Skip to main content
Download PDF

Prognostic, oncogenic roles, and pharmacogenomic features of AMD1 in hepatocellular carcinoma

Cancer Cell International volume 24, Article number: 398 (2024) Cite this article

Abstract

Background

AMD1 is the gene encoding S-adenosylmethionine decarboxylase 1. How AMD1 affects the prognosis of hepatocellular carcinoma (HCC) patients is unclear.

Methods

Using the Cancer Genome Atlas (TCGA) liver hepatocellular carcinoma datasets, gene enrichment and immunological traits were compared between groups with high and low AMD1 expression. After altering AMD1 expression in HCC cells, cell viability, the clonal formation rate, and migration and invasion ability were detected. Univariate Cox regression analysis and Pearson correlation were used to screen for AMD1-related genes (ARGs). Multidimensional bioinformatic algorithms were utilized to establish a risk score model for ARGs.

Results

AMD1 expression was notably increased in the majority of cancer types. High AMD1 expression was associated with adverse outcomes and poorer immunotherapy response in HCC patients. AMD1 exhibited higher expression levels in HCC cell lines. The efficient inhibition of HCC cell proliferation, migration, and invasion in vitro can be achieved through the downregulation of AMD1. The AMD1-related risk score was constructed with the expression of 9 ARGs, and demonstrated high predictive efficacy in multiple validation cohorts. Patients with high risk scores exhibited greater resistance to classical chemotherapy drugs. The nomogram, which consists of age, stage, and the AMD1-related risk score, was used to calculate the probability of survival for each individual.

Conclusion

The present study indicates that AMD1 functions as a potential role in HCC progression and may serve as a therapeutic target in HCC. This study constructed a novel AMD1-related scoring system for predicting the prognosis and treatment responsiveness of patients with HCC, enabling the prediction of prognosis and identification of potential treatment targets.

Introduction

In the 21st century, cancer will overtake all other causes of mortality worldwide. The International Agency for Research on Cancer reported that there were 9.7 million cancer deaths and 20 million new cases of cancer globally, with hepatocellular carcinoma (HCC) having the third highest mortality rate. Among male deaths, the mortality rate of HCC reached 10.2%, second only to that of lung cancer [1]. Surgical resection is currently the main therapeutic approach for treating liver cancer, with additional treatments such as immunotherapy, chemo/radiotherapy, molecular targeted therapy, and interventional therapy being regarded as supplemental treatments. Patients with HCC now have a comparatively better prognosis because of advancements in comprehensive treatment [2, 3]. However, with early signs that are often hidden, local recurrence, and distant metastases, the overall prognosis for patients with liver cancer remains inadequate [4]. Therefore, identifying a reliable marker for HCC prognostication is vital.

AMD1 is the gene encoding S-adenosylmethionine decarboxylase 1, which catalyzes the decarboxylation of S-adenosylmethionine and is involved in regulating polyamine metabolism [5]. Polyamines play a wide range of roles in the synthesis of amino acids and proteins, as well as in cell development, proliferation, and differentiation [6,7,8]. Polyamines are also more prevalent in many types of cancer [9, 10]. According to a recent study, AMD1 upregulates the expression of several pluripotency proteins, which may support the self-renewal of embryonic stem cells and preserve their stemness [11, 12]. Prostate, gastric, colorectal, and non-small cell lung cancers all exhibit elevated levels of AMD1, which is strongly linked to cancer recurrence [13,14,15]. In prostate cancer, AMD1’s effect on proliferation depends on the mTORC1 signalling pathway. Comprehensive metabolomics results indicate that mTORC1 is associated with S-adenosylmethionine decarboxylase activity and decarboxylated adenosylmethionine levels in mouse and human prostate tumours. The mechanism by which mTORC1 increases AMD1 expression may involve the stabilization of S-adenosylmethionine decarboxylase through the phosphorylation of residue S298 [13, 16]. Gupta demonstrated how AMD1 silencing can prevent MCF7 breast cancer cells from proliferating and invading [17]. However, reports of AMD1 in HCC are limited. The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC) databases served as the foundation for this investigation of the expression, clinical relevance, and impact of AMD1 on the biological function of HCC cells.

Materials and methods

Pancancer analyses

The online application Tumour Immune Estimation Resource (TIMER; version 2.0) was utilized for investigating AMD1 expression in 33 various types of cancer [18]. The TIMER online analytical tools normalize the expression data for the pancancer analyses and visualize the results. Furthermore, AMD1 expression in HCC was investigated via the Integrative Molecular Database of Hepatocellular Carcinoma, a database of HCC expression atlases that includes specifically selected public HCC expression datasets with an extensive set of clinical samples [19].

Data collection and processing

Using TCGA database and the ICGC database, we obtained an expression matrix and associated clinical data of HCC patients. Consequently, we included 371 liver hepatocellular carcinoma (LIHC) patients from TCGA cohort and 212 LIHC patients from the ICGC cohort for our enrolment. Samples lacking comprehensive information on survival were excluded. A total of 340 patients from the TCGA database were assigned to two groups at random: 204 patients (60%) in the training cohort and 136 patients (40%) in the testing cohort. Similarly, a total of 188 patients from the ICGC database served as the external validation cohort.

Functional enrichment and gene set variation analysis (GSVA)

Database for Annotation Visualization and Integrated Discovery bioinformatics resources consist of an integrated biological knowledgebase and analytic tools [20]. The most critical AMD1 genes, or the distinctive genes of the cell cluster, were submitted to the database. Homo sapiens was chosen as the species, and the official gene symbol was chosen as the identifier. The enrichment results were ultimately acquired from Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. The GO categories included biological process, cellular component, and molecular function. In this study, the top five outcomes were arranged according to increasing P values (P < 0.05). The AmiGO2 portal provides a list of genes related to the cell cycle. The functional enrichment score for each HCC sample was calculated via the default settings and the R package. A heatmap of the enrichment results was created via the pheatmap package. Pearson correlation analysis was used to evaluate the relationship between AMD1 and the cell cycle. The circos plot was obtained using the R package “Circlize”. The correlation matrix of AMD1 and cycle-related metagenes was plotted using the R package “Corrgram”.

Tumor Immune Dysfunction and Exclusion (TIDE) and single-cell RNA-sequencing data analysis

TIDE analysis [20] was employed to investigate the variations in immune checkpoint inhibitor medication response among various HCC patient types. TIDE scores, T-cell dysfunction scores and exclusion scores of each patient were acquired by TIDE algorithm. Microsatellite instability (MSI) is a form of hypermutation caused by defects in the mismatch repair mechanism. MSI algorithm was used to evaluate the relationship with immune response. From the LIHC_GSE98638 dataset, approximately 5059 high-quality cells were filtered and retrieved. Tumor Immune Single-cell Hub 2 is a single-cell RNA-seq database that integrates biological data and analytical tools to provide systematic and comprehensive annotation of tumor microenvironment. The AMD1 expression pattern was visualized using Tumor Immune Single-cell Hub 2 [21].

