Perturbed CD81 in lung-cancer-derived extracellular vesicles modifies its function in cancer pathophysiology

Abstract

Lung cancer (LC) remains a leading cause of cancer-related fatalities, necessitating an urgent need for potent treatment strategies. This study investigates the role of extracellular vesicle (EV) cargo: CD81 in lung cancer progression and its potential as a therapeutic target. CD81 is a tetraspanin protein that has garnered considerable interest in regulating angiogenesis in various cancer types, including LC. Our study unveiled the presence of elevated levels of CD81 in EVs derived from LC cell lines and patient-derived tumoroids compared to disease-free counterparts, suggesting a potential role in cancer pathogenesis. EVs derived from CD81-silenced LC cells (EV^-siCD81) show an enhanced cellular uptake by the recipient cancer cells. More importantly, treatment with EV^-siCD81 resulted in a significant reduction in colony formation and inhibited the migratory capabilities of LC cells, suggesting therapeutic potential of CD81-bearing EVs. Notably, our study unveiled an abundance of tissue inhibitor of metalloproteinase 2 (TIMP-2) within EV^-siCD81, a key player in inhibiting matrix remodeling and promoting anti-tumoral effects in LC models. In conclusion, this study demonstrates the pivotal role of CD81 in lung cancer progression and highlights the potential of EV^-siCD81 as an innovative therapeutic agent, offering promising prospects for research and lung cancer management.

Graphical abstract

Paramanantham and colleagues illustrated the pathophysiological role of the CD81 protein found in the cargo of extracellular vesicles (EVs) in the progression of lung cancer. EVs derived from lung cancer cells with silenced CD81 (EV^-siCD81) demonstrate an impact on cargo packaging, highlighting their therapeutic potential in cancer treatment.

Introduction

Lung cancer (LC) is the leading cause of cancer-related deaths worldwide, making it imperative to prioritize efforts for prevention, early detection, and effective treatment. With an estimated 2 million new cases and 1.76 (81%) million deaths per year, lung cancer has become a growing concern that requires immediate attention (Globocan 2024).¹ LC is a complex and heterogeneous disease that is histopathologically categorized into two major subtypes: non-small cell lung cancer (NSCLC), which accounts for approximately 70% of LC cases, and small cell lung cancer (SCLC), a more aggressive but less prevalent form. The advent of targetable molecules has transformed medical therapies, particularly in cancer treatment. Targeted therapy offers precise interventions, improving effectiveness and minimizing collateral damage. Intercellular communication is a significant factor in the development of resistance. Consequently, there is an urgent need to explore alternative therapeutic targets, such as novel molecular pathways or genetic alterations.

Extracellular vesicles (EVs; formerly known by the umbrella term of exosomes) have been identified as important mediators of intercellular communication in various physiological and pathological processes. EVs have emerged as key players in cancer development² and progression^2,3 with multifaceted roles in cancer metabolism⁴ and physiology.⁵ They are characterized by their unique membrane structure and enriched protein markers, including CD81, CD63, CD9, and TSG101, and due to their prevalence, they are often used for EV detection. CD81 is a tetraspanin protein that has been implicated in regulating angiogenesis, immune system,^6,7 and development of several types of cancer.^8,9 Functional correlation between CD81 and angiogenesis has been actively explored in many studies on solid tumors.^{9,10,11,12,13} CD81, a tetraspanin protein enriched in extracellular vesicles (EVs), has been implicated in the regulation of angiogenic processes through its role in modulating vascular endothelial growth factor (VEGF) receptor signaling and exosomal cargo composition. Notably, elevated CD81 expression has been associated with increased microvessel density and poor prognosis in NSCLC, suggesting its potential involvement in tumor-driven angiogenesis.^14,15,16,17 Studies also have shown that tetraspanin families seem to alter the cargo of EVs.¹⁸ However, the specific role of EV-CD81 in cancer pathophysiology remains elusive despite its presence as an established EV marker. Understanding the role of CD81 in cancer EV, tumor initiation, and metastasis is crucial for developing effective therapeutic strategies.

This study investigates the role of CD81 in EVs in lung cancer progression. Elevated CD81 levels were observed in EVs from lung cancer cells and patient-derived tumoroids. Silencing CD81 in cell lines produced EVs with altered protein/molecular cargo, affecting angiogenesis-related genes. These EVs (EV^-siCD81) enhanced uptake, reduced colony formation, suppressed migration, and increased TIMP-2 levels, highlighting their therapeutic potential.

Results

Analysis of the expression of CD81 and other proliferative proteins in cell lysates and their corresponding EVs

Western blot analysis was performed to assess CD81 expression in two lung cancer cell lines and their corresponding EVs, using an equal total EV protein concentration of 30 μg (Figure 1A). CD81 was highly expressed in lung cancer EVs, whereas the expression levels were considerably less in the normal lung fibroblast cell line MRC-9. Densitometry analysis further revealed that H1299 EVs exhibited approximately a 9-fold increase in CD81 protein compared to their corresponding cell lysates; similarly, A549 EVs showed a 6.7-fold increase. However, normal fibroblast MRC9 EVs showed reduced levels of CD81 compared to the MRC9 cells (Figure 1C). A similar pattern was observed in patient-derived tumoroids, where CD81 expression in tumoroid-derived EVs was elevated compared to their matched normal tissues (Figure 1B). This result suggests that the presence of CD81 in cancer-derived EVs influences cancer growth and pathogenesis. To assess the impact of CD81 in two lung cancer cell lines (A549 and H1299), expression of CD81 was silenced (Figure S1) using small interfering RNA (siRNA). The EVs collected from silenced cells will be referred in this manuscript as EV^-siCD81.

