ACPA prevents lung fibroblast-to-CAF transformation by reprogramming the tumor microenvironment through NSCLC-derived exosomes.

Introduction
Lung cancer (LC) is the most frequently diagnosed malignancy at a late stage, accounting for 12.4% of all cancer cases1. The exceptionally high global mortality rate results directly from the fact that only 15% of patients are diagnosed early, while approximately 70% are identified at a late metastatic stage2. Non-small cell lung cancer (NSCLC) makes up 87% of all LC cases3. In cases of advanced-stage NSCLC where surgery is considered unsuitable, management strategies4 often encounter significant limitations such as systemic adverse effects, autoimmune-like reactions, or immune-related toxicity5,6. This underscores the need for the development of novel therapeutics that effectively target the tumor microenvironment (TME)7,8, with a tolerable side effect profile at lower therapeutic concentrations.
Cancer-associated fibroblasts (CAFs) are widely distributed stromal cells within the solid TME9, stimulating cancer cell proliferation, metastasis and angiogenesis via secretomes (Interleukin-6/-8/-11 (IL-6/-8/-11)), chemokines and transforming growth factor beta (TGF-β)10, including NSCLC11,12. Unlike healthy fibroblasts, CAFs, highly express alpha smooth muscle actin (α-SMA), phosphorylated Smad2 (pSmad2), fibroblast activation protein (FAP) and podoplanin (PDPN)13, contributing to regulate immunosuppression, inducing cancer proliferation, invasion and migration, remodeling the extracellular matrix, promoting angiogenesis11,14 and inhibiting apoptosis by PI3K/Akt/mTOR and Hedgehog activation15. Moreover, the interaction of CAFs with cancer/cancer stem cells induces tumor progression by regulating cell plasticity and ensures the preservation of the cancer stem cell pool11, thereby develops resistance to chemotherapeutic agents16. Therefore, targeting CAFs is crucial for suppressing progression within the TME.
Cannabinoids (CB) reduce cell proliferation and metastasis17 and induce apoptosis7 in NSCLC via CB1/2 receptors. We further reported a dose- and time-dependent anti-proliferative and apoptotic therapeutic window for synthetic CB1 receptor agonist Arachidonylcyclopropylamide (ACPA) on NSCLC via Akt/PI3K, JNK, glycolysis, citric acid cycle and urea cycle pathways in vitro and in vivo18–20, as well as other CB agonists on solid cancers including endometrial cancer21 and osteosarcoma22. However, targeting tumor cells alone may not be sufficient for the translational therapeutic efficacy of cannabinoid derivatives. CAFs, which facilitate the spread of cancer cells in the TME, are reported to be among the critical treatment targets of NSCLC11. As a potential TME regulator, ACPA can strengthen the therapeutic potential by reducing the tumor-inducing behavior of CAFs.
NSCLC exosomes promote tumor development by inducing the transformation of healthy lung epithelial cells and fibroblasts within the TME23. Cannabidiol-containing exosomes diminished cell viability, led to cell cycle arrest at G0/G1 phase and triggered apoptosis in hepatocellular carcinoma24. Co-application of cannabidiol and temozolomide (TMZ) reduced pro-oncogenic miR-21 and increased anti-oncogenic miR-126 in glioblastoma exosomes25. Preserving healthy fibroblast populations may be a therapeutic target to prevent peritumoral invasion. Therefore, we aimed to protect the healthy fibroblast population in the tumor stroma and prevent their transformation into CAFs by examining the effect and mechanism of action of ACPA on the release profile of NSCLC cell-derived exosomes and the potency of transformation of healthy lung fibroblasts (HLFs) into CAFs for the first time.

