You are viewing the site in preview mode

Skip to main content
Download PDF

Comparative evaluation of methods for isolating extracellular vesicles from ICC cell culture supernatants: Insights into proteomic and glycomic analysis

Cell Communication and Signaling volume 23, Article number: 207 (2025) Cite this article

Abstract

Background

Extracellular vesicles (EVs) are nanoscale structures involved in intercellular communication and play a key role in cancer pathology. Intrahepatic cholangiocarcinoma (ICC) is a highly invasive malignancy marked by abnormal sialylated glycosylation. Analyzing proteins and glycans in EVs provides insights into ICC molecular subtyping and mechanisms. Optimizing EV isolation methods for ICC-derived EVs enables comprehensive proteomic and glycomic analysis.

Methods

We systematically evaluated five EV isolation methods—Ultracentrifugation (UC), exoEasy, Total Exosome Isolation (TEI), EVtrap, and ÄKTA—by analyzing the biophysical properties, proteomic profiles, and glycomic structures of EVs. Subsequently, we applied TMT-based quantitative proteome and light/heavy methylamine labeling for the quantification of sialylated N-glycan linkage isomers to investigate alterations in proteins and N-glycans within EVs secreted by HuCCT1 and HCCC-9810 cells with overexpressing ST6 β‑galactoside α2,6‑sialyltransferase 1 (ST6GAL1).

Results

By evaluating the biophysical properties, proteome, and N-glycome of EVs extracted using five different methods, UC was identified as the optimal approach for this study, as it offered a balance between operational complexity, cost-effectiveness, and the preservation of EVs activity. In this study, a total of 1,928 high-confidence proteins and over 84 high-confidence glycans were quantified. EVs secreted by HuCCT1 and HCCC-9810 cells overexpressing ST6GAL1 exhibited consistent upregulation of 16 proteins, consistent downregulation of 10 proteins, as well as consistent upregulation of 3 glycans and consistent downregulation of 3 glycans.

Conclusions

Quantitative proteomic and glycomic analysis of ICC-derived EVs revealed that ST6GAL1 overexpression led to significant alterations in proteins involved in cancer cell adhesion and glycosylation pathways, along with specific changes in N-glycan structures. Notably, these modifications extended beyond α2,6-sialylation, suggesting that interactions between glycosyltransferases and glycans may drive these alterations.

Background

Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver malignancy, arising from bile duct epithelial cells and affecting ducts from secondary branches to the smallest divisions [1]. Its global incidence has risen in the past decade, posing a significant challenge. ICC is characterized by silent onset, high heterogeneity, strong invasiveness, and poor prognosis [2, 3]. Ongoing omics studies on ICC-related biological samples aim to enhance understanding of its biology and advance precision treatment strategies [4,5,6].

Proteins serve as the primary executors of a vast array of biological activities, functioning as the key molecules that drive cellular processes. Glycosylation, an important post-translational modification of proteins, plays a wide range of roles in various functions, from structural integrity to involvement in molecular transport, self-recognition, and removal [7]. Aberrant glycosylation patterns, particularly altered terminal sialylation, are frequently observed in tumor cells and are closely linked to enhance tumor cell proliferation, invasion, and metastasis [8,9,10]. Multiple glycoproteins have been reported or are already used as tumor markers in clinical practice, including CA19 - 9 and mucin [11]. Notably, protein glycosylation is a universal phenomenon and is not confined to a specific type of tumor [7]. This underscores the significance of studying glycosylation patterns across various cancer types. Recent advancements in quantitative multi-omics approaches, particularly those integrating proteomic and glycomic analysis, offer novel and comprehensive insights into the molecular underpinnings of cancer [4, 12, 13]. These multi-omics strategies enable a more accurate diagnosis, facilitate the identification of potential therapeutic targets, and improve prognostic evaluations, thereby enhancing the overall understanding and management of cancer.

Extracellular vesicles (EVs) are vesicular structures released by various cells, encapsulating proteins, DNA, lipids, and other components [7]. EVs play essential roles in pivotal processes such as cell communication, cell migration, angiogenesis, and tumor cell growth [14]. It is better to isolate EVs from complex samples before conducting omics analysis of their intralumenal cargo. The study of isolating and characterizing EVs can be traced back to the 1970 s [15]. Subsequently, various methods for separating EVs have been developed based on the fundamental principles of ultracentrifugation, chromatography, membrane affinity, precipitation, and chemical affinity [7, 16,17,18]. Ultracentrifugation (UC) is widely recognized as the gold standard for EVs isolation. Despite reports indicating that high shear forces may compromise EVs integrity, this method remains extensively employed in proteomics and glycomics research, facilitating the characterization of protein and glycan profiles across various biological samples [19,20,21]. Chromatographic techniques provide versatile approaches for EVs isolation, offering flexibility in application. Notably, the integration of multimodal chromatography columns with the ÄKTA system enables the development of automated and scalable workflows, improving isolation efficiency and reproducibility [22, 23]. Membrane affinity-based methods, particularly the exoEasy Maxi Kit, have been widely adopted for EVs extraction from biological fluids such as blood and sputum [24, 25]. Precipitation-based methods are commonly utilized to isolate EVs from complex biological samples, including cell culture supernatants and organoids, facilitating the exploration of their proteomic landscapes [26,27,28]. Chemical affinity-based approaches, exemplified by EVtrap, have been employed for EVs isolation from sperm and urine, enabling comprehensive proteomic and glycomic analyses [29,30,31]. Each of these methods has its own advantages and limitations. While all are suitable for comprehensive proteomic and glycomic analyses, their selection should be carefully tailored to the specific sample type and research objectives to ensure optimal performance. Despite the publication of several Minimal Information for Studies of Extracellular Vesicles guidelines by the International Society for Extracellular Vesicles [32,33,34,35], it is not clear which isolation method is the optimal choice. As natural carriers of biomolecules, EVs possess unique advantages in effectively enriching low-abundance biomolecules from complex biological fluids [6]. These biomolecules often become mediators causing alterations in cell proliferation, invasion, and migration, and quantitative multi-omics analysis can help identify potential regulators that exert these functions [36, 37]. It is crucial to evaluate methods for isolating and characterizing ICC-derived EVs and their performance in multi-omics analysis.

In this study, EVs were isolated using five distinct methods: UC, exoEasy, Total Exosome Isolation (TEI), EVtrap, and ÄKTA. These methods are based on different separation principles, including centrifugal force, membrane affinity, precipitation, chemical affinity, and multi-modal chromatography. We aimed to evaluate the performance of these isolation techniques by comparing the morphology, purity, yield, biological activity, and cargo (proteins and N-glycans) of EVs derived from ICC cell lines. Characterization of the EVs was performed using transmission electron microscopy (TEM), flow cytometry (FCM), a Keyence BZ-X810 microscope, and mass spectrometry (MS). Following a comprehensive evaluation, UC was identified as the optimal method for relative quantitative analysis of the proteome and N-glycome in EVs derived from ICC cells. Furthermore, using UC, a total of 1,928 high-confidence proteins and over 84 high-confidence glycans were quantified from EVs secreted by two ICC cell lines (HuCCT1 and HCCC- 9810). EVs derived from ST6GAL1-overexpressing cells showed consistent upregulation of 16 proteins, downregulation of 10 proteins, as well as upregulation of 3 glycans and downregulation of 3 glycans.

Methods

Materials and regents

The detailed information related to the materials and regents was provided in the Additional file 1.

Cell culture

Cell culture was maintained in the humidified incubator with 5% CO2 at a constant temperature of 37 °C. The 293 T cells were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin–streptomycin. The ICC cell lines such as HuCCT1 and HCCC- 9810 were grown in RPMI‑1640 medium supplemented with 10% FBS and 1% penicillin/streptomycin. For the collection of EVs, the cells were allowed to reach approximately 80% confluence before being switched to serum-free conditioned medium. Following this, the cells were incubated for an additional 48 h to facilitate EVs secretion. After this incubation period, the culture supernatant was carefully collected using a sterile pipette to minimize disruption of the cell monolayer. The collected conditioned cell culture medium (CCCM) was then processed for subsequent downstream applications.

