Immunoconjugated Magnetic Graphene for Exosome Capture in SARS-CoV-2 Pseudovirus-Infected Cells

TL;DR

Researchers developed a new graphene-based method to isolate exosomes from cells infected with a SARS-CoV-2 pseudovirus, which could help study virus-host interactions and improve diagnostics.

Contribution

The study introduces MAGU-anti-CD9, a novel magnetic graphene-based tool for selective exosome isolation in viral research.

Findings

MAGU-anti-CD9 efficiently isolated CD9+ exosomes with canonical markers like ALIX, CD147, TSG101, and Flotillin-1.

Exosomes from SARS-CoV-2 pseudovirus-infected cells showed enrichment in CD147, a potential cofactor in viral entry.

The method supports molecular profiling of exosome subpopulations for virus-host interaction studies and diagnostics.

Abstract

Graphene-based nanomaterials exhibit exceptional physicochemical properties that facilitate a range of diverse biomedical applications, including liquid biopsy. In this study, graphene-based magnetic units, termed MAGU (MAGnetic Units), were specifically engineered for the selective isolation of exosomes. Total extracellular vesicles were first enriched using ultracentrifugation, followed by immunomagnetic capture of CD9+ exosomes. MAGU functionalized with anti-CD9 antibody (MAGU-anti-CD9) efficiently recovered a CD9-positive exosome subpopulation expressing canonical markers ALIX, CD147, TSG101, and Flotillin-1, thereby confirming selective isolation performance. To investigate viral associated signaling, 293T cells were transduced with SARS-CoV-2 spike pseudovirus. This pseudovirus was engineered to express the SARS-CoV-2 spike protein, enabling simulation of viral entry and assessment of potential alterations in the exosomal profile induced by viral binding. Exosomes released by pseudovirus-transduced 293T cells were analyzed and compared to those from non-transduced controls. The MAGU-anti-CD9 complex selectively isolated a defined subset of CD9-positive vesicles enriched in the multifunctional transmembrane glycoprotein CD147, which has been proposed as a cofactor in SARS-CoV-2 entry. Comprehensive molecular profiling of selectively captured exosome subpopulations is expected to further support the application of MAGU technology in virus–host interaction research and liquid-biopsy-based diagnostics.

Figures

Full text

1. Introduction

In recent decades, liquid biopsy has emerged as a minimally invasive approach for detecting molecular biomarkers, including extracellular/cell-free nucleic acids, circulating tumor cells, and extracellular vesicles (EVs) such as exosomes, in blood and other biological fluids [1]. This innovative technology has demonstrated significant clinical value in the early detection of a wide range of diseases and in the real-time monitoring of therapeutic response, thereby enabling optimized treatment strategies for patients [2]. Furthermore, unlike conventional invasive biopsies, which often require several weeks for processing and analysis, liquid biopsy offers a rapid diagnostic alternative [3].

Exosomes are pivotal biomarkers in liquid biopsy. They are ubiquitous in biological fluids and play essential roles in intercellular communication, immune regulation, and disease progression [4]. These vesicles contain a diverse cargo of proteins critical for cellular functions, including the evolutionarily conserved Rab proteins, multivesicular body (MVB) synthesis-associated proteins (e.g., Alix, TSG101), heat shock proteins (HSC70, HSC90), and tetraspanins (CD9, CD82, CD81, CD63) [5]. Among these, CD9, CD63, and CD81 are the most widely utilized markers due to their ubiquitous expression on exosomal membranes [6,7].

Viral infections can alter the expression profiles of host cell mRNAs, non-coding RNAs (ncRNAs), thereby modifying exosomal composition [8,9]. In the context of viral pathogenesis, exosomes can act as carriers of viral components, facilitating immune evasion or, conversely, enhancing antiviral immunity [10]. This dual role has been observed in infections caused by HIV, HBV, HCV, and SARS-CoV-2 [11,12]. Specifically, the “Trojan exosome hypothesis” suggests that retroviruses hijack the exosome pathway to package and disseminate viral material independently of traditional receptor-mediated entry [13].