Cell culture and transfection

HCC cell lines Huh7 (RRID: CVCL_U443), HepG2 (RRID: CVCL_A8BS), and PLC/PRF/5 (RRID: CVCL_0485), QGY7703 (RRID: CVCL_6715) and a human normal hepatocyte cell line LO2 (RRID: CVCL_6926) were purchased from the Cell Bank of Type Culture Collection (China Academy of Sciences, Shanghai, China). All the cell lines were cultured in high-glucose DMEM medium (Thermo Fisher, USA) supplemented with 10% FBS (CellMax, China), 100 U/ml penicillin, and 100 mg/ml streptomycin (Sangon, China). The cells were cultivated in a humidified incubator (Thermo Fisher, USA) at 37 °C under 5% CO2. Short hairpin RNA (shRNA) lentivirus for AMD1 knockdown and negative control (NC) lentivirus were designed and synthesized by HANBIO (Shanghai, China), shRNA sequences are shown in Supplementary Table 1. The pHBLV-AMD1 vector was constructed by HANBIO (Shanghai, China). The vector was digested by an enzyme, and the interference fragment was connected by the AMD1 interference target and the synthesized primer. The lentiviral vector system consisted of shRNA vector plasmid, viral packaging helper plasmids (psPAX2 and pMD2G vectors). The three plasmids were co-transfected into 293T cells via LipofiterTM reagent. The transfected virus supernatant was collected, a high-titer lentiviral solution was obtained by ultracentrifugation. After 24 h, medium containing 2 µg/ml puromycin (Beyotime, China) was added for culture at 37 °C. After screening, AMD1-87-shRNA (Sh-1) and AMD1-312-shRNA (Sh-2) Huh7 and HepG2 cells were obtained. The sources of cells and experimental reagents are shown in Supplementary Table 2.

Real-time quantitative polymerase chain reaction

The cells in each group were collected, and the cells were lysed with TRIzol (Invitrogen, USA). Total RNA was extracted with chloroform and isopropyl alcohol, and the OD values and concentrations of the samples were determined with a spectrophotometer. The TB Green Premix DimerEraser Kit (TaKaRa, China) instructions were followed for PCR and reverse transcription of RNA into cDNA. The amplification conditions were as follows: 95 °C for 5 s, 55 °C for 30 s, and 72 °C for 34 s for a total of 40 cycles. After the reaction, a standard curve and amplification curve were automatically generated by the PCR instrument, and the 2-ΔΔCt sequence detection system (SDS) 2.4 was used to calculate expression. The primer sequences were as follows: AMD1 forward primer, 5’-CATCACTCCAGAACCAGAAT-3’; AMD1 reverse primer, 5’-TAACAAACAAGGTGGTCACA-3’; GAPDH forward primer, 5’-GAGTCAACGGATTTGGTCGT-3’; and GAPDH reverse primer, 5’-GACAAGCTTCCCGTTCTCAG-3’.

Cell counting kit-8 assay

Trypsin digestion of the cells produced a single-cell mixture with a regulated cell density of 2 × 104/ml. The cell suspension was inoculated into 96-well plates (2000 cells/well), and cultured in an incubator for 24, 48, 72, or 96 h. The cells were incubated with 10% cell counting kit-8 reagent (Beyotime, China) for 2 h. Using a multifunctional microplate reader, the sample’s absorbance value at 450 nm was determined, and the cell viability was computed.

Wound healing assay

The transfected cells were seeded in a 6-well plate and cultured for 12 h. The cells were wounded with a 10-µl pipet tip scraped across the monolayer and incubated after being washed with PBS. The speed of the wound recovery was photographed at 0, 24, and 48 h. Images were processed via ImageJ for data analysis.

Plate colony-forming assay

Single-cell transfected suspensions were added to Petri dishes (1000 cells/well), and cultured for 2 weeks to determine the growth state of the cells. After being fixed in 4% paraformaldehyde (Beyotime, China) for 15 min, the colonies were stained with 0.5% crystal violet (Beyotime, China) dye for 20 min, and the number of colonies (≥ 50 cells) was calculated after washing with distilled water.

Cell invasion assay

The cell invasion assay was performed using 24-well Corning Costar Transwell chambers (NY, USA). The transfected cells were resuspended in serum-free medium and inoculated into the upper chambers with Corning Matrigel, and 600 µL of culture medium containing 20% FBS was added to the bottom chamber. After incubation for 24 h, the invading cells were fixed in 4% paraformaldehyde for 20 min. Then, the cells were washed with PBS and stained with 0.5% crystal violet for 15 min. The cells were counted under a microscope.

Construction and validation of the prognostic signature using AMD1-related genes (ARGs)

Pearson’s correlation analysis revealed 32 ARGs (|FoldChange (FC)| > 2, P < 0.001). To construct a risk model based on ARGs, the Least absolute shrinkage and selection operator (LASSO) regression analysis and multivariate Cox analysis were performed. In the end, 9 genes were found to be vital for constructing the model. The risk score was determined as follows: risk score = ∑(Ci*Ei), where i refers to an individual ARG, C is the ARG coefficient, and E is the ARG expression. A median risk score was applied to divide HCC patients into low- and high-risk groups. The “survival” and “survminer” R packages were used to plot the survival status of patients with different risk patterns. The “pheatmap” R package was used to depict the expression of 9 genes in the risk model. The Kaplan‒Meier method was used to analyse the overall survival (OS). Using receiver operating characteristic (ROC) curves and the “timeROC” R package, the accuracy of the prognostic signature for prediction was evaluated. The clinical application value of the risk model was validated using C-Index and clinically relevant ROC curves.

Gene set enrichment analysis

The screening criteria for the differentially expressed mRNAs (DEmRNAs) were |log2FC| ≥ 1 and P < 0.05. We used the R package ‘limma’ for analysis. DEmRNAs were subjected to GO and KEGG enrichment analysis via the R packages “RColorBrewer,” “ggpubr,” “org.Hs.eg.db,” and “clusterProfiler.” P < 0.05 was used to classify KEGG pathways and GO terms as significantly enriched. 5 GO terms and 5 KEGG pathways with the lowest P values were selected for the study and displayed.

