than 80 had higher ABCC1 expression (p = 0.029). In LIHC, ABCC1 expression also displayed a notable positive correlation with the histological grade (p = 0.011), clinical stage (p = 0.031), and pathological T stage (p = 0.014), suggesting that higher expression levels of ABCC1 are linked to an unfavorable prognosis among LIHC patients, which is consistent with the analysis of OS (Fig. 5B). In UVM, ABCC1 was associated with new tumor events (p = 0.0025) and neoplasm cancer status (p = 0.044) (Fig. 5C). Taken together, the data indicate that high expression of ABCC1 is related to unfavorable clinical characteristics.

Fig. 5.

The relationship of ABCC1 expression in (A) LGG, (B) LIHC and (C) UVM with clinical features.

Genes and signaling pathways related to ABCC1 in various cancers

By analyzing the TCGA database, we generated heat maps and volcano plots of differentially expressed genes between groups separated by the median ABCC1 level in LIHC, LGG, and UVM (Fig. 6A) and generated a co-expression circle diagram of ABCC1 (Fig. 6B). In addition, we visualized the PPI network of ABCC1 using the Gene MANIA database (Fig. 6C). We further identified several ABCC1-binding proteins, including VIM, FLNA, ENPP2, B2M, etc., in LIHC, LGG, and UVM using the STRING database (Fig. 6D). Subsequently, in order to delve into the cancer-related roles of ABCC1, we conducted GO-BP and KEGG enrichment analyses of ABCC1-associated genes in LIHC, LGG, and UVM (Fig. 6E). GO-BP analyses showed that most of the ABCC1-related signaling pathways in LGG, LIHC, and UVM are immune-related, including B cell differentiation, lymphocyte activation involved in the immune response, mononuclear cell differentiation, regulation of leukocyte and lymphocyte differentiation, interleukin-1 production, negative regulation of cytokine production, etc. The results of KEGG analyses were consistent with those of GO-BP, as ABCC1-related signaling pathways include cytokine receptor-cytokine interaction, T cell receptor signaling pathway, leukocyte transendothelial migration, natural killer cell-mediated cytotoxicity, etc. We further explored the immunologic signature database using GSEA software to confirm the relationship between ABCC1 and immunity (Supplementary Fig. 2). These findings imply a link between ABCC1 expression and the immune response. Consequently, we proceeded to explore the interconnection between ABCC1 and the antitumor immune responses.

Fig. 6.

Genes and signaling pathways related to ABCC1. (A) Heat maps and volcano plots of differentially expressed genes between groups separated by the median expression level of ABCC1 in LIHC, LGG, and UVM. (B) Co-expression circle diagram of ABCC1. (C, D) Determination of genes interacting with ABCC1 in tumors based on a PPI network constructed using the GeneMANIA and STRING databases. (E) GO and KEGG enrichment analysis of ABCC1-related differentially expressed genes in LIHC, LGG and UVM.

ABCC1 expression exhibited a strong correlation with immune infiltration across various cancers

Research findings indicate that tumor-infiltrating immune cells (TIICs), integral to the tumor immune microenvironment (TIME), may influence tumor behavior and response to cancer therapy [24]. Given that the above results demonstrated that ABCC1 may regulate the TIME in various cancers, we delved into the connection between ABCC1 expression and the diversity of TIIC across cancers. Our initial focus was on the association between ABCC1 expression and ESTIMATE scores, encompassing immune and stromal scores, in various tumors. These scores reflect the proportions of infiltrating immune cells and stromal cells within tumor tissues, with higher scores indicating lower tumor purity. Using the ``Estimate'' R package, we assessed the association of ABCC1 expression with and immune and stromal scores in LGG, LIHC, and UVM using TCGA data (Fig. 7A), and also analyzed other cancer types using the SangerBox database (Supplementary Figs. 3 and 4). There was unveiled a consistent inverse correlation between ABCC1 expression and immune\stromal scores across most cancer types. These results illustrate a notable link between elevated ABCC1 expression and diminished immune cell infiltration, signifying heightened tumor purity in diverse cancers.