Figure 1.

Comparative CD81 expression in patient tumor and normal tissues with corresponding EVs

(A) CD81 and other proliferative protein expression analysis of cell lysates and their respective EVs. (B) CD81 and other proliferative protein expression analysis of patient tumor samples and their respective normal tissues and EVs. (C) The densitometry analysis of EVs CD81 protein was performed by normalizing it with the total protein content and measuring it against their respective cell lysates. (D) The densitometry analysis of EVs CD81 protein involved normalization with total protein and comparison with their respective normal lung tissues. The values expressed as mean ± standard deviation (SD) (n = 3) (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.00001).

Analysis of EV^-siCD81 packaging and characterization of EVs from naive and siRNA-mediated CD81 knockdown in H1299 and A549 cells

EV characterization was performed through nanoparticle tracking analysis (NTA) and TEM imaging. The CD81-silenced EVs (EV^-siCD81) did not exhibit any discernible differences compared to the unmodified EVs. The size remained nearly the same, in the average size range of 150 nm, and there was no significant alteration in the concentration (approx nos). TEM images also revealed an almost spherical bilayer structure in both the EV^-siCD81 and EV^-Control (Figure 2A). EVs derived from CD81-silenced cells (siCD81 EVs) were analyzed for EV-specific markers using western blot. The negative marker GRP94 and the positive marker CD9 were assessed to evaluate EV purity, confirming the successful isolation of pure EV populations (Figure 2B). Next, to assess whether the cargo of EV^-siCD81 had changed, we conducted RT² Profiler PCR Arrays to analyze angiogenesis-related genes, as CD81 has propensity toward angiogenesis in LC.^19,20,21 The results revealed a significant increase in the expression of certain angiogenic-related mRNAs, such as tissue inhibitor of metalloproteinase 2 (TIMP-2) with a fold change of 24,998.1 ± 22 and E-cadherin or cadherin 1 (CDH1) with a fold change of 98,991.11 ± 25 (Figure 2C). TIMP-2 plays a crucial role in promoting an anti-tumoral transcriptional profile in lung cancer, including the upregulation of E-cadherin. CDH1 is involved in mechanisms regulating cell-cell adhesion, epithelial cell mobility, and proliferation, and it serves as a potent suppressor of invasion.²² In addition to TIMP-2 and CDH1, several mRNAs related to angiogenesis, including COL11A1, COL4A2, CTNND1, ITGB2, MMP8, TIMP2, B2M, COL6A2, PECAM1, and SELP, exhibited differential expression in the PCR array analysis. The results from the high-throughput RT² Profiler PCR Array experiment were validated by RT-qPCR analysis for TIMP2 and CDH1 genes using specific primers. The results from RT-qPCR showed similar pattern of expression profile as observed in the RT² Profiler array study, further substantiating the observation that enhanced confidence to our findings (Figure 2C). The alteration in TIMP2 and CDH1, along with the reduction in CD81, suggests a potential direct relationship between CD81, TIMP2, and CDH1.

Figure 2.

Characterization of EVs and changes in cargo packaging of CD81 silenced EV

Characterization of EVs from H1299- and A549-cell-isolated naive and siRNA-mediated CD81 knockdown and analyses of si-CD81 EVs package (A) Nanoparticle analysis of EVs and TEM image of EV shows the bilayer structure of EV. (B) Western blot analysis was performed to evaluate the membrane and intraluminal proteins characteristic of EVs. The markers analyzed included CD63, CD9, and the endoplasmic reticulum marker GRP94. (C) Selected genes from array analysis that were confirmed by PCR with their fold change values. Protein-protein interaction between differentially expressed genes and CD81. (D) String analysis showed the protein-protein interaction (PPI) network of 12 proteins. The lines connecting the proteins depict “known” or “predicted” interactions. The thickness of the line corresponds to the strength of the interaction between the proteins. (E) Confirmation of CD81-protein interaction with TIMP2 through co-immunoprecipitation.

To validate the direct relationship between CD81 and other differentially expressed molecules, they were subjected to protein-protein interaction network analysis using the STRING database.²³ This analysis produced an interconnected protein network with a medium confidence level of 0.04, resulting in a single module. The protein-protein interaction (PPI) network analysis of the differentially expressed proteins revealed a single module comprising 12 proteins, including B2M, SELP, ITGB2, COL4A2, PECAM1, CTNND1, COL11A1, COL6A2, TIMP2, MMP8, and CD81(Figure 2D). We conducted co-immunoprecipitation (Co-IP) experiments to demonstrate the interaction between CD81 and TIMP2, using immunoglobulin G (IgG) as a negative control to ensure specificity. The results of the Co-IP analysis revealed a direct relationship between the levels of CD81 and TIMP2, confirming the interaction between CD81 and TIMP2 (Figure 2E). This result highlights the important connection between the alteration of CD81 and the change in EV cargo.