Immune-related functional analysis

The tumour purity score was obtained using the R package “Estimate”. CIBERSORT is a method for characterizing cell composition from gene expression profiles of complex tissues [22]. We evaluated the immune cell enrichment score via CIBERSORT, a tool that illustrates the relative abundance of infiltrating immune cells.

Cancer drug sensitivity analysis

The 50% inhibitory concentration (IC50) of each medication was assessed via the R program Oncopredict to calculate the value of the risk score in predicting drug sensitivity and chemotherapy based on the Genomics of Drug Sensitivity in Cancer drug data source. Patient sensitivity to medicine increases with decreasing IC50 values. Comparisons were conducted between the low- and high-risk groups in terms of drug sensitivity.

Nomogram construction

We conducted univariate and multivariate Cox analyses according to age, gender, stage, and risk score. A nomogram was created to predict HCC patients based on the risk score, age, and stage. Calibration curves can also be used to assess the degree to which expected and actual outcomes align. The R packages “survival,” “regplot,” and “rms” were used to display the aforementioned images.

Statistical analysis

All bioinformatic analyses were performed out via R 4.3.1. Differences in expression levels between groups were compared via the Wilcoxon rank sum test. For the correlation analysis, Pearson correlation was used. Statistics were considered significant if P < 0.05. Throughout this study, “*” represents P < 0.05, “**” represents P < 0.01, and “***” represents P < 0.001.

Results

AMD1 expression across the various cancer and HCC databases

A flowchart describes the procedure used in the present study (Fig. 1). First, pancancer analyses based on the TIMER database suggest that high AMD1 expression is noted in various types of cancer (Fig. 2A). Compared with that noted in equivalent normal tissues, AMD1 expression was substantially higher in the majority of cancer types, including cholangiocarcinoma, colon adenocarcinoma, esophageal carcinoma, and stomach adenocarcinoma. On the other hand, AMD1 downregulation was noted in bladder urothelial carcinoma, breast invasive carcinoma, glioblastoma multiforme, kidney chromophobe, kidney renal clear cell carcinoma, kidney renal papillary cell carcinoma, lung adenocarcinoma, and thyroid carcinoma. Moreover, AMD1 expression was upregulated in HCC tissues from the HCC datasets GSE36376, GSE14520, and OEP000321 (Fig. 2B, P < 0.001).

Fig. 1
figure 1

The workflow. The analysis flowchart of this study

Full size image
Fig. 2
figure 2

AMD1 expression. (A) Tumour Immune Estimation Resource analysis was used to identify differences in AMD1 expression between diverse cancer types and matched normal samples. (B) Analysis of AMD1 expression in hepatocellular carcinoma (HCC) cohorts. The expression values were scaled by Integrative Molecular Database of Hepatocellular Carcinoma analytical tools. Box plots displaying AMD1 expression in datasets. The red area represents tumor, whereas the green area represents adjacent tissue

Full size image

Gene enrichment analysis and GSVA

GO and KEGG functional enrichment analyses as well as GSVA were performed to investigate the biological processes and roles associated with AMD1 expression. Following the identification of the genes most closely associated with AMD1 expression, GO and KEGG analyses were conducted on these genes. In the TCGA and ICGC cohorts, GO analysis revealed a significant association between AMD1 and mRNA processing and cell division (Supplementary Fig. 1A‒C, E‒G). AMD1 expression was significantly correlated with endocytosis and the cell cycle in both cohorts according to KEGG analysis (Supplementary Fig. 1D, H). Cell cycle disorders may lead to the occurrence of malignant tumours. Thus, we investigated how AMD1 affects cytokine profiles and cell cycle processes. Gene set enrichment analysis of the TCGA and ICGC databases were used to calculate the enrichment score of the cell cycle process. AMD1 expression was positively associated with the majority of cell cycle processes, including the mitotic cell cycle, cell division, and cytokinesis, according to a correlation study between the enrichment score and AMD1 expression (Supplementary Fig. 2A). This finding was confirmed in the ICGC database (Supplementary Fig. 2B). In addition, AMD1 was positively correlated with the cell cycle molecules CDK1, CDK2, CDK2AP1, CDK4, CK19, CCNB1 and CCNB2 in TCGA and ICGC cohorts (Supplementary Fig. 2C). Moreover, the results of this study revealed that AMD1 was positively correlated with cell cycle signalling factors, such as c-MYC, RB, E2F, and TRET, in TCGA and ICGC cohorts (Supplementary Fig. 2D).

Prognostic significance of AMD1, its association with immunotherapy response and its expression pattern in immunocytes

We focused on the significance of AMD1 in HCC. Survival analysis revealed that patients with higher AMD1 expression had worse OS than those with lower AMD1 expression (Fig. 3A). The TIDE score of the low-risk group was lower than that of the high-risk group, and the immunotherapy response rate of the low-risk group was greater than that of the high-risk group (Fig. 3B). Furthermore, compared with the high-risk group, increased dysfunction was noted in the low-risk group. Compared with the high-risk group, there were fewer exclusions in the low-risk group. To gain additional insight into AMD1 expression patterns inside the tumour microenvironment (TME) of HCC, we conducted single-cell profile-based experiments. Different cell types in the TME exhibit different AMD1 expression levels. Tprolif cells exhibited the highest levels of AMD1 expression (Fig. 3C‒D).

Fig. 3
figure 3

Prognostic significance and expression pattern of AMD1. (A) The overall survival (OS) curves for high- and low-AMD1 expressing subpopulations from The Cancer Genome Atlas (TCGA) liver hepatocellular carcinoma datasets. (B) Tumor Immune Dysfunction and Exclusion (TIDE) algorithm was used to predict the clinical response of immune checkpoint inhibitors. TIDE scores, Dysfunction scores and Exclusion scores of each patient were acquired by TIDE algorithm. Microsatellite instability (MSI) algorithm was used to evaluated the relationship with immune response. Violin plot to show differences of TIDE, Dysfunction, Exclusion and MSI scores between high-risk and low-risk groups. (C) Annotation of all cell types in GSE98638 and the percentage of each cell type. (D) AMD1 expression in each cell type

