Proteomic and metabolomic analysis of platelet related samples reveals energy metabolism disorders in hepatocellular carcinoma

Abstract

Background

Hepatocellular carcinoma (HCC) is one of the most prevalent cancers in the world. Platelets play an important role in thrombosis, inflammation, and tumors. This study tries to gain a pathway-based view of platelets-related samples for understanding metabolic disorders and identifying novel biomarkers in HCC by combining proteomics and metabolomics.

Methods

Forty-five HCC patients and thirty-four healthy controls were included in the study. We performed label-free proteomic analysis of platelets from 14 HCC patients and 14 healthy controls. Target metabolomics analysis was performed on platelet-rich plasma (PRP) from 31 HCC patients and 20 healthy controls. Western blotting was performed to validate the results of proteomics. Glutamine (Gln) deprivation assay was conducted to evaluate the effect of Gln metabolism on the platelet-induced proliferation, migration, and invasion of HCC cells.

Results

Proteomics analysis revealed dysregulation of platelets function and energy metabolism in HCC patients compared with healthy controls. Target metabolomics analysis showed widespread dysregulation of amino acids in HCC patients. Integrating analysis of differential proteins and metabolites revealed five significant dysregulated pathways in HCC patients. Western blotting validation results showed that the expression levels of SDHB, CISY, and FUMH were significantly up-regulated in HCC patients compared to healthy controls, which was consistent with the proteomics findings. Biological function revealed that Gln-free weakens the ability of platelet-induced proliferation, migration, and invasion of HCC cells. Diagnostic evaluation demonstrated superior discriminatory power for SDHB (AUC = 0.929) compared to alpha-fetoprotein (AFP) in platelet proteomics. Furthermore, a triad of tricarboxylic acid cycle intermediates (succinic acid, fumaric acid, and malic acid) significantly enhanced the AUC, specificity, and sensitivity of distinguishing HCC patients from healthy controls compared to AFP.

Conclusions

Through multi-omics characterization of platelets-related samples, the network of altered proteins and metabolites provides a comprehensive view of altered metabolism in the peripheral circulation of HCC patients. Gln deprivation inhibited the ability of platelet-induced malignant biology function in HCC cells. Collectively, proteomics and metabolomics provide evidence for possible future non-invasive or minimally invasive biopsies for patients with HCC.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-025-06694-x.

Background

Hepatocellular carcinoma (HCC) is the most prevalent subtype of liver cancer and ranks as the third leading cause of cancer-related deaths worldwide [1]. The risk factors for HCC include chronic hepatitis B/C virus infection, nonalcoholic fatty liver disease, smoking, obesity, and metabolic disorders [2]. China has the heaviest burden of HCC worldwide owing to the prevalence of hepatitis B. Despite tremendous advances in diagnosis using molecular biomarkers, the 5-year survival rate for HCC remains unsatisfactory [3]. Therefore, identifying new biomarkers and exploring the underlying pathological mechanisms are essential for improving the detection and treatment of HCC.

Most studies have focused on identifying cancer biomarkers or assessing therapeutic efficacy using serum or plasma. Serum alpha-fetoprotein (AFP) is the most commonly used biomarker for HCC diagnosis. However, only 20% of patients with early-stage HCC have elevated AFP [4]. Recently, platelets, which are rich in bioactive substances, could serve as a novel source for discovering cancer biomarkers [5, 6]. Platelets are small pieces of cytoplasm shed from megakaryocytes in the bone marrow, with a complete cell membrane on the surface but no nucleus. Platelets are also the smallest peripheral blood cells besides red blood cells and white blood cells. Some studies have shown that platelets can promote the growth, immune escape, and metastasis of various tumor cells [7–9]. As a mediator of information transmission, platelets interact with tumor cells by secreting transforming growth factor beta 1 (TGF-β1) to activate the TGF-β/Smad and NF-kappaB pathways [10]. Additionally, exposure of HCC cells to platelets lysates has been shown to counteract the effects of sorafenib and regorafenib, suggesting that platelets may contribute to chemoresistance through the release of epidermal growth factor (EGF) and insulin-like growth factor-I (IGF-I) [11].

With the rapid development of high-throughput technology, the field of identifying potential disease biomarkers has also been vigorously promoted. In the past, most studies explored differential RNA profiles of tumor platelets based on RNA-seq technology [12]. Moreover, proteomics has emerged as a powerful tool for exploring the complexity of platelets function in both healthy individuals and patients, uncovering potential diagnostic biomarkers, and identifying novel targets for antiplatelet therapies [13, 14]. The metabolomic is a supplement to the genomic and proteomic that provides metabolic profiles associated with disease. Some studies have explored the protein expression changes of platelets in early cancer based on proteomics [14], and used targeted metabolomics to reveal the dysregulation of platelet-rich plasma (PRP) related metabolites in neuroendocrine tumors [15]. Pablo et al. have used multi-omics analysis to find that phospholipase and linolenic acid metabolism were dysregulated in PRP, and the expression of genes related to platelet activation and arachidonic acid metabolism was up-regulated in allergic diseases [16]. A recent study revealed that platelets mediated circulating HCC cells evade natural killer cell killing through the immune checkpoint CD155-TIGIT [17]. However, the protein and metabolic profiles of platelets in patients with HCC are still poorly understood.

In this study, we performed label-free proteomic analysis of platelets from HCC patients and healthy controls to investigate the molecular characteristics and functional regulation of HCC. Our results indicate that the differentially expressed proteins (DEPs) were related to dysregulation of fatty acid degradation and amino acid metabolism. Then, targeted metabolomics studies of PRP were performed in HCC patients and healthy controls to validate the enrichment results of proteomics. Additionally, we integrated proteomics and metabolomics data to gain new insights into the pathological mechanisms underlying HCC. Besides, we validate the effect of glutamine (Gln) metabolism on platelet-incubated HCC cells in vitro. Finally, we evaluated the diagnostic efficacy of these differential proteins and metabolites. This study may contribute to characterizing the protein and metabolic profiles of platelets in HCC, and provide a theoretical basis for identifying new biomarkers and developing new treatment strategies.