Overexpression of ST6GAL1 in ICC cells

Following previous methods [38], we established four cell lines for subsequent experiments, HuCCT1 with control vector (HuCCT1-Vector), HuCCT1 with overexpression ST6GAL1 (HuCCT1-ST6GAL1), HCCC- 9810 with control vector (HCCC- 9810-Vector), HCCC- 9810 with overexpression ST6GAL1 (HCCC- 9810-ST6GAL1). In brief, the transfer plasmid (ST6GAL1 overexpression plasmid or corresponding control plasmid), envelope plasmid, and packaging plasmids were co-transfected into 293 T packaging cells following the manufacturer's instructions (Hanbio Biotechnology Co., Ltd.). After a 16-h incubation period, the culture medium was replenished to maintain optimal transfection conditions. The supernatant containing the lentiviral particles was subsequently harvested and centrifuged for virus purification. The resulting lentivirus, which either expressed the ST6GAL1 gene or contained the control vector, was resuspended in complete medium and used to infect the ICC cells HuCCT1 and HCCC- 9810. Following infection, the cells were cultured for one week under standard conditions, after which successfully transfected cells were selected using puromycin (1 µg/mL).

Pre-preparation of CCCM

All pre-preparation steps of the CCCM were conducted at 4 °C. The CCCM underwent sequential centrifugation at 500 × g for 10 min, followed by 3000 × g for another 10 min, and finally at 10,000 × g for 30 min. After each centrifugation step, the precipitate was carefully discarded, retaining only the supernatant. Subsequently, filtration was performed using a 0.22 μm membrane, after which the CCCM was concentrated through dialysis.

To prepare the dialysis membranes, they were first treated at 100 °C for 10 min in a solution containing 2% NaHCO3 and 1 mM EDTA (pH 8.0). After three thorough rinses with distilled water to remove any residual chemicals, the membranes were treated again in a boiling water bath for 10 min in a solution containing 1 mM EDTA (pH 8.0). The prepared membranes were then set aside for later use. The samples were dialyzed using a dialysis buffer composed of 150 g PEG 8000, 4.385 g NaCl, and 1 L of H2O, with the buffer being replaced every 4 h to maintain the concentration gradient. Through this process, the CCCM solution could be concentrated about tenfold.

Isolation of EVs

Total EVs were isolated from pretreated CCCM using five distinct isolation methods. UC remains one of the most widely used techniques for EVs isolation. The exoEasy utilizes a membrane affinity-based enrichment approach, while TEI represents a precipitation-based method. EVtrap, developed by Professor Andy Tao's group at Purdue University, is a chemical affinity-based method for EV enrichment [31]. The ÄKTA pure 25 system, equipped with a HiScreen Capto™ Core 700 column, employs a multimodal enrichment strategy that combines hydrophobic interaction, charge interaction, and size exclusion, offering a sophisticated alternative to traditional size exclusion chromatography resins. The detailed procedures for these extraction methods are provided in the Additional file 1.

Characterization of EVs using TEM

The isolated EVs were carefully applied to a 200-mesh grid and incubated at room temperature (RT) for 10 min to allow the sample to adhere. Excess liquid was gently removed, and the grid was allowed to air dry slightly. The grid was then negatively stained with 2% phosphotungstic acid for 3 min. Any residual liquid was removed using filter paper, and the grid was dried under incandescent light. Finally, the EVs were observed and imaged using a JEM1400 transmission electron microscope.

Characterization of EVs using NanoFCM

EVs size, concentration, and phenotype were analyzed using the NanoAnalyzer N30E instrument equipped with dual lasers (488/640 nm) and single-photon counting avalanche photodiode detectors (SPCM APDs). Prior to testing the size distribution and concentration of EVs, the NanoAnalyzer was calibrated using 250 nm silica nanoparticles of known concentration and a proprietary four-modal silica nanosphere cocktail (S16M-Exo) from NanoFCM Inc. Filtered 1 × PBS solution was used as a blank control. Sampling was conducted for 1 min while maintaining a sampling pressure of 1.0 kPa.

For EVs tetraspanin phenotyping, three aliquots of the isolated EV fractions were incubated with 10 µL of Alexa Fluor 488 Mouse anti-Human CD9, FITC Mouse anti-Human CD81, and Alexa Fluor 488 Mouse anti-Human CD81, respectively, for 30 min at 37 °C. After incubation, an appropriate amount of 1 × PBS was added and mixed, followed by ultrafiltration at 4 °C using a 100 kDa ultrafiltration tube to remove free dyes. The samples were then washed with PBS five times. Finally, the samples were collected by inverting the tube and centrifuging at 3000 × g for 10 min at 4 °C. The samples were diluted to 1 × 10⁸–1 × 10⁹ particles/mL in PBS for immediate phenotypic analysis.

All data were processed using NanoFCM software (version V2.0).

Characterization of EVs activity

The isolated EVs were resuspended in PBS and mixed with PKH67 staining solution in a 1:1 volume ratio. After a 20-min incubation, unbound dye was removed using a 100 kDa ultrafiltration tube, and the PKH67-labeled EVs were resuspended in RPMI- 1640 medium. The labeled EVs were then incubated with HuCCT1 cells for 4 h at 37 °C in a dark environment with 5% CO2. Following incubation, the cells were washed three times with PBS, fixed with 4% paraformaldehyde, and stained with DAPI before imaging.

The activity of the EVs was further compared by FCM. After co-incubation of the cells with PKH67-labeled EVs, we released the cells from adhesion and resuspended them in 200 µL of PBS. Detection was performed using a BD FACSLyric instrument controlled by BD FACSSuite RUO software (version 1.5). Experimental data were analyzed using FlowJo software (version 10.8.1), and the activity of the EVs was compared based on the mean fluorescence intensity.

Proteins pre-treatment

In the initial methodological evaluation, EV proteomics analysis was performed without TMT labeling, while a TMT-based approach was later employed for protein comparison in the ST6GAL1 cell study.

Proteins were extracted from EVs using the"7 + 2"lysis buffer, and protein concentration was quantified using the Bradford assay. Following impurity removal with an ultrafiltration tube, the EVs proteins were reduced with 10 mM DTT for 1 h at 56 °C, then alkylated with 20 mM IAA for 45 min at RT in the dark. Trypsin was subsequently added at a protein:trypsin mass ratio of 50:1, and digestion was performed overnight at 37 °C. After digestion, TFA was added to adjust the pH to below 3, thereby halting the enzymatic reaction. The solution was then desalted and lyophilized for storage.

The TMT10plex Isobaric Mass Tagging Kit was equilibrated at RT for 15 min, followed by centrifugation. After opening the tubes, 41 μL of ACN was added to each tube containing 0.8 mg of labeling reagent, and the contents were vortexed to ensure complete dissolution. For each EV-derived proteolytic peptide sample, 25 μg was dissolved in 100 mM TEAB buffer to a final concentration of 1 μg/μL. Proteolytic peptides from the four EV-derived cell lines (HuCCT1-ST6GAL1, HuCCT1-Vector, HCCC- 9810-ST6GAL1 and HCCC- 9810-Vector) were labeled using the four channels of the TMT 10-plex reagent, following the manufacturer’s instructions. Specifically, 25 μL of the peptide solution was mixed with 10.25 μL of the corresponding TMT reagent and incubated at RT with shaking for 2 h. The reaction was quenched by adding 2 μL of 5% hydroxylamine, resulting in a final concentration of 0.4%. The TMT-labeled peptides were then pooled in equal proportions, desalted, and lyophilized for storage.

N-glycans pre-treatment

Proteins extracted from EVs were reduced and alkylated, followed by the addition of PNGase F, and the reaction was carried out at 37 °C with shaking at 1100 rpm for 16 h. The N-glycan solution was obtained via ultrafiltration.

A small piece of cotton (8.75–9.25 mg) was placed at the bottom of a 10 μL pipette tip, which was then mounted on a 2.0 mL centrifuge tube. Prior to sample loading, the tip was washed twice with 80 μL water and 80 μL 80% ACN, respectively. The N-glycan solution was prepared with 80% ACN and loaded into the pipette tip, which was subsequently washed four times with 80 μL 80% ACN. During the enrichment process, each solution was allowed to stand for 1 min before centrifugation to ensure complete wetting of the cotton. The glycans were eluted from the cotton three times with 50 μL of ultrapure water, and the eluates were pooled and subsequently lyophilized for storage.

