High nuclear ABCG1 expression is a poor predictor for hepatocellular carcinoma patient survival

Bin Xi, Fng-Zhou Luo , Bin He , Fng Wng , Ze-Kun Li, Ming-Chun Li,Shu-Sen Zheng ,c,?

a Division of Hepatobiliary and Pancreatic Surgery, Department of Surgery, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310 0 03, China

b Department of Radiotherapy, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310 0 03, China

c Key Laboratory of Combined Multi-Organ Transplantation, Ministry of Public Health, Zhejiang University, Hangzhou 310 0 03, China

Keywords:ATP-binding cassette transporter G1 Hepatocellular carcinoma Overall survival Prognostic factor Progression-free survival

ABSTRACT Background: ATP-binding cassette transporter G1 (ABCG1) regulates cellular cholesterol homeostasis and plays a significant role in tumor immunity.But, for hepatocellular carcinoma (HCC), the role of ABCG1 has not been investigated.Thus, the aim of this study was to evaluate the prognostic value and clinicopathological significance of ABCG1 in HCC.

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors, and the fourth most common cause of cancerrelated death worldwide[1].Despite progress in the diagnosis and treatment of HCC, patient prognosis remains poor because of the high risk of recurrence and metastasis after surgery [2,3].The mechanism underlying HCC development remains elusive and requires further investigation[4].Therefore, there is a pressing need for research to better understand the regulatory mechanisms underlying the development of HCC and identify new therapeutic targets.

ATP-binding cassette transporter G1 (ABCG1) has been demonstrated to play an important role in tumor resistance, promotion of drug efflux and enhancement of the chemical resistance to anticancer drugs in cancer cells [5,6].MutatedABCgenes can influence the proliferation, differentiation, migration, and invasion in cancer cells[7].The main function of ABCG1 is to promote the efflux of excess cholesterol from cells to high-density lipoprotein (HDL)particles; it thus plays an important role in intracellular cholesterol transport and reverse cholesterol transport [8–11].Further-more, ABCG1 has also been reported to be highly expressed and predicts poor outcome in different tumors [7,12-14].However, the biological role and molecular mechanism of ABCG1 in HCC has not been explored.

The prognostic significance of ABCG1 has been confirmed in a variety of tumors, but not in HCC; thus, in this study, we investigated the prognostic value and the correlated clinical characteristics of ABCG1 expression in patients with HCC, to further enhance our understanding of the molecular mechanisms underlying HCC development.

Methods

Patients and follow-up

One hundred and four adult patients with HCC were enrolled.HCC and matched tumor-adjacent tissue samples were collected from the First Affiliated Hospital, Zhejiang University School of Medicine between January 2011 and December 2013.All patients underwent surgical resection.HCCs were pathologically confirmed.Clinical characteristics of the patients, including demographics, degree of tumor differentiation, tumor number, tumor size, laboratory data, and follow-up data were collected.For tumor differentiation assessment, Edmondson-Steiner grading was used.

All patients were followed up until the end of December 2020,and the median follow-up time was 35.5 months.The primary endpoints of the study were overall survival (OS) and recurrencefree survival (RFS).OS was calculated as the duration between the date of surgery and the date of death or the date of last follow-up,while RFS was calculated as the duration between date of surgery and the date of first confirmed tumor recurrence.This study was approved by the Ethics Committee of The First Affiliated Hospital,Zhejiang University School of Medicine.

Immunohistochemical (IHC) staining

To examine ABCG1 expression, tissues were embedded in paraffin.The slices were antigen-retrieved and incubated with primary antibody (Novus, cat number: NB400-132, Littleton, CO, USA).Then, the staining intensity and extent score were assessed and multiplied to generate a final IHC score[15].The score of staining intensity was evaluated as follows: 0, unstained; 1, weakly stained;2, moderately stained; and 3, strongly stained.The extent score was rated according to the extent of positive staining: 1, ≤25%;2, 26%-50%; 3, 51%-75%; and 4, ≥76%.If the final score of samples was less than 7, the expression of ABCG1 was considered lowexpression, while the score of 7 or higher was considered high expression[16].The nuclear and cytosolic expression levels were calculated separately, and the total expression was defined as the average of the nuclear and cytosolic expression levels.

ABCG1 mRNA expression and patient outcomes

The relationship between ABCG1 mRNA and patient outcomes was analyzed by online platform, Kaplan-Meier Plotter ( http://kmplot.com/analysis/ ).The mRNA expression, OS and RFS were obtained from The Cancer Genome Atlas (TCGA) database via Kaplan-Meier Plotter.In total, 322 patients were enrolled, and the followup time of all these patients was more than 5 years after operation.