Full size image

Silencing AMD1 inhibits HCC cell activity, proliferation and invasion in vitro

The RT-PCR results revealed that the AMD1 expression values in Huh7, HepG2 and PLC/PRF/5 cells were 13.19 ± 0.01, 6.55 ± 0.01 and 6.26 ± 0.01, respectively, which were greater than those in LO2 cells (1.00 ± 0.00) (Fig. 4A, P < 0.001). Compared with that in Huh7 cells (13.28 ± 0.13) and the NC group (13.05 ± 0.70), AMD1 expression values in Sh-1 and Sh-2 Huh7 cells were 2.66 ± 0.58 and 3.18 ± 0.13, respectively (Fig. 4B, P < 0.001). Compared with the HepG2 cell (6.55 ± 0.06) and NC groups (6.50 ± 0.22), the Sh-1 and Sh-2 HepG2 cell groups presented values of 1.17 ± 0.14 and 1.23 ± 0.05, respectively (Fig. 4C, P < 0.001). Huh7 and HepG2 cell viability was reduced after infection with lentivirus. With the increase of interference time, the inhibition of its proliferation ability gradually became obvious and showed a time-gradient dependence (Fig. 4D). Following AMD1 silencing, Huh7 and HepG2 wound healing rates were significantly slower than control group (Fig. 4E). According to the results of the cell plate colony-forming assay, there were fewer cell clones in the silent group (Fig. 4F). Transwell invasion assays showed that the numbers of cells passing through the microporous membrane in the AMD1-knockdown group were decreased than the NC group (Fig. 4G). The experimental results of changes in cell viability, scratch area rate, colony formation and invasion ability of AMD1-knockdown HCC cells can be obtained in Supplementary Table 3.

Fig. 4
figure 4

In vitro validation experiments. (A) qRT‒PCR validation of AMD1 expression in human normal liver cells (L02) and HCC cell lines. (B-C) qRT‒PCR validation of transfection efficiency in Huh7 and HepG2 cell. (D) The effect of silencing AMD1 on cell proliferation was explored via the cell counting kit-8 assay. (E) The wound healing assay was used to assess the effect of silencing AMD1 on cell migration. (F) The colony formation assay was conducted to detect HCC cell proliferation after silencing AMD1. (G) Matrigel invasion assay was used to assess the effect of silencing AMD1 on HCC cell invasion. The error bars indicate standard deviation. **p < 0.01, and ***p < 0.001

Full size image

The AMD1-derived genomic model for HCC prognosis

The Pearson correlation coefficient test identified 32 genes associated with AMD1. The selection of genes was based on two criteria: FC > 2 and P < 0.001. LASSO analysis was then used to select 9 genes from the 32 genes for subsequent analyses in order to precisely construct the prognostic model (Supplementary Fig. 3A‒C). Ultimately, 9 ARGs were identified via multivariate Cox proportional hazards regression analysis and subsequently utilized to construct a prognostic model. Risk score = (3.037429× EXP CREG2) + (-0.335394 × EXP GPR171) + (0.001656× EXP HMGA1) + (0.184712× EXP MTMR2) + (0.008732× EXP NCF2) + (0.801766× EXP OTOGL) + (0.000586× EXP PLXNA3) + (0.548550× EXP PPFIA4) + (0.005455× EXP STX3).

Verification of the prognostic significance of the AMD1-derived genomic model in HCC

The training cohort (60%, n = 204) and testing cohort (40%, n = 136) were randomly assigned to the 340 HCC patients (Supplementary Table 4). The training and testing cohorts were split into high- and low-risk groups on the basis of the median risk score. The distributions of the HCC patients in the training cohort are displayed in the upper panel of Supplementary Fig. 4A. The associations among patient survival time, survival status, and the risk score in the training cohort are shown in the lower panel of Supplementary Fig. 4A. Supplementary Fig. 4D displays the expression of 9 ARGs of the training cohort. The distributions of HCC patients according to risk score, survival duration, survival status, and 9 AMD1-related gene expression in the testing cohort are displayed in Supplementary Fig. 4B, E. The external validation cohort was also subject to corresponding analyses (Supplementary Fig. 4C, F). Compared with patients in the low-risk cohort, those in the high-risk cohort had significantly shorter OS (Fig. 5A, P < 0.001). Both the testing cohort and the external validation cohort verified the aforementioned results (Fig. 5B‒C, P < 0.01). Area under curve (AUC) values of 0.751, 0.734, and 0.750 were obtained for the training cohort, predicting OS at 1, 3, and 5 years, respectively (Fig. 5D). AUC values of 0.770, 0.788, and 0.680 were detected for the testing cohort, predicting OS at 1, 3, and 5 years, respectively (Fig. 5E). The AUC values for the external validation cohort that predicted OS at 1, 3, and 5 years were 0.802, 0.735, and 0.845, respectively (Fig. 5F). On the basis of the training, testing, and external validation cohorts, we generated clinical ROC curves and C-index curves and determined that the risk score was the most useful predictor of prognosis (Fig. 5G‒L).

Fig. 5
figure 5

Validation of the prognostic signature. (A-C) The OS rates of the high-risk and low-risk groups are shown for the training, testing, and external validation cohorts. (D-F) Receiver operating characteristic (ROC) curve showing the predictive accuracy of the risk score for the prognosis of the patients in the training, testing, and external validation cohorts. (G-I) ROC curves were generated to confirm the superiority of the risk score in clinical application in the training, testing, and external validation cohort. (J-L) The C-index curves for the risk score, age, gender and stage in the training, testing, and external validation cohorts

Full size image

Molecular pathways associated with the genomic signature derived from AMD1

Furthermore, we examined the molecular processes underlying the genetic signature derived from AMD1. The high-risk subpopulation had significantly activated pathways, as illustrated in Supplementary Fig. 5A‒B. These pathways included the alpha amino acid catabolic process, cellular amino acid catabolic process, drug metabolism cytochrome p450, fatty acid metabolism, and peroxisome, which are suggestive of the subpopulation’s adverse survival outcomes. Moreover, the low-risk subpopulation presented increased activity in the adaptive immune response, positive regulation of cell activation, regulation of immune effector process, and cytokine-cytokine receptor interaction (Supplementary Fig. 5C‒D).

Evaluation of immune infiltration on the basis of risk signature groupings

After assigning scores, including stromal, immune, and estimation scores, to immunotherapy, we evaluated how these scores varied between the two risk groups (Fig. 6A). According to these findings, the low-risk group outperformed the high-risk group in terms of immune infiltration, stromal, and estimation scores. Furthermore, the risk score was positively correlated with the infiltration of resting mast cells, activated NK cells, and monocytes but negatively correlated with the infiltration of native B cells and CD8 + T cells (Fig. 6B‒G). These findings suggest that the immunological profile of HCC patients can be predicted using our risk model.