Methods

Reagents and antibodies

Regents included Dulbecco’s Modified Medium (DMEM) (Gibco, USA), L-Glutamine (Glu)-free DMEM (Percell, China, Cat. No PM150213), fetal bovine serum (FBS, Gibco, USA), Penicillin/Streptomycin antibiotic (Solarbio, China), and Edu kit (Beyotime, China). Primary antibodies for DLDH (Zenbio, China, Cat#R389268), SDHB (Zenbio, China, Cat#R381845), FUMH (Zenbio, China, Cat#R24309), CISY (Zenbio, China, Cat#R26602), MDHM (Zenbio, China, Cat#R389116), β-Actin (Abmart, Cat#T40104S), and secondary antibody Goat anti-rabbit IgG-HRP (Zenbio, Cat#511,203) and Goat anti-mouse IgG-HRP (Zenbio, Cat#511,103) were used for western blotting.

Patient cohort

This study included 45 patients with HCC and 34 healthy controls, and all blood samples were collected from the First Affiliated Hospital of Guangxi Medical University between August 2023 and July 2024. The inclusion criteria for patients were as follows: (1) A clear diagnosis by pathological examination; (2) No anti-tumor treatments were received before surgery. The exclusion criteria were: (1) Received anti-tumor therapy (such as radiotherapy, chemotherapy, or targeted therapy). (2) Patients with co-infection, pneumonia, blood system diseases, previous tumors, and other serious complications. At the same time, healthy control samples were obtained from the physical examination center. The Medical Ethics Committee of First Affiliated Hospital of Guangxi Medical University has approved the study (No. 2023-E254-01). Informed consent was obtained from all participants.

Human PRP and platelet preparation

Blood was collected in purple-capped EDTA-coated tubes and centrifuged for 20 min at 120 g to obtain PRP. Then, the PRP was collected in a 1.5 mL tube and centrifuged for 20 min at 360 g. Removed the platelet-depleted plasma, the platelet pellet is resuspended in PBS and centrifuged for 20 min at 360 g.

Proteins extraction, digestion, and quantification

Protein extraction of platelet samples conducted in RIPA lysis solution (Thermo Scientific) with 1 × protease inhibitor cocktail at 4 ℃ for 30 min. Supernatants were obtained after centrifugation (14,000 rpm, 30 min, 4 °C), and protein concentrations in the supernatant were determined by the BCA Protein Assay Kit (Thermo Scientific, USA). A total of fifty μg of protein samples were reduced at 37 °C for 1 h with 0.6 μL 0.5 M Tris(2-carboxyethyl) phosphine (TCEP) and alkylated with 1.2 μL 1 M chloroacetamide (CAA) at room temperature in the dark for an additional 30 min. Then, samples were precipitated overnight at −20 °C with five times the volume of pre-cooled acetone. After centrifugation at 14,000 rcf for 10 min, the precipitation was washed by 250 μL of 90% pre-cooled acetone aqueous solution once. Proteins were then digested with trypsin at 1:50 (w/w) ratios in 50 mM tetraethylammonium bromide (TEAB) at 37 °C overnight and quenched with 10% trichloroacetic acid (TFA). The peptide mixture was desalted by Sep-Pak C18 cartridges (Waters, USA), quantified by Pierce™ quantitative colorimetric peptide assay kit (Thermo Scientific, USA), and lyophilized. The lyophilized mixture of peptides was then resuspended in 20 μL of 0.1% formic acid (FA) for LC − MS/MS analysis.

LC–MS/MS analysis for proteomics

We loaded the sample onto a Nano-Elute liquid chromatograph (Bruker) with a 75 μm × 25 cm (1.6 μm id) long column at 200nL/min and the chromatograph was coupled to a TimsTOF mass spectrometer using PASEF (Bruker Daltonics). The gradient was set as follows: from 2 to 22% acetonitrile (ACN) in 45 min, from 22 to 37% ACN in 5 min, from 37 to 80% ACN in 5 min, holding at 80% ACN for 5 min. Mass Range 100 to 1700 m/z, Capillary Voltage 4500 V, Dry Gas 3 L/min, Dry Temp 180 °C. PASEF settings: 4 MS/MS scans (total cycle time 0.53 s), charge range 0–5, active exclusion for 0.4 min, Scheduling Target intensity 20,000, Intensity threshold 2500, CID collision energy started 27 eV, end 45 eV.

Proteomic data analysis

All the raw data acquired in DDA mode were analyzed by Peaks online (Bioinformatics Solutions Inc.). All the data were searched against the Swiss-Prot human database (20,375 entries). The search parameters were set as follows: precursor mass tolerance was set at 20 ppm, and fragment mass tolerance was set at 0.05 Da; cysteine carbamido-methylation was set as a fixed modification and N-terminal acetylation and methionine oxidations were set as variable modifications. The false discovery rate was set to 0.01 for both proteins and PSM with a minimum length of six amino acids. A maximum of three missed cleavages was allowed for the database search. The 50% missing value principle filtered out the proteins with more missing values, and the KNN imputer was used to fill in the missing values in the group data. Upregulated and downregulated proteins were defined as P value < 0.05 and fold change (FC) ≥ 2 or FC < 0.5. Next, R tools “ggpubr” and “ggplot2” were used to assess the volcano plot and the “pheatmap” package was used to generate the heatmap. Finally, we used the Gene Ontology (GO) term to classify the upregulated and downregulated proteins into three categories, then used the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis and Gene Set Enrichment Analysis (GSEA) to identify enriched pathways.

Metabolite extraction

One millilitre of pre-cooled 80% methanol was mixed with 100 μL PRP and then placed at −30℃ for overnight extraction. Subsequently, the supernatant was collected at 14,000 rpm for 30 min at 4 ℃ and dried using a SpeedVac centrifuge (Eppendorf, USA). Then, the extract was resuspended in 50 μL 80% methanol for liquid chromatography tandem mass spectrometry (LC–MS/MS) analysis.