Statistical analysis

SPSS v23.0 software (SPSS, Chicago, IL, USA) and Prism 7.0(GraphPad Software, La Jolla, CA, USA) software were used to analyze the data.The associations between ABCG1 expression and clinicopathological parameters were analyzed by Chi-square test,while comparison of OS and RFS in relation to ABCG1 expression was analyzed by Kaplan-Meier and log-rank test.The Cox proportional hazard regression model was used to identify independent factors associated with RFS in patients with HCC.Statistical significance was set atP<0.05.

Results

Patient characteristics

The clinicopathological characteristics of the 104 patients with HCC are shown in Table 1.Among these patients, 96 were male and 8 were female, with a median age of 54 years; 62 (59.6%)patients received transcatheter arterial chemoembolization (TACE)before operation.Ninety-five (91.3%) patients developed HCC underlying hepatitis B infection, and 102 (98.1%) had liver cirrhosis.According to the pathological results, the resected specimens included 42 (40.4%) well- or moderately-differentiated tumors and 62 (59.6%) poorly-differentiated tumors.In total, 41 (39.4%) patients had more than 2 tumors and 63 (60.6%) had 1 or 2 tumors.Over the follow-up period, tumor recurrence was confirmed in 60(57.7%) patients.

Expression of ABCG1 in HCC samples

The total ABCG1 expression in HCC samples was higher than that in matched tumor-adjacent samples (P<0.001,Fig.1A and B).As shown inFig.1A, ABCG1 was present in both the nucleus and the cytosol.The IHC results showed that nuclear ABCG1 expression was significantly upregulated in HCC samples compared to that in matched tumor-adjacent samples (P<0.001,Fig.1C);however, no significant differences were observed in cytosolic expression between HCC samples and matched tumor-adjacent samples (Fig.1D).

Clinical significance of ABCG1 expression in HCC cases

Patients were divided into the high expression groups and lowexpression groups based on the median value of ABCG1 mRNA expression, and no statistical difference in OS and RFS was observed between the two groups (Fig.2A and B), according to the data from TCGA database.

To explore the effect of total, cytosolic, and nuclear ABCG1 expressions on prognosis, all patients were divided into two groups according to the relative expression score of the total, cytosolic,and nuclear ABCG1 expression.As shown inFig.2C-F, there were no significant differences in OS and RFS between the high total or cytosolic ABCG1 expression group and the low total or cytosolic ABCG1 expression group.

In contrast, OS (33.2% vs.57.2%,P= 0.012) and RFS (26.3% vs.52.1%,P= 0.020) were substantially reduced in patients with high nuclear ABCG1 expression compared to those in patients with low nuclear ABCG1 expression (Fig.3B and C).Furthermore, tumor size≥5 cm (74.5% vs.47.2%,P= 0.004) and tumor recurrence (68.6%vs.47.2%,P= 0.027) were more frequent in patients with high nuclear ABCG1 expression (Table 1).

Although the cytosolic expression of ABCG1 could not predict the prognosis in patients with HCC, we found a heterogenous ABCG1 expression pattern.We divided patients into 4 groups based on the nuclear and cytosolic ABCG1 expression ratio:Lownucleus/Lowcytosol(low nuclear expression and low cytosolic expression), Lownucleus /Highcytosol (low nuclear expression and high cytosolic expression), and Highnucleus /Lowcytosol (high nuclear expression and low cytosolic expression), and Highnucleus /Highcytosol(high nuclear expression and high cytosolic expression).Amongthe 4 groups, patients with Lownucleus /Highcytosol had a superior OS and RFS (Fig.4and Table 2), followed by the Lown ucleus/Lowc ytosoland Highnucleus/Highcytosolgroups.Patients in the Highnucleus /Low cytosol group showed the worst OS and RFS.

Table 1Clinicopathological variables of patients with HCC based on nuclear expression of ABCG1.

Table 2The overall survival and recurrence-free survival of the Low nucleus /High cytosol groups compared with other three groups.

Fig.1.Correlations between ABCG1 expression and HCC samples.A: Representative images of IHC ( n = 3) showing the expression of ABCG1 in HCC samples and matched tumor-adjacent tissues.Upper row: original magnification × 40, lower row: original magnification × 200.B: Total ABCG1 expression in 104 pairs of HCC samples and matched tumor-adjacent tissues.C: Nuclear ABCG1 expression in 104 pairs of HCC samples and matched tumor-adjacent tissues.D: Cytosolic ABCG1 expression in 104 pairs of HCC samples and matched tumor-adjacent tissues.ABCG1: ATP-binding cassette transporter G1; HCC: hepatocellular carcinoma; IHC: immunohistochemical staining; N:tumor-adjacent non-tumor tissue; T: tumor tissue.