Results

NSCLC exosomes exhibited a homogeneous particle size range, optimum polydispersity index and zeta potential values
Control and ACPA-treated NSCLC cell exosomes exhibited protein concentrations of 312.44 and 460.96 µg/ml, respectively, by BCA. Their particle size (PS), polydispersity index (PDI) and zeta potential (ZP) were 250.97 ± 85.45 and 317.33 ± 277.34 nm; 0.50 ± 0.23 and 0.56 ± 0.20; and − 11.66 ± 5.83 and − 23.13 ± 2.06 mV, respectively, by Zeta Sizer. Control exosomes presented CD9, CD63 and CD81 labelling as 96.93%, 98.35% and 77.33% by flow cytometry (FCM), respectively. ACPA-treated NSCLC cell exosomes labelled with CD9, CD63 and CD81 were 95.75%, 95.86% and 70.16%, respectively (Fig. 1c). Quantitative TEM revealed the mean diameter of control A549 cell exosomes as 147 ± 39 nm; and ACPA-treated ones as 132 ± 46 nm (Fig. 1d-f). Control NSCLC cell exosomes presented a mean PS and concentration of 150.0 ± 47.3 nm and 9.69 × 109 ± 1.50 × 108 particles/ml (Fig. 1g, h) by nanoparticle tracking analysis (NTA). ACPA-treated NSCLC cell exosomes exhibited a mean PS and concentration of 268.3 ± 69.8 nm and 7.21 × 1010 ± 2.94 × 1010 particles/ml (Fig. 1i, j) by NTA.

ACPA suppresses oncogenic miRNAs in NSCLC exosomes and reduces fibroblast viability
IC50 ACPA significantly reduced miR-21 (p = 0.007) and miR-23a (p = 0.041) expression in NSCLC exosomes compared to control (Fig. 2a). ACPA-applied NSCLC exosomes at 100 µg/ml significantly decreased HLF viability on day 1 (p < 0.05), whereas 10 µg/ml of exosomes increased fibroblast viability compared to control (Fig. 2b). Similarly, ACPA-applied NSCLC exosomes at 10 and 50 µg/ml diminished fibroblast viability compared to control exosomes at 100 µg/ml, control exo-free medium and A549:LL24 co-culture groups on day 1 by RTCA (p < 0.05, Fig. 2e). ACPA-applied NSCLC exosomes at 10 and 50 µg/ml did not diminish cell viability on days 2–3 (p < 0.05, Fig. 2c-d). ACPA-applied NSCLC exosomes at 10 µg/ml dose reduced cell viability compared to A549:LL24 co-culture groups on days 2 and 3 (p < 0.05); conversely, it lost its anti-proliferative activity on HLFs compared to controls (p > 0.05, Fig. 2e). IC50 of control and ACPA-applied NSCLC exosomes were calculated as 106.8 (R2 = 1; Fig. 2f) and 45.1 (R2 = 1; Fig. 2g) µg/ml, respectively following 12 h of application. Co-culture of A549:HLFs with 5:1 ratio exhibited a superior anti-proliferative effect on HLFs when compared to 1:1 and 1:5 A549:LL24 co-culture groups (Fig. 2h) and the treatment responses of 50% and 80% of fibroblasts were at 25 and 56 h following the initiation of co-culture model (Fig. 2i). The treatment responses of 50% of the fibroblasts reached 28 and 38 h following the initiation of 1:1 and 1:5 co-culture models (Fig. 2i).

Treatment of HLFs with ACPA-administered NSCLC cell-derived exosomes (50 µg/ml) resulted in a marked reduction in mean PDPN immunolabeling, which decreased to 62.36% on day 1 (p = 0.0164, Fig. 3a, d). In contrast, fibroblasts cultured in control medium exhibited 73.04% PDPN labeling. Mean FAP expressions were 96.50 and 82.42% in control exosomes-treated and negative control fibroblasts, and 94.28% in ACPA-treated exosome-applied fibroblasts (p = 0.0208, Fig. 3b, e). Mean α-SMA labeling was 87.93% and 89.43% in fibroblasts treated with control exosomes and exosome-depleted medium, respectively, whereas ACPA-treated exosomes significantly reduced α-SMA expression to 44.76% (p = 0.0013, Fig. 3c, f).

HLFs treated with ACPA-administered NSCLC cell-derived exosomes at 10, 50 and 100 µg/ml exhibited increased IL-6 and IL-8 secretion on days 1–3 compared to control medium (p < 0.05, Fig. 3g, h). Control NSCLC exosomes at 100 µg/ml showed a reduction in IL-6 release from the fibroblasts on day 1 compared to ACPA-administered NSCLC exosomes or control medium (Fig. 3g). ACPA-administered NSCLC cell-derived exosomes at 50 and 100 µg/ml increased IL-6 release in HLFs on days 2–3 (p < 0.05), while they decreased IL-8 release like the control (Fig. 3h).

