AS was synthesized, purified, and structurally characterized by modern spectroscopic techniques, according to a previously reported procedure [27].

All solvents were dried and purified according to standard procedures prior to use. Melting points were determined with a Buchi SMP-20 apparatus (Buchi, Flawil, Switzerland) and are uncorrected. When required, reactions were carried out under an inert atmosphere (dry Ar) in preflamed glassware. Anhydrous Na₂SO₄ was used for drying solutions, and the solvents were then routinely removed at ca. 40°C under reduced pressure using a rotary vacuum evaporator. All reagents employed in the present work were commercially available and used without further purification. Flash column chromatography (FCC) was performed on silica gel (70–230 and 230–400 mesh, Merck, Darmstadt, Germany) and analytical thin layer chromatography (TLC) on silica gel 60-F₂₅₄ precoated aluminum foils (0.2 mm film, Merck, KGaA, Darmstadt, Germany). Spots on the TLC plates were visualized with UV light at 254 nm and using p-anisaldehyde or charring solution. ¹H NMR spectra were recorded in CDCl₃ at 600.13 MHz and ¹³C spectra at 150.9 MHz on a Bruker AVANCEIII HD spectrometer. Chemical shifts (δ) are referenced with respect to the residual CHCl₃ proton of the solvent CDCl₃ at δ = 7.26 ppm. ¹³C NMR spectra were fully decoupled and are referenced to the middle peak of the solvent CDCl₃ at δ = 77.0 ppm. Splitting patterns are designated as: s, singlet; d, doublet; t, triplet; q, quadruplet; dd, double doublet, etc. Coupling constants (J) are reported in Hertz. Electrospray Ionization High-Resolution Mass Spectrometry spectra (ESI-HRMS spectra) were recorded at 30 eV on a HESI-ORBITRAP MS, EXPLORIS 120 (Thermo Fischer Scientific, BRE725531, Waltham, MA, USA) using MeOH as the solvent.

To an ice-cooled (0°C) suspension of artemisinin (200 mg, 0.708 mmol) in dry MeOH (1.2 mL), NaBH₄ (67 mg, 1.77 mmol) was added portion-wise and the reaction mixture was stirred for 2 h at 0°C. Then, the emulsion was neutralized (pH~6) with addition of a solution 30% CH₃COOH in MeOH, and concentrated under reduced pressure to dryness. The residue was, then extracted twice with EtOAc and the combined organic layers were washed with water and brine, dried over anhydrous Na₂SO₄, filtered and concentrated under vacuum. The white residue was subjected to FCC using PhMe/EtOAc 7:3 as eluting system and the afforded intermediate (dihydroartemisinin) was used directly to the next step. Rf (PhMe/EtOAc 7:3) = 0.28, m.p.: 150–151°C.

A solution of succinic anhydride (97.1 mg, 0.969 mmol) in DCM (591 μL) was treated under Ar with 60 μL Et₃N. To this solution, dihydroartemisinin (197 mg, 0.693 mmol) was added portion-wise, and the resulting mixture was stirred at room temperature for further 6 h. Then, it was quenched with cold aqueous citric acid solution (5%), and the organic layer was washed with water. The aqueous phase was extracted thoroughly with EtOAc, and the combined organic layers were dried over anhydrous Na₂SO₄, filtered and concentrated to dryness. The pure product (253 mg, 93% yield) was afforded in the form of white needles after FCC purification using PhMe/EtOAc 8:2 as eluting system. Rf (PhMe/EtOAc 8:2) = 0.27, m.p.: 131–133°C.

¹H NMR (CDCl₃, 600 MHz) δ 5.78 (dd, J = 9.9, 1.7 Hz, 1 H), 5.43 (d, J = 1.38 Hz, 1 H), 2.76–2.69 (m, 3 H), 2.68–2.62 (m, 1 H), 2.55 (tdd, J = 12.4, 6.0, 3.7 Hz, 1 H), 2.36 (m, 1 H), 2.02 (dt, J = 14.5, 4.1 Hz, 1 H), 1.88 (ddd, J = 14.4, 6.6, 3.5 Hz, 1 H), 1.76 (dt, J = 12.5, 3.7 Hz, 1 H), 1.71 (m, 1 H), 1.61 (dtd, J = 13.9, 4.6, 1.4 Hz, 1 H), 1.48 (ddd, J = 13.5, 11.6, 5.0, 1.8 Hz, 1 H), 1.42 (d, J = 1.73 Hz, 3 H), 1.39–1.24 (m, 3 H), 1.01 (m, 1 H), 0.95 (dd, J = 6.2, 1.5 Hz, 3 H), 0.83 (dd, J = 7.2, 1.6 Hz, 3 H);

¹³C NMR (151 MHz) δ 178.0, 171.3, 104.7, 92.5, 91.7, 80.3, 51.8, 45.5, 37.5, 36.4, 34.3, 32.0, 29.2, 28.9, 26.1, 24.8, 22.2, 20.4, 12.2.