For the methodological evaluation, sialylated N-glycans were labeled with methylamine. To analyze the sialylated N-glycan isomers in EVs derived from ICC cells with ST6GAL1 overexpression, a sequential selective derivatization method was applied [39]. The procedure was as follows: 10 μL of reaction solution (625 mM EDC·HCl, 125 mM HOBt, and 2 M d3 MA·HCl) was added to the lyophilized samples, following by incubation at 37 °C for 2 h. After the addition of 40 μL of ACN and 110 μL of 80% ACN, the sample was desalted using cotton and then lyophilized. Next, 10 μL of 5 M d0 MA·HCl (dissolved in 1 M TEAB) was added, and the mixture was incubated at 37 °C for 1 h. After lyophilization, 20 μL of 125 mM PyAOP (prepared from 4-NMM and DMSO in a 3:7 volume ratio) was added, and the reaction was incubated at 37 °C for an additional 1 h. Finally, 80 μL of ACN and 300 μL of 80% ACN were added, and the sample was desalted with cotton, lyophilized, and stored for future use.

LC–MS/MS analysis

The peptides were dissolved in solvent A (0.1% FA), separated via nano-liquid chromatography (NanoLC), and analyzed using online electrospray tandem mass spectrometry. The LC–MS/MS analysis was conducted on an EASY-NanoLC 1200 system (Thermo Fisher Scientific, MA, USA) connected to an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific, MA, USA) with field asymmetric waveform ion mobility spectrometry (FAIMS). FAIMS was operated at compensation voltages of − 45 V and − 65 V. Peptide separation employed a 75 μm × 25 cm column (2 μm inner diameter). Solvent B consisted of 80% ACN and 0.1% FA.

The elution time for tryptic peptides from unlabeled EVs proteins was 90 min. The gradient progressed from 2 to 35% solvent B over 75 min, 35% to 55% in the next 9 min, followed by a rapid increase to 100% within 1 min, which was maintained for 5 min. The flow rate was consistently set at 300 nL/min. The data were acquired within an m/z range of 350–1600 using data-dependent acquisition (DDA) with higher-energy collisional dissociation (HCD) at 30% collision energy. The automatic gain control (AGC) target was set at 300%, with a maximum injection time of 50 ms. MS1 and MS/MS spectra were recorded at resolutions of 120,000 and 15,000 respectively. To minimize redundant sequencing, dynamic exclusion was applied with a 30 s exclusion window. A 1.6 m/z isolation window was employed for MS/MS, with precursor ions fragmented at a normalized collision energy (NCE) of 30% and a fixed first mass of 110 m/z.

TMT-labeled peptides was eluted using a 120 min gradient. The gradient initially progressed from 2 to 33% solvent B over 110 min, followed by an increase from 33 to 55% over the next 4 min, and a rapid rise to 100% within 1 min, which was maintained for 5 min. The flow rate was consistently set at 300 nL/min. Data were acquired within an m/z range of 400–1600 using DDA with HCD. The AGC target was set at 300%, with a maximum injection time of 50 ms. MS1 and MS/MS spectra were recorded at resolutions of 120,000 and 30,000 respectively. To minimize redundant sequencing, dynamic exclusion was applied with a 30 s exclusion window. A 2 m/z isolation window was employed for MS/MS, with precursor ions fragmented at a NCE of 33% and a fixed first mass of 120 m/z.

MALDI-TOF MS analysis

The Super-DHB solution was prepared by dissolving the mixture in 50% (v/v) ACN containing 0.1% TFA to a final concentration of 10 μg/μL. Similarly, CHCA was dissolved in 50% (v/v) ACN containing 0.1% TFA at a concentration of 5 mg/mL, with 0.75 μg/μL ammonium citrate added. For calibration, the peptide calibration standard mono (Bruker Daltonics) was deposited onto the calibration zone of the target plate. CHCA was then applied to the same spot and allowed to dry at RT.

Samples dissolved in 5 μL of ultrapure water were deposited on the same spot and air-dried. Mass spectra were acquired using a rapifleX matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) MS instrument (Bruker Daltonics, USA) in positive ion mode. Data collection was performed in triplicate for each sample, with each spectrum generated by accumulating 60,000 laser shots.

Data analysis

The raw LC–MS/MS proteomic data were analyzed using PEAKS Online (Bioinformatics Solutions Inc.) and searched against the human proteome database (Homo sapiens, 20,435 entries, https://www.uniprot.org/, updated July 30, 2024). Mass tolerances for precursor and fragment ions were set to 10 ppm and 0.02 Da, respectively. Trypsin was specified as the digestion enzyme, allowing up to two missed cleavages. Carbamidomethylation of cysteine (+ 57.021 Da) was set as a fixed modification, while methionine oxidation (+ 15.995 Da) and N-terminal acetylation (+ 42.010 Da) were included as variable modifications. A 1% false discovery rate threshold was applied at both the peptide-spectrum match and protein levels. TMT- 10-plex quantification was performed using MS2 reporter ions, with all other parameters set to default.

MALDI-TOF MS data for N-glycome analysis were processed using flexAnalysis (version 4.0) with peak detection set to"Snap"and a signal-to-noise (S/N) threshold of 3. Isotope peaks were merged, and quantitative data were extracted based on peak areas and S/N ratios. N-glycans were identified using GlyHunter [40], developed by Lu's group at Fudan University (https://glyhunter.streamlit.app/), and further confirmed by manually matching theoretical m/z values with those in the peak list. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were conducted using Database for Annotation, Visualization, and Integrated Discovery (DAVID) V6.8 (https://david.ncifcrf.gov), while statistical analysis was performed with GraphPad Prism (version 9, Dotmatics).

Results

Evaluation of EVs isolation methods

To evaluate the effectiveness of the EV isolation methods used in this study, a comprehensive characterization of the EV samples isolated from HuCCT1 cells was conducted using multiple complementary techniques (Fig. 1). TEM was employed to assess EVs morphology and size, while NanoFCM analyzed particle populations, size distributions, concentration distributions and biological activity. The biological activity of EVs was further confirmed through fluorescence imaging using a Keyence BZ-X810 microscope. Proteomic analysis of EV-associated proteins was performed using LC–MS/MS, and glycomic profiling of EV-associated glycans was carried out via MALDI-TOF MS. This integrated analytical strategy ensured a robust and thorough assessment of the isolated EVs.

Fig. 1
figure 1

Schematic overview of the experimental evaluation workflow for EV isolation methods

Full size image

Biophysical characterization of EVs

TEM and NanoFCM confirmed successful EV isolation by all five methods, with significant differences observed. TEM showed the characteristic cup-shaped or spherical morphology of EVs, and NanoFCM analysis revealed a particle size distribution between 45 and 120 nm (Fig. 2A). Statistical comparison of size distributions (Fig. 2B) indicated that exoEasy/EVtrap methods isolated larger EVs. Particle concentrations were higher with TEI and ÄKTA methods (Fig. 2C). Protein content showed no direct correlation with particle count (Fig. 2D). For example, TEI yielded the highest particle count, ÄKTA isolated the most protein, and UC performed intermediate in both metrics (Fig. 2C-D). The particle-to-protein ratio was not a reliable indicator for assessing the purity of the isolated EVs and could only provide a rough preliminary evaluation, as previously reported [41]. Rinsing EVs from the dialysis bag with PBS and ultracentrifugation showed negligible protein in the rinse. It indicated a high recovery rate of EVs during the dialysis process. NanoFCM analysis revealed that most methods, except EVtrap, isolated CD9-positive subpopulations (Fig. 2E-F), with EVtrap showing fewer tetramer-positive subpopulations, possibly due to antibody inactivation by ammonium ions in the elution buffer.