The malignant phenotype of lung cancer cells is reduced after treatment with EV^-siCD81

EV^-siCD81 caused alternation in the anti-tumor gene landscape. To assess its therapeutic potential, we conducted colony formation and wound healing assays using A549 and H1299 cells against their unmodified counterparts. To evaluate the uptake efficiency of EV^-siCD81 by NSCLC cells, the vesicles were initially labeled with PKH67 dye and administered to the cells at a concentration of approximately 200 μg of EV protein. Following a 24-h incubation period, the fluorescence imaging and quantitation of fluorescence indicated an enhanced uptake of EV^-siCD81 by both cancer cells compared to their unmodified counterparts (EV^-control) (Figures 3A and 3B). H1299 EV^-siCD81 treated showed about a 219% increase in uptake while A549 EV^-siCD81 treated showed an enhancement of 188%. This result suggests that the modification of CD81 did not impair the uptake of EVs by the cells. The 24-h incubation period was selected based on a time-dependent study (Figure S4), which showed that uptake peaked at 24 h. Subsequently, to assess the pathophysiological role of EV^-siCD81, we tested their ability to inhibit lumen formation, colony formation, and wound healing in lung cancer cell lines, H1299 and A549. Treatment with both A549- and H1299-derived EV^-siCD81 showed almost 50% decrease in the branch formation (Figures 3C and 3D). In colony formation assay, the cells showed a notable reduction in survival efficiency after 12 days post EV^-siCD81 treatment. A staggering 27.7% reduction in H1299 cells and 37.7% in A549 cells (Figures 3E and 3F) was noted, suggesting the critical importance of CD81 in cancer-derived EVs needed for the progression of the disease. To corroborate our finding, we further conducted a wound healing assay to assess cell migration ability following treatment with EV^-siCD81. The results indicated a substantial increase in the wound gap area in EV^-siCD81 treated compared to that in EV^-Control treated. This trend was observed in a time-dependent manner, which was about 103%, 197%, and 557% for 0, 24, and 49 h in H1299 and 95%, 116%, and 185% for 0, 24, and 49 h in A549 (Figures 3G, 3H, and 3I). These findings highlight the importance of EV-CD81 in cancer pathogenesis and therefore establish the therapeutic potential of EV^-siCD81 in lung cancer cell lines.

Figure 3.

Treatment of EV^-siCD81 attenuates the malignant phenotype of lung cancer cells

(A) Uptake of PKH67-labeled EVs. (B) Intensity of PKH67 was measured using ImageJ, and the percentage was plotted. (C) The lumen formation by HUVECs with EV^-siCD81 treatments. (D) The comparison of the number. of branches formed by HUVECs between EV-Control and EV^-siCD81. (E) Ability of the single cells to form a colony was greatly reduced by the treatment of EV^-siCD81. (F) The survival efficiency of cells was plotted. (G) Migratory potential lung cancer cells were also reduced by treatment of EV^-siCD81, which was shown using scratch assay. (H and I) The percentage of wound width was plotted in a graph. The values expressed as mean ± standard deviation (SD) (n = 3) (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

Western blot analysis revealed that EV^-siCD81 treated exhibited an increase in TIMP2 protein levels (Figure 4A). TIMP2 is well established to play a crucial role in regulating various cellular processes such as matrix remodeling, cell growth, differentiation, angiogenesis, and apoptosis, both in vitro and in vivo.¹⁴ Moreover, β1 integrin’s primary function is to facilitate the formation of focal adhesions between cancer cells and the extracellular matrix (ECM).¹⁵ This adhesion is pivotal for the survival of cancer cells and is closely linked to their migratory and metastatic capabilities, all of which were dysregulated following EV^-siCD81 treatment. Furthermore, other adhesion-related proteins like MMP2, MMP9, GSK3β, and ERK 1/2 were also diminished with EV^-siCD81 treatment (Figure 4). A similar pattern was observed in tumoroids derived from patients (Figure 4). These results highlight the therapeutic potential of CD81-modified EVs at the molecular level.

Figure 4.

The effect of treatment with EV^-siCD81 on lung cancer cells

(A) Western blot analysis of various metastasis-related proteins in lung cancer cell lines (A549 and H1299) and patient-derived tumoroids (MU381 and MU410). The blots were normalized using TPN.

The suppression of EVs did not alter the therapeutic efficacy of EV^-siCD81

GW4869 is a compound known for inhibiting the ceramide-mediated inward budding of multivesicular bodies, effectively preventing the release of EVs. In our study, we applied 5 μM of GW4869 to both H1299 and A549 cells. Subsequent NTA analysis showed a reduction in EV concentration due to GW4869 treatment, with no notable change in the size of these EVs (Figures 5A and 5B). The GW4869-treated NSCLC cells incapacitated of EV production were administered EV^-siCD81 to demonstrate the exclusive effect EV^-siCD81 in absence of autogenic production EVs by the treated cells. Our results demonstrated that the cell’s ability to proliferate and form colonies was reduced with EV^-siCD81 treatment (Figures 5C and 5D) (Figures S2A and S2D). Likewise, the wound healing assay indicated that the migratory potential of EV^-siCD81-treated cells remained unaffected by the EV inhibitor (Figures 5E and 5F). Collectively, the findings from our experiments strongly suggest that the observed effects following EV^-siCD81 treatment are primarily attributed to si-CD81 EVs themselves, rather than being influenced by additive factors.