ACPA-applied NSCLC cell exosomes diminish fibroblast viability through carbohydrate, lipid and amino acid metabolic pathways
PCA graphs of the LL24 fibroblast metabolites following ACPA-applied exosome, control exosome or control medium evidently present significant differences (Fig. 4a, b). ACPA-applied NSCLC cell exosomes at 50 µg/ml inhibited fibroblast growth through suppressing erythrose 4-phosphate, glucose and glucuronic acid levels and increasing mannitol level related to carbohydrate metabolism (Fig. 4c, d, e). ACPA-applied NSCLC exosomes also diminished cholesterol, alanine, glycine, isoleucine and norleucine levels in fibroblasts but promoted malonic acid level (Fig. 4c, d, f, g). Control NSCLC exosomes decreased flavin adenine dinucleotide (FAD) level in fibroblasts (Fig. 4c, d, h). ACPA-applied NSCLC exosomes did not alter the levels of glyceraldehyde, maltose, maltitol, gluconic acid, gluconic acid lactone, N-acetyl-D-glucosamine and glucosaminic acid related to carbohydrate metabolism (Fig. 4e); arachidic acid, heptadecanoic acid, glycerol 1-phosphate, beta-glycerolphosphate regarding the lipid metabolism (Fig. 4f); glutamic acid and hypotaurine related to amino acid metabolism (Fig. 4g); guanosine-5’-monophosphate regarding nucleotide metabolism (Fig. 4h) and citric acid (crebs cycle), glycolic acid (cori cycle) and lactic acid (glyoxylate pathway) (Figs. 4i and 5).