Fig. 2
figure 2

valuation of EV isolation methods. A Particle size distribution and representative TEM image of EVs (scale bar: 100 nm). B EV size distributions. C EV concentrations. D Particle count to protein ratio. E NanoFCM analysis of CD9, CD81, and CD63 tetraspanins. F Quantification of tetraspanin markers by NanoFCM. Data were analyzed by one-way ANOVA with Tukey’s HSD for multiple comparisons (n = 3) [16]. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Full size image

Qualitative analysis of the EVs proteome

Due to the limitations of NanoFCM and the Bradford assay, EVs can’t be accurately distinguished from contaminating particles, rendering particle count and protein concentration unreliable for evaluating EV purity and yield. Proteins extracted from CCCM-derived EVs using five isolation methods were analyzed by LC–MS/MS, identifying over 1,677 proteins, with EVtrap detecting the most, followed by UC (Fig. 3A, Table S1). Over 60% of proteins were consistently detected across three replicates (Fig. 3B). A Venn diagram revealed 893 shared proteins, confirming their reliable identification (Fig. 3C). Comparison with Vesiclepedia and ExoCarta showed over 90% overlap with entries in these databases, and UC, exoEasy, and ÄKTA captured over 87% of the top 100 proteins (Fig. 3D) [42,43,44]. Principal component analysis (PCA) of the proteins displayed five distinct clusters (Fig. 3E), with no significant differences in protein subpopulations across replicate, validating the MS approach. However, method-dependent biases were observed. Proteins consistently identified by all five methods were considered reliable EV proteins. GO and KEGG analysis of the 893 shared proteins revealed significant enrichment in extracellular components and pathways related to cell proliferation, adhesion, and metastasis (Fig. 3F-G). Proteomics indicated over 90% of proteins were localized in EVs, supporting further research on their glycosylation modifications.

Fig. 3
figure 3

Proteomic analysis of EVs isolated by different methods. A Protein identification comparison across methods. B Stacked bar chart of proteins identified in three technical replicates, color-coded by replicate count. C Venn diagram showing overlap of EV proteins identified by different methods. D Venn diagram comparing identified proteins with Vesiclepedia, ExoCarta, and top 100 proteins. E 3D-PCA of proteomes from EVs showing five clusters based on three principal components. F GO analysis of shared proteins by biological process, cellular component, and molecular function. G KEGG pathway analysis of shared proteins

Full size image

Qualitative analysis of the EVs N-glycome

N-glycans extracted from EVs using different isolation methods were analyzed by MALDI-TOF MS to assess glycans identification efficiency and specificity. A total of 35 distinct N-glycans were consistently identified across all five methods, confirming their reliability (Fig. 4A-D), with the full list provided in Table S2. The TEI identified 45 N-glycans (Fig. 4E), slightly fewer than the other methods: UC (54), exoEasy: (50), EVtrap: (60), ÄKTA: (46). This variation may be attributed to differences in extraction efficiency or the selectivity of the TEI protocol.

Fig. 4
figure 4

Glycomic analysis of EVs isolated by different methods. MALDI-MS spectra of glycans from EVs isolated by UC (A), exoEasy (B), TEI (C), EVtrap (D), and ÄKTA (E). Glycan peaks are marked with “*”. Red squares represent glycans identified from UC, blue squares from exoEasy, orange squares from TEI, green squares from EVtrap, purple squares from ÄKTA, and white squares indicate unidentified glycans

Full size image

Analysis of EVs internalization

EVs are internalized by recipient cells to facilitate communication, making their ability to maintain internalization crucial for understanding their functional mechanisms. To evaluate the suitability of different isolation methods for functional studies, we assessed the activity of EVs extracted using five techniques (Fig. 5). EVs quantification was performed using NanoFCM, with particle numbers normalized before further analysis. Fluorescent labeling was applied to the EVs, which were then co-incubated with recipient cells. After incubation, the samples were divided into two portions: one was analyzed by fluorescence imaging to visualize and quantify EV internalization by cells, enabling a comparative assessment of EVs activity across isolation methods (Fig. 5A). The second portion was analyzed by FCM, measuring mean fluorescence intensity for a quantitative comparison of EV activity (Fig. 5B). Both techniques yielded consistent results: EVs isolated by the exoEasy method exhibited the highest activity, followed by UC and ÄKTA, while the TEI and EVtrap methods showed the lowest activity.

Fig. 5
figure 5

Comparison of EVs internalization by HuCCT1 cells isolated by different methods. A Fluorescence images of EVs labeled with DAPI (blue, 0.5 μg/mL) and PKH67 (green, 100 μM) after incubation with HuCCT1 cells. B Mean fluorescence intensity of PKH67-labeled EVs after 6 h of incubation with HuCCT1 cells. **p < 0.01, ***p < 0.001, ****p < 0.0001

Full size image

Selection of EVs isolation methods

To establish a comprehensive and high-performance EVs extraction method suitable for mechanistic studies, we performed an in-depth comparative analysis of five different extraction techniques, focusing on key parameters such as biophysical characteristics, cargo identification comprehensiveness, and operational feasibility. These factors were systematically evaluated to ensure that the selected method not only yielded high-quality EVs but also demonstrated broad applicability in omics and mechanistic research. A summary of the advantages and limitations of each method was provided in Table 1. Based on this comprehensive evaluation, UC, exoEasy, and ÄKTA were identified as the most favorable techniques. Among these, exoEasy does not require specialized equipment but is relatively expensive, while ÄKTA reduces manual handling and supports automation, though it requires specialized equipment and entails higher operational complexity. Ultimately, UC was selected as the optimal method for EVs extraction in subsequent experiments, offering a well-balanced combination of performance, reproducibility, and feasibility. However, this method does have certain limitations. While UC is considered the gold standard for EVs extraction, it faces challenges in removing copious lipoprotein contamination from serum [7, 18, 45, 46], significantly limiting its application in isolating EVs from serum-containing culture media. Addressing concerns about the potential damage to EVs caused by high shear forces during UC, we performed activity comparisons using endocytosis experiments and confirmed through microscope and FCM that EV activity was preserved under our experimental conditions and could be internalized by recipient cells.

Table 1 Comparison summary of five extracellular vesicles extraction methods
Full size table

Cell lines construction

Abnormal sialylation is strongly associated with cancer progression, including key processes such as proliferation, migration, and invasion [47]. The Golgi-resident sialyltransferase ST6GAL1 catalyzes the addition of negatively charged α2,6-linked sialic acid to the termini of N-glycans and has been found to be upregulated in various malignancies [48]. However, research on the relationship between ST6GAL1 and ICC remains limited. We established stable ST6GAL1-overexpressing HuCCT1 (HuCCT1-ST6GAL1) and HCCC- 9810 (HCCC- 9810-ST6GAL1) cell lines, as well as their respective vector controls (HuCCT1-Vector and HCCC- 9810-Vector), for quantitative proteomics and glycomics analysis of EVs. The methods for cell line construction were consistent with those described in our previous study [38]. The overexpression of ST6GAL1 in HCCC9810 and HuCCT1 cells was confirmed by RT-qPCR and western blotting, with the relevant data already published in our previous research [38].

Quantitative analysis of EV proteins

To assess the impact of ST6GAL1 overexpression on the protein composition of EVs derived from ICC cells, we conducted TMT-based quantification to analyze alterations in EV proteins from HuCCT1 and HCCC- 9810 cells upon ST6GAL1 overexpression. A correlation analysis of the proteins identified in EVs derived from ICC cells was presented in Figure S1 A. The overlap between three triplicate experiments was illustrated in Figure S1B-E. We observed that the correlation of proteins from EVs derived from the same ICC cell line was consistently high, with a value greater than 0.92 and an average of 0.95, and the overlap between the three technical replicates was above 85.2%. This strong correlation and high overlap reflected stable MS performance throughout the testing period. In contrast, the correlation between proteins from EVs of the two different cell lines was lower, with values above 0.59 and an average of 0.73, suggesting significant differences in the protein compositions of EVs isolated from HuCCT1 and HCCC- 9810 (Fig. S1 A).