Figure 5.

The treatment with EV^-siCD81 attenuates the malignant phenotype of lung cancer cells, both with and without an EV inhibitor, in H1299 and A549 cell lines

Nanoparticle analysis of EVs from (A) H1299 and (B) A549 cells was conducted. The ability of individual cells to form colonies was significantly reduced by EV^-siCD81 treatment in (C) H1299 and (D) A549 cells. Additionally, the migratory potential of lung cancer cells was reduced following EV^-siCD81 treatment, as demonstrated by a scratch assay in (E) H1299 and (F) A549 cells. The values are expressed as mean ± standard deviation (SD) with n = 3 (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

Impact of EV^-siCD81 on mouse lung tumor xenografts

To evaluate the effectiveness of CD81-silenced EVs in subcutaneous lung tumor models, we administered both EV^-siCD81 and control EVs directly via intra-tumoral injection. Figure 6A illustrates the tumor growth patterns in animals bearing subcutaneous A549 tumors over an 18-day period following the initial treatment with either EV^-control or EV^-siCD81. Mice treated with EV^control exhibited uncontrolled tumor growth, whereas those treated with EV^-siCD81 showed significantly reduced tumor volumes (Figure 6A). Tumors in the EV^-siCD81-treated group, harvested on the 18th day after measurement, demonstrated a notable reduction in volume (63% inhibition) with an average size of 93 mm³ (p < 0.05) compared to 362.7 mm³ in the EV^-control group. Western blot and immunohistochemistry (IHC) analyses confirmed decreased expression of metastasis-associated proteins in the EV^-siCD81 group, consistent with our in vitro findings. Notably, EV^-siCD81 treatment led to increased TIMP2 levels and reduced expression of adhesion-related proteins, including MMP2, MMP9, GSK3β, and ERK1/2 (Figures 6B and 6C). Similar results were observed in patient-derived tumoroids.

Figure 6.

The impact of EV^-siCD81 treatment on xenograft mouse models

The impact of EV^-siCD81 treatment on xenograft mouse models was assessed. (A) Tumor volume (%) was reduced following siCD81 treatment. (B) Western blot analysis was performed to examine various metastasis-related proteins in mouse tumors. All blots were normalized using TPN method (C). Immunohistochemistry (IHC) was conducted to evaluate the expression of metastasis-related proteins in the tumors. (D) The relative density of these proteins was quantified using ImageJ. The values are expressed as mean ± standard deviation (SD) with n = 3 (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001,∗∗∗∗p < 0.00001).

Discussion

CD81 is involved in various aspects of lung cancer biology, including angiogenesis, tumor initiation, progression, and immune modulation.^24,25 Research in this area aims to uncover the precise mechanisms by which EV CD81 influences lung cancer and leverage this knowledge for the development of innovative diagnostic and therapeutic strategies. Tetraspanins, including CD9, CD37, CD81, CD82, and CD151, have long been linked to cancer progression. More recently, they have regained attention as potential targets for cancer therapy.^21,26 Among these tetraspanins, CD81, which has been extensively studied, has emerged as particularly influential in malignant cellular transformation.²⁷ Research has shown that the downregulation of CD81 in various cancers is associated with diminished tumor growth, reduced migration, and decreased cell invasion.^7,28,29

Conversely, the overexpression of CD81 in melanoma has been linked to increased metastasis.³⁰ Most of the studies conducted so far have primarily focused on cells and tissues. Considering that CD81 is widely recognized as a key marker for EVs, we became intrigued by the idea of investigating the impact of EV-CD81 in the context of cancer progression.

Considering this, we investigated the CD81 expression in different lung cancer cell lines, as well as in patient-derived tumoroids, across cells, tissues, and EVs. Surprisingly, we observed significantly elevated levels of CD81 in cancer-derived EVs in comparison to their normal counterparts. This captivating finding motivated us to further explore the inhibition of CD81 expression in EVs and examine any alterations in their cargo composition and potential therapeutic applications. There have been suggestions regarding the significance of tetraspanin family members in the formation and cargo recruitment of EVs.¹⁸ However, it remains unclear whether modifications in the tetraspanin expression of EVs would have any impact on their functionality and composition. Given the observed variations in CD81 abundance on EVs that correlated with cellular levels and considering the proposed influence of tetraspanins on EV cargo recruitment, we conducted an analysis of the EV content after the inhibition of CD81, particularly focusing on angiogenic-related genes, using a quantitative RT² Profiler PCR assay.