Recent research has highlighted that exosomes are active participants in the pathogenesis and progression of coronavirus infections, including SARS-CoV-2 [14]. Exosomes can promote the dissemination of viral components and modulate innate immune response, contributing to the progression and explaining the clinical severity and diverse symptoms associated with COVID-19 [15,16]. The identification of novel exosome-based viral biomarkers is essential to mitigate the impact of the SARS-CoV-2 pandemic on healthcare services and address the emergence of future viral variants [17].

Current exosome isolation methods include ultracentrifugation, filtration, and immuno-affinity, often used in combination. Circulating exosomes can be released from various cell types, including circulating blood cells and other cells in close contact with the circulation. Thus, blood contains a highly heterogeneous exosome population. While ultracentrifugation remains the most widely used technique, it typically yields a heterogeneous mixture of subpopulations [18]. Therefore, improving isolation techniques is a pivotal step toward identifying specific biomarkers or therapeutic agents based on exosomes, a challenge that may be addressed through nanotechnological approaches [3,19,20,21,22].

Among the various strategies explored for liquid biopsy, magnetic-based technologies are particularly promising. These methods combine precise control over target analytes via an external magnetic field with high-sensitivity detection. The sensitivity of the technique is mainly attributed to the low background noise in biological fluids, which minimizes interference and enhances measurement accuracy [23,24]. Consequently, magnetic units, comprising a magnetic core, a functional surface coating, and specific binding ligands, have been proposed for the capture of diverse biomarkers, including exosomes, in liquid biopsy applications [19,22,25,26].

Drawing on our previous expertise in developing innovative graphene-based nanomaterials for liquid biopsy [22], we designed fine-tuned graphene-based magnetic units to selectively capture circulating exosome subpopulations (Figure 1). The MAGnetic Units (MAGU) consist of a graphene core engineered with a suitable linker to entrap Fe3O4 (magnetite) nanoparticles. Magnetic graphene (MG) was synthesized from graphene oxide (GO) via a facile, highly reproducible one-step chemical coprecipitation method. Subsequently, pyrenacetic acid (linker) was anchored onto the MG surface by exploiting π−π interactions between the aromatic rings of the linker and the sp2 network of graphene [22].

Finally, bioconjugation between the carboxylic functionalities of the MAGU and the primary amines of monoclonal antibodies (Abs) yielded the functionalized MAGU-Ab complex (Figure 1). Anti-CD9 antibody was selected for the selective capture of exosome subtypes in the context of SARS-CoV-2 infection. To achieve this, we simulated the virus–host interaction using a pseudovirus model. SARS-CoV-2-pseudoviruses are engineered to express the spike glycoprotein (S) on the surface and simulate the viral attachment to the host cells. While these pseudoviruses can enter target cells like live viruses, they lack a viral genome and therefore pose no infection risk [27]. Exosomes derived from cells transduced with the SARS-CoV-2-pseudovirus were pre-enriched via ultracentrifugation and subsequently subjected to selective capture using the immune-conjugated magnetic units.

2. Results

2.1. Chemistry of MAGU

Magnetic graphene (MG) was prepared by a slightly modified co-precipitation strategy using iron ions. The Fe^3+^ and Fe^2+^ ions were sequestered by carboxylate anion groups on the GO surface, followed by the slow introduction of ammonia solution to cause the precipitation of Fe_3_O_4_ and simultaneously a partial reduction of GO (Figure 2A) [28]. Subsequently, the surface of MG was derivatized with pyrenacetic acid (PA), which serves as an anchoring site on MAGU for antibody immobilization (Figure 2A).