LC–MS/MS analysis for metabolomics

After extraction of metabolites from PRP, all samples were performed by Triple Quad™ 7500 hybrid triple quadrupole/linear ion trap mass spectrometry (AB SCIEX, CA) interfaced with HPLC systems (SHIMADSU, Tokyo) based on multiple reaction monitoring (MRM) analysis. For positive ionisation mode analysis, chromatographic separation of extracts was implemented on the Welch Ultimate AQ-C18 column at a flow rate of 200 μL/min using a 20 min gradient: with 100% solution A (0.1% formic acid in water) for 1 min, 0% to 90% solution B (0.1% formic acid in acetonitrile) for 9 min, 90% solution B for 2 min, 90% to 0% solution B for 0.5 min, and 100% solution A for 7.5 min. For negative ionisation mode analysis, extracts of each sample were separated on a Waters ACQUITY BEH HILIC (2.1 × 100 mm, 1.7 µm id) at a flow rate of 200 μL/min. A gradient of 15 min was shown as follows: with 20% solution A (10 mM ammonium formate, 0.2% ammonia) for 4.9 min, 80% to 50% solution B (100% ACN) for 3 min, 50% solution B for 7 min, 50% to 80% solution B for 0.1 min. The ion pair information, Entrance Potential (EP), Collision Energy (CE), and Collison cell exit potential (CXP) were listed in Table S1.

Metabolomics data analysis

The MRM data were processed using Skyline software (v 22.2.0.255). The data were normalized and uploaded to the MetaboAnalyst website for further analysis. Next, PCA was used to explore global metabolic variations in each group. The screening criteria for differential metabolites were a P value < 0.05 and fold change (FC) > 1.3 or FC < 0.67. R tools “ggpubr”, “ggplot2”, and “pheatmap” were used to assess the volcano plot and heatmap. The KEGG enrichment analysis of differential metabolites was performed on MetaboAnalyst.

Western blotting

The total protein was extracted from PLTs from healthy controls and HCC patients at 4 ℃. The protein lysates were subsequently subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. Following transfer, membranes were blocked with 5% (w/v) skim milk (Solarbio, China) in Tris-buffered saline with Tween-20 (TBST) for 1 h at room temperature to prevent nonspecific binding, followed by overnight incubation with primary antibodies at 4 ℃. Finally, the membranes were incubated with a secondary antibody for 1 h at room temperature, and used an e-BLOT imaging system (e-BLOT, China) to visualize the proteins. After detection of CISY, FUMH, and DLDH antigen, the membrane was incubated with membrane regeneration solution (Solarbio, China) for 60 min for stripping and incubated with β-ACTIN on the same membrane.

Cell culture and treatments

The HCC cells MHCC-97H and Huh7 cells were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) under standard culture conditions (5% CO₂, 37 °C), with experimental groups cultured in Gln-free DMEM. All media were supplemented with 10% FBS and 100 IU/mL penicillin–streptomycin. For platelet-treated groups, the washed platelets were resuspended in DMEM or Gln-free DMEM and incubated at 37 ℃ for 1 h. These pretreated platelets were subsequently introduced into HCC cell cultures at platelet-to-cell ratios (1000:1), establishing a 24 h co-culture system for functional assays.

Cell proliferation assay

HCC cells were seeded into 96-well culture plates at an initial density of 4 × 10³ cells/well allowed to adhere for 16–18 h under standard culture conditions (5% CO₂, 37 °C) before platelet treatment. Proliferation was assessed based on fluorescence-conjugated Edu into newly synthesized DNA according to the manufacturer’s instructions (BeyoClick^™ Edu-555 kit, Beyotime, Cat#C0075S). Cells were imaged with fluorescence microscopy, and the Edu-positive cells were quantified by ImageJ software (v 1.53t).

Wound-healing assay

The HCC cells were seeded in 6-well plates at a density of 5 × 10⁵ cells/well and cultured until the cell density reached 100%. Following serum starvation for 12 h, standardized wound generation was performed using sterile 200 μL pipette tips perpendicular to pre-marked grid lines. Floating cells were removed with three PBS washes. Experimental groups received: (1) serum-free DMEM control, (2) serum-free DMEM with platelets, (3) Gln-free serum-free DMEM with platelets. Wound closure was photographed at 0 h and 24 h post-treatment, with quantitative analysis performed using ImageJ software (v 1.53t).

Transwell invasion assay

Cell invasion assays were conducted using 8-μm pore transwell chambers pre-coated with Matrigel^® Basement Membrane Matrix (Corning, USA). Following 24 h pretreatment with experimental conditions, 5 × 10⁴ cells in 200 μL serum-free DMEM or Gln-free DMEM were plated in the upper chamber. The lower compartment received 600 μL DMEM or Gln-free DMEM with 10% FBS. After 48 h incubation, non-invading cells on the upper membrane surface were removed. Invasived cells were fixed, stained, and counted.

Statistical analysis

SPSS 26.0 software (SPSS Inc., Chicago, IL, USA), Graphpad software (version 8.0), ImageJ software (v 1.53t), and MedCalc software (version 11.3.8.0) were used for data analysis and image generation. We conducted parametric (Student’s t-test) or non-parametric tests (Chi-Square Test or Mann–Whitney U Test) to compare differences between the groups. The receiver operating characteristic (ROC) curve, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were calculated through Medcalc software to evaluate the diagnostic value. Data are expressed as mean ± SD. P value < 0.05 is considered statistically significant.

Results

Clinical characteristics of participants

A total of 45 HCC patients and 34 healthy controls were enrolled in the study. Quantitative label-free proteomic analysis was conducted on platelets isolated from 14 HCC patients and 14 age-/sex-matched healthy controls. To further validate the proteomic results, metabolomics analysis was performed on PRP from an expanded cohort (31 HCC patients and 20 healthy controls) since the quantity of platelets obtained from HCC patients was insufficient to meet the sample requirements for metabolomics. The workflow chart of this study was shown in Fig. 1. The clinical characteristics of all participants are shown in Table 1. Age and sex were not significantly different between HCC and control in both groups (P > 0.05). The pathologic TNM stages and CNLC stage of the HCC patients were consistent in both groups.