Thus, we established a risk stratification system based on the distribution of ABCG1 in the nucleus and cytoplasm: high risk,indicated by the Highnucleus/Lowcytosolpattern; moderate risk, indicated by the Highnucleus/Highcytosolor Lownucleus/Lowcytosolpatterns; and lowrisk, indicated by the Lownucleus/Highcytosolpattern.This ABCG1-based risk stratification pattern could distinguish between OS (23.6%, 48.6% and 84.6%, respectively,P= 0.005) and RFS (15.8%, 42.1% and 76.9%, respectively,P= 0.009;Fig.5A andB), and the ABCG1 high risk group had more patients with tumor size ≥5 cm (P= 0.004) and tumor recurrence (P= 0.004)(Table 3).

Fig.2.Correlations between ABCG1 mRNA, total and cytosolic ABCG1 expression in HCC samples and patient OS and RFS.A: No significant differences in OS were observed between the high ( n = 161) and low ( n = 161) expression groups.B: No significant differences of RFS were observed between the high ( n = 148) and low ( n = 148) expression groups.C: No significant differences in OS were observed between the high ( n = 42) and low ( n = 62) total ABCG1 expression groups.D: No significant differences in RFS were observed between the high ( n = 42) and low ( n = 62) total ABCG1 expression groups.E: No significant differences in OS were observed between the high ( n = 37) and low ( n = 67) cytosolic ABCG1 expression groups.F: No significant differences in RFS were observed between the high ( n = 37) and low ( n = 67) cytosolic ABCG1 expression groups.ABCG1: ATP-binding cassette transporter G1; HCC: hepatocellular carcinoma; OS: overall survival; RFS: recurrence-free survival.

Fig.3.Correlations between nuclear ABCG1 expression and OS and RFS.A: Representative images of HCC samples with high and low nuclear ABCG1 expression.Left: original magnification × 40, right: original magnification × 200.B: OS was significantly lower in the high nuclear ABCG1 expression group ( n = 51) than in the low nuclear ABCG1 expression group ( n = 53).C: RFS was significantly lower in the high nuclear ABCG1 expression group ( n = 51) than in the low nuclear ABCG1 expression group ( n = 53).ABCG1: ATP-binding cassette transporter G1; HCC: hepatocellular carcinoma; OS: overall survival; RFS: recurrence-free survival.

Fig.4.Correlations between different ABCG1 nuclear/cytosolic expression in HCC samples and patient OS and RFS.OS ( A) and RFS ( B) were significantly different among the 4 groups.ABCG1: ATP-binding cassette transporter G1; HCC: hepatocellular carcinoma; OS: overall survival; RFS: recurrence-free survival.

Univariate and multivariate Cox proportional hazard analyses

Univariate Cox regression analysis indicated that age ≥50 years old [hazard ratio (HR) = 0.44,P= 0.002], AFP ≥400 ng/mL (HR = 3.86,P<0.001), tumor number>2 (HR = 2.89,P<0.001), tumor size ≥ 5 cm (HR = 2.23,P= 0.006),poor tumor differentiation (HR = 1.99,P= 0.013), high nuclear ABCG1 expression (HR = 1.83,P= 0.022), and ABCG1 high risk (HR = 1.86,P= 0.022) were significantly associated with RFS in patients with HCC (Table 4).Multivariate Cox regression analysis indicated that AFP ≥400 ng/mL (HR = 3.48,P<0.001), tumor number>2 (HR = 2.88,P<0.001), and ABCG1 high risk (HR = 2.03,P= 0.015) were independent factors for poor RFS, and age ≥50 years old (HR = 0.50,P= 0.012)was an independent factor for better RFS in patients with HCC(Table 5).

Table 3Clinicopathological characteristics of patients with HCC based on ABCG1 risk groups.

Table 4Univariate Cox regression analysis of RFS characteristics in 104 patients with HCC.

Table 5Multivariate Cox regression analysis of the clinicopathological characteristics significantly related to RFS.

Fig.5.OS ( A) and RFS ( B) of HCC patients based on ABCG1 risk groups.ABCG1: ATP-binding cassette transporter G1; HCC: hepatocellular carcinoma; OS: overall survival;RFS: recurrence-free survival.

Discussion

In this study, we assessed ABCG1 expression levels in HCC and explored the association between ABCG1 expression and prognosis in patients with HCC.We first discovered that the total and nuclear expression of ABCG1 is significantly higher inHCC samples compared to that in matched tumor-adjacent samples.Results also demonstrated that ABCG1 nuclear expression and Highnucleus/LowcytosolABCG1 expression were positively associated with tumor size and tumor recurrence in HCC.Moreover, Highnucleus/LowcytosolABCG1 expression was correlated with poorer prognosis and could be considered an independent prognostic factor for RFS in HCC.