Specific evidence of graphene surface derivatization was provided by thermal gravimetric analysis (TGA). The GO thermal profile shows two main weight-loss processes: the first from 100° to 200° C due to the CO, CO_2_, and steam release from the most labile functional groups; and the second from 200° to 340° C related to the degradation of the most stable functional groups. The subsequent slow degradation is ascribed to the pyrolysis of the carbon skeleton [29]. In contrast, MG shows much higher thermal stability with a first weight-loss step from 100° to 230° C, followed by a constant degradation up to 800° C. The thermogram of MG is comparable with that of the reduced graphene platform obtained in our previous works [30,31], confirming the success of the co-precipitation strategy in partially restoring the sp2 graphene network. Additionally, the MAGU sample exhibits higher thermal stability between 100 °C and 650 °C, likely due to PA absorption onto the graphene surface via π-π stacking interactions [22] (Figure 2B). MAGU were conjugated to Ab using carbodiimide chemistry. Specifically, EDC in the presence of NHS activated the carboxyl groups on the MAGU surface for subsequent conjugation with the amine functionalities on antibodies (Figure 3A). Therefore, MAGU were weighed, activated, and subsequently incubated overnight at 4 °C with 5 μg of anti-CD9 antibodies under continuous rotation (Figure 3B). The next day, MAGU-anti-CD9 were washed with PBS and analyzed by SDS-PAGE as detailed in Section 4. Free anti-CD9 antibody (5 μg) was used as a control. Coomassie Brilliant Blue staining revealed a clear antibody band in the MAGU-anti-CD9 lane, confirming the successful covalent attachment of the anti-CD9 to MAGU (Figure 3B). Densitometric quantification, using the free anti-CD9 antibody bands as a control, indicated that approximately 1.8 µg of antibodies, corresponding to ~36% of the total 5 µg loaded, was effectively bound to the magnetic carrier (Figure 3C).

2.2. Specific Binding of Immuno-Conjugated MAGU to Exosome Subpopulations

Immunomagnetic approaches employ antibody-coated magnetic particles to selectively isolate exosomes based on the expression of specific surface markers. Accordingly, we evaluated the ability of antibody-functionalized magnetic units to selectively isolate exosomes expressing the surface tetraspanin CD9. As shown in Figure 4A, THP-1 monocytic cells were cultured under hypoxic conditions for 24 h to stimulate exosome release. Exosomes were subsequently isolated from the conditioned medium via ultracentrifugation, yielding a heterogeneous vesicle population. To assess capture specificity and recovery efficiency, isolated exosomes were incubated overnight at 4 °C with MAGU-anti-CD9, MAGU-IgG (non-specific control), and unconjugated MAGU. Following magnetic separation, the particles were washed, lysed, and analyzed by Western blot. The membrane was probed for Flotillin-1, a well-established exosomal marker, to evaluate the exosome capture efficiency. As shown in Figure 4B, an accumulation of Flotillin-1 was observed in the MAGU-anti-CD9 lane (Figure 4B lane 3). No signal was observed in the MAGU-IgG (Figure 4B lane 4) or unconjugated MAGU lanes (Figure 4B lane 5), confirming that the observed exosome binding is antibody-specific and not due to non-specific interactions with the MAGU surface. Collectively, these results demonstrate that MAGU-anti-CD9 selectively recovers a subpopulation of exosomes expressing Flotillin-1.

2.3. Selective Recovery of Exosome Subpopulations by MAGU-Anti-CD9 Following SARS-CoV-2-Pseudovirus-Mediated Signaling

After validating the specificity and efficiency of MAGU-anti-CD9 for isolating exosome subpopulations, we applied this system to examine exosomal changes associated with SARS-CoV-2-related binding signals. To simulate virus–host interactions, we used pseudoviruses expressing the SARS-CoV-2 spike protein. The pseudoviruses were successfully produced and characterized, confirming effective expression and packaging (Figure S1). 293T cells were transduced with the SARS-CoV-2 pseudovirus for 2 h, after which the inoculum was removed and replaced with a growth medium. After 72 h, the supernatant was removed, and the sample was ultracentrifuged to obtain the total exosome population. The resulting total exosome population was incubated overnight at 4 °C with MAGU-anti-CD9 to isolate a CD9^+^ subpopulation. The experimental workflow is schematically illustrated in Figure 5A. Exosomes bound to magnetic units were lysed and analyzed by Western blot. Membranes were probed for a panel of exosome markers: ALIX, CD147, Flotillin-1, CD9, and TSG101 (Figure 5B). Data shown in Figure 5B (lanes 1 and 2) and quantitatively reported in Figure 5C, illustrated the abundance of exosomal markers in whole-cell lysates from untreated and SARS-CoV-2-pseudovirus-exposed cells. A modest but detectable increase in ALIX, CD147, TSG101, Flotillin-1, and CD9 was observed following pseudovirus transduction. This slight induction suggests that signaling pathways triggered by spike-mediated receptor engagement may activate components of the endosomal and vesicle-trafficking machinery. Comparative analysis (Figure 5B, lanes 5–6) demonstrated that MAGU-anti-CD9 platform recovered a subpopulation of exosomes with a distinct protein profile compared to the total EV pool (lanes 3–4). Indeed, as shown in Figure 5 panel D, quantification of marker levels in isolated exosome fractions indicates that the total exosome populations from both mock-transduced and pseudovirus-transduced cells exhibit heterogeneous distributions of canonical markers. In contrast, the MAGU-anti-CD9-captured vesicles display a distinct profile, particularly for CD147. Notably, following SARS-CoV-2 pseudovirus transduction, MAGU-anti-CD9 selectively retrieved a subset of exosomes that were strongly enriched for CD147 (Figure 5D). Densitometric analysis of protein bands was reported in Tables S1 and S2.