Fig. 1.

Schematic illustration of integrating platelets proteomic and platelet-rich plasma metabolomic characterization from HCC patients

Table 1.

The clinical characteristics features of participants

	Proteomics set			Metabolomics set
Parameters	Healthy	HCC	P value	Healthy	HCC	P value
Patients	14	14		20	31
Age (mean ± SD)	57.14 ± 11.43	57.14 ± 13.13	1.00	55.15 ± 12.75	55.00 ± 12.00	0.97
Gender (male/female)	12/2	12/2	1.00	16/4	24/7	0.83
CEA (μg/L)	3.32 ± 2.90	2.68 ± 1.08	0.45	1.61 ± 0.94	3.5 ± 2.39	< 0.01
AFP (ng/ml)
< 250	14 (100.0)	5(35.71)	< 0.01	20(100.0)	16(51.61)	< 0.01
> = 250	0(0.00)	9(64.29)		0(0.00)	15(48.39)
Platelet count (10^9/L) (mean ± SD)	249.53 ± 43.93	190.14 ± 54.49	< 0.01	226.87 ± 41.27	212.33 ± 91.08	0.44
Mean platelet Volume (FL) (mean ± SD)	8.13 ± 0.77	9.4 ± 0.80	< 0.01	8.10 ± 1.10	8.92 ± 1.69	0.06
PIVKA-II (mAU/mL)
< 40		3(21.43)			7(22.58)
> = 40		11(78.57)			24(77.42)
Tumor size(cm)
< 5 cm		11(78.57)			17(54.84)
> = 5 cm		3(21.43)			14(45.16)
The number of tumors
Solitary		13(92.86)			30(96.77)
Multiple		1(7.14)			1(3.23)
Microvascular invasion
−		10(71.43)			22(70.97)
+		4(28.57)			9(29.03)
Tumor Invasion (T stage)
T1 + T2		14(100.0)			30(96.77)
T3 + T4		0(0.00)			1(3.23)
Lymph node metastasis (N stage)
N0		14(100.0)			31(100.0)
N1-N3		0(0.00)			0(0.00)
Distant metastasis (M stage)
M0		14(100.0)			31(100.0)
M1		0(0.00)			0(0.00)
CNLC stage
I-II		13(92.86)			30(96.77)
III-IV		1(7.14)			1(3.23)

PLT Platelet count, MPV Mean platelet Volume, CNLC stage China Liver Cancer Staging

Proteomic profiling of platelets in patients with HCC

The workflow chart of proteomics was shown in Fig. 2A. A total of 4846 proteins were identified in platelets samples. After data normalization, OPLS-DA was performed on the proteomic data. The OPLS-DA model showed obviously separation between the HCC group and the healthy controls group, suggesting changes in the proteomics of platelets in HCC patients (Fig. 2B). The proteins with |log₂FC|> 1 and P value < 0.05 were considered as differentially expressed proteins (DEPs). Compared to healthy controls, there were 316 upregulated proteins and 231 downregulated proteins in the platelets of patients with HCC (Fig. 2C).

Fig. 2.

Label-free proteomic analysis with platelets. A Overview of the experimental setup for platelets proteomics profiling. B OPLS-DA analysis result of platelets proteins in HCC and controls. C The volcano plot showed the dysregulated proteins in HCC patients compared with healthy controls. The COG/KOG functional classification distribution of the upregulated (D) and downregulated (E) proteins in the platelet proteomics of HCC patients compared with healthy controls

In order to further understand the functions of differential proteins, COG/KOG functional analysis, GO functional analysis, KEGG enrichment analysis, and GSEA enrichment analysis were performed. According to the annotation results of COG/KOG, the energy production and conversion was the most distributed pathway in the up-regulated DEPs (Fig. 2D), and the most distributed pathway in the down-regulated DEPs was signal transduction mechanisms (Fig. 2E). Next, GO functional analysis was used to analyze the potential roles of differential molecules in cellular components (CC), molecular functions (MF), and biological processes (BP). Similarly, upregulated proteins in platelets of HCC patients were enriched in pathways related to energy metabolism, including fatty acid beta-oxidation using acyl-CoA dehydrogenase, mitochondrial transmembrane transport, fatty acid catabolic process, generation of precursor metabolites and energy, oxidoreductase activity, and electron transfer activity (Fig. 3A–C). Moreover, upregulated proteins were significantly enriched in specific cellular components such as the outer membrane, matrix, and inner membrane of mitochondria (Fig. 3B). In addition, down-regulated proteins in platelets of HCC patients were significantly enriched in BP including cholesterol transport and complement activation, CC including high-density lipoprotein particle and focal adhesion, and MF including cadherin binding and enzyme inhibitor activity (Fig. 3A–C). The KEGG enrichment analysis results showed that pathways related to metabolic processes such as valine, leucine and isoleucine (BCAAs) degradation, fatty acid metabolism, and platelet activation pathways were significantly enriched in the upregulated proteins of platelets in HCC patients. Complement and coagulation cascades, cholesterol metabolism, and T cell receptor signaling pathway were overrepresented in down-regulated proteins (Fig. 3D). Meanwhile, the GSEA enrichment revealed that the pathways in cancer, Rap1 signaling pathway were overrepresented in upregulated proteins, and the complement and coagulation cascades pathways were overrepresented in down-regulated proteins (Fig. 3E). Heatmaps visualize differences in protein expression across samples. As shown in Fig. 3 F–H, compared with the control group, proteins expression in amino acid metabolism, fatty acid metabolism, and platelet activation pathways was elevated in platelets of HCC patients. Together, the above results indicated significant changes in the proteome of platelets in patients with HCC, especially in the metabolism processes of various substances.

Fig. 3.