Fig. 6
figure 6

Immunological abnormalities between risk score groups. (A) Evaluation of the differences in the stromal score, immune score, and tumour purity between the high-risk and low-risk groups. (B) Comparison of the infiltration of immune cells between the high-risk and low-risk groups. (C-G) Correlations between the risk score and resting mast cells, activated NK cells, monocytes, native B cells and CD8 + T cells. Comparisons between two groups were assessed using the Wilcoxon test, spearman test was used to the correlation analysis. Values less than Q1 (Lower quartile) -1.5IQR (Interquartile range) or greater than Q3 (Upper quartile) + 1.5IQR in the box plot were identified as outliers. *p < 0.05, **p < 0.01, ***p < 0.001

Full size image

The AMD1-derived genomic model forecasts the medication response of HCC patients

The management of HCC is significantly affected by targeted medicines, and responses to these therapies are associated with genetic abnormalities that contribute to HCC heterogeneity. There is still a severe dearth of accurate predictive biomarkers for focused treatments. To determine whether the medication responses of HCC patients can be predicted using this genetic model established from AMD1, more investigations are needed. The “Oncopredict” R package was used to calculate the predicted IC50 based on the Genomics of Drug Sensitivity in Cancer drug data source and gene expression profile data (Supplementary Table 5). Figure 7 suggests that high-risk subpopulations were more sensitive to dasatinib and gefitinib. Specifically, the high-risk group had considerably lower estimated IC50 values of dasatinib and gefitinib compared to the low-risk group. Axisplatin, cisplatin, cytarabine, fludarabine, gemcitabine, oxaliplatin, and sorafenib exhibited significantly lower estimated IC50 values in the low-risk group than in the high-risk group, suggesting that low-risk subpopulations are more likely to respond to these treatments. These findings suggest a potential link between susceptibility to the aforementioned drugs and genes originating from AMD1.

Fig. 7
figure 7

The “Oncopredict” R package was used to calculate the predicted 50% inhibitory concentration (IC50) based on the Genomics of Drug Sensitivity in Cancer drug data source and the training cohort gene expression profile data, boxplot to show differences of predicted drug IC50 value between risk score groups. (A) Axitini. (B) Cisplatin. (C) Cytarabine. (D) Dasatinib. (E) Fludarabine. (F) Gemcitabine. (G) Gefitinib. (H) Oxaliplatin. (I) Sorafenib. Comparisons between two groups were assessed using the Wilcoxon test

Full size image

Construction of a credible nomogram for predicting the prognosis of HCC patients

The univariate Cox regression analysis in the upper panel of Fig. 8A-C revealed a substantial correlation between the stage and the risk score derived from AMD1 and the prognosis of HCC in the training, testing, and external validation cohorts. In the training and external validation cohorts, the stage and AMD1-derived risk score, as shown in the lower panel of Fig. 8A‒C, functioned as independent predictive predictors of HCC. Age, stage, and the risk score were combined to construct a nomogram (Fig. 8D). We also assessed the nomogram’s prediction performance using calibration curves. Our findings showed that the nomogram’s 1-, 3-, and 5-year predictions were fairly accurate with respect to the actual survival duration (Fig. 8E–G). The statistics above demonstrate the high prediction ability of this nomogram.

Fig. 8
figure 8

Univariate and multivariate Cox regression analyses and the nomogram. (A-C) Univariate and multivariate Cox regression models were constructed to determine the associations of clinical features and the AMD1-derived risk score with HCC survival outcomes in the training, testing, and external validation cohorts. (D) A prognostic nomogram is exploited through the integration of prognostic indicators (age, stage, and AMD1-derived risk score) to estimate the 1-, 3-, and 5-year survival probabilities. (E-G) Calibration plots showing the associations of the predicted 1-, 3-, and 5-year OS with the actual survival duration in the training, testing, and external validation cohorts

Full size image

Discussion

HCC is a solid tumour that rapidly progresses and has a poor prognosis. Owing to their low sensitivity and specificity, tumour-node-metastasis staging, alpha-fetoprotein, and other indicators can be used to predict the prognosis of HCC patients but cannot serve as independent indicators of poor prognosis. During the gradual development of HCC, numerous genetic alterations lead to phenotypic abnormalities in hepatocytes. These changes result in the formation of intermediate cells and various monoclonal populations, which ultimately develop into HCC. The identification of critical targets and novel predictive signatures associated with cancer diagnosis and prognosis has been made possible by the combination of high-throughput sequencing analysis and bioinformatics tools [23, 24].

The gene encoding human AMD1 mRNA is located on chromosome 6q21, which is a linear structure containing 10 exons [25]. AMD1 is involved in the rate-limiting step of polyamine synthesis and is closely related to the malignant transformation of cells [26, 27]. AMD1 is involved in cell cycle regulation, protein translation, epigenetic regulation and various cancer signal transduction pathways, affecting tumour prognosis [28]. Our findings revealed that AMD1 expression is markedly upregulated in the majority of cancer types. AMD1 expression is upregulated in HCC and considerably greater than that in normal liver tissues in this study, which involved the analysis of HCC gene sequencing data from several tumour databases. These findings suggest that AMD1 expression may potentially impact the disease course. HCC patients with high AMD1 expression have increased disease malignancy and poor survival. As a result, AMD1 is a significant proto-oncogene in HCC and a molecular marker for a poor prognosis in this disease. Our additional investigations revealed a correlation between AMD1 overexpression and the cancer cell cycle. The information we gathered suggested that AMD1 could be a valuable predictor. In this study, silencing AMD1 effectively inhibited HCC cell proliferation. AMD1 polyamine catalytic products are involved in MAPK phosphorylation; promote the expression of the proto-oncogenes c-myc, c-fos and c-jun [29]; regulate downstream protein translation; and control the growth of proliferating tumour cells. AMD1 is involved in the synthesis and translation of spermine, which is a hydroxyl-containing protein that is a necessary precursor of eIF5A synthesis [30]. Some scholars have noted that AMD1 expression is upregulated in the G1‒S phase of cell division and that cell growth, differentiation and apoptosis are periodically regulated by the regulation of cyclins/CDKs [31, 32]. In addition, AMD1 is involved in the S-adenosine methionine metabolic pathway, which is a cofactor of chromatin methylation that maintains epigenetic inheritance. S-adenosine methionine is downregulated in tumours, cell proliferation is stopped, and cells are arrested at the G1 phase.