Collectively, these data support the notion that SARS-CoV-2-related binding signaling preferentially modulates the release of specific CD9^+^ exosomal subpopulations. Furthermore, the MAGU-anti-CD9 system demonstrates a high selectivity for discriminating and recovering these specific vesicles.

3. Discussion

Immunomagnetic separation is a powerful strategy that utilizes antibodies conjugated to magnetic beads to selectively isolate extracellular vesicles, particularly exosomes, based on specific surface markers [25]. Traditional exosome isolation methods, such as ultracentrifugation, filtration, or precipitation, routinely yield mixed EV populations that often obscure the functional differences associated with discrete surface phenotypes. Immunomagnetic bead-based capture overcomes this limitation by enabling selective enrichment of vesicles expressing specific markers (e.g., CD9, CD81, ALIX), thereby allowing for the isolation of distinct subpopulations [32,33,34]. Earlier immunocapture methods using latex beads have been largely replaced by magnetic systems, which offer superior separation efficiency and reproducibility. Advances in this field have demonstrated the potential of antibody-coated magnetic particles to isolate functionally distinct exosomes derived from diverse biological sources [35]. For instance, CD81^+^ vesicles have been isolated from T lymphocyte cultures [36], Tim4-coated beads can recognize phosphatidylserine [37], and HLA-class II-enriched exosomes can be obtained from antigen-presenting cells [38]. In this pilot study, we extend this concept by employing graphene-based magnetic nanostructures functionalized with antibodies targeting CD9 to achieve selective exosome capture. By conjugating magnetic units with anti-CD9 antibodies, we demonstrated the ability of the MAGU-anti-CD9 system to recover a distinct subpopulation of Flotillin-1^+^ vesicles (Figure 4B). Notably, no signal was detected in the MAGU-IgG or unconjugated MAGU controls. This confirmed the specificity of antibody-mediated capture rather than nonspecific binding to the magnetic units. This selective recovery validates the coupling chemistry and the functional competence of our immunomagnetic platform.

Following the confirmation of the efficiency and specificity of the MAGU-anti-CD9 system, pseudoviruses expressing the SARS-CoV-2 spike protein were used to simulate virus–host interactions. Our findings demonstrate that MAGU-anti-CD9 selectively isolates a defined subset of vesicles enriched in CD147, a multifunctional transmembrane glycoprotein implicated in extracellular matrix remodeling and inflammatory signaling, which has also been proposed as a cofactor in SARS-CoV-2 entry [39].

Additionally, we verified that the same conjugation strategy can be successfully extended to other antibodies. As shown in Figure S2, we confirmed the efficient coupling of anti-5mC antibodies to MAGU using the same experimental protocol. SDS-PAGE analysis revealed the absence of unbound anti-5mC in the eluate (Figure S2, lane 3). Fragmentation of unbound anti-5mC produced bands at approximately 25 kDa and 70 kDa, corresponding to light and heavy chains, respectively, along with a prominent ~100 kDa band representing a heavy-chain–single-chain fragment (Figure S2, lane 4). Consistently, lanes 1 and 2, representing anti-5mC conjugated to MAGU, displayed a 25 kDa light-chain band and a faint ~130 kDa dimerized fragment. These results further support the robustness of our coupling strategy and suggest that the MAGU can be readily adapted to capture additional antibody classes.