Through a PCR-based array analysis, we identified an abundance of TIMP2 within the cargo of EVs. TIMPs play a critical role in regulating invasion and angiogenesis by both inhibiting matrix metalloproteinases (MMPs) and directly influencing endothelial cells. Additionally, they contribute to the activation of specific MMPs. Numerous studies have consistently reported elevated expressions of MMPs and TIMPs in NSCLC.³¹ Notably, elevated levels of preoperative serum or plasma MMPs and TIMP-1 have emerged as robust predictive indicators for unfavorable prognoses in colorectal cancer (CRC) patients.^32,33,34 Furthermore, TIMPs have been shown to interact with the tetraspanins CD63.³⁵ Interestingly, when we silenced CD81, we observed a reduction in CD63 levels using ONI, a super-resolution microscope (Figure S3). CD63 itself is part of the tetraspanin web and often co-localizes with CD81.³⁶ Our observation of decreased CD63 levels following CD81 silencing suggests possible disruption of this tetraspanin network, which may alter TIMP1 localization or signaling capacity. While the exact mechanism requires further study, these results imply that CD81 may indirectly influence TIMP activity through modulation of CD63.

Another noteworthy observation was the significantly elevated uptake of EV^-siCD81 compared to unaltered EVs. This was contrary to our initial expectation, as CD81 is a well-known protein associated with cellular uptake processes. Huang et al. postulated that tetraspanins groups, including CD9/CD81/TSPAN2 and CD37/CD82, could have arisen via en bloc duplications.^37,38 As a result, it is conceivable that tetraspanins could compensate for the deficiency of one another. However, the distinct phenotypic outcomes observed upon CD81 knockdown in this study, such as altered cargo profiles and downstream signaling effects, suggest that CD81 performs specific, non-redundant roles, particularly in EV cargo sorting and interaction with target cells.

We also assessed the therapeutic potential of EV^-siCD81 through lumen formation, colony formation, and wound healing assays. The treatment with EV^-siCD81 resulted in a significant reduction in lumen formation colony formation and hindered the ability of cancer cells to migrate, emphasizing the promising therapeutic effects of EV^-siCD81 in lung cancer. Additionally, the downregulation of adhesion-related proteins and signaling molecules like β1 integrin, MMP2, MMP9, GSK3β, and ERK 1/2 further highlights the molecular impact of CD81 modification at the cellular level. The downregulation of these adhesion-related proteins was also observed in patient-derived tumoroids and xenograft mouse models, demonstrating that the effects of EV^-siCD81 are not restricted to cancer cell lines.

Lastly, we explored the potential influence of an EV inhibitor, GW4869, on the therapeutic efficacy of EV^-siCD81. Notably, the presence of GW4869 did not diminish the observed effects of EV^-siCD81 in terms of colony formation and cell migration. This outcome strengthened the notion that the therapeutic benefits observed were primarily attributed to EV^-siCD81 themselves, rather than being influenced by external factors such as EVs produced by the cells. Additionally, the exact mechanisms by which tetraspanins like CD81 influence EV cargo recruitment and functionality remain unclear and warrant further investigation.

In conclusion, our study provides compelling evidence for the significance of CD81 modification in lung cancer progression and its potential as a therapeutic target. The comprehensive analysis presented here sheds light on the intricate relationship between CD81, EV cargo composition, and the therapeutic effects of EV^-siCD81, offering valuable insights for future research and clinical applications.

Materials and methods

Cell culture

Human lung cancer cell lines NCI-A549 and NCI-H1299 and normal lung fibroblast cells (NCI-MRC9) were obtained from the American Type Culture Collection (ATCC). NCI-MRC-9 was cultured in Minimum Essential Medium (MEM), NCI-A549 was cultured in Dulbecco’s modified Eagle medium (DMEM) (Hyclone, Marlborough, MA, USA), and NCI-H1299 was grown in Roswell Park Memorial Institute (RPMI) containing 10% of heat-inactivated (v/v) 10% exosome-free fetal bovine serum (FBS) (Systems biosciences, Cat. No. EXO-FBSHI-250A-1), 1 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (Thermo Fisher Scientific, Cat. No. P4333) at 37°C in a humidified atmosphere of 95% air and 5% CO₂. Cells were frequently³⁹ evaluated for mycoplasma contamination using PCR and particular oligonucleotides. GW4869 (Selleck, Cat. No. S7609) was used to inhibit the production of EVs.

Patient-derived tumoroids

Tumor samples from resected lungs of NSCLC patients (MU381, MU402, MU409, and MU410) were collected from University of Missouri Hospital using approved IRB no 2010166 after obtaining written informed consent. PDTs were generated as described in the manuscript.³⁹ Briefly, once the samples were received, they were washed with 1× PBS (phosphate buffered saline) three times, and then samples were cut into small pieces (1–2 mm³) using sterile instrument, and mixed with 20 mL of digestion medium (DMEM/F12, 0.4% fungizone, 1% antibiotics, 500 μg/mL collagenase I, 25 μg/mL DNase I, 25 μg/mL elastase, 100 μg/mL hyaluronidase) for 1 h at 37°C, 5% CO₂ in the incubator with agitation (200 rpm). After incubation, the suspension was passed through 70-μm cell strainers (Corning, Cat. No. 352350), and the strained cells were centrifuged at 240 × g for 4 min. Then the pellet was resuspended in DMEM/F12 and Nutrient Mixture F-12 (Thermo Fisher Scientific, Cat. No. 12634010) containing bFGF 20 ng/mL (Invitrogen, Cat. No. RP-8627), EGF 50 ng/mL (Invitrogen Cat. No. RP-10927), B-27 Supplement (Invitrogen, Cat. No. 12587010), 1 × ROCK Inhibitor (Y-27632) (Medchem, Cat. No.: HY-10071), and Penicillin-Streptomycin-Amphotericin B Suspension followed by centrifugation at 240 × g for 4 min. Patient-derived cells were suspended in ECM Gel (Sigma, Cat. No. E1270-5ML) 10⁷ cells/mL. The ECM gel was liquefied at 4°C prior to use. The suspension was seeded in the 24 well culture plate and kept at 37°C for gelation. Then, 500 μL of DMEM/F12 was added to each well. The media was changed every 4 days. After tumoroids were formed, they were passaged as follows: the tumoroids with ECM gels were scooped out using 1 mL tip and kept in ice for a few minutes and then mixed with media, followed by centrifugation 800 × g for 4 min. The step was repeated to remove the ECM gel. Prechilled PBS was added and washed to remove any residual gel. The isolated tumoroids were completely washed and centrifuged at 800 × g for 5 min at 4°C. The steps were repeated until the residual gel was entirely removed. Purified tumoroids were reseeded in the ECM gel in a new culture plate.