Our research verified that AMD1 potentially promotes HCC cell invasion and migration in vitro. Adenosine methionine decarboxylation in cells is involved in the hydroxylation and lysine modification of eIF5A. Inhibitors of AMD1 effectively reduce eIF5A levels [33]. Research indicates that eIF5A2 regulates epithelial‒mesenchymal transition in gastric, breast, liver, and other cancers [34,35,36]. By stimulating the RhoA/Rac1 pathway [34, 37] and downregulating cadherin and β-catenin [38], eIF5A2 contributes significantly to HCC invasion. AMD1 activates the spermidine-eIF5A hypusination-TCF4 axis, and contributes significantly to the prognosis of basal-like breast cancer [39]. Constitutive activation of JNKs and c-jun phosphorylation, accompanied by downregulation of TSP-1 and upregulation of MMP-2, occur in highly invasive tumour cells overexpressing AMD1; these processes mediate tumour invasion by degrading vascular walls [40]. Furthermore, AMD1 expression is directly correlated with T-cell proliferation in the TME and influences the effectiveness of immunotherapy in patients with HCC. Multiple immune-related pathways are highly correlated with polyamine metabolism [41, 42]. The increased uptake of polyamines by immune cells leads to the production of tumour killer cytokines and a reduction in the number of adhesion molecules, such as CD11a and CD56, which weakens the cytotoxic ability of immune cells, suppresses host immunity [43], and encourages tumour cell growth, migration, and invasion.

We constructed an AMD1-derived genomic model with HMGA1, GPR171, PPFIA4, CREG2, OTOGL, NCF2, STX3, MTMR2, and PLXNA3 using LASSO-Cox analysis. The genetic model accurately and independently predicted the prognosis of the patients after verification in a validation cohort. The model also has the ability to predict the effects of medications such as sorafenib, oxaliplatin, gefitinib, cisplatin, cytarabine, fludarabine, and gemcitabine.

By combining many risk indicators, a nomogram is an effective tool for estimating a person’s risk in a clinical environment [44]. In this investigation, age, stage, and an AMD1-derived genomic model for HCC patients were combined with the nomogram to predict the probability of 1-, 3-, and 5-year OS probability. A score was given to each factor based on its contribution to the risk of survival. Furthermore, calibration curves demonstrated that the anticipated survival duration from the nomogram and the actual survival duration were consistent. The predictive efficacy of AMD1 needs to be confirmed in larger cohorts of HCC patients.

The study’s shortcomings and future directions are discussed in more detail below. There is a dearth of clinical liver cancer tissue and adjacent normal liver tissue. We need to validate the level of protein expression of AMD1 in liver and HCC. Furthermore, the precise role of AMD1 in HCC and its pathogenesis remain unclear. Further investigations into the molecular mechanism and in vivo evidence pertaining to AMD1-silenced HCC cells are needed to provide a deeper understanding of potential therapeutic targets.

Conclusions

In summary, using gene sequencing data from HCC tissues, this study examined AMD1 expression in HCC and its impact on disease prognosis. The risk model that was created in this study is anticipated to be used to predict the prognosis and treatment outcome of HCC patients. Furthermore, AMD1-silenced HCC cells were generated via a lentiviral vector, revealing the effects of AMD1 on HCC cell growth, migration, and invasion of HCC cells in vitro. This information is helpful in identifying new mechanisms and intervention strategies for molecular targeted therapy of HCC.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

ARGs:

AMD1-related genes

AUC:

Area under curve

FC:

FoldChange

GO:

Gene Ontology

GSVA:

Gene set variation analysis

HCC:

Hepatocellular carcinoma

IC50 :

50% inhibitory concentration

ICGC:

International Cancer Genome Consortium

KEGG:

Kyoto Encyclopedia of Genes and Genomes

LASSO:

Least absolute shrinkage and selection operator

LIHC:

Liver hepatocellular carcinoma

MSI:

Microsatellite instability

NC:

Negative control

OS:

Overall survival

ROC:

Receiver operating characteristic

shRNA:

Short hairpin RNA

TCGA:

The Cancer Genome Atlas

TME:

Tumour microenvironment

TIDE:

Tumor Immune Dysfunction and Exclusion

TIMER:

Tumour Immune Estimation Resource

References

  1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, Jemal A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229–63.

    Article  PubMed  Google Scholar 

  2. Lee DH, Kim D, Park YH, Yoon J, Kim JS. Long-term surgical outcomes in patients with hepatocellular carcinoma undergoing laparoscopic vs. open liver resection: a retrospective and propensity score-matched study. Asian J Surg. 2021;44(1):206–12.

    Article  PubMed  Google Scholar 

  3. Orcutt ST, Anaya DA. Liver Resection and Surgical strategies for management of primary Liver Cancer. Cancer Control. 2018;25(1):1073274817744621.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Forner A, Reig M, Bruix J. Hepatocellular carcinoma. Lancet. 2018;391(10127):1301–14.

    Article  PubMed  Google Scholar 

  5. Bale S, Ealick SE. Structural biology of S-adenosylmethionine decarboxylase. Amino Acids. 2010;38(2):451–60.

    Article  CAS  PubMed  Google Scholar 

  6. Miller-Fleming L, Olin-Sandoval V, Campbell K, Ralser M. Remaining mysteries of Molecular Biology: the role of polyamines in the cell. J Mol Biol. 2015;427(21):3389–406.

    Article  CAS  PubMed  Google Scholar 

  7. Li J, Zhang L, Xiong J, Cheng X, Huang Y, Su Z, Yi M, Liu S. Polyamines disrupt the KaiABC Oscillator by inducing protein denaturation. Molecules. 2019; 24(18).

  8. Evageliou NF, Haber M, Vu A, Laetsch TW, Murray J, Gamble LD, Cheng NC, Liu K, Reese M, Corrigan KA, et al. Polyamine Antagonist therapies inhibit neuroblastoma initiation and progression. Clin Cancer Res. 2016;22(17):4391–404.

    Article  CAS  PubMed  Google Scholar 

  9. Murray-Stewart TR, Woster PM, Casero RA. Jr. Targeting polyamine metabolism for cancer therapy and prevention. Biochem J. 2016;473(19):2937–53.

    Article  CAS  PubMed  Google Scholar 

  10. Kaminski L, Torrino S, Dufies M, Djabari Z, Haider R, Roustan FR, Jaune E, Laurent K, Nottet N, Michiels JF, et al. PGC1alpha inhibits polyamine synthesis to suppress prostate Cancer aggressiveness. Cancer Res. 2019;79(13):3268–80.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Zhang D, Zhao T, Ang HS, Chong P, Saiki R, Igarashi K, Yang H, Vardy LA. AMD1 is essential for ESC self-renewal and is translationally down-regulated on differentiation to neural precursor cells. Genes Dev. 2012;26(5):461–73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Zhao T, Goh KJ, Ng HH, Vardy LA. A role for polyamine regulators in ESC self-renewal. Cell Cycle. 2012;11(24):4517–23.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Zabala-Letona A, Arruabarrena-Aristorena A, Martín-Martín N, Fernandez-Ruiz S, Sutherland JD, Clasquin M, Tomas-Cortazar J, Jimenez J, Torres I, Quang P, et al. mTORC1-dependent AMD1 regulation sustains polyamine metabolism in prostate cancer. Nature. 2017;547(7661):109–13.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Xu L, You X, Cao Q, Huang M, Hong LL, Chen XL, Lei L, Ling ZQ, Chen Y. Polyamine synthesis enzyme AMD1 is closely associated with tumorigenesis and prognosis of human gastric cancers. Carcinogenesis. 2019.