In conclusion, this study serves as a proof-of-concept pilot designed to establish the performance parameters of the MAGU platform, including antibody coupling efficiency, target specificity, and exosome recovery yield. The graphene-based magnetic units demonstrated high reproducibility and selectivity, laying the foundation for subsequent optimization.

Future developments may involve designing magnetic-based isolation platforms functionalized with multiplexed antibody arrays, enabling the simultaneous capture of exosome subpopulations expressing distinct surface markers, followed by comprehensive molecular profiling of the isolated vesicle subsets through quantitative proteomic and transcriptomic analyses. These investigations will enable more comprehensive molecular characterization of CD9^+^/CD147-enriched exosome subpopulations. They will also help clarify signaling roles and potential contributions to viral pathogenesis or host defense mechanisms.

4. Materials and Methods

4.1. Chemical Reagents and Antibodies

Graphene oxide was purchased from Graphenea (San Sebastián, Spain). Copper (II) sulfate, Copper (III) chloride, ammonia solution 28–30%, 1-Pyreneacetic acid, (2-(N-morpholino) ethanesulfonic acid) (MES), N-Hydroxysuccinimide (NHS), N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide (EDC), and 3 mM KCl, tablet pH 7.4 at 25 °C were purchased from Merck (Darmstadt, Germany). PBS Dulbecco’s w/o Calcium w/o Magnesium (Code: ECB4004) was purchased from Euroclone (Euroclone S.p.A. Pero, MI, Italy). The antibodies anti-CD9 (ab-236630), anti-TSG-101 (ab83), anti-ALIX (ab-186429), anti-Flotillin (ab41927), anti-5-methylcytosine (5-mC) (ab73938), and anti-rabbit IgG were purchased from Abcam (Cambridge, UK), and anti-BASIGIN (#13287) from Cell Signaling Technology (Beverly, MA, USA). The anti-mouse IgG (AP124P) was purchased from Merck (Darmstadt, Germany).

4.2. Cell Cultures

VERO (African green monkey kidney), 293T (human embryonic kidney), and THP-1 (human acute monocytic leukaemia) cell lines were originally obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). VERO cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Euroclone, Italy) supplemented with 6% fetal bovine serum (FBS). 293T cells were cultured in DMEM supplemented with 10% FBS, and THP-1 cells were cultured in RPMI-1640 medium supplemented with 10% FBS 1mM Sodium Pyruvate, 10 mM Hepes buffer and 4.5 g/L D-glucose (Merck, Darmstadt, Germany). All cell lines were incubated at 37° C with 5% CO_2_. All cell lines were confirmed to be mycoplasma-free via Lonza MycoAlert™ PLUS Mycoplasma Detection kit (Basel, Switzerland).

4.3. Preparation of Graphene Magnetic Unit

Magnetic graphene (MG) was obtained by the addition of Fe2+/Fe3+ iron salt to a dispersion of graphene oxide (GO) in water [28]. Briefly, 75 mL of FeSO4/FeCl3 water solution (3.84 mg/7.44 mg/mL) was added to 30 mL of GO (5 mg/mL) that had been stirred to 80 °C in an oil bath. Then, the temperature was increased to 85 °C, and a solution of ammonia (28–30%) was added dropwise until the pH reached 10. The reaction mixture was stirred for 45 min and then cooled to room temperature. The magnetic graphene was collected using an external magnet, washed twice with water, and dried at 100 °C to yield 365 mg of MG [28].

MAGU were prepared by treating 1 mL of MG water dispersion (10mg/mL) with 5 mL of a methanol solution of 1-Pyreneacetic acid (3 mg/mL) under sonication for 30 min. The reaction mixture was stirred vigorously at room temperature for 24 h, and then the precipitate was recovered using an external magnet. The product was washed twice with water/methanol (9:1), once with water/ethanol (9:1), and finally with water/ethanol (7:3). The MAGU were then dried at 100 °C, yielding 103 mg of product.