Overexpression of ST6GAL1 led to significant changes in the protein composition of EVs at the protein level (Fig. 6A-B). Proteins that were significantly upregulated or downregulated in EVs from the ST6GAL1-overexpressing cell lines tended to cluster together (Fig. S2 A-B). Specifically, 16 proteins were consistently upregulated, while 10 proteins were consistently downregulated (Fig. S2 C-D). In total, 2,293 quantifiable proteins were identified in EVs derived from ICC cells, with 1,928 of these considered high-confidence proteins, defined as those with fewer than 50% missing values in the MS quantification data across all experiments (Table S3). Further analysis revealed the absence of Calnexin, while EV marker proteins such as CD9, CD81, CD63, HSP70, and TSG101 were detected, confirming the lack of significant Calnexin contamination in the extracts. Notably, CD63 was found to be significantly overexpressed in EVs derived from ST6GAL1-overexpressing cells (Fig. S2E). Gene Ontology (GO) analysis of the quantified proteins revealed a predominant localization in the extracellular exosome, which aligned with expectations (Fig. 6C-D). KEGG pathway analysis suggested that these proteins might be involved in cancer cell adhesion and migration, with significant enrichment in pathways such as glycolysis/gluconeogenesis and proteoglycans in cancer, highlighting the need for further investigation into the glycosylation of EV proteins (Fig. 6E-F).

Fig. 6
figure 6

Quantitative proteomics and bioinformatics analysis of EV proteins. A Volcano plot comparing the EV proteins between HuCCT1-ST6GAL1 and HuCCT1-Vector. B Volcano plot comparing the EV proteins between HCCC- 9810-ST6GAL1 and HCCC- 9810-Vector. C GO analysis of EV proteins significantly upregulated or downregulated in HuCCT1-ST6GAL1. D GO analysis of EV proteins significantly upregulated or downregulated in HCCC- 9810-ST6GAL1. E KEGG pathway analysis of EV proteins significantly upregulated or downregulated in HuCCT1-ST6GAL1. F KEGG pathway analysis of EV proteins significantly upregulated or downregulated in HCCC- 9810-ST6GAL1

Full size image

Quantitative analysis of EV N-glycans

The fingerprint profile of EVs derived from ICC cell lines was established using MALDI-TOF MS to analyze the impact of ST6GAL1 overexpression on the relative abundance of N-glycans in these EVs. This analysis allowed us to investigate how ST6GAL1 overexpression altered the glycosylation patterns of EVs, providing insights into the potential functional changes induced by these alterations.

For HuCCT1, 84 N-glycans were identified in EVs derived from HuCCT1- ST6GAL1 across three technical replicates, while 67 N-glycans were identified in EVs from HuCCT1- Vector (Fig. 7A, Table S4). The overlap between the three replicates was 70.7% and 80.4%, respectively (Fig. S3 A-B). Among these, 84 glycans with quantitative information were detected in at least two of the three technical replicates. Further analysis revealed significant changes in the relative abundance of 20 N-glycans, including 6 α2,6- sialylated glycans that were significantly upregulated and 1 α2,6-sialylated glycan that was significantly downregulated (Fig. S3 C-D). At the overall glycan type level, ST6GAL1 overexpression induced alterations in the distribution of glycan types in EVs. Specifically, the proportions of high-mannose and sialylated glycans increased in EVs derived from HuCCT1-ST6GAL1, while the proportions of neutral glycans, fucosylated glycans, and complex glycans were reduced. Notably, the increase in sialylated glycans was attributed to both α2,6-sialylated and α2,3-sialylated glycans, while the proportion of hybrid glycans remained unchanged (Fig. 7B, Table S5).

Fig. 7
figure 7

Quantitative glycomics analysis. A MALDI-TOF MS analysis of sialylated N-glycan isomers in EVs derived from HuCCT1-ST6GAL1 and HuCCT1-Vector, with an m/z range of 1000–4000. Selected N-glycan structures are highlighted with their corresponding MS peaks, where red indicates upregulated N-glycans and green indicates downregulated N-glycans. B The distribution of abundance proportions of different N-glycan types

Full size image

For HCCC- 9810, 95 N-glycans were identified in EVs derived from HCCC- 9810- ST6GAL1 across three technical replicates, while 65 N-glycans were identified in EVs from HCCC- 9810-Vector (Fig. S4 A, Table S4). The overlap between the three replicates was 64.7% and 79.8%, respectively (Fig. S5 A-B). Among these, 86 glycans with quantitative data were identified in at least two of the three technical replicates. Further analysis revealed significant changes in the relative abundance of 18 N-glycans, including 2 α2,6-sialylated glycans that were significantly upregulated and 7 α2,6-sialylated glycans that were significantly downregulated (Fig. S4 C-D). At the overall glycan-type level, ST6GAL1 overexpression induced alterations in the relative abundance of several glycan types in HCCC- 9810-derived EVs. Specifically, the proportions of high-mannose and sialylated glycans increased, while the proportions of neutral glycans, fucosylated glycans, and complex glycans decreased. No significant changes were observed in the proportions of hybrid glycans or α2,3-sialylated glycans. In contrast to EVs derived from HuCCT1 cells, the increase in sialylated glycans in ST6GAL1-overexpressing HCCC- 9810-derived EVs was primarily due to the increase in α2,6-sialylated glycans.

A comprehensive analysis of the EV glycome revealed that complex glycans were the predominant N-glycan type in EVs, followed by a smaller proportion of high-mannose glycans, while hybrid glycans were present at the lowest levels-findings consistent with previous reports [49]. Detailed structural analysis of the EV glycome derived from ICC cells revealed significant upregulation of α2,6-sialylated glycans, such as H5 N4 F1S2 and H5 N4S2, alongside the neutral glycan H6 N2. In contrast, several neutral glycans, including H4 N5 F1, H6 N5, and H7 N6 F1, were found to be downregulated. This quantitative glycomic analysis highlighted the advantageous of MS-based structural approaches over lectin array-based methods by providing a more comprehensive and detailed understanding of glycan structures, as well as facilitating the identification of potential targets. Notably, this study represents the first glycomics analysis of EVs derived from ICC cells, offering new insights into glycan alterations in EVs secreted by ICC cell lines overexpressing ST6GAL1.

ST6GAL1 is a glycosyltransferase responsible for catalyzing the α2,6-sialylation of N-glycans [50], playing a crucial role in glycosylation processes, especially in the regulation of protein glycosylation within EVs. In this study, we employed light and heavy methylamidation labeling techniques, which generated a 3.0186 Da mass difference between α2,6-sialic acid and α2,3-sialic acid linkage isomers, enabling a comprehensive analysis of the glycome of EVs derived from ICC cell lines. Our analysis revealed significant alterations in the relative abundance of glycan species within EVs (Fig. 7, Fig. S4), with a notable increase in α2,6-sialylated linkage isomers in EVs from both ST6GAL1-overexpressing ICC cell lines. These findings suggest that ST6GAL1 overexpression not only alters sialic acid linkage patterns but also induces substantial modifications in the overall glycosylation landscape of EVs, particularly promoting the accumulation of α2,6-sialylated glycans. This underscores the pivotal role of ST6GAL1 in glycosylation regulation and indicates that its overexpression may influence intercellular signaling and modulate the cellular microenvironment by altering the glycan composition and distribution in EVs. Consistent with the above analysis, several neutral N-glycans exhibited significant changes in relative abundance in EVs from ICC cells, and a single α2,3-sialylated glycan showed decreased abundance in EVs derived from HCCC- 9810-ST6GAL1. These findings suggested that ST6GAL1 overexpression not only promoted the synthesis of α2,6-sialylated glycans but may also influenced the synthesis and degradation of other glycan types through multiple pathways. One possible explanation is that N-glycans may modulate the activity and processing of glycosyltransferases [51]. We propose that the overexpression of ST6GAL1 could induce alterations in α2,6-sialylated glycans, which may, in turn, influence the function and processing of other glycosyltransferases, ultimately resulting in modifications to other glycan types, including neutral glycans. This hypothesis was further supported by differential and pathway analysis of N-glycans in EVs secreted by ST6GAL1-overexpressing cells (Fig. 8).