Dysregulated proteins and pathways were identified by proteomic analysis. A–C Barplot colored by different categories showed enriched biological processes (BP), cellular components (CC), and molecular functions (MF) of dysregulated proteins in HCC patients compared with healthy controls. D KEGG pathways and (E) GSEA results of dysregulated proteins in HCC patients compared with healthy controls. (F) Heatmap showed the expression of proteins in seven pathways. The barplot showed the quantification of proteins related to metabolism (G) and proteins related to platelet function (H). *P < 0.05, **P < 0.01, ***P < 0.001

Metabolomic profiles and function change related to HCC

To further validate the proteomic results, metabolomics analysis was performed on PRP from 31 HCC patients and 20 healthy controls since the platelets from HCC patients were insufficient to meet the sample quantity requirements for metabolomics (Fig. 4A). A total of 130 metabolites were identified and quantified by LC–MS/MS technology based on MRM mode. PCA was performed on the metabolomic data, indicating HCC group and healthy controls group can be clustered into two individual groups (Fig. 4B). The metabolites with FC > 1.3 or < 0.67 and P value < 0.05 were considered as differential metabolites, of which 13 were down-regulated and 39 were up-regulated in HCC patients compared with healthy controls (Fig. 4C–D). The KEGG enrichment results of differential metabolites showed that alanine, aspartate and glutamate metabolism, TCA cycle, glyoxylate and dicarboxylate metabolism, BCAAs biosynthesis, etc. were significantly enriched (Fig. 4E). Together, metabolic disorders occur in PRP of HCC patients, especially energy metabolism related metabolites.

Fig. 4.

Targeted metabolomic profiles and functional alterations associated with HCC platelet-rich plasma. A Overview of the experimental setup for plasma metabolomic profiling. B PCA analysis result of plasma metabolomics in HCC and healthy controls. The volcano plot (C) and heatmap (D) showed the differential metabolites in HCC patients compared with healthy controls. E Pathway analysis of all differential metabolites

Pathway analysis of integrating proteomic and metabolomic data in platelet and PRP samples

To better understand the correlation between differential proteins and metabolites, an integrated analysis was performed on the results of proteomics and metabolomics. The pathway enrichment (Supplementary Fig. 1) showed several proteins and metabolites that were both involved in central carbon metabolism in cancer, including TCA cycle and amino acid metabolism. Metabolic pathways are composed of protein-catalyzed metabolite reactions, therefore the pathway for direct transformation between differential metabolites or direct interaction between differential metabolites and differential proteins was used for joint analysis. There were 5 pathways selected, including glycine, serine and threonine metabolism; glycerolipid metabolism; valine, leucine and isoleucine degradation; TCA cycle; alanine, aspartate and glutamate metabolism. The 5 selected pathways were used for detailed illustration by labeling up- and down-regulated proteins and metabolites (Fig. 5A). Changes in specific proteins are shown in the heatmaps in Fig. 5A, and changes in metabolites are shown in bar plots (Fig. 5B).

Fig. 5.

Illustration of the significant pathways relating to amino acid metabolism connected to the TCA cycle. A Pathway overview of amino acid metabolism and glycerolipid metabolism. Rectangles represent proteins. Red-filled indicates that the metabolites or proteins are upregulated in HCC group compared with healthy control group, and blue-filled indicates downregulated. Solid lines indicate direct reactions between metabolites, while dashed lines indicate multi-step reactions. B Normalized intensity of metabolites involved in the illustrated pathways. *P < 0.05, **P < 0.01, ***P < 0.001

Compared to healthy controls, choline, betaine, dimethylglycine, serine, alanine, and creatine were upregulated in glycine, serine and threonine metabolism in HCC patients. Glycine amidinotransferase (GATM) and guanidinoacetate N-methyltransferase (GAMT) regulated the synthesizing of creatine from glycine and arginine [18]. Serine hydroxymethyltransferase (GLYM) can enhance serine-glycine-one-carbon (SGOC) metabolism [19]. Consistent with upregulated metabolites, GATM and GLYM were also upregulated in platelets of HCC patients compared with controls. In glycerolipid metabolism, lactic acid, glyceric acid, glycerol, and glycerol-3-phosphate were upregulated in HCC patients compared to healthy controls. For metabolic enzymes, 4-trimethylaminobutyraldehyde dehydrogenase (AL9A1), lysophospholipid acyltransferase 2 (MBOA2), 1-acyl-sn-glycerol 3-phosphate acyltransferase alpha (PLCA) were up-regulated in platelets of HCC patients, while aldo–keto reductase family 1 member A1 (AK1A1), aldo–keto reductase family 1 member B1 (ALDR), lysocardiolipin acyltransferase 1 (LCLT1) were down-regulated in platelets of HCC patients. Therefore, there was an accumulation of metabolites of glycerolipid metabolism and dysregulation of metabolic enzymes in HCC patients. BCAAs include leucine, isoleucine, and valine. In valine, leucine and isoleucine degradation, the enzymes of many key steps were up-regulated in platelets of HCC patients compared with healthy controls. And among the metabolites, leucine and isoleucine were down-regulated in HCC patients, and their downstream metabolites 3-hydroxyisovalerate and methylmalonyl-CoA were up-regulated in HCC patients. This indicated that higher enzymatic activity enhances BCAAs catabolism in HCC patients. In TCA cycle, the majority of TCA intermediates were up-regulated in HCC patients, including citrate, isocitrate acid, α-ketoglutarate, succinic acid, fumaric acid, and malic acid. Isocitrate dehydrogenase (including IDH3A, IDH3B and IDHP) catalyze the reactions from isocitrate acid to α-ketoglutarate. Fumarate hydratase (FUMH) catalyzes the reactions from fumaric acid to malic acid. In general, fumarate accumulation affects the conversion of succinate to fumarate by succinate dehydrogenase (SDH) in the TCA cycle, thereby reducing SDH-dependent mitochondrial respiration [20]. TCA circulating metabolites were primarily considered as by-products of cellular metabolism and were important for the biosynthesis of macromolecules such as nucleotides, lipids, and proteins.