4.4. Bioconjugation of MAGU to Antibodies

The bioconjugation of antibodies to Magnetic Units (MAGU) was carried out using a two-step carbodiimide crosslinking strategy to form stable covalent amide bonds between the carboxyl groups on the MAGU surface and the primary amines of the antibody. Briefly, 1 mg of MAGU was washed three times with 0.1 M MES buffer (pH 5.0) and resuspended in a reaction volume of 500 μL of the same buffer. The surface carboxyl groups were activated by adding EDC (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) and NHS (N-hydroxysuccinimide) at a weight ratio of 1:4:6 (MAGU:EDC:NHS [40,41]. The activation was carried out for 15 min at 37 °C under continuous rotation. Following activation, the MAGU-NHS were magnetically separated to remove excess reagents, then immediately mixed with 500 μL of anti-CD9 antibody diluted in 10 mM PBS buffer (pH 7.4). The conjugation was performed using a standardized mass-to-mass ratio of 1 mg MAGU to 5 μg of anti-CD9 antibody (final antibody concentration: 10 μg/mL). and the mixture was incubated overnight at 4 °C with continuous rotation. The resulting MAGU-Ab conjugates were magnetically separated, and the supernatants were collected and lyophilized. The MAGU-Ab units were washed three times with 10 mM PBS buffer (pH 7.4). To verify the success of the immune-coupling of MAGU-anti-CD9 and 5 µg of free CD9 antibody (control), the samples were analyzed by SDS-PAGE followed by Coomassie Brilliant Blue staining. Densitometric analysis was performed with ImageJ 1.54g software by measuring protein band intensities and calculating antibody binding by multiplying the loaded antibody quantity (5 µg) by the intensity ratio.

4.5. Pseudovirus Production

A pseudovirus system was developed to reproduce SARS-CoV-2 infection in a BSL-2 laboratory. The system utilizes a lentiviral vector carrying a luciferase gene reporter, enabling the production of SARS-CoV-2 pseudoviruses and subsequent transduction assay. To generate SARS-CoV-2 pseudoviruses, a three-plasmid co-transfection approach was employed using HEK-293T cells, as previously reported [42]. The following plasmids were used: (i) pcDNA3.1(-) SARS-Swt-C9 for the α-SARS-CoV-2 variant, generously provided by Jean K. Millet; (ii) pCMV-MLVgag-pol encoding the MLV gag-pol proteins; (iii) pTG-Luc encoding the luciferase reporter gene. The methodology used for the pseudovirus production was adapted from a previously reported protocol for producing pseudotyped particles in a BSL-2 setting [43]. HEK-293T cells were cultured to an appropriate confluence and then transfected with Lipofectamine reagent, using the following plasmid quantities: 300 ng of pCMV-MLVgag-pol, 400 ng of pTG-Luc and 300 ng of the spike-encoding vector. We have used a 1 μg DNA:3 μL Lipofectamine 2000 ratio for the transfections as previously reported by Millet and coauthors [43]. In parallel, negative control particles lacking viral envelope glycoproteins (Δenv particles) were produced by replacing the spike-encoding plasmid with an empty vector. As a positive control, pseudovirus bearing the vesicular stomatitis virus G glycoprotein (VSV-Gpp), which is known to efficiently mediate entry into a wide range of cell types, was generated by substituting the spike plasmid with a VSV-G expression vector [42]. Seventy-two hours post-transfection, supernatants containing the assembled pseudoviral particles were harvested and centrifuged at 3200× g for 15 min at 4 °C to remove cellular debris. The cleared supernatant was then collected and subjected to polyethylene glycol (PEG) precipitation (Abcam) by adding 2.5 mL of 5X PEG solution to 10 mL of viral supernatant. This mixture was incubated overnight at 4 °C under continuous rotation. The following day, the samples were centrifuged at 3200× g for 30 min at 4 °C, after which the supernatant was carefully removed by aspiration. The resulting beige or white viral pellet was then resuspended in 100 μL of 1X Virus Resuspension Solution. This concentrated viral stock was used to infect Vero cell monolayers. Production efficiency was evaluated 72 h post-transduction by quantifying luciferase gene expression using the Luciferase Assay System (Promega, Madison, WI, USA) (Figure S1).