Fig. 7.

Correlation of ABCC1 expression with the tumor immune microenvironment and immune cells in LGG, LIHC and UVM. (A) Correlation of ABCC1 expression with the immune score and stromal score. (B) Correlation of ABCC1 expression with checkpoint genes. (C) The relevance of ABCC1 to immune cells.

Subsequently, we analyzed TCGA data using the ``CIBERSORT'' R package to delineate the composition of infiltrating immune cells in LGG, LIHC, and UVM (Fig. 7B). In LGG, ABCC1 expression exhibited a positive correlation with M2 macrophages, rested memory CD4⁺ T cells, M1 macrophages, T regulatory cells (Tregs), and neutrophils, while exhibiting a negative correlation with monocytes, naive B cells, follicular helper T cells, naive CD4⁺ T cells, and plasma cells. In LIHC, ABCC1 expression exhibited a positive association with M0 macrophages, Tregs, neutrophils, resting dendritic cells, activated memory CD4⁺ T cells, M2 macrophages, and eosinophils, yet displayed a negative correlation with rested memory CD4⁺ T cells, rested NK cells, CD8⁺ T cells, naive B cells, and monocytes. In UVM, ABCC1 expression was positively associated with follicular helper T cells, CD8⁺ T cells, resting mast cells, activated NK cells, M1 macrophages, and gamma delta T cells, but negatively correlated with monocytes, resting NK cells, plasma cells, and naive B cells. Further details about the relationship between ABCC1 expression and immune cells in LGG, LIHC, and UVM were also gleaned from the TIMER2.0 database (Supplementary Fig. 5). Other cancer types were also analyzed using the SangerBox database (Supplementary Fig. 6).

Correlation of ABCC1 expression with genes involved in immunity

Because of their regulatory effect on immune cells within the tumor microenvironment, immune checkpoint blockade (ICB) proteins have become a promising target for cancer immunotherapy [25,26]. Consequently, we delved into the potential correlation between the expression of immune checkpoint (ICP) genes and ABCC1 across various cancers. We used the ``corrplot'' R package to analyze TCGA data, and the findings revealed positive associations of ABCC1 expression with 35 out of 37 ICP genes in LGG, 36 out of 37 ICP genes in LIHC, and all ICP genes in UVM (Fig. 7C). In addition, our inquiry was extended to other cancer types using the SangerBox database (Supplementary Fig. 7), which indicated a consistent positive correlation between ABCC1 expression and the majority of ICP genes in diverse cancers. Moreover, there was a link between ABCC1 expression and factors impacting the efficacy of immune checkpoint inhibitors, namely tumor mutation burden (TMB) [27] (Supplementary Fig. 8A) and microsatellite instability (MSI) [28] (Supplementary Fig. 8B) in multiple cancers. In conclusion, all the results suggest a potential influence of ABCC1 on the responsiveness of a wide spectrum of cancers to immune checkpoint inhibitor therapies, suggesting its promise as a valuable immunological biomarker.

Encouraged by these findings, we further extended our investigation by delving into the co-expression patterns of ABCC1 with a spectrum of immune-related genes. These encompassed genes associated with immune activation (Fig. 8A), immunosuppressive functions (Fig. 8B), chemokine signaling (Fig. 8C), chemokine receptor activity (Fig. 8D), as well as genes encoding major histocompatibility complex (MHC) molecules (Fig. 8E). We proceeded with an in-depth analysis of TCGA data to unveil the intricate connection between ABCC1 expression and immune cell markers (Table 1, Supplementary Fig. 9). This investigation utilized R packages such as ``limma'', ``reshape2′', ``ggpubr'', and ``ggExtra''. Taking LGG, LIHC, and UVM as examples, the analysis accentuated that ABCC1 exhibited significant co-expression with genes involved in immune modulation across a diverse range of cancers. These results collectively imply that ABCC1 might play a pivotal role in immune regulation and response modulation. In summary, the intricate co-expression relationships between ABCC1 and immune-related genes, coupled with its connection to immune cell markers, position ABCC1 as a potential target in the realm of cancer immunotherapy.