  15. Chen K, Liu H, Liu Z, Luo S, Patz EF Jr., Moorman PG, Su L, Shen S, Christiani DC, Wei Q. Genetic variants in RUNX3, AMD1 and MSRA in the methionine metabolic pathway and survival in nonsmall cell lung cancer patients. Int J Cancer. 2019;145(3):621–31.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Gomes AP, Schild T, Blenis J. Adding polyamine metabolism to the mTORC1 Toolkit in Cell Growth and Cancer. Dev Cell. 2017;42(2):112–4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Gupta ED, Pachauri M, Ghosh PC, Rajam MV. Targeting polyamine biosynthetic pathway through RNAi causes the abrogation of MCF 7 breast cancer cell line. Tumour Biol. 2016;37(1):1159–71.

    Article  CAS  PubMed  Google Scholar 

  18. Li T, Fu J, Zeng Z, Cohen D, Li J, Chen Q, Li B, Liu XS. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 2020;48(W1):W509–14.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Lian Q, Wang S, Zhang G, Wang D, Luo G, Tang J, Chen L, Gu J. HCCDB: a database of Hepatocellular Carcinoma expression Atlas. Genomics Proteom Bioinf. 2018;16(4):269–75.

    Article  Google Scholar 

  20. Sherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, Imamichi T, Chang W. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50(W1):W216–21.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Han Y, Wang Y, Dong X, Sun D, Liu Z, Yue J, Wang H, Li T, Wang C. TISCH2: expanded datasets and new tools for single-cell transcriptome analyses of the tumor microenvironment. Nucleic Acids Res. 2023;51(D1):D1425–31.

    Article  PubMed  Google Scholar 

  22. Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Bian G, Cao J, Li W, Huang D, Ding X, Zang X, Ye Y, Li P. Identification and validation of a Cancer-Testis Antigen-related signature to predict the prognosis in stomach adenocarcinoma. J Cancer. 2024;15(11):3596–611.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Gui Z, Ye Y, Li Y, Ren Z, Wei N, Liu L, Wang H, Zhang M. Construction of a novel cancer-associated fibroblast-related signature to predict clinical outcome and immune response in cervical cancer. Transl Oncol. 2024;46:102001.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Rubina T, Alok J, Ganesh C, Om Prakash D, Saurabh G, Marwaha RK, Nikhil T, Dwaipayan B. Genetic variant of AMD1 is associated with obesity in urban Indian children. PLoS ONE. 2012;7(4):e33162.

    Article  Google Scholar 

  26. Casero RA Jr., Murray Stewart T, Pegg AE. Polyamine metabolism and cancer: treatments, challenges and opportunities. Nat Rev Cancer. 2018;18(11):681–95.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Yordanova MM, Loughran G, Zhdanov AV, Mariotti M, Kiniry SJ, O’Connor PBF, Andreev DE, Tzani I, Saffert P, Michel AM, et al. AMD1 mRNA employs ribosome stalling as a mechanism for molecular memory formation. Nature. 2018;553(7688):356–60.

    Article  CAS  PubMed  Google Scholar 

  28. Lim HK, Rahim AB, Leo VI, Das S, Lim TC, Uemura T, Igarashi K, Common J, Vardy LA. Polyamine Regulator AMD1 promotes Cell Migration in epidermal Wound Healing. J Invest Dermatol. 2018;138(12):2653–65.

    Article  CAS  PubMed  Google Scholar 

  29. Arruabarrena-Aristorena A, Zabala-Letona A, Carracedo A. Oil for the cancer engine: the cross-talk between oncogenic signaling and polyamine metabolism. Sci Adv. 2018;4(1):eaar2606.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Shin BS, Katoh T, Gutierrez E, Kim JR, Suga H, Dever TE. Amino acid substrates impose polyamine, eIF5A, or hypusine requirement for peptide synthesis. Nucleic Acids Res. 2017;45(14):8392–402.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Yamashita T, Nishimura K, Saiki R, Okudaira H, Tome M, Higashi K, Nakamura M, Terui Y, Fujiwara K, Kashiwagi K, et al. Role of polyamines at the G1/S boundary and G2/M phase of the cell cycle. Int J Biochem Cell Biol. 2013;45(6):1042–50.

    Article  CAS  PubMed  Google Scholar 

  32. Lin DW, Chung BP, Kaiser P. S-adenosylmethionine limitation induces p38 mitogen-activated protein kinase and triggers cell cycle arrest in G1. J Cell Sci. 2014;127(Pt 1):50–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  33. Tang D-J, Dong S-S, Ma N-F, Xie D, Chen L, Fu L, Lau SH, Li Y, Li Y, Guan X-Y. Overexpression of eukaryotic initiation factor 5A2 enhances cell motility and promotes tumor metastasis in hepatocellular carcinoma. Hepatology. 2010;51(4):1255–63.

    Article  CAS  PubMed  Google Scholar 

  34. Fujimura K, Wright T, Strnadel J, Kaushal S, Metildi C, Lowy AM, Bouvet M, Kelber JA, Klemke RL. A hypusine-eIF5A-PEAK1 switch regulates the pathogenesis of pancreatic cancer. Cancer Res. 2014;74(22):6671–81.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Liu RR, Lv YS, Tang YX, Wang YF, Chen XL, Zheng XX, Xie SZ, Cai Y, Yu J, Zhang XN. Eukaryotic translation initiation factor 5A2 regulates the migration and invasion of hepatocellular carcinoma cells via pathways involving reactive oxygen species. Oncotarget. 2016;7(17):24348–60.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Wang FW, Guan XY, Xie D. Roles of eukaryotic initiation factor 5A2 in human cancer. Int J Biol Sci. 2013;9(10):1013–20.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Muramatsu T, Kozaki KI, Imoto S, Yamaguchi R, Tsuda H, Kawano T, Fujiwara N, Morishita M, Miyano S, Inazawa J. The hypusine cascade promotes cancer progression and metastasis through the regulation of RhoA in squamous cell carcinoma. Oncogene. 2016;35(40):5304–16.