EV isolation

Cell culture media was used for the isolation of EVs <200 nm released from NSCLC cell lines, while PDT media was used for isolating EVs produced by patient-derived tumoroid cells.⁴⁰ H1299 and A549 cells were seeded at a density of 1×10⁶ cells/dish in T75 flasks (Cellstar Cat. No. F010013) and cultured on a conditioned medium made up of RPMI and DMEM with exosome-free-FBS. After 48 h, the medium was removed from the cells. Prior to the isolation of EVs, the media was clarified by centrifuging at 2,500 × g for 30 min (Sorvall ST4 Plus Centrifuge Series), followed by filtration through 0.22 μm (Thermo Fisher Scientific, MA, USA, Cat. No. SLGVV255F) syringe filters. This procedure was performed to remove any detached cells, large debris, and large particle (microvesicles [also known as ectosomes]) contaminants. The same procedure is followed for PDT culture as well. The supernatant collected is then ultracentrifuged (UC) at 110,000 × g at 4°C for 120 min, with a 70.2 Ti rotor (Beckman Coulter Optima XPN-80) to pellet EVs. The pellets were carefully separated from the supernatant, resuspended in 1× PBS, (Corning, Cat. No. 21-040-CV), and stored at −80°C.

Nanoparticle tracking analysis system

The size distribution and number of EVs isolated through UC were characterized by NTA system-NanoSight NS300 (NanoSight, Malvern, Worcestershire, UK). One microliter of purified small EVs (sEVs) were injected into the system using an automated syringe pump system (Harvard apparatus, Cat. No. 98-4730) with a constant flow rate of 100. The samples were diluted to a final volume of 1 mL in PBS. Optimum measurement concentrations were found by performing preliminary testing on the ideal particle per frame value (20–100 particles/frame). The following adjustments were performed in accordance with the software handbook provided by the manufacturer (NanoSight NS300 User Manual, MAN0541-01-EN-00, 2017). Camera level was increased to a point where all particles were clearly seen, but without going above a particle signal saturation of over 20%. The moving samples in the flow cell were illuminated using 530 nm laser, and the scattering track made by each particle was recorded in a 60-s video using sCMOS camera. A total of three similar runs were performed to obtain statistically significant data. Data acquired were analyzed using NTA 3.2 software to give the final estimate of the particle size and concentration of EVs. The NTA capture and analysis parameters were kept constant for all EV characterization runs.

Transmission electron microscopy

EVs were put onto 300 hex mesh, formvar coated, glow-discharge copper grids using the single drop method. Staining solutions and deionized water were contained in a syringe equipped with a fresh 0.2 mm filter, and then 10 mL of the sample was allowed to settle on the grid for 5 min. The unsettled sample was removed by wicking it with filter paper followed by washing with deionized water for 10 s. The water was removed using filter paper, and 10 mL 2% uranyl acetate was deposited on the grid for 20 s. The grid was allowed to dry in a desiccator and then placed in a grid storage box and was ready for viewing. Grids were viewed on a JEOL JEM-1400 transmission electron microscope at 80 kV equipped with a 2k AMT digital camera (JEOL JEM-1400, “JEOL,” Tokyo, Japan).

PKH67 labeling

Purified EVs derived from H1299 and A549 cells were labeled with a PKH67 green fluorescent labeling kit⁴¹ (Sigma, Cat.No. PKH67GL-1KT). The labeling was done according to the manufacturer protocol that includes briefly, 6 μL of PKH67 were added to 1 mL of Diluent C in microfuge tubes for each EV sample, as well as one tube for a media-only control. The dye/diluent mixture was added to the ultracentrifugation⁴¹ tube for each sample, followed by gentle pipetting for 30 s. To quench the reaction, 2 mL of 10% BSA in PBS were added, and the volume was adjusted to 8.5 mL with serum-free media. A 0.971 M sucrose solution was prepared using a 2.5 M sucrose stock solution, and 1.5 mL of this solution was carefully added to the bottom of the tube, creating a sucrose cushion beneath the sEV/PKH67 solution. After centrifugation at 110,000 × g for 2 h at 4°C, EV was found in the pellet, and most of the excess dye remained in the interface layer. The media and interface layer were aspirated, and the EV pellet was gently resuspended through pipetting.