Fig. 8.

Pan-cancer analysis of the correlation between ABCC1 expression and immune related genes. (A-E) Correlation of ABCC1 with (A) immune activation, (B) immunosuppression, (C) chemokines, (D) chemokine receptor genes, and (E)genes encoding major histocompatibility complex (MHC) molecules. (*, p < 0.05; **, p < 0.01; ***, p < 0.001).

Table 1.

Correlation between the expression of ABCC1 and immune cell markers in pan-cancer.

Immune cell	Marker gene	LGG		LIHC		UVM
		Cor	pValue	Cor	pValue	Cor	pValue
B cell	CD19	0.258	1.71E−09	0.321	1.96E−10	−0.060	0.5945
	CD79A	0.264	8.21E−10	0.238	3.11E−06	−0.082	0.4685
CD8+ T cell	CD8A	0.102	0.0193	0.286	2.09E−08	0.659	0
	CD8B	0.244	1.23E−08	0.238	3.26E−06	0.599	4.45E−09
	CD4	0.359	8.09E−18	0.290	1.36E−08	0.335	0.0025
M1 macrophage	NOS2	−0.089	0.0404	0.091	0.0785	0.266	0.0174
	IRF5	0.325	2.27E−14	0.282	3.46E−08	0.647	0
	PTGS2	0.061	0.1599	0.448	6.77E−20	0.636	2.32E−10
M2 macrophage	CD163	0.401	0	0.343	1.27E−11	0.619	7.24E−10
	VSIG4	0.336	2.37E−15	0.466	0	0.606	4.38E−09
	MS4A4A	0.410	0	0.411	0	0.578	3.74E−08
Neutrophil	CEACAM8	0.003	0.9481	0.122	0.01867	0.391	0.0004
	ITGAM	0.302	1.69E−12	0.656	0	\	\
	CCR7	0.242	1.67E−08	0.301	3.49E−09	0.363	0.0009
DC cell	HLA-DPB1	0.362	2.19E−18	0.441	0	0.595	1.07E−08
	HLA-DQB1	0.347	2.38E−16	0.353	2.98E−12	0.563	9.90E−08
	HLA-DRA	0.389	0	0.463	0	0.665	0
	HLA-DPA1	0.383	0	0.452	0	0.652	0
	CD1C	0.294	5.44E−12	0.257	4.84E−07	0.300	0.0069
	NRP1	0.416	0	0.460	0	0.712	0
	ITGAX	0.280	6.65E−11	0.560	0	0.695	0

Single-cell sequencing in HCC

To investigate hepatocellular carcinoma (HCC) at the single-cell level, we obtained an extensive transcriptional profile from ten HCC samples sourced from the GEO database (GSE149614). Quality control procedures were conducted (Supplementary Fig. 10A and B), followed by PCA dimensionality reduction (Supplementary Fig. 10C-E) to identify anchor points. Utilizing the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm for cell clustering analysis, we classified the cells into 33 distinct clusters (Fig. 9A-C). Subsequent annotation of the subclusters allowed us to identify seven discernible cell types: T/NK cells, B cells, cancer-associated fibroblasts (CAFs), macrophages, tumor-associated endothelial cells (TECs), HCC tumor cells, and plasma cells (Fig. 9D). Remarkable differential expression of marker genes was observed across various cell clusters (Fig. 9E). We computed and visualized the proportions of each cell type for each patient (Fig. 9F). Finally, we found a positive correlation between ABCC1 expression in HCC tumor cells and the proportions of tumor-infiltrating macrophages (Fig. 9G). These results underscore the substantial correlation between ABCC1 expression and macrophage infiltration in the context of HCC.