The integrated analysis of proteome and metabolome revealed that energy metabolism was accelerated in HCC patients, especially in TCA cycle, which serves as the primary pathway for energy production in the body. Therefore, we selected proteins that were identified as upregulated in TCA cycle and exhibited a Log₂FC greater than 1.5 (SDHB, CISY, FUMH, DLDH, MDHM) for Western blotting validation in platelets from HCC patients and healthy controls. Our findings revealed that the expression levels of SDHB, CISY, and FUMH were significantly elevated in HCC patients compared to healthy controls, which was consistent with the proteomics results (Fig. 6A). Together, we found and validated that the energy metabolism was accelerated in HCC patients.

Fig. 6.

Validation of proteomics and metabolomics findings. A Protein level validation of differentially expressed proteins related to the TCA cycle of PLT in healthy controls and HCC patients (n = 6). B Edu assay showed that Gln-free treatment weakens the proliferation ability of PLT induced MHCC-97H and Huh7 cells. Scale bar, 100 μm. C Wound healing showed that Gln-free treatment weakens the migration ability of PLT induced MHCC-97H and Huh7 cells. Scale bar, 200 μm. D Transwell assay showed that Gln-free treatment weakens the invasion ability of PLT induced MHCC-97H and Huh7 cells. Scale bar, 100 μm. All experiments were performed three times, and data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

Glutamine-free weakens the ability of platelets to promote proliferation, migration, and invasion of HCC cells

The integrated analysis revealed that the increase of glutamate promotes the accumulation of α-KG into the TCA cycle (Fig. 5). To elucidate the impact of Gln metabolism on the co-culture systems of platelets and HCC cells, we established Gln deprivation models to evaluate HCC cell proliferation, migration, and invasive potential. Edu assay demonstrated that co-culture with platelets significantly enhanced proliferative rates in both MHCC-97H and Huh7 cell lines compared to the control group. Notably, Gln deprivation effectively attenuated platelet-induced proliferation (Fig. 6B). To further investigate the role of Gln availability in metastatic potential, we conducted wound healing and transwell invasion assays. The platelet-treated groups exhibited markedly increased migration and invasive capacities relative to the control group. Importantly, Gln deprivation substantially suppressed both platelet-enhanced migration (Fig. 6C) and invasion (Fig. 6D) phenotypes in both cell lines. These findings suggested that Gln availability is a critical modulator of platelet-driven HCC progression.

Candidate biomarker proteins and metabolites for diagnosing patients with HCC

Serum AFP remains a cornerstone biomarker for HCC diagnosis. To confirm the feasibility of using protein to distinguish between individuals diagnosed with HCC from healthy controls. In the proteomic dataset, the AUC, sensitivity, and specificity of AFP in diagnosing patients with HCC were 0.885 (95% CI 0.708–0.974), 92.86%, and 71.43%, respectively. We evaluated the diagnostic performance of SDHB based on the criteria: the highest FC in proteomic (FC: 5.70, p < 0.001) and western blotting validation (1.73-fold increase vs. controls, p = 0.007). Compared with AFP, SDHB showed superior diagnostic performance, achieving higher AUC values of 0.929 (95% CI 0.765–0.991) and specificity of 92.86% in distinguishing HCC patients from healthy controls (Fig. 7A, B).

Fig. 7.

ROC curve analysis of proteins and metabolites in platelets and platelet-rich plasma of HCC patients. A, B The diagnostic efficacy of SDHB and AFP in distinguishing HCC patients from healthy controls. C, D ROC curve analysis of AFP, α-Ketoglutarate, succinic acid, fumaric acid, malic acid, citrate, isocitrate acid, GABA, and glutamate in PRP of HCC patients. E The area under the curve (AUC), sensitivity, and specificity of α-Ketoglutarate, succinic acid, fumaric acid, malic acid, citrate, isocitrate acid, GABA, and glutamate in distinguishing HCC patients from healthy controls. The P values indicated results from the Delong-method. CI, confidence interval

In the metabolomic dataset, the AUC, sensitivity, and specificity of AFP in diagnosing patients with HCC were 0.868 (95% CI: 0.743–0.946), 83.87%, and 90.00%, respectively (Fig. 7C). Then, we evaluated 8 PRP metabolites associated with the TCA cycle as potential biomarkers for distinguishing HCC patients from healthy controls. Compared with AFP, the three metabolites (succinic acid, fumaric acid, and malic acid) increased the AUC, sensitivity, and specificity of distinguishing patients with HCC from healthy controls (Fig. 7D, E). These results suggest that SDHB, succinic acid, fumaric acid, and malic acid are promising as potential biomarkers to distinguish HCC patients from healthy controls.

Discussion

Omics-based analytical approaches have garnered increasing interest in the field of cancer research. Researchers have studied the diagnosis [21–23], risk factors [24], pathogenesis [25], and therapeutic targets [26] of HCC based on metabolomics, proteomics, and microbiology. These studies have yielded insights that are instrumental in the diagnostic and therapeutic management of HCC. They encompassed the quantification of differential metabolites, the construction of potential diagnostic models, and the identification of key signaling pathways that are pivotal in the tumor progression. However, the majority of clinical omics research pertaining to oncological samples has been predicated on plasma or serum. Although there are some platelets proteomic studies investigated the functional changes of platelets in other cancer patients, such as lung cancer and head of pancreas cancer [14], colon cancer [27], and myeloproliferative neoplasms [28], the platelets proteomics of HCC patients has been overlooked. Consequently, there persists a significant gap regarding proteomic and metabolomic investigations focused on platelets in the context of HCC. In the present study, we undertook a comprehensive analysis of the proteomic and targeted metabolomic profiles of platelet-related samples derived from HCC patients and healthy controls. This analysis aimed to elucidate the DEPs and differential metabolites associated with platelets in HCC, thereby potentially uncovering novel biomarkers and pathophysiological mechanisms pertinent to HCC.