Silencing CD81 by RNA interference

The siRNAs targeting CD81(Dharmacon, CO, USA, Cat. No. L-017257-00-0020) were obtained from Dharmacon/horizon discovery. siRNA stocks were aliquoted and stored at −20°C and, then before use, resuspended in 1 mL of Dharmacon siRNA buffer. Conditions from the siRNA screen were scaled up to a 12-well plate size to extract protein and/or RNA for siRNA validation (quantifying mRNA knockdown and figuring out the impact of knockdown on transactivation). In a 200 mL amount of RPMI basal media, 25 nM Dharmacon SMARTpool deconvoluted siRNA (individual duplex) were complexed for 20 min at room temperature with a final concentration of 1 mL of DharmaFECT transfection (Dharmacon, CO, USA, Cat. No. T-2005-01) lipid per well. After 4 h of incubation, the medium was replaced with RPMI (complete media). For the protein analysis, the cells were cultured for 48 h.

RT² profiler PCR arrays test

According to the manufacturer’s directions,⁴² total RNA was extracted from the EVs using exoRNeasy Midi Kit (QIAGEN, Cat. no. 77144). A NanoDrop (Thermo Scientific, USA) was used to read RNA quality against the RNA quality control parameters OD_260/280 were between 1.8 and 2.0. Reverse transcription was performed by using the RT² PreAMP cDNA Synthesis Kit (QIAGEN, Cat. no. 330451). The cDNA was used on the real-time RT² Profiler PCR Array (QIAGEN, Cat. no. PAHS-013Z) in combination with RT² SYBR Green qPCR Mastermix (Cat. no. 330,529). The PCR array was performed in triplicate. The delta threshold cycle (Ct) was calculated by subtracting Ct for the average of six housekeeping genes, including β-actin (H1), β-2-macroglobulin (H2), GAPDH (H3), lactate dehydrogenase A (H4), non-POU domain-containing, octamer-binding-like (H5), and Rabbit Genomic DNA Contamination (H6) from Ct for genes of interest. The human profile PCR array profiles the expression of 84 key genes associated with the cell-cell and cell-matrix interaction (Table S1). In brief, a reaction mixture was prepared by successively adding 550 μL of 2× RT² SYBR Green qPCR Mastermix, 102 μL of the diluted first-strand cDNA synthesis reaction, and 448 μL of ddH₂O. The mixture was then administrated to the PCR array, and PCR was performed on thermocycler (QuantStudio 3, Applied biosystems, USA) at 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min. Ct values were derived from an Excel file to build a table of Ct values, which were then uploaded onto the data analysis web portal at http://www.qiagen.com/geneglobe. Ct values were normalized based on automatic selection from full panel of reference.

qPCR validation

Total RNA was isolated from both CD81-silenced and non-silenced EVs using the exoRNeasy Midi Kit⁴³ (Qiagen Cat no. 77144), and RNA quality was assessed via NanoDrop spectrophotometry. Complementary DNA (cDNA) was synthesized from 100 ng of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Cat. No. 4374967). For RT-qPCR analysis, 1 μg of synthesized cDNA (quantified by NanoDrop) was used per reaction, performed on a QuantStudio 3 system with SYBR Green RT-qPCR reagents (Invitrogen). Primer sequences are listed in Table S2. The thermal cycling conditions were as follows: 95°C for 30 s, followed by 40 cycles of 95°C for 20 s, 58.5°C for 20 s, and 72°C for 20 s. Ct values were normalized to human 18S rRNA, and relative mRNA expression was calculated as fold change compared to control, with standard deviation (SD) included. All reactions were performed in triplicate, and each experiment was repeated at least three times for reproducibility, with data subjected to statistical analysis.

Migration assay

The cells were grown in 6-well plates to 100% confluent monolayer and then scratched with 1 mL sterile pipette tip to form a “wound.” After the wound formation, the cells were incubated in serum-free media for 0, 18, and 24 h at 37°C in a CO₂ incubator. The scratch was viewed using a photomicroscope (Olympus, CX 44). In a scratch assay, the distance between wound edges at the assay’s start (0 h) and end (24, 48 h) was measured, and the initial distance was subtracted to determine the migration distance, assessing cell or particle movement. The percentage of the migration distance is calculated and plotted.

Colony formation assay

The cells were seeded in a 6-cm dish with the cell seeding capacity of 500 cells/plate and starved for 12 h in serum-free RPMI media. After 12 h, the serum-free media was discarded, and RPMI with 10% heat-inactivated FBS was added. The cells were incubated for 10 days. The media were replaced every 3 days. After 12 days, the plates were washed with 1× PBS and then the cells were fixed with 4% formaldehyde for 30 min. After fixation, the cells were stained with 0.6% Giemsa stain (Sigma, Cat. No. 32884-250ML) for 30 min. The stain was washed with distilled water and then the pictures were taken using a camera. The colony count was determined⁴⁴ using ImageJ software, and the plating efficiency (PE) was calculated using the Equation 1:

$$PE=\frac{Numberofcolonies}{Numberofcellsplated}\times 100$$ (Equation 1)