Fig. 9.

Single cell sequencing analysis of HCC. (A) The t-SNE plot, showing cell origins by color for each patient. (B) t-SEN plot of 10 HCC samples with 33 cell clusters from GSE149614. (C) A heat map showcasing the top-10 marker genes with differential expression patterns across the 33 clusters. (E) t-SNE plot, demonstrating cell type annotation and color codes representing various cell types within the HCC ecosystem. (F) A bubble chart visually represents the expression levels of marker genes across the 7 cell types. (G) Correlation between the expression of ABCC1 in HCC cells and the proportions of immune cells.

Relationship between ABCC1 expression and drug sensitivity

Subsequently, we further analyzed the LIHC dataset sourced from TCGA, employing the ``pRRophetic'' R package to investigate how the level of ABCC1 influences the responsiveness to both chemotherapy and targeted therapy. Our analysis brought to light a robust relationship between ABCC1 expression and the susceptibility to a range of drugs in the context of LIHC. Most chemotherapeutic and targeted agents, including sorafenib, vinorelbine, sunitinib, saracatinib, pyrimethamine, lenalidomide, pazopanib, temsirolimus, cyclopamine, bexarotene, imatinib, and elesclomol, showed higher efficacy in cancers with low levels of ABCC1 expression, except for erlotinib, where low ABCC1 expression levels represented higher sensitivity (Fig. 10, Supplementary Fig. 11).

Fig. 10.

Correlation between ABCC1 expression levels and sensitivity to representative cancer drugs in LIHC.

ABCC1 promotes the progression of hepatocellular carcinoma and influences macrophage recruitment

We conducted Western blot analysis to assess ABCC1 expression across various HCC cell lines. Notably, ABCC1 protein levels were markedly elevated in MHCC97H, HCCLM3, and HCCLM6 cells compared with other HCC cell lines (Fig. 11A). Leveraging the inherently high ABCC1 expression in these HCC cell lines, we achieved stable ABCC1 knockdown in MHCC97H and HCCLM3 cells by lentivirus transfection (Fig. 11B). In vitro analyses, including CCK-8, EdU, wound healing, and Transwell assays, revealed a substantial reduction in cell proliferation, migration, and invasion of MHCC97H and HCCLM3 cells upon ABCC1 knockdown (Fig. 11C–F). Moreover, ABCC1 knockdown in MHCC97H and HCCLM3 cells led to a noticeable decrease of macrophage infiltration (Fig. 11G). To explore the impact of ABCC1 on tumorigenicity, we xenografted HCC cells into nude mice. Three weeks post-inoculation, the mice were sacrificed to evaluate tumor growth (Fig. 11H). Tumor volumes and weights were notably diminished in mice injected with ABCC1-knockdown MHCC97H cells compared to controls. Immunohistochemistry staining for CD68 revealed fewer infiltrating macrophages in tumors with lower ABCC1 levels (Fig. 11I). Collectively, these findings underscore that ABCC1 plays a significant role in enhancing HCC progression both in vitro and in vivo.

Fig. 11.

ABCC1 promotes the proliferation and migration of HCC cells and affect the recruitment of macrophages. (A) Endogenous expression of ABCC1 in HCC cell lines. (B) Verification of knockdown efficiency of ABCC1 in MHCC97H and HCCLM3 cell lines by western blotting. The biological effects of ABCC1 on HCC cell lines were verified by CCK-8 (C), EdU (D), wound healing (E) and Transwell (F) assays. (G) Homing of macrophages toward MHCC97H and HCCLM3 cells transfected with vector, shABCC1#1, shABCC1#2; Scale bar: 100 μm. (H) Macroscopic images of subcutaneous tumors from nude mice (n = 6) xenografted with ABCC1-knockdown MHCC97H cells. The tumor volumes and weights were quantified. (I) Representative hematoxylin and eosin (H&E) staining as well as IHC staining for CD68 in the subcutaneous xenografts. (*, p < 0.05; **, p < 0.01; ***, p < 0.001).