Discussion
In this study, we demonstrated that effectively isolated ACPA-treated NSCLC exosomes expressed high levels of CD9, CD63, and CD81. ACPA suppressed oncogenic miR-21 and miR-23 in NSCLC cell exosomes. ACPA-treated NSCLC exosomes decreased HLF viability within the first 12 h, reduced expressions of PDPN, FAP, and α-SMA, and ultimately lowered the transformation potential into CAFs by inhibiting the pentose phosphate pathway and lipid and amino acid metabolism. ACPA-administered NSCLC exosomes did not alter IL-6 and IL-8 secretion profiles of HLFs between 24 and 72 h. Our group has previously shown the dose- and time-dependent anti-tumor effect of ACPA on NSCLC in vitro and in vivo (Turkish Patent and Trademark Office App.No: TR2019/12451; Patent Cooperation Treaty App.No: PCT/TR2020/050618)18–20,26. Building on this, HLFs treated with NSCLC exosomes underwent transformation and highly expressed CAF markers. Conversely, ACPA suppressed the capacity of HLFs to transform into CAFs at a notably low dose (1 pM), in this study.
IC50 ACPA-treated NSCLC exosomes were effectively isolated by ultracentifugation. Control and ACPA-treated NSCLC cell exosomes exhibited PS as 250.967 ± 85.451 and 317.333 ± 277.338 nm. The PDI values were 0.498 ± 0.232 and 0.563 ± 0.205, and the ZP values were − 11.663 ± 5.830 and − 23.133 ± 2.060 mV, respectively. As ultracentrifugation is regarded as the gold standard method for exosome isolation27, it was employed in this study to obtain large-scale exosomes. NSCLC exosomes had a PDI value of 0.4–0.5, an average PS of 110–140 nm, and a ZP of -10 mV28. When Docetaxel was loaded into NSCLC exosomes, PDI, PS, and ZP values varied as 0.5–0.6, 120–160 nm and 21.1 mV, respectively. The PDI values of control NSCLC exosomes in the present study are consistent with the characterization data reported for the control NSCLC exosomes in the literature. Herein, ACPA was not loaded into NSCLC cell exosomes and applied directly to NSCLC cells. However, its exposure to parental NSCLC cells might have modulated intracellular pathways regarding vesicle formation and secretion, leading to alterations in exosome characteristics. Alternatively, the tumor cells are known to metabolize anti-cancer drugs through exosome release29. Once ACPA is applied to NSCLC cells, it could either bind to the CB1 receptor or be internalized through endocytosis and may subsequently be exported within exosomes regarding cell-based loading30 and released into the extracellular space via exocytosis. The difference in PS between the control and ACPA groups could be attributed to ACPA being encapsulated in NSCLC cell exosomes. Control and ACPA-applied A549 cell exosomes expressed CD9, CD63, and CD81 markers as 96.93%, 98.35% and 77.33%; and 95.75%, 95.86% and 70.16%, respectively, in our research. Plasma-derived exosomes exhibited average CD9, CD63, and CD81 expression levels of 35%, 7% and 15%, respectively, by FCM in NSCLC patient plasma samples31. In this study, NSCLC exosomes displayed higher expression of CD9, CD63 and CD81 by FCM compared to the previous research, which ensures the purity of the population. High CD9, CD63, CD81, Hsp90 and/or Tsg101 protein levels were also detected in A549 NSCLC cell exosomes by Western blot32,33. FCM allows direct membrane labeling for triple-positive exosome verification, while Western blot has a total CD9, CD63, and CD81 labeling capability in cell/exosome lysates. Whole characterization data confirm the successful isolation of a highly pure exosome population that ensure the reliability of subsequent experiments.
Herein, IC50 ACPA reduces miR-21 and miR-23a levels by 20- and 10-fold and miR-23b level by 1.1-fold in NSCLC exosomes compared to control, which are known to be elevated in NSCLC and their exosomes compared to healthy lung cells, making them potential prognostic markers34. MiR-23a level in hypoxic LC cells increased 5-fold compared to normoxic cells35. Co-administration of cannabidiol at 1–5 µM and TMZ diminished the miR-21 expression 5- and 1.82-fold and promoted anti-oncogenic miR-126 expression 2.5- and 6-fold in LN18 and CRL-2611 glioblastoma cell-derived exosomes25 which were lower than that of lower IC50 ACPA. These original findings indicate that ACPA is a potent suppressor of the oncogenic properties of miR-21 and miR-23 in NSCLC cell exosomes, which might improve the prognosis of NSCLC.
We demonstrated the effectiveness of ACPA-treated and control NSCLC exosomes at 50 and 100 µg/ml, respectively, using IC50 in HLFs within the first 12 h with the MTT assay and RTCA. Our experimental approach was systematic, starting with determining the optimal therapeutic window via MTT and pinpointing the specific dose and timing for maximum anti-fibrotic effect through real-time proliferation analysis (RTCA), which served as initial viability assessments. We also confirmed that MTT and RTCA results were consistent, showing the best therapeutic response at 12 h- a key time for capturing metabolic changes before full cellular phenotype shifts. Doses of 50 and 100 µg/ml for ACPA-treated and control exosomes were chosen based on RTCA curves and min-max cell indices. The anti-proliferative and CAF activation-inhibiting effects of ACPA-exosomes were strongest within the first 12–24 h, likely involving rapid cargo delivery that triggers cellular shock responses, temporarily suppressing activation pathways. Lower doses may cause mild stress, showing a biphasic, hormetic-like response36. Within this context, low concentrations of ACPA-treated exosomes did not fully inhibit transformation or cause cytotoxicity in fibroblasts, suggesting mild stress37. However, the reduced effect of