Non-invasive diagnosis with platelets samples has been reported. platelets are prevalent in the peripheral circulation and are repositories of a multitude of proteins. They possess the capacity to engage in bidirectional communication with tumor cells, a process that can lead to the transformation of platelets into"tumor-educated platelets."This interaction renders platelets a promising source for"liquid biopsy"applications. Their RNA and proteins have been extensively studied for the diagnosis of breast cancer [29] and colorectal cancer [30]. However, research on the proteomics of platelets in HCC patients remains relatively scarce. P-selectin (sP-selectin) is a recognized marker of platelets activation in vivo. Previous studies have found that sP-selectin is elevated in patients with cirrhosis and is strongly associated with disease stage [31]. Similarly, elevated levels of sP-selectin/platelet ratio have been observed in HCC patients, demonstrating increased platelet activation [32]. Moreover, Ras-associated protein (RAP) is an abundant signaling molecule in the platelet/megakaryocyte lineage and play a critical role in cell adhesion and normal hemostasis in vivo [33]. Consistent with previous research, our proteomics data reveal that proteins involved in platelet activation are significantly upregulated in HCC patients, including RAP1A, RAP1B, R2RX1, and GP1BB. Traditionally, the complement system has been recognized for its pivotal roles in the elimination of pathogens, the recruitment of immune cells, and the initiation of adaptive immune responses. Nevertheless, burgeoning evidence indicates that the complement system exerts context-dependent functions in the realm of oncology, being capable of either orchestrating the destruction of neoplastic cells or, conversely, fostering tumorigenesis and tumor progression. This duality underscores the intricate interplay between the complement cascade and malignant cells [34]. Serum proteomics in patients with alcohol-associated hepatitis (AH) have shown reduced levels of proteins involved in hepatocyte function, such as coagulation factors, complement cascade components, and hepatocyte growth activator [35]. A bioinformatics analysis also found that the down-regulated differentially expressed genes were mainly enriched in complement and coagulation cascades [36] in HCC patients. Our results indicate that complement activation-related C1s, MASP, and C4b binding protein (C4BPA) are downregulated in platelets of HCC patients, which are consistent with previous studies. Interestingly, activation of the complement system can promote cancer progression by modulating immunosuppressive cells [37], our results are consistent with the observed downregulation of the T cell receptor signaling pathway. Furthermore, elevated cyclic guanosine monophosphate (cGMP) activates the corresponding protein kinases and protein kinase G (PKG), which phosphorylates a variety of substrates responsible for platelet inhibition [38]. Our findings show that proteins involved in the cGMP-PKG signaling pathway were significantly upregulated, suggesting potential inhibition of platelet function in HCC patients.

In addition, we observed a significant upregulation of BCAAs metabolism and fatty acid metabolism in platelets derived from HCC patients. Conversely, cholesterol metabolism exhibited downregulation, indicating a potential metabolic shift within the platelet compartment in the context of HCC. In patients with type 2 diabetes mellitus, BCAAs levels in platelets were elevated, and positive regulators of BCAAs catabolism promoted platelet activation [39]. Ghatge M et al. [40] investigated the overall metabolic changes during the transition of human platelets from the resting state to the activated state and found significant upregulation of amino acid metabolism pathways, lipid metabolism, and glycolysis/gluconeogenesis, etc. Similarly, studies have demonstrated that platelets utilize both glucose/glycogen and endogenous fatty acids to support glycolysis, relying on glycolysis for energy regardless of nutrient availability [41]. Moreover, the degree of fibrosis in the surrounding liver cancer tissues showed a negative correlation with serum levels of cholesterol, low-density lipoprotein, ApoA1, and ApoB [42]. PRP, which is rich in platelets and easily obtained from circulating peripheral blood, has shown great potential in regenerative medicine [43], but metabolomics studies of PRP in HCC patients are still lacking. To delve deeper into these metabolic alterations, we employed targeted metabolomic analysis to scrutinize the metabolic profiles of PRP in HCC patients. Our study found that several metabolic pathways, including alanine, aspartate and glutamate metabolism, TCA cycle, glycerolipid metabolism, BCAAs metabolism, and others were significantly enriched in HCC patients.

By integrating proteomic alterations into the metabolite network, a more discernible map of metabolomic pathway perturbations can be delineated. The present investigation employs an integrated proteomic and metabolomic approach to reveal pivotal energy metabolism pathways that are dysregulated in HCC patients, thereby providing a comprehensive view of the metabolic landscape associated with this disease. Metabolites derived from glucose, amino acid, and lipid metabolisms are crucial substrates that fuel cellular proliferation and the biosynthesis of macromolecules essential for the neoplastic growth and survival. Our findings indicate that 94 proteins related to energy production and conversion were upregulated, constituting 29.75% of all upregulated proteins. Notably, BCAAs metabolism and the TCA cycle were significantly enhanced, both crucial for generating intermediates in cellular energy and biosynthesis pathways. As fuel and important nutrients, the catabolism of BCAAs plays an important role in tumor progression. In pancreatic ductal adenocarcinoma (PDAC), BCAAs accumulate in plasma and catabolism is reduced [44]. However, HMGCL is downregulated in HCC tissues [45]. Similarly, ACADS is down-regulated in HCC tissue samples and shows lower expression in HCC cell lines compared to normal cells [46]. Besides, a low serum BCAAs/AAA ratio reduces the biosynthesis and secretion of albumin in hepatocytes, leading to hypoproteinemia [47], which may contribute to albumin depletion in HCC patients [48]. In non-small cell lung cancer (NSCLC), enhanced BCAAs uptake leads to circulating BCAA depletion [44], and HCC tumors show accumulation of BCAAs in tumor tissues [49]. In addition, dysregulated expression of pivotal enzymes and the consequent accumulation of TCA cycle intermediates are implicated in the etiology and progression of a spectrum of cancers. α-Ketoglutarate (α-KG) is produced by isocitrate oxidative decarboxylation mediated by IDH or from glutamate via oxaloacetate transamination. Tumor cells often use glutamate–oxaloacetate transaminase (GOT2) rather than glutamate dehydrogenase (GLUD1) to produce α-KG [50]. Then, we performed western blotting assay to validate the DEPs we chose in the TCA cycle. We found that the expression levels of SDHB, FUMH, and CISY were significantly upregulated in HCC patients compared to healthy controls, which was consistent with our proteomics results. Besides, there was no significant difference in DLDH and MDHM in our validated samples, which may be due to individual differences between patients and the small number of platelets samples. Collectively, our findings indicate that PRP serves as an effective mirror of the metabolic perturbations occurring in HCC patients, providing a window into the disease's metabolic fingerprint.