4.6. Exosome Purification

Exosome purification was performed by differential ultracentrifugation, which involves a series of centrifugation cycles at different centrifugal forces and durations to isolate exosomes based on their density and size [1]. The initial centrifugation steps include spinning the samples at 300× g for 10 min at 4 °C to remove cells and large debris. Then, the supernatant was transferred to a new tube and centrifuged at 2000× g for 10 min at 4 °C to remove dead cells and larger vesicles. The supernatant was collected and centrifuged at 10,000× g for 30 min at 4 °C. Lastly, to collect the exosome fraction, the final supernatant was then loaded into ultra-clear centrifuge tubes (Beckman Coulter, Inc., Brea, CA, USA) and centrifuged at 100,000× g for 180 min at 4 °C using a Beckman Coulter Optima L-80 equipped with an SW 55 Ti swinging-bucket rotor (K-factor 13,04; Beckman Coulter, Brea, CA, USA). To increase the purity of the preparation and remove co-precipitated proteins, the supernatant was carefully discarded, and the pellet was washed by resuspension in 10 mL of sterile phosphate-buffered saline (PBS, pH 7.4), followed by a second ultracentrifugation step at 100,000× g for 60 min. The final pellet was resuspended in 500 μL of sterile PBS. Isolated exosomes were characterized by Western blot analysis of protein markers commonly found in the extracellular vesicles (i.e., Flotillin, ALIX) in accordance with the MISEV 2023 criteria. To rule out significant cellular contamination, samples were screened for Calnexin, an endoplasmic reticulum-resident protein; the absence of Calnexin in the EV fractions confirmed successful purification (Figure S3).

4.7. Protein Extractions and Immunoblot Analysis

Immunoblot analysis was performed to assess the accumulation of exosome markers. Cells, exosomes and magnetic units were lysed using RIPA buffer (50 mM Tris-HCl, pH 7.4–8.0; 150 mM NaCl; 1% NP-40; 0.5% sodium deoxycholate; and 0.1% SDS) supplemented with 1× protease inhibitor cocktail (Roche) and held at 100 °C for 5 min, to obtain total protein extraction. Total protein concentration in cell lysates was quantified using the Qubit Protein Assay Kit. For cellular lysates, 5 µg of total protein were loaded per lane and band intensities were normalized to an internal loading control (β-actin). For extracellular vesicle analysis, equal volumes of EV lysates were loaded per lane and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), in accordance with the MISEV 2023 guidelines for Western blotting in EV research [44]. Proteins were transferred onto nitrocellulose membranes (Bio-Rad Life Science Research, Hercules, CA) for 1 h at 100 V cost in ice, and the membranes were incubated in blocking buffer consisting of 5% non-fat dry milk in Tris-buffered saline with 0.01% Tween-20 (TBST) for 1 h at room temperature to minimize non-specific binding. Membranes were then incubated overnight at 4 °C with primary antibodies against CD9, TSG-101, ALIX, Flotillin, and BASIGIN (CD147), diluted 1:500 in TBST containing 5% BSA. Following primary incubation, the membranes were washed three times with TBST (TBS + 0.01% Tween-20). Subsequently, membranes were incubated for 1 h at room temperature with HRP-conjugated goat anti-mouse or anti-rabbit IgG secondary antibodies diluted 1:2000 in 5% milk-TBST. After secondary incubation, membranes underwent three 10 min washes in TBST. Protein bands were visualized using chemiluminescent detection according to the manufacturer’s instructions. For cellular lysates, band intensities were normalized to beta-actin as an internal loading control. Densitometric analysis was performed using ImageJ, and results were expressed as fold changes relative to the control samples.

5. Conclusions

This work presents a proof-of-concept pilot study describing the development of graphene-based magnetic units and their functionalization with anti-CD9 antibodies for the selective capture of CD9^+^ exosome subpopulations. The system enables selective isolation of CD9^+^ exosomes released under controlled experimental conditions, including exposure to SARS-CoV-2 spike-pseudotyped particles. The combination of ultracentrifugation and magnetic bead-based capture provides a robust method for isolating a specific exosome subtype. This system enables detailed molecular analysis of exosomes and paves the way for potential diagnostic and therapeutic innovations in viral infections.