Fig. 8
figure 8

Differential and pathway analysis of all N-glycans in EVs from ICC cells. N-glycans with different numbers of monosaccharides are displayed in separate columns. The color scheme in the figure conveys specific meanings: dark blue represents N-glycans uniquely detected in EVs derived from ST6GAL1-overexpressing cells; light blue indicates N-glycans with higher expression levels in ST6GAL1-overexpressing EVs; dark red denotes components exclusively found in EVs from vector-transfected cells; and light red indicates N-glycans with lower expression levels in ST6GAL1-overexpressing EVs. Lines connecting different N-glycans signify monosaccharide additions: the red line indicates the addition of NeuAc (N-acetylneuraminic acid), the green line represents dHex (fucose), the blue line denotes HexNAc (N-acetylhexosamine), and the orange line corresponds to Hex (hexose)

Full size image

Discussion

EVs, a heterogeneous group of phospholipid bilayer-enclosed vesicles, are widely present across living organisms and are generally believed to play a role in intercellular communication. The membranes of EVs encapsulate biomolecules, including proteins, lipids, and DNA, which serve as relatively stable cargo [7]. Aberrant protein glycosylation is a hallmark of cancer, and the composition and structure of EVs, along with their protein glycosylation patterns, can reflect the physiological state of the parent cells [52, 53]. By combining the analysis of these two factors, we can gain insights into the mechanisms underlying EVs function. In our previous study, we found that reduced expression of ST6GAL1 mRNA and protein inhibited the proliferation and migration of ICC cells [38], suggesting that ST6GAL1 may play a role in mediating pathological processes, such as proliferation, migration, and invasion, in ICC. These effects could be mediated by intercellular communication via EVs as signaling messengers. Given that EVs are present in a complex biological environment and that the abundance of proteins and glycosylation varies across a wide range [54], it is essential to identify an appropriate method for EVs extraction.

This study systematically evaluated five EV isolation methods—UC, exoEasy, TEI, EVtrap, and ÄKTA—by analyzing the biophysical properties, proteomic profiles, and glycomic structures of EVs. UC is widely regarded as the gold standard for EVs isolation [7]. In this study, UC successfully isolated structurally intact EVs (Fig. 2A) while maintaining internalized activity, which was confirmed through receptor cell uptake (Fig. 5A). In proteomic and glycomic analysis, EVs isolated by UC demonstrated superior performance compared to those isolated by exoEasy, TEI, and ÄKTA (Fig. 4, Table S2). Additionally, while EVtrap proved effective for analyzing EV-associated proteins and N-glycans, it failed to preserve EVs activity (Fig. 2E, Fig. 5). This limitation made EVtrap particularly suitable for liquid biomarker screening [31], but less effective for mechanistic EV studies. Based on these findings, UC was selected as the optimal method in this study due to its balance of operational complexity, cost-effectiveness, and ability to preserve EV activity. However, the application of the UC method may be limited. As is well known, UC faces challenges in removing contaminating lipoproteins from serum [7, 18, 45, 46]. Consequently, for cell lines that are not amenable to serum-free medium culture, achieving high-purity EVs using UC becomes problematic. These limitations highlight the importance of selecting an EVs isolation method that aligns with the specific experimental subject and research objectives.

ICC is a malignancy characterized by an insidious onset, high invasiveness, and poor prognosis. In recent years, research on ICC has increasingly focused on proteins and glycosylation, aiming to elucidate the mechanisms driving its development [4, 13]. EVs play a pivotal role in key biological processes, including cell communication, migration, angiogenesis, and tumor progression. To further under these mechanisms, this study explored the optimal isolation method for EVs and focused on analyzing the protein composition and detailed N-glycan structures of EVs derived from ICC cells. High-resolution MS was employed to comprehensively characterize the proteomic and N-glycomic alterations in EVs derived from ICC cells with ST6GAL1 overexpression, providing robust data to support future mechanistic investigations.

To explore the preliminary impact of ST6GAL1 overexpression on the cargo of EVs derived from ICC cells, proteomic and N-glycomic analysis was performed on the isolated EVs. In the proteomic quantification analysis, 2,292 proteins were successfully quantified, including EV markers such as CD9, CD81, CD63, and TSG101 (Table S3). Notably, CD63 was identified as the only marker that was consistently upregulated in EVs derived from ST6GAL1-overexpressing cells (Fig. S2E). The absence of calnexin further validated the successful extraction of EVs. Compared to EVs from HuCCT1-Vector cells, EVs from HuCCT1-ST6GAL1 cells exhibited 214 upregulated proteins and 57 downregulated proteins (Fig. 6A, Fig. S2 C). Similarly, EVs derived from HCCC- 9810-ST6GAL1 cells showed 108 upregulated proteins and 186 downregulated proteins compared to their vector controls (Fig. 6B, Fig. S2D). Notably, 16 proteins were consistently upregulated, and 10 proteins were consistently downregulated in EVs from both cell lines with ST6GAL1 overexpression (Fig. S2 C-D). Among these proteins, Gelsolin (GELS), whose expression in the serum of cholangiocarcinoma patients has been reported to be significantly reduced compared to healthy individuals and those with primary sclerosing cholangitis [55], was also downregulated in EVs from ST6GAL1-overexpressing cells. Conversely, Tissue-type plasminogen activator (TPA), which is associated with ICC metastasis in TP53/KRAS co-mutated tissues through the integrin FAK-SRC signaling pathway [13], was upregulated in these EVs. Consistent with these findings, the quantification in this study further supports the involvement of EV-associated proteins, such as GELS and TPA, in ICC progression. Additionally, this analysis identified EV-specific proteins distinct from those found in other biological samples, emphasizing the need for further investigation into their functions and characteristics. These findings will form a major focus of future research.

Quantitative proteomic analysis revealed that differentially expressed proteins were significantly enriched in pathways related to cell adhesion, glycolysis/gluconeogenesis, and cancer-associated protein glycosylation (Fig. 6). Given the critical role of glycosylation in regulating cellular functions, we conducted a comprehensive characterization of the N-glycan fine structures in EVs, focusing specifically on the two most common sialic acid linkage isomers: α2,3- and α2,6-linked isomers. These isomers are known to influence key physiological and pathological processes, including cell recognition, signal transduction, and cancer progression [8,9,10]. High-resolution glycomic profiling revealed that ST6GAL1 overexpression led to a significant increase in α2,6-linked isomers, while also affecting the levels of α2,3-linked isomers. Furthermore, we examined the broader impact of ST6GAL1 on the overall glycosylation landscape and observed systematic alterations in other glycan modifications. These changes may arise from the intricate regulatory interactions among glycosyltransferases and the dynamic adaptation of glycans within the cellular microenvironment. By integrating proteomic and glycomic analyses, our study highlights the central role of ST6GAL1 in modulating the glycosylation network, providing new insights into the functional significance of glycosylation in ICC and other malignancies. Further N-glycan analysis revealed that most N-glycans in EVs exhibited significant differences upon ST6GAL1 overexpression, with specific N-glycan structures being profoundly altered (Fig. 7, Fig. S3–S5). Notably, these changes were not restricted to α2,6-sialylated glycans, indicating a more complex interplay between glycans and glycosyltransferases, a concept supported by previous studies [51]. Pathway analysis further confirmed that ST6GAL1 overexpression influenced the expression of multiple glycosyltransferases, reinforcing the notion that its effects extend beyond α2,6-sialylation (Fig. 8).

Conclusion

This study systematically investigated the biophysical properties and composition of EVs utilizing FCM, TEM, and MS. Among the various isolation techniques evaluated, UC was identified as the most effective method for isolating EVs from cells cultured in serum-free medium. For the first time, a combination of LC–MS/MS and MALDI-TOF–MS was employed to comprehensively characterize the proteomic and glycomic alterations in EVs derived from ST6GAL1-overexpressing ICC cells. The results revealed that ST6GAL1 overexpression caused significant changes in the protein composition of EVs, with differentially expressed proteins being closely associated with glycolysis/gluconeogenesis and proteoglycans in cancer. In addition, specific glycan structures within the EVs were significantly altered following ST6GAL1 overexpression. Notably, these changes were not confined to α2,6-sialylated glycans, indicating that glycosyltransferases may interact with glycans to drive these modifications. These findings provide a crucial theoretical basis for further investigation into the molecular mechanisms underlying ICC. Future research will aim to explore the relationship between EV-mediated alterations in biological pathways and the pathological progression of ICC, potentially identifying new biomarkers and therapeutic targets for cancer diagnosis and treatment.