Subsequently, the surviving fraction (SF) was calculated by Equation 2:

$$SF=\frac{PEoftreatedsample}{PEofcontrol}\times 100$$ (Equation 2)

Lumen formation assay

Lumen formation was evaluated in 24-well plates coated with growth-factor-reduced Matrigel. HUVECs (2 × 10⁵ cells/well) were seeded onto the Matrigel-coated wells (100 μL/well) in various types of medium (HUVEC), with or without si-CD81 EV treatment. Matrigel organization was observed after 24 h at 37°C, using calcein AM stain (Invitrogen, Cat.No. C1430) and a Keyence microscope (BZ-X810). The total lumen area was quantified as the mean pixel density through ImageJ software. Specifically, the software was used to trace and measure the length of lumen lines created around HUVECs. The results are presented as means ± SD of tube length per field.

Western blot analysis

The cells were lysed using RIPA buffer (Sigma, Cat. No. R0278-500ML) with sonication. EVs were directly employed for estimating protein concentration. Protein levels in cell lysate and EV samples were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Cat. No. 23225). EV and cell lysate samples were heated at 95°C for 10 min with a dye (4 × Laemmli Sample Buffer, BioRad, Cat. No. 1610747). Subsequently, proteins were separated via SDS-PAGE and transferred onto polyvinylidene fluoride (PVDF) membranes from Sigma-Aldrich (Cat. No. IPVH00010). After blocking with a 3% non-fat powdered milk solution in Tris-buffered saline with 0.1% Tween 20 (TBST), they were probed with specified primary antibodies detailed in Table S3. Secondary antibodies, HRP-conjugated (Jackson Immuno Research, Cat. No. 111-035-144 and 115-035-146), were applied. The novel total protein normalization (TPN) method uses Invitrogen No-Stain Protein Labeling (NSPL)⁴⁵ Reagent (Thermo Fisher Scientific, Cat. No. A44449), enhancing the detection sensitivity and reliability of protein quantification in sEVs. Blots were developed using a chemiluminescence kit (BioRad, Irvine, Cat. No. 1705061) and imaged using the Gbox mini system by Syngene in CA, USA. Band intensity was quantified using ImageJ.

Subcutaneous tumor generation

Female nude mice (Nu-Nu), aged 4 to 6 weeks, were obtained from Charles River Laboratories. All experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Missouri School of Medicine Vivarium and conducted in strict accordance with their guidelines. To establish subcutaneous tumors, the mice were anesthetized using 2%–3% isoflurane and then injected with 2.6 × 10⁶ A549 tumor cells in 100 μL of PBS subcutaneously into the lower right flank. Once the tumors reached a size of approximately 50–100 mm³, the mice were randomized for further studies.

Immunohistochemical staining of tumor tissues

Xenograft tumors were paraffin-embedded and immunostained with antibodies against TIMP2 (sc-21735, 1:100; Santa Cruz), integrin β1 (mouse, sc-9970, 1:100; Santa Cruz), ERK 1/2 (sc-514302, 1:100; Santa Cruz), GSK-3 α/β (sc-7291, 1:100; Santa Cruz), MMP9 (sc-393859, 1:100; Santa Cruz), and MMP2 (sc-13594, 1:100; Santa Cruz). The tissue sections underwent deparaffinization, rehydration, heat-mediated antigen retrieval, permeabilization, and blocking. They were then incubated with primary antibodies overnight at 4°C. HRP-conjugated secondary antibodies were applied for 1 h at room temperature. Staining was developed with DAB, followed by counterstaining with hematoxylin and analyzed using ImageJ. The results for each marker in the two treatment groups were analyzed statistically.

Statistical analysis

The results were expressed as means ± SEM from at least three independent experiments. Significant differences were determined by the one-way analysis of variance (ANOVA) with Newman-Keuls post hoc test for comparison of at least five treatment groups and Student’s t test for two groups. Statistical significance was defined as p < 0.05.

Data availability

All data generated or analyzed during this study, if not included in this article and its supplemental information files, are available from the corresponding author on reasonable request.

Acknowledgments

The authors thank Dr. Pawan K Singh for providing access to fluorescen microscopes, Molecular Cytology Core, and to the Electron Microscopy (EM) Core at the University of Missouri, Columbia, Missouri, USA, for their assistance with confocal imaging and EM imaging respectively. This study was supported, in part, by the University of Missouri School of Medicine (UM-SoM) startup funds (A.S.) and the Department of Defense (DoD) through the Lung Cancer Research Program (LCRP) award W81XWH-22-1-0016 (A.S.).

All experiments conducted received approval from the University of Missouri School of Medicine’s Institutional Biosafety Committee (IBC), and Patients samples used in the study were obtained with written consent following the guidelines of the Institutional Review Board (IRB) of the University of Missouri (MU) through an approved protocol (IRB#: 2010166).

Author contributions

A.P. and A.K. designed the study and wrote the manuscript. A.P., S.D., and R.A. performed the experiments and reviewed the manuscript. G.G. and G.M. analyzed the data. Y.M., H.H., and J.K. contributed to the study design and information collection. All authors read and approved the final manuscript.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the author(s) used Grammarly in order to improve the sentence structure and to check grammar. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.