HR-HESI-ORBITRAP-MS (30 eV) _m/z_: [M-H]-Calcd for C19H28O8-383.1706, Found: 383.1708.

MDA-MB-231 (HTB-26, RRID: CVCL_0062), MDA-MB-436 (HTB-130, RRID: CVCL_0623), A549 (CCL-185, RRID: CVCL_0023), H1299 (CRL-5803, RRID: CVCL_0060), and HT-29 (HTB-38, RRID: CVCL_0320) cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) Glutamax (10566016, Thermo Fisher Scientific, Waltham, MA, USA) which contained 10% fetal bovine serum (FBS) (PAN-Biotech, P40-37500, Aidenbach, Germany), and 50 U/mL penicillin/50 μg/mL streptomycin (15140122, Thermo Fisher Scientific). MCF-7 (HTB-22, RRID: CVCL_0031) cells, were cultured as described above with the addition of 0.01 mg/mL human recombinant insulin (12585014, Thermo Fisher Scientific). DMS 454 (95062832, RRID: CVCL_2438), SW-620 (CCL-227, RRID: CVCL_0547) and CTC-MCC-41 (RRID: CVCL_0I26) cells were cultured in Roswell Park Memorial Institute 1640 Medium (RPMI 1640) (21875034, Thermo Fisher Scientific) containing 1% L-Glutamine and supplemented with 10% FBS, with the latter cell line being also supplemented with 1% insulin, transferrin and selenium (51300044, Thermo Fisher Scientific), epidermal growth factor (EGF; 20 ng/mL) (Qkine, QK011-0100, Cambridge, UK) and fibroblast growth factor (FGF; 10 ng/mL) (QK025-0050, Qkine). CTC-MCC-41 cells were grown in ultra-low attachment flasks (Corning, 3814, New York, NY, USA).

BEAS-2B (CRL-9609, RRID: CVCL_0168) were grown in the designated Bronchial Epithelial Cell Growth Basal Medium (Lonza Bioscience, CC-3171, Walkersville, MD, USA) and their supplements/growth factors, without serum and in flasks/plates lacking any substrate coating [28]. Cells were maintained at 37°C in 5% CO₂. Cells were sub-cultivated with 0.25% trypsin-EDTA (25200056, Thermo Fisher Scientific).

Colon cancer cell lines and the colon cancer patient-derived CTC-MCC-41 cell line were kindly provided by Professor Alix-Panabières (University Medical Centre of Montpellier, France). The breast MDA-MB-436 cancer cell line was kindly provided by Dr. Martin (University of Maryland School of Medicine, Baltimore, MD, USA). The DMS 454 cell line was obtained from Sigma-Aldrich (Merck KGaA, St. Louis, MO, USA). All other cancer cell lines were obtained from the American Type Culture Collection (ATCC) (Manassas, VA, USA). ATCC’s curated citations support culture conditions used in this study [29,30,31], and numerous peer-reviewed publications from our group report successful growth and experimental use of these cell lines in DMEM or RPMI 1640 media [22,32,33]. The cell lines were authenticated using short tandem repeat (STR) profiling, and all experiments were performed with mycoplasma-free cells.

To achieve non-adherent conditions, 48-well plates were pre-coated with 10 mg/mL poly(2-hydroxyethyl methacrylate) [polyHEMA] (529257-1G, Sigma-Aldrich) in 95% EtOH, following the manufacturer’s instructions. polyHEMA was left to air-dry overnight in a laminar flow hood and plates were sterilized using the cabinet’s built in Ultraviolet-C (UV-C) germicidal lamp (254 nm) for 20 min before use [5]. In polyHEMA conditions (hydrophilic), cells formed substratum (ECM)-detached 3D spheroids, as previously described for lung, breast, and colon cells [5,34,35]. Spheroid formation was defined by the presence of dense, 3D cellular aggregates with well-defined borders. Integrity and viability were verified by microscopic inspection, where spheroids maintained their spherical morphology without signs of cellular dissociation or debris throughout the experimental period.