Discussion

In the past decade, tumor immunotherapy has made impressive clinical advancements, but it has yet to reach its full potential [29,30]. Notably, the clinical utilization of tumor immunotherapy is greatly constrained due to its low response rate and considerable adverse effects. Therefore, the development of new immunological biomarkers and targets is of significant importance for the diagnosis and treatment of tumors [31]. This research presents substantial evidence that ABCC1 may be a novel immunological marker, influencing macrophage infiltration as well as facilitating tumor proliferation, migration, and metastasis.

ABCC1, a constituent of the ABC transporter superfamily, is composed of two NBDs and two TMDs [3] and its primary role lies in MRP function [8,9]. In recent years, ABCC1 was shown to be significantly linked with diverse human malignancies, exerting pivotal functions in cancer progression and metastasis [12], [13], [14], [15], [16]. However, the underlying mechanism remains incompletely understood, particularly its potential interrelation with antitumor immunity. The abundance and composition of TIICs within the tumor environment profoundly affect tumor behavior and immunotherapy efficacy [24]. Following a comprehensive pan-cancer investigation, we noted a significant link between ABCC1 expression and immune cell infiltration across different cancers. Furthermore, ABCC1 exhibited a strong co-expression pattern with immunomodulatory and immune checkpoint inhibitor (ICP) genes. Based on the findings, we propose ABCC1 as a promising candidate for targeted cancer immunotherapy.

We conducted a thorough analysis of ABCC1 expression and its functional implications using diverse databases. Initially, we assessed ABCC1 expression across normal and cancerous tissues. Consistent with previous reports [32,33], the results underscored comparatively elevated ABCC1 expression in most tumors, implying its role as an oncogenic gene in numerous malignancies. Furthermore, Kaplan-Meier survival curves and Cox proportional hazard models, revealed a significant relationship of elevated ABCC1 expression with an unfavorable prognosis in BLCA, KIRP, LGG, LIHC, MESO, PAAD, THCA, and UVM. In sum, our analysis underscores the potential of ABCC1 as a prognostic biomarker across various cancers. In light of the distinct expression and prognostic patterns of ABCC1, we selected LGG, LIHC, and UVM as prototypical cancer types for in-depth exploration.

Then, we investigated the link between ABCC1 expression levels and clinical attributes. In LGG, LIHC, and UVM, elevated ABCC1 expression was consistently correlated with worse clinical characteristics. Notably, heightened ABCC1 expression was correlated with higher histological grade, clinical stage, and pathological T stage in LIHC. These results further confirmed that ABCC1 is a promising prognostic marker for various cancers.

To investigate the functional aspects of ABCC1, we conducted GO and KEGG analyses of genes associated with ABCC1. The results of these analyses demonstrated a significant correlation between elevated ABCC1 expression and the activation of diverse immune signaling pathways in LGG, LIHC, and UVM. These pathways encompassed processes such as B cell differentiation, lymphocyte activation, cytokine receptor interaction, and T cell receptor signaling, among others. There are reports indicating that ABCC1 has the capability to augment the secretion of S1P by mast cells, consequently influencing their functionality [34]. However, research on the role of ABCC1 in antitumor immunity is currently limited. Therefore, our focus shifted towards the impact of ABCC1 on immune functions within the tumor context.