    Article  CAS  PubMed  Google Scholar 

  38. Wang C, Ruan P, Zhao Y, Li X, Wang J, Wu X, Liu T, Wang S, Hou J, Li W, et al. Spermidine/spermine N1-acetyltransferase regulates cell growth and metastasis via AKT/beta-catenin signaling pathways in hepatocellular and colorectal carcinoma cells. Oncotarget. 2017;8(1):1092–109.

    Article  PubMed  Google Scholar 

  39. Liao R, Chen X, Cao Q, Bai L, Ma C, Dai Z, Dong C. AMD1 promotes breast cancer aggressiveness via a spermidine-eIF5A hypusination-TCF4 axis. Breast Cancer Res. 2024;26(1):70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Paasinen-Sohns A, Kaariainen E, Yin M, Jarvinen K, Nummela P, Holtta E. Chaotic neovascularization induced by aggressive fibrosarcoma cells overexpressing S-adenosylmethionine decarboxylase. Int J Biochem Cell Biol. 2011;43(3):441–54.

    Article  CAS  PubMed  Google Scholar 

  41. Chamoto K, Zhang B, Tajima M, Honjo T, Fagarasan S. Spermidine - an old molecule with a new age-defying immune function. Trends Cell Biol. 2024;34(5):363–70.

    Article  CAS  PubMed  Google Scholar 

  42. Li B, Kong Z, Liu Y, Xu B, Liu X, Li S, Zhang Z. A polyamine metabolism risk signature for predicting the prognosis and immune therapeutic response of kidney cancer. Transl Cancer Res. 2023;12(10):2477–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Yue F, Li W, Zou J, Jiang X, Xu G, Huang H, Liu L. Spermidine Prolongs Lifespan and prevents liver fibrosis and Hepatocellular Carcinoma by activating MAP1S-Mediated autophagy. Cancer Res. 2017;77(11):2938–51.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Miao Y, Li Q, Wang J, Quan W, Li C, Yang Y, Mi D. Prognostic implications of metabolism-associated gene signatures in colorectal cancer. PeerJ. 2020;8:e9847.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We would like to thank the public databases for data availability.

Funding

Not applicable.

Author information

Author notes

  1. Youliang Zhou and Yi Zhou contributed equally to this work.

Authors and Affiliations

  1. Department of Medical Insurance, Affiliated Hangzhou First People’s Hospital, School of Medicine, Westlake University, Hangzhou, 310015, Zhejiang, China

    Youliang Zhou, Yao Xiao & Liping Yu

  2. Department of Orthopedic, Yuhuan Hospital of Traditional Chinese Medicine, Taizhou, 317600, Zhejiang, China

    Yi Zhou

  3. Department of Surgery, Yuhuan Hospital of Traditional Chinese Medicine, Taizhou, 317600, Zhejiang, China

    Jiabin Hu

  4. Department of Pediatrics, The Yuhuan Branch of The First Affiliated Hospital with Wenzhou Medical University, Taizhou, 317600, Zhejiang, China

    Yan Zhou

Authors
  1. Youliang Zhou
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yi Zhou
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Jiabin Hu
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Yao Xiao
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Yan Zhou
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Liping Yu
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

LPY and YZ designed the study and provided financial support for the experiments in personal capacity. YLZ drafted the initial manuscript and performed experiments. YZ acquired the data and performed the bioinformatics analyses. JBH and YX performed the statistical analysis and revised the manuscript. All authors approved the final manuscript. All authors contributed to the article and approved the submitted version.

Corresponding authors

Correspondence to Yan Zhou or Liping Yu.

Ethics declarations

Ethical approval

The present study used secondary data which are available in public domain. The dataset has no identifiable information of the survey participants. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Fig. 1

: Enrichment of differentially expressed genes. (A-C) Biological process, cellular component, and molecular function are mostly related to AMD1 in the The Cancer Genome Atlas (TCGA) database. (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of AMD1 in the TCGA database. (E-G) Biological process, cellular component, and molecular function are mostly related to AMD1 in the International Cancer Genome Consortium (ICGC) database. (H) KEGG pathway analysis of AMD1 in the ICGC database.

Supplementary Fig. 2

: Correlation analysis between AMD1 expression and the cell cycle. (A, B) Heatmap showing AMD1 expression and the enrichment scores of the cell cycle of each patient in the TCGA and ICGC databases. The samples were arranged in ascending order of AMD1 expression. The column graph and line graph on the right show the R-value and P-value of the correlation analysis. (C) Circos plot displaying the interconnectivity among cell cycle genes related to AMD1. The thickness and colour of the ribbons correlate with the correlation of gene expression in TCGA and ICGC datasets. (D) Correlation matrix of AMD1 and cell cycle-related metagenes. The bottom left shows the correlation coefficient. The correlation coefficients are presented as proportions of the pie charts. The red area represents a positive correlation, whereas the green area represents a negative correlation. The correlation was tested via Pearson correlation analysis

Supplementary Fig. 3

: Screening of key genes associated with AMD1. (A) The forest plot shows the AMD1-related genes (ARGs) chosen by univariate Cox regression analysis. (B) The Least absolute shrinkage and selection operator (LASSO) regression analysis revealed that the trajectory of each independent variable varied. (C) ARGs were selected via LASSO regression

Supplementary Fig. 4

: Patient risk score in relation to risk and survival and the expression analysis of 9 genes in the high- and low-risk groups. (A-C) The upper panel shows the distributions of high-risk and low-risk HCC patients in the training, testing, and external validation cohorts. The lower panel displays the relationships between patient survival time, survival status, and the risk score in the training, testing, and external validation cohorts. (D-F) Heatmap displaying 9 ARGs ' expression in the training, testing, and external validation cohorts

Supplementary Fig. 5

: Gene set enrichment analysis. (A-B) Biological activities enriched in the high-risk group in the training cohort. (C-D) Biological activities enriched in the low-risk group.

Supplementary Table 1

: Sequence data

Supplementary Table 2

: Key resources table

Supplementary Table 3

: Cell viability, scratch area rate, colony formation and invasion of HCC cells with AMD1-silenced or not

Supplementary Table 4

: Baseline data sheet for the cohort of TCGA liver hepatocellular carcinoma datasets

Supplementary Table 5

: The Drug Sensitivity in Cancer data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhou, Y., Zhou, Y., Hu, J. et al. Prognostic, oncogenic roles, and pharmacogenomic features of AMD1 in hepatocellular carcinoma. Cancer Cell Int 24, 398 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12935-024-03593-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12935-024-03593-x

Keywords

Cancer Cell International

ISSN: 1475-2867

Contact us