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identifier PXD062160.

Abbreviations

ICC:

Intrahepatic cholangiocarcinoma

EVs:

Extracellular vesicles

UC:

Ultracentrifugation

TEI:

Total exosome isolation

TEM:

Transmission electron microscopy

FCM:

Flow cytometry

MS:

Mass spectrometry

ST6GAL1:

ST6 β‑galactoside α2,6‑sialyltransferase 1

CCCM:

Conditioned cell culture medium

FBS:

Fetal bovine serum

NanoLC:

Nano-liquid chromatography

FAIMS:

Field asymmetric waveform ion mobility spectrometry

S/N:

Signal-to-noise

GO:

Gene ontology

KEGG:

Kyoto encyclopedia of genes and genomes

RT:

Room temperature

MALDI-TOF:

Matrix-assisted laser desorption ionization time-of-flight

PCA:

Principal component analysis

NeuAc:

N-acetylneuraminic acid

dHex:

Fucose

HexNAc:

N-acetylhexosamine

Hex:

Hexose

References

  1. Moris D, Palta M, Kim C, et al. Advances in the treatment of intrahepatic cholangiocarcinoma: An overview of the current and future therapeutic landscape for clinicians. CA Cancer J Clin. 2022;73(2):198–222.

    Article  PubMed  Google Scholar 

  2. Yang JQ, Wang XG, Wu B. Incidence trend and prognosis of intrahepatic cholangiocarcinoma: a study based on the SEER database[J]. Translational Cancer Research. 2023;12(11):3007–15.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Banales JM, Cardinale V, Carpino G, et al. Current knowledge and future perspectives consensus statement from the european network for the study of cholangiocarcinoma (ENS-CCA)[J]. Nat Rev Gastroenterol Hepatol. 2016;13:261–80.

    Article  PubMed  Google Scholar 

  4. Ochoa-Rios S, Blaschke CRK, Wang M, et al. Analysis of N-linked glycan alterations in tissue and serum reveals promising biomarkers for intrahepatic cholangiocarcinoma[J]. Cancer Res Commun. 2023;3(3):383–94.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Liu ZH, Lian BF, Dong QZ, et al. Whole-exome mutational and transcriptional landscapes of combined hepatocellular cholangiocarcinoma and intrahepatic cholangiocarcinoma reveal molecular diversity. Biochim Biophys Acta Mol Basis Dis . 2018;1864(6):2360–8.

    Article  CAS  PubMed  Google Scholar 

  6. Sirica AE, Gores GJ, Groopman JD, et al. Intrahepatic cholangiocarcinoma: continuing challenges and translational advances[J]. Hepatology. 2019;69(4):1803–15.

    Article  PubMed  Google Scholar 

  7. Wu LL, Gao CF. Comprehensive overview the role of glycosylation of extracellular vesicles in cancers[J]. ACS Omega. 2023;8:47380–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Cheng M, Shu H, Yang M, et al. Fast discrimination of sialylated N-glycan linkage isomers with one-step derivatization by microfluidic capillary electrophoresis-mass spectrometry[J]. Anal Chem. 2022;94(11):4666–76.

    Article  CAS  PubMed  Google Scholar 

  9. Pally D, Pramanik D, Hussain S, et al. Heterogeneity in 2,6-Linked sialic acids potentiates invasion of breast cancer epithelia[J]. ACS Cent Sci. 2021;7(1):110–25.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Mereiter S, Balmana M, Campos D, et al. Glycosylation in the era of cancer-targeted therapy: where are we heading?[J]. Cancer Cell. 2019;36(1):6–16.

    Article  CAS  PubMed  Google Scholar 

  11. Rahnemai-Azar AA, Weisbrod A, Dillhoff M, et al. Intrahepatic cholangiocarcinoma: molecular markers for diagnosis and prognosis[J]. Surg Oncol. 2017;26(2):125–37.

    Article  PubMed  Google Scholar 

  12. Liu W, Wang H, Zhao Q, et al. Multiomics analysis reveals metabolic subtypes and identifies diacylglycerol kinase α (DGKA) as a potential therapeutic target for intrahepatic cholangiocarcinoma[J]. Cancer Commun. 2024;44:226–50.

    Article  Google Scholar 

  13. Dong L, Lu D, Chen R, et al. Proteogenomic characterization identifies clinically relevant subgroups of intrahepatic cholangiocarcinoma. Cancer Cell. 2022;40(1):70-87 e15.

    Article  CAS  PubMed  Google Scholar 

  14. Asao T, Tobias GC, Lucotti S, et al. Extracellular vesicles and particles as mediators of long-range communication in cancer: connecting biological function to clinical applications[J]. Extracellular Vesicles and Circulating Nucleic Acids. 2023;4(3):461–85.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Ali SY, Sajdera SW, Anderson HC. Isolation and characterization of calcifying matrix vesicles from epiphyseal cartilage[J]. Proc Natl Acad Sci U S A. 1970;67(3):1513–20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Veerman RE, Teeuwen L, Czarnewski P, et al. Molecular evaluation of five different isolation methods for extracellular vesicles reveals different clinical applicability and subcellular origin[J]. Journal of Extracellular Vesicles. 2021;10(9): e12128.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Hinzman CP, Singh B, Bansal S, et al. A multi-omics approach identifies pancreatic cancer cell extracellular vesicles as mediators of the unfolded protein response in normal pancreatic epithelial cells[J]. Journal of Extracellular Vesicles. 2022;11(6): e12232.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Li Y, Wang J, Chen W, et al. Comprehensive review of MS-based studies on N-glycoproteome and N-glycome of extracellular vesicles[J]. Proteomics. 2023;24:e2300065.

    Article  PubMed  Google Scholar 

  19. Harada Y, Kizuka Y, Tokoro Y, et al. N-glycome inheritance from cells to extracellular vesicles in B16 melanomas[J]. FEBS Lett. 2019;593(9):942–51.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Huang L, Bockorny B, Paul I, et al. PDX-derived organoids model in vivo drug response and secrete biomarkers[J]. JCI Insight. 2020;5(21): 135544.

    Article  PubMed  Google Scholar 

  21. Kondo K, Harada Y, Nakano M, et al. Identification of distinct N-glycosylation patterns on extracellular vesicles from small-cell and non-small-cell lung cancer cells[J]. J Biol Chem. 2022;298(6): 101950.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Corso G, Mager I, Lee Y, et al. Reproducible and scalable purification of extracellular vesicles using combined bind-elute and size exclusion chromatography[J]. Sci Rep. 2017;7(1):11561.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Diaz G, Bridges C, Lucas M, et al. Protein digestion, ultrafiltration, and size exclusion chromatography to optimize the isolation of exosomes from human blood plasma and serum[J]. J Vis Exp. 2018;134: e57467.

    Google Scholar 

  24. Shi C, Zhou X, Yang W, et al. Proteomic analysis of plasma-derived extracellular vesicles from mice with echinococcus granulosus at different infection stages and their immunomodulatory functions[J]. Frontiers in Cellular and Infection Microbiologyl. 2022;12: 805010.

    Article  CAS  Google Scholar 

  25. Kotsiou OS, Katsanaki K, Tsiggene A, et al. Detection and characterization of extracellular vesicles in sputum samples of COPD patients[J]. Journal of Personalized Medicine. 2024;14(8):820.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Vitucci ECM, Carberry CK, Payton A, et al. Characterizing the extracellular vesicle proteomic landscape of the human airway using in vitro organotypic multi-cellular models[J]. iScience. 2023;26(11):108162.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Balusu S, van Wonterghem E, de Rycke R, et al. Identification of a novel mechanism of blood-brain communication during peripheral inflammation via choroid plexus-derived extracellular vesicles[J]. EMBO Mol Med. 2016;8(10):1162–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Capaci V, Kharrat F, Conti A, et al. The deep proteomics approach identified extracellular vesicular proteins correlated to extracellular matrix in type one and two endometrial cancer[J]. Int J Mol Sci. 2024;25(9):4650.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Li Y, Fu B, Li Y, et al. O-glycoIsoQuant: A novel O-glycome quantitative approach through superbase release and isotopic Girard’s P labeling[J]. Anal Chem. 2024;96(18):7289–96.