Our study demonstrates that elevated glutamate levels enhance α-KG incorporation into the TCA cycle, and a large number of studies have also shown that the intake of Gln promotes tumor development [51, 52]. Exogenous Gln deprivation significantly reduced intracellular glutamate concentrations. Our results showed that the proliferation, migration, and invasion of MHCC-97 h and Huh7 cells were significantly promoted by platelets. Gln deprivation can significantly weaken the platelet-induced proliferation, migration, and invasion of MHCC-97 h and Huh7 cells. Similarly, prior studies report that Gln deprivation suppressed pancreatic cancer growth and induced ferroptosis in vitro and in vivo [50, 53]. Gln deprivation inhibited cholesterol synthesis to enhanced cell death of HCC cells [54]. These results collectively suggest that Gln plays a critical role in mediating platelet-induced HCC growth and metastasis.

Non-invasive or minimally invasive specimen collection strategies for disease diagnosis are undeniably promising, but they require extensive theoretical and practical validation. Compared to traditional biomarkers (AFP), SDHB demonstrated higher AUC and specificity of distinguishing patients with HCC from healthy controls, with an AUC of 0.929 (95%CI 0.765–0.991). A study showed that the diagnostic model of chitinase-3-like protein 1(CHI3L1), growth/differentiation factor 15 (GDF15), interleukin-1 receptor antagonist protein (IL1RN), and E-selectin (SELE) can improve the diagnostic efficiency of distinguishing HCC patients from healthy controls compared with the clinical variable model [55]. Another study also showed that the Retinol-binding protein 4 (RBP4) can improve the AUC of distinguishing HCC from healthy controls compared with AFP, with an AUC of 0.879 [56]. Additionally, succinic acid, fumaric acid, and malic acid increased the AUC and sensitivity of distinguishing patients with HCC from healthy controls compared with AFP. A study showed that phenylalanyl‐tryptophan and glycocholate can improve the diagnostic efficiency of distinguishing HCC patients from healthy controls [21]. Plasma metabolomics studies have shown that HCC is characterized by heightened activity in the citric acid cycle and oxidative phosphorylation [57]. To develop a minimally invasive and reliable diagnostic approach for HCC, future research should focus on proteomic and metabolomic studies to identify robust biomarkers.

While RNA sequencing has become the predominant approach for characterizing transcriptomic alterations in tumor-educated platelets, our study advances the field by implementing an integrated proteomic and metabolomic profiling strategy to systematically investigate platelet protein signatures and PRP metabolic reprogramming in HCC patients. This multi-omics approach reveals several key advantages over existing single-omics studies. Nevertheless, our study has several limitations. First, proteomic and metabolomic analyses of platelets were not conducted on the same participants. To achieve the platelets concentration required for both omics studies, a larger volume of blood samples from patients would be necessary. Second, the limited sample size in this study underscores the need for further multicenter and longitudinal investigations to evaluate the diagnostic efficacy of these biomarkers, particularly in the early stages of HCC.

Conclusion

In summary, this study has unveiled the pivotal roles of dysregulated energy metabolism, including disturbances in the TCA cycle, amino acid metabolism, and glycerolipid metabolism, in platelets of patients with HCC through proteomics and targeted metabolomics analysis. Importantly, our findings delineate that SDHB, succinic acid, fumaric acid, and malic acid have high AUC values and specificity, indicating their superior diagnostic accuracy in distinguishing HCC patients from healthy controls. We anticipate that future research incorporating more comprehensive multi-omics approaches will further elucidate the underlying mechanisms of HCC and contribute to the development of simpler and more effective diagnostic methods.

Abbreviations

HCC

Hepatocellular carcinoma

PRP

Platelet-rich plasma

SVM

Support vector machine

AUC

Area under the curve

ROC

Receiver operating characteristic

AFP

Alpha-fetoprotein

DEPs

Differentially expressed proteins

TCEP

Tris(2-carboxyethyl) phosphine

CAA

Chloroacetamide

TEAB

Tetraethylammonium bromide

TFA

Trichloroacetic acid

LC − MS

Liquid chromatography mass spectrometry

ACN

Acetonitrile

GO

Gene ontology

CC

Cellular components

MF

Molecular functions

BP

Biological processes

KEGG

Kyoto encyclopedia of genes and genomes

GSEA

Gene set enrichment analysis

EP

Entrance potential

CE

Collision energy

CXP

Collison cell exit potential

OPLS-DA

Orthogonal partial least squares-discriminant analysis

TNM

Tumor node metastasis classification

CNLC

China liver cancer staging

BCAAs

Valine, leucine and isoleucine

TCA cycle

Tricarboxylic acid cycle

Author contributions

Huabing Wang and Xue Qin designed the experiments. Zuojian Hu and Fenglin Shen performed the experiments. Xuelian Ruan analyzed and interpreted the proteomics and metabolomic data. Xuelian Ruan and Zuojian Hu wrote the manuscript and performed the validation experiments. Yongling Chen and Ziqing Zhong, and Guiyong Ou collected clinical samples and patient information. Xing Luo participated in cell experiments. All authors reviewed and/or edited the manuscript.

Funding

This work was supported by National Natural Science Foundation of China (grant number: 32460228 and 82260419), Joint Project on Regional High-Incidence Diseases Research of Guangxi Natural Science Foundation (grant number: 2023GXNSFAA026030 and 2023GXNSFDA026001), Ministry of Science and Technology of the People's Republic of China (grant number: G2023033003L), and Innovation Project of Guangxi Graduate Education (grant number: YCBZ2024127).

Availability of data and materials

Data will be made available on request.

Declarations

Ethics approval and consent to participate

The Medical Ethics Committee of First Affiliated Hospital of Guangxi Medical University has approved the study (No. 2023-E254-01). Informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing Interests

No potential conflict of interest was reported by the authors.