exosomes at low concentration after the initial 24 h might be due to lysosomal degradation of the internalized exosomes38,39 or possible upregulation of new pro-fibrotic proteins following the initial stress40. Additionally, recipient fibroblasts internalize exosomes through rapid endocytic pathways such as clathrin-mediated endocytosis, caveolin-dependent endocytosis, lipid raft endocytosis, macro-pinocytosis, phagocytosis, or direct membrane fusion41,42. This process results in the rapid release of their cargo, such as reduced pro-oncogenic miRNAs like miR-21/2343, into the cytoplasm, thereby reducing fibroblast activation. The release of pro-oncogenic miRNAs, particularly in exosomes from untreated NSCLC cells (controls), might be sufficient to rapidly trigger the translation of key fibrotic proteins. Here, ACPA exhibited a superior anti-proliferative effect on HLFs compared to co-culture with A549 cells in vitro. Co-culture of NSCLC and HLFs with a ratio of 1:5 presented a better proliferative effect on HLFs when compared to 1:1 and 5:1 A549:LL24 co-culture groups, suggesting the importance of the existence of a greater population of HLFs in the microenvironment. ACPA also diminished PDPN, FAP and α-SMA CAF expressions in the first 24 h compared to control. ACPA when solely applied at lower IC50 (1 pM) previously suppressed NSCLC proliferation and induced apoptosis through CB1 agonism via PI3K/Akt, p38/JNK, Ras/MEK/ERK, glycolysis, citric acid cycle and urea cycle within 24 h in vitro and in vivo18,19. ACPA-treated NSCLC cell exosomes at higher doses could be able to suppress the HLF transformation into CAFs. This might be attributed to the higher CB1 receptor expression on tumor cells that enables direct targeting by ACPA or the internalization of a higher amount of exosomes by fibroblasts for an equal anti-tumoral activity of ACPA. NSCLC-derived exosomes significantly increased the activation of WI-38 and IMR90 fibroblasts and α-SMA level via miR-142-3p compared to control44. NSCLC exosomes significantly increased the proliferation and α-SMA level of HLF-1 at 72 h compared to control by immunofluorescence23. The high PDPN, FAP, and α-SMA markers detected in the control group in our study align with the previous findings23,44. Moreover, the CAF transformation capability was suppressed by the potent effect of ACPA. PDPN is known to be raised in CAFs, triggering tumor cell migration and progression45,46. FAP and SMA tend to serve as cell surface markers in activated myofibroblasts46 and predominantly distinguished close to epithelial tumor cells47 generally correlates with poor survival. We demonstrated, for the first time, that a cannabinoid-mediated effect of NSCLC exosomes act as a barrier to transformation into CAF-like cells.
We demonstrated that ACPA reduced erythrose 4-phosphate, glucose, glucuronic acid, cholesterol, alanine, glycine, isoleucine, norleucine and FAD levels, whereas it promoted mannitol and malonic acid levels in HLFs on day 1. Our metabolomics data highlights the fundamental anti-proliferative pathways of HLFs in the tumor niche. The reduction in erythrose 4-phosphate and glucose might indicate the impaired energy metabolism and anabolic pathways (pentose phosphate pathway) which might make the fibroblasts susceptible to damage48,49. Glucuronic acid has a key role in detoxification and proteoglycan synthesis and its significant decrease in CAFs might impair extracellular matrix maintenance leading to a reduction in the proliferative capacity of the tumor cells50. Moreover, the decrease in cholesterol could affect the cell membrane fluidity and survival51, which might indicate the weakened CAF membrane network. CAFs heavily rely on protein synthesis to reshape TME, contributing to tumor progression52.
Our metabolomics data revealed changes in lipid metabolism and amino acids like glycine, which are key to ferroptosis. Ferroptosis is a controlled cell death process that manages lipid peroxidation and redox balance through amino acid antiporter systems, affecting both cancer and stromal cells53. These metabolic shifts may indicate ferroptotic activation in fibroblasts, likely due to the anti-tumor effects of ACPA-treated NSCLC cell exosomes. Alternatively, the very low amino acid levels in fibroblasts exposed to these exosomes suggest that ACPA may exert anti-proliferative and stress-inducing effects by increasing amino acid breakdown. Decreased FAD levels could also indicate mitochondrial dysfunction and reduced energy production54,55. Conversely, higher mannitol levels after ACPA treatment might show an attempt by fibroblasts to combat oxidative stress, while elevated malonic acid could reflect impaired mitochondrial respiration and increased autophagy, as it inhibits succinate dehydrogenase in the Krebs cycle56. These metabolic changes, compared with untreated controls, occurred alongside early CAF activation markers, highlighting exosomal signalling’s role in metabolic reprogramming and inhibiting fibroblast activation. Overall, ACPA might disrupt key metabolites involved in the pentose phosphate pathway, lipid metabolism, and amino acid pathways.