ABCG1 plays a critical role in the unidirectional efflux of cholesterol, and further promotes the transportation of intracellular cholesterol and oxysterols away from the endoplasmic reticulum.Dysregulation of cholesterol metabolism has been confirmed to play oncogenic roles in multiple tumors, including breast cancer[17], lymphoma[18], lung cancer [19,20]and pancreatic cancer[21].Dietary cholesterol can promote the development of HCC related to steatohepatitis by dysregulating metabolism and calcium signaling[22].ABCG1 has been confirmed to facilitate tumorige-nesis in metastatic cancer, while depletion of ABCG1 causes tumor regression[23].Elevated ABCG1 expression in human colorectal cancer is further correlated with worse prognosis[24].Other researchers found that patients with an early-stage HCC subtype characterized by disrupted cholesterol homeostasis had a worse OS and poorer prognosis compared with other subtypes[25].Sag et al.found that the absence of ABCG1 in mice remarkably inhibited MB49-bladder carcinoma and B16-melanoma growth and prolonged survival[26].In addition, Tian et al.found that ABCG1 overexpression can promote proliferation by regulating the proliferation, apoptosis, and associated markers in HKULC4 lung cancer cells[12].ABCG1 also downregulates the expression of several miRNAs and promotes cell migration and invasion of lung cancer[12].Another team found that lung cancer can be inhibited by betulinic acid nanoparticles via downregulation of ABCG1 in HKULC2 cells[27].Chen et al.identified that knockdown of ABCG1 in mesenchymal glioblastoma enhanced apoptosis by increasing CHOP endoplasmic reticulum stress protein expression, reduced tumor growth and increased mouse survivalinvivo[28].

The present study revealed that nuclear ABCG1 expression,rather than total expression, was significantly associated with OS and RFS in HCC.We further analyzed the correlation of nuclear ABCG1 expression with the clinical characteristics of patients with HCC, and identified a positive correlation of nuclear ABCG1 expression with tumor size and tumor recurrence rate.Subgroup analy- ses have found that the Highnucleus/Lowcytosolgroup of ABCG1 had the worst OS and RFS, while patients with Highn ucleus/Highc ytosolhad a relatively better prognosis.We reviewed similar literatures focusing on the localized expression of transmembrane proteins.One research found that aberrant expression and subcellular distribution of Bag-1 play an important role in gastric cancer and can be considered a prognostic biomarker[29].Similarly, another study confirmed that nuclear expression of S100A4 is associated with aggressive behavior of epithelial ovarian carcinoma[30].We speculate that one possible explanation is that there is a mechanism that drives the translocation of ABCG1 into the nucleus, and the abnormal distribution of ABCG1 promotes the progression of HCC.

Multivariate Cox regression analysis revealed that only the Highn ucleus/Lowc ytosolgroup, not ABCG1 nuclear expression was significantly correlated with RFS of HCC.Our results support that Highnucleus/LowcytosolABCG1 expression is a potential biomarker of HCC.

This investigation has some limitations.First, it is still unclear why the Highnucleus/Lowcytosolgroup, instead of high total, nuclear or cytosolic expression, was associated with poor RFS in patients with HCC.Second, the underlying molecular mechanism of ABCG1 in HCC remains unexplored, although it could be hypothesized that modulation of cholesterol metabolism or macrophage activity may play a role.Third, the existence of a potential upstream molecules promoting the translocation of ABCG1 into the nucleus warrants further consideration and exploration.

In conclusion, our study suggested thatHighn ucleus/Lowc ytosolABCG1 expression in patients with HCC was significantly correlated with poor prognosis and independently predicted RFS.However,further research is required to fully understand the detailed underlying mechanisms of ABCG1 in HCC.

Acknowledgments

None.

CRediT authorship contribution statement

Bin Xi:Conceptualization, Investigation, Project administration,Writing – original draft, Writing – review & editing.Fang-Zhou Luo:Data curation, Formal analysis, Investigation, Writing –review & editing.Bin He:Formal analysis, Validation.Fang Wang:Formal analysis, Investigation.Ze-Kuan Li:Software, Validation.Ming-Chun Lai:Software, Investigation.Shu-Sen Zheng:Conceptualization, Funding acquisition, Project administration, Writing –review & editing.

Funding

This study was supported by a grant fromNatural Science Foundation of Zhejiang Province(LQ21H160031).

Ethical approval

This study was approved by the Ethics Committee of The First Affiliated Hospital, Zhejiang University School of Medicine.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.