Subsequently, we delved into the interplay between ABCC1 expression and TIICs as well as immune-related genes. The results revealed a consistent negative correlation of ABCC1 expression with and stromal and immune scores across most cancers. In LGG, ABCC1 expression exhibited positive links with M2 macrophages, resting memory CD4⁺ T cells, M1 macrophages, Tregs, and neutrophils, while showing negative associations with monocytes, naive B cells, follicular helper T cells, naive CD4⁺ T cells, and plasma cells. In LIHC, ABCC1 expression was positively linked to M0 macrophages, Tregs, neutrophils, resting dendritic cells, activated memory CD4⁺ T cells, M2 macrophages, and eosinophils, while showing negative correlations with resting memory CD4⁺ T cells, resting NK cells, CD8⁺ T cells, naive B cells, and monocytes. In UVM, ABCC1 expression demonstrated positive associations with follicular helper T cells, CD8⁺ T cells, rested mast cells, activated NK cells, M1 macrophages, and gamma delta T cells, alongside negative correlations with monocytes, rested NK cells, plasma cells, and naive B cells. Remarkably, ABCC1 expression exhibited positive correlations with multiple immune checkpoint genes across diverse cancers. Single-cell RNA. sequencing analysis also demonstrated a positive correlation between the expression of ABCC1 in HCC cells and macrophage infiltration. These insights underscored the pivotal role of ABCC1 in shaping the TIME and its potential relevance to immunotherapy. Notably, both in vitro and in vivo experiments using HCC cells recapitulated the oncogenic role of ABCC1 in promoting HCC cell growth, migration, and invasion. Moreover, targeted knockdown of ABCC1 in HCC97H and HCCLM3 cells resulted a reduction of macrophage infiltration.

Previous studies delineated the oncogenic role of ABCC1 plays in various cancers by promoting tumor cell proliferation and migration [14,15], inducing drug resistance [12,35,36], and modulating lipid-signaling pathways [37]. In addition, ABCC1 also plays a crucial role in regulating immune cell functions [34,38] and autoimmune diseases [39,40]. However, little is known about the role of ABCC1 in the TIME. Here, we conducted a comprehensive pan-cancer analysis of ABCC1 and corresponding experimental validation in HCC cells, demonstrating that ABCC1 promotes macrophage migration, thus participating in the regulation of the TIME. Our data on the role of ABCC1 in the TIME may offer a new perspective for targeting ABCC1 in cancer treatment.

Conclusions

In conclusion, the above findings suggest a general increase of ABCC1 expression across diverse cancers and its significant association with a worse cancer prognosis. The intricate connection between ABCC1 and TIICs within the tumor microenvironment as well as its impact on immunotherapeutic responses in various cancers are notable. This positions ABCC1 as a potential immunotherapeutic biomarker, aiding in the identification of tumor patients who might benefit from ICB therapy. Both in vitro and in vivo experiments reaffirm the role of ABCC1 as an oncogene in HCC. Based on this, we propose ABCC1 not only as a potential prognostic biomarker, but also as a promising indicator of immunotherapeutic responsiveness across diverse malignancies. Nonetheless, our current comprehension of the ABC superfamily merely scratches the surface, warranting future exploration to unravel additional functions, mechanisms, and plausible therapeutic targets.

Funding

Research was supported by grants from the National Natural Science Foundation of China No. 82372917 (W.H.), No. 82173313 (W.H.), No. 82273310 (L.X.), the Natural Science Foundation of Hubei Province 2022CFA016 (L.X.), and the Basic Research Support Program of Huazhong University of Science and Technology 2023BR038 (L.X.).

Ethics approval and consent to participate

The animal study was approved by the Ethics Committee of Tongji Hospital affiliated with Huazhong University of Science and Technology Animal Care and Use Committee.

Availability of data and material

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

Consent for publication

Not applicable.

CRediT authorship contribution statement

Tiantian Wang: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Dean Rao: Methodology, Supervision. Chenan Fu: Data curation, Investigation. Yiming Luo: Data curation, Software. Junli Lu: Validation, Visualization. Huifang Liang: Data curation, Funding acquisition, Supervision, Writing – review & editing. Limin Xia: Conceptualization, Data curation, Funding acquisition, Methodology, Supervision, Writing – review & editing. Wenjie Huang: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.