Herein, ACPA-treated NSCLC exosomes did not alter the IL-6/-8 secretion profile of HLFs at 24–72 h compared to the control. IL-6 is a pleiotropic cytokine that directly influences the induction of the CAF phenotype by promoting α-SMA57, and IL-8 promotes fibroblast activation and angiogenesis58. Given their importance in targeting fibroblasts, IL-6 and -8 are considered promising biomarkers for demonstrating CAF transformation and tumor growth potential. Exosomes derived from Osimertinib-resistant NSCLC cells increased IL-6 and IL-8 mRNA levels in MRC-5 HLFs at 48 h compared to the control59. Since IL-6 and IL-8 are vital in acute-phase responses, their levels during early inflammatory stages should be analysed at multiple time points. Conversely, the IL-6 and IL-8 secretion profiles demonstrate transient increases at specific intervals, despite the overall inhibitory effects of ACPA-treated NSCLC cell exosomes on fibroblast activation. The rise in IL-6/IL-8 levels caused by ACPA-treated NSCLC cell exosomes after 24 h may reflect a temporary stress response or an inflammatory reaction following exosome internalization, rather than a tumor-promoting effect, as these cytokines also contribute to remodelling processes60. Therefore, the dual functions of these cytokines could be activated concurrently with the anti-fibrotic effect60,61.
Within the scope of this preliminary study, the profiles of miR-21 and miR-23 in control and ACPA-treated NSCLC cell exosomes, and their influence on fibroblasts, were evaluated solely. However, a functional analysis of these miRNAs in reducing fibroblast proliferation and suppressing secretion profiles remains unexplored and represents a significant limitation of the study. Our study reveals a highly specific correlation that functionally aligns with established literature showing miRNAs whose high expression in lung tumor cell exosomes is already documented62,63. We showed that ACPA selectively inhibits the protumorigenic effects mediated by CAFs via NSCLC cell exosomes by reducing the exosomal levels of miR-21 and miR-23 within the TME, which are two of the most well-documented pro-tumorigenic miRNAs. Moreover, the subsequent reversal of CAF markers (α-SMA, FAP) in the fibroblasts provides strong correlative functional evidence that these two miRNAs as potent regulators of CAF activation64,65, clearly illustrating how the mechanism operates at this stage of the research. Consequently, these findings persist in adding to the literature, providing innovative insights for future research66,67. This study lacks direct exosome uptake by lung fibroblasts, which may prevent a definitive distinction between differential uptake kinetics and a true biological response. However, our primary aim was initially to determine the optimal therapeutic concentration that maximally suppresses fibroblast activation without significant cytotoxicity. The slight viability increase at 10 µg/ml, which suggests a genuine hormetic response, was observed but not further examined mechanistically. Additionally, a low dose of exosomes may not be enough to fully inhibit, instead serving as a mild stressor to the fibroblasts37. Therefore, we interpret the data for the 10 µg/ml dose as a typical response of fibroblasts to a lower exosome concentration. Our subsequent metabolomic analysis was intentionally limited to one highly effective dose compared to untreated control. This choice aimed to clearly define the metabolic signature linked to the maximum therapeutic effect of ACPA, focusing the mechanistic discussion on the most relevant anti-tumor action. Herein, we did not assess additional cytokines such as IL-33, TGF-β, stromal cell derived factor-1 (SDF-1), and CAF-derived cardiotrophin-like cytokine factor 1 (CLCF1)68. Our analysis of IL‑6, IL‑8, and classical CAF markers including α‑SMA, FAP, and PDPN still provides robust evidence for NSCLC exosome-mediated fibroblast activation, supporting the validity of our findings. Our experimental design did not involve direct loading of ACPA into exosomes, however, future proteomic or mass spectrometry-based analyses would be valuable to determine whether ACPA is physically associated with exosomes or acts indirectly by modulating exosome-producing cells. The observed effects on recipient cells still support the functional relevance of exosome-mediated signaling. The output is limited to in vitro conditions. These findings, therefore, require further validation through additional functional ex vivo and in vivo studies. However, a high-impact molecular dataset with a large sample size and a normal distribution outline for the first time, demonstrating the potency of CB agonistic ACPA targeting the TME by CAFs.
Cannabinoids are emerging as promising candidates in oncology, extending their established role in preventing nausea69. Various ongoing and completed clinical trials are exploring the benefits of cannabinoids in managing cancer-related symptoms for patients with NSCLC (ClinicalTrials.gov IDs: NCT06418204, NCT02675842, NCT02802540, and NCT04155008). By modulating the TME through CAF interactions, cannabinoids provide a novel approach to hinder tumor progression. Therefore, this research highlights the realization of the anti-tumor effects of ACPA on CAF transformation, paving the way for their inclusion in clinical phase trials.
In conclusion, ACPA emerges as a potent therapeutic due to its ability to target both the tumor and accessory cells. Furthermore, it could suppress the protumorigenic effects mediated by CAFs via NSCLC cell exosomes, acting as prospective carriers, within the TME. Most CAF-targeted cancer therapies directly act on fibroblastic but tumor cell behaviors (4 ongoing/completed clinical trials; ClinicalTrials.gov IDs NCT06107608, NCT06024538, NCT01976741 and NCT01593995). This brings up the possibility of ACPA as a potential chemotherapeutic not only in NSCLC but also in other solid tumors. As such, ACPA is a promising candidate that warrants further investigation in vivo or ex vivo models, and, subsequently, clinical trials for its potency in modulating tumor cells along with the stromal cells and reprogramming the TME of NSCLC and other respiratory system and solid tumors.