    Article  CAS  PubMed  Google Scholar 

  30. Li Y, Fu B, Wang M, et al. Urinary extracellular vesicle N-glycomics identifies diagnostic glycosignatures for bladder cancer[J]. Nat Commun. 2025;16(1):2292.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Wu X, Li L, Iliuk A, et al. Highly efficient phosphoproteome capture and analysis from urinary extracellular vesicles[J]. J Proteome Res. 2018;17(9):3308–16.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Welsh JA, Goberdhan DCI, O’Driscoll L, et al. Minimal information for studies of extracellular vesicles (MISEV2023): From basic to advanced approaches[J]. Journal of Extracellular Vesicles. 2024;13(2): e12404.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Welsh JA, Goberdhan DC, O’Driscoll L, et al. MISEV2023: An updated guide to EV research and applications[J]. Journal of Extracellular Vesicles. 2024;13(2): e12416.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Witwer KW, Goberdhan DC, O’Driscoll L, et al. Updating MISEV: Evolving the minimal requirements for studies of extracellular vesicles[J]. Journal of Extracellular Vesicles. 2021;10(14): e12182.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Witwer KW, Soekmadji C, Hill AF, et al. Updating the MISEV minimal requirements for extracellular vesicle studies: building bridges to reproducibility[J]. Journal of Extracellular Vesicles. 2017;6(1):1396823.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Zhou Q, Niu X, Zhang Z, et al. Glycan profiling in small extracellular vesicles with a SERS microfluidic biosensor Identifies early malignant development in lung cancer[J]. Advanced Science. 2024;11:e2401818.

    Article  PubMed  Google Scholar 

  37. Vyhlidalova Kotrbová A, Gomoryova K, Mikulova A, et al. Proteomic analysis of ascitic extracellular vesicles describes tumour microenvironment and predicts patient survival in ovarian cancer[J]. J Extracell Vesicles. 2024;13(3):e12420.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Zhan Y, Wang R, Huang C, et al. Digitoxin inhibits ICC cell properties via the NF-kappaB/ST6GAL1 signaling pathway[J]. Oncol Rep. 2024;52(2):103.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Peng Y, Wang L, Zhang Y, et al. Stable isotope sequential derivatization for linkage-specific analysis of sialylated N-glycan isomers by MS[J]. Anal Chem. 2019;91(24):15993–6001.

    Article  CAS  PubMed  Google Scholar 

  40. Li YY, Zhai YJ, Fu B, et al. A comprehensive N-glycome map of porcine sperm membrane before and after capacitation[J]. Carbohyd Polym. 2024;335: 122084.

    Article  CAS  Google Scholar 

  41. Takov K, Yellon DM, Davidson SM. Comparison of small extracellular vesicles isolated from plasma by ultracentrifugation or size-exclusion chromatography: yield, purity and functional potential[J]. Journal of Extracellular Vesicles. 2019;8(1):1560809.

    Article  CAS  PubMed  Google Scholar 

  42. SAI V CHITTI, SRIRAM GUMMADI, TAEYOUNG KANG, Vesiclepedia, et al. an extracellular vesicles and extracellular particles repository[J]. Nucleic Acids Res. 2024;2024(52):D1694–8.

    Google Scholar 

  43. Keerthikumar S, Chisanga D, Ariyaratne D, et al. ExoCarta: a web-based compendium of exosomal cargo[J]. J Mol Biol. 2016;428(4):688–92.

    Article  CAS  PubMed  Google Scholar 

  44. Mathivanan S, Fahner CJ, Reid GE, et al. ExoCarta 2012: database of exosomal proteins, RNA and lipids[J]. Nucleic Acids Res. 2012;40(Database issue):D1241-1244.

    Article  CAS  PubMed  Google Scholar 

  45. Hu A, Zhang J, Zhang L, et al. Efficient cancer biomarker screening and multicancer detection enabled by a multidimensional serum proteomic strategy[J]. Anal Chem. 2024;96(49):19294–303.

    Article  CAS  PubMed  Google Scholar 

  46. Zhang M, Jin K, Gao L, et al. Methods and technologies for exosome isolation and characterization[J]. Small Methods. 2018;2(9):1800021.

    Article  Google Scholar 

  47. Gc S, Bellis SL, Hjelmeland AB. ST6Gal1: Oncogenic signaling pathways and targets[J]. Front Mol Biosci. 2022;9: 962908.

    Article  PubMed  PubMed Central  Google Scholar 

  48. Garnham R, Scott E, Livermore K, et al. ST6GAL1: A key player in cancer[J]. Oncol Lett. 2019;18(2):983–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  49. Costa J, Gatermann M, Nimtz M, et al. N-glycosylation of extracellular vesicles from HEK-293 and glioma cell lines[J]. Anal Chem. 2018;90(13):7871–9.

    Article  CAS  PubMed  Google Scholar 

  50. Hugonnet M, Singh P, Haas Q, et al. The distinct roles of sialyltransferases in cancer biology and onco-immunology[J]. Front Immunol. 2021;12: 799861.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Mikolajczyk K, Kaczmarek R, Czerwinski M. How glycosylation affects glycosylation: the role of N-glycans in glycosyltransferase activity[J]. Glycobiology. 2020;30(12):941–69.

    Article  CAS  PubMed  Google Scholar 

  52. Jankovic T, Jankovic M. Extracellular vesicles and glycans: new avenue for biomarker research[J]. Biochemistry and Medicine (Zagreb). 2024;34(2): 020503.

    Article  Google Scholar 

  53. Gomes J, Gomes-Alves P, Carvalho SB, et al. Extracellular vesicles from ovarian carcinoma cells display specific glycosignatures[J]. Biomolecules. 2015;5(3):1741–61.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Schjoldager KT, Narimatsu Y, Joshi HJ, et al. Global view of human protein glycosylation pathways and functions[J]. Nat Rev Mol Cell Biol. 2020;21(12):729–49.

    Article  CAS  PubMed  Google Scholar 

  55. Lapitz A, Azkargorta M, Milkiewicz P, et al. Liquid biopsy-based protein biomarkers for risk prediction, early diagnosis, and prognostication of cholangiocarcinoma[J]. J Hepatol. 2023;79(1):93–108.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We express our gratitude to the Core Facility of Shanghai Medical College, Fudan University. Special thanks to all reviewers who provided valuable corrections to this paper.

Funding

This research was supported by the National Science Foundation of China under grant number 82372321 and 22404113.

Author information

Author notes

  1. Linlin Wu and Jiao Wei contributed equally to this work.

Authors and Affiliations

  1. Clinical Laboratory Medicine Center, Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, 200437, China

    Linlin Wu, Jiao Wei, Yueping Zhan, Xiao Xiao, Xuewen Xu, Chenjun Huang, Tian Long, Mengmeng Wang & Chunfang Gao

  2. Shanghai Cancer Center and Institutes of Biomedical Sciences and Department of Chemistry and NHC Key Laboratory of Glycoconjugates Research, Fudan University, Shanghai, 200030, China

    Yueyue Li & Bin Fu

Authors
  1. Linlin Wu
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Jiao Wei
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Yueping Zhan
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Xiao Xiao
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Xuewen Xu
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Chenjun Huang
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Tian Long
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Yueyue Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Bin Fu
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Mengmeng Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Chunfang Gao
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

L.W: Experiments, Data analysis, Writing – review & editing, funding support (22,404,113). J.W: Experiments, Data analysis. Y.Z: Cell lines construction. X.X, X.X, C.H, T.L, Y.L, B.F and M.W: Resources. C.G: Writing – review & editing, funding support (82,372,321). All authors reviewed the manuscript.

Corresponding author

Correspondence to Chunfang Gao.

Ethics declarations

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.

Supplementary Information

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

Wu, L., Wei, J., Zhan, Y. et al. Comparative evaluation of methods for isolating extracellular vesicles from ICC cell culture supernatants: Insights into proteomic and glycomic analysis. Cell Commun Signal 23, 207 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12964-025-02207-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12964-025-02207-x

Keywords

Cell Communication and Signaling

ISSN: 1478-811X