Materials and methods

Cell culture and drug application
A549 non-small cell lung adenocarcinoma cell line A549 (CCL- 185) and LL24 healthy lung fibroblasts (CCL-151) (all from ATCC, USA) were cultured with DMEM High Glucose and Ham’s F12K, respectively, supplemented with 10–15% fetal bovine serum (FBS, Capricorn, USA), or 2% L-glutamine (Capricorn, USA) and 1% penicillin-streptomycin (Capricorn, USA). Cells were cultured at 37 °C in an incubator containing 5% CO2.
IC50 ACPA (1.39 × 10− 12 M)18 in media containing exosome-depleted FBS was applied to A549 cells for 24 h as previously determined.

Exosome purification and characterization
Supernatants obtained from drug-free and drug-treated A549 cells were centrifuged at 10 000 x g for 10 min to remove cells and debris. After centrifugation, the supernatants were filtered through 0.22 μm pore diameter filters, followed by ultracentrifugation at 220 000 x g for 90 min at + 4 °C. After repeating the ultracentrifugation step twice, the resulting pellet was dissolved in PBS and stored in -80 °C refrigerator. A BCA assay kit (#23225, Thermo Scientific, USA) assessed the total protein concentration according to the manufacturer’s instructions. Malvern Nano ZS (Malvern Instruments, Ltd., Worcestershire, UK) device measured the PS, PDI and ZP; NTA (Nanosight NS300, Malvern Analytical, UK) detected the particle numbers and particle size dispersity27. Exosomes were diluted to 1:100 and 1:1000 and measured at 23.6 °C with a detection threshold of 20 (n = 3).
The presence of surface markers CD9, CD63, and CD81 was determined by FCM for further characterization. Briefly, diluted and centrifuged aldehyde/sulfate latex beads (#2198595, Invitrogen, USA) in PBS were incubated at room temperature on a shaker for 30 min, followed by the addition of unconjugated anti-CD9 (#312102, Biolegend) and incubation overnight at 4 °C on a shaker. The samples were centrifuged at 12 000 × g for 10 min, and the pellet was resuspended in 5% BSA and incubated for 2 h. 5 µg of exosome sample was added to the mixture and incubated for an additional 2 h on a shaker. After washing, anti-human CD9-APC (#E-AB-F1086E, Elabscience, USA), anti-human CD63-FITC (#353006, Biolegend, USA), and anti-human CD81-PE (#349506, Biolegend, USA) antibodies were added at a 1:100 dilution and incubated for 2 h at 4 °C on a shaker. The samples were then washed by centrifugation at 12 000 × g for 10 min at 4 °C. Finally, the samples were resuspended in 100 µl of PBS and analyzed using a NovoCyte flow cytometer (ACEA, USA)27,70.
For the ultrastructural characterization, 5 µl of the exosome samples were dropped onto 100–200 mesh formwar-coated copper grids and kept for 20 min for the exosomes to settle on the grid. The grids were then kept in phosphotungustic acid (PTA, Sigma, Germany) for 10 min and in uranyl acetate (UA, Sigma, Germany) for 15 min and washed with distilled water. The grids were dried by removing excess water from the edges of the grid with filter paper. The exosome diameters were measured under a transmission electron microscope (Jeol, JEM1400, Japan). All characterization studies were performed trice.

Quantitative Real-Time polymerase chain reaction (qRT-PCR)
Gene expression analysis was performed to detect miR-21 and miR-23 expression in exosomes obtained from control and ACPA-treated A549 NSCLC cells as previously done71. Briefly, the cell-trizol mixture was centrifuged at 12 000 x g at 4 °C for 10 min. After adding chloroform to the samples, they were centrifuged for 15 min. The mixture was passed through the kit columns (#305 − 101, GeneAll, South Korea) according to manufacturer’s instructions, the purified miRNA samples were reconstituted in nuclease-free water and purities and concentrations were determined by NanoDrop spectrophotometer (NanoDrop 1000, ThermoScientific, Massachusetts, USA) at 260–280 nm wavelength. Following cDNA synthesis by using stem-loop reverse transcriptase primers with a reaction series of incubation at 25 °C for 10 min, 37 °C for 120 min, 85 °C for 5 min and 4 °C for 1 h, the cDNAs were mixed with iTAG Universal Sybr Green master mix (#17225124, Biorad), forward and reverse primers, and RNase-free water, followed by qRT-PCR in the CFX Connect instrument (Biorad, California, USA). U6 was used as a housekeeping gene. The Cq values ​​of miRNAs in the ACPA-applied group were compared with those of the control (n = 4). The experimental results were calculated with the “2−ΔΔCT”71.

Cell viability by MTT assay
ACPA-treated or control A549 NSCLC exosomes were diluted to 0, 10, 50 and 100 µg/ml in LL24 fibroblast media containing exosome-depleted FBS72,73. LL24 cells were seeded in 96-well plates at 1 × 105 cells/well, followed by exosome application (n = 5). Absorbances at 24, 48 and 72 h were measured at 590 nm wavelength by VersaMax Microplate Reader (Molecular Device, USA) after adding 0.5 mg/ml MTT solution and incubation for 2 h.

Real-Time proliferation analysis (RTCA)
IC50 of ACPA-treated or control A549 NSCLC exosomes were assessed by xCELLigence (ACEA, Roche Applied System)18. LL24 cells were seeded at a concentration of 5 × 103 cells/well in 96-well plates coated with gold microelectrodes enabling the recording of impedance as a “cell index” and once the cell index exceeds 1.0, 10, 50 and 100 µg/ml ACPA-treated or control A549 NSCLC exosomes were applied to cells every 24 h for 3 days (n = 3) and cell indices were compared with control medium- and medium including exosome-depleted FBS-applied cells (n = 6). To fully elucidate the effect of A549 cells on LL24 healthy fibroblasts, A549 cells were seeded on inserts at a ratio of 1:1, 1:5 or 5:1 in 96-well plates (n = 6). ACPA at a dose range of 10− 9-10− 12 M18 were also applied to LL24 cells to observe the sole effect of ACPA on healthy fibroblasts (n = 6).

Flow cytometry
PDPN, FAP and α-SMA cell surface markers were evaluated for CAF activation in exosome-treated LL24 cells by FCM74. 3-5 × 105 cells were fixed with 3% paraformaldehyde (PFA) and incubated for 15 min at + 4 °C, then washed twice with PBS and centrifuged at + 4 °C for 5 min at 500 x g. Following the incubation with anti-human FAP (#BMS168, Invitrogen), anti-human α-SMA (#ab150301, Abcam) and anti-human podoplanin (#ab10288, Abcam)75,76 primary antibodies in 1% bovine serum albumin (BSA) and 0.1% sodium azide/PBS for 1 h at + 4 °C, the cells were washed and incubated with IgG secondary antibody (#ab7086, Abcam) in the dark for 30 min. After washing, the resulting cell pellet was resuspended in PBS and analyzed on a flow cytometry device (Novocyte 2000R Flow Cytometer System, Agilent, USA) (n = 3).

ELISA
The IL-6 (#D6050, R&D Systems) and IL-8 (#D8000C, R&D Systems) secretome profile of NSCLC exosome-treated CAFs was evaluated by ELISA. The supernatants were collected, centrifuged at 12,000 rpm for 5 min and the labeling was carried out according to the instructions. The absorbances were measured at 450 nm wavelength in a microplate reader (n = 3).

Metabolomics analysis
Metabolomic profiling of 100 µg/ml control and 50 µg/ml ACPA-applied A549 cell exosomes-administered LL24 cells for 12 h was conducted as previously reported18,77. Briefly, following the administration, LL24 cells were extracted using methanol: water mixture (9:1, v/v) and analyzed using total ion count (TIC) normalization by GC-MS after methoxyamination and derivatization (n = 3).

Statistical analysis
All data followed normal distribution per Shapiro-Wilk test. Independent t-test compared control and ACPA-treated NSCLC exosome miRNA levels by qRT-PCR. MTT, proliferation, metabolomics and ELISA data were analyzed with one-way ANOVA and post-hoc Duncan’s test. Results are shown as mean ± SD, evaluated within 95% CI. All statistical analyses were performed using SPSS v25 and Graphpad Prism 8.