Cells were seeded in 48-well plates coated with (3 × 10⁴ cells/well) or without polyHEMA (15 × 10³ cells/well) and were grown for 24 h. Cells were serum-starved for 18 h before the addition of AS (4-oxo-4-[[(1R,4S,5R,8S,9R,10S,12R,13R)-1,5,9-trimethyl-11,14,15,16-tetraoxatetracyclo [10.3.1.04,13.08,13]hexadecan-10-yl]oxy]butanoic acid) and/or 5-fluorouracil (5-FU) (TCI Chemicals, F0151, Tokyo, Japan). Viability was examined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide (MTT) assay, following 48 h (or 24 h when indicated) treatments [36]. Briefly, 1/10 of the medium’s volume of MTT (5 mg/mL in PBS) was added to each well, resulting in a final concentration of 0.5 mg/mL. Plates were incubated for 2 h at 37°C, protected from light. Formazan crystals were solubilized with acidified isopropanol and the absorbance was measured at 550 nm using a Dynex MRX microplate reader (Dynex Technol. Chantilly, VA, USA). Background absorbance was corrected by subtracting the mean optical density of cell-free (culture medium-containing) wells from all experimental readings before calculating viability.

Full concentration-response analyses were performed in representative cell lines of each cancer type, while additional cell lines were evaluated at selected concentrations of AS (10 μM and 100 μM) and 5-FU (10 μM) to allow comparisons across models. DMSO served as the vehicle for both AS and 5-FU, with final concentrations of 0.1% for 100 μM AS and 0.05% for 10 μM 5-FU. The number of replicates and independent experiments conducted is explicitly mentioned in the text, figure legends, and Table S1.

A total of 5 chemotherapy-naive SCLC patients were enrolled in the study. Patients’ median age was 66 years (range 61–72). The inclusion criteria were; confirmed SCLC diagnosis and no prior anticancer treatment. Exclusion criteria included prior treatment for SCLC, presence of another active malignancy, or severe comorbid conditions. This study adheres to the Declaration of Helsinki guidelines, as updated in 2013, and was approved by the Ethics and Scientific Committees of the University General Hospital of Larissa, 41334 Larissa, Greece (32710/3-8-20). All patients gave their informed written consent for having their blood collected and for their clinical follow-up data to be used for research purposes. Peripheral blood (10 mL), from all patients was collected in EDTA K2 tubes. Peripheral blood mononuclear cells (PBMCs) were isolated from SCLC patients’ blood samples using Ficoll-Hypaque (P04-60500, PAN-Biotech) density gradient centrifugation at 360× g for 30 min at room temperature (RT) (NUVE NF 200, Ankara, Turkey). PBMCs were washed twice with PBS and centrifuged at 250× g for 10 min at RT. Aliquots of 500,000 cells/500 μL were centrifuged at 450× g for 2 min at RT on Superfrost glass slides (22-265446, Thermo Fisher Scientific) using a Hettich Rotofix 32 A centrifuge (Hettich GmbH, Tuttlingen, Germany). After immunofluorescence (IF) staining, one slide from each patient was analyzed for the identification of CTCs and evaluation of the examined molecules, by observation with the VyCAP Puncher system version 1.5.1 (VyCAP B.V., Enschede, The Netherlands). CTC enumeration was performed after IF staining on the cytospin preparations.

PBMCs’ fractions (including CTCs) (5 × 10⁶ cells/well) were cultured for 4–5 days, in 6-well plates (SPL Life Sciences, 30006, Pocheon-si, Republic of Korea), in a final volume of 1.5 mL. Cells were cultured in RPMI 1640 (21875034, Thermo Fisher Scientific) containing 1% L-Glutamine and supplemented with 10% FBS (P40-37500, PAN-Biotech), 1% insulin, transferrin and selenium (51300044, Thermo Fisher Scientific), EGF (20 ng/mL) (QK011-0100, QKine) and FGF (10 ng/mL) (QK025-0050, Qkine). The plates were pre-coated with 10 mg/mL polyHEMA as described in [37]. Samples from individual patients were processed independently. Cells were cultured as reported in [37], with medium renewal occurring every two days. Cells capable of surviving under non-adherent conditions, including CTCs, were then characterized with immunofluorescence experiments using the TetherChip platform, a microfluidic device designed to immobilize non-adherent tumor cells on a transparent, cell-repellent surface via lipid-based tethering. This approach permits stable cell retention and compatibility with standard formaldehyde fixation for subsequent high-resolution imaging, as previously reported by Ju et al. (25). Cytokeratin (CK), an epithelial marker, was used to characterize a cell as a CTC. Regarding the (CK/PD-L1/CD45) and (CK/M30/CD45) stainings, the leukocyte marker CD45 was used as a negative marker for CTC selection. In cases where CK expression was low, the cytomorphological criteria, such as high nuclear-cytoplasmic ratio, were also applied for CTCs’ identification, as described in [17]. The identification of CTCs was performed blindly to clinical data. In contrast, most of the PBMCs in this distinct cell culture medium and conditions after five days of culture did not survive.

Before AS treatment, cells were incubated in serum-free medium for 24 h to deplete serum factors. Subsequently, cells were treated with 10 μM AS in serum-free medium for 24 h, while cells treated with 10 μM DMSO (AS carrier) served as controls. This concentration was kept constant across all experimental groups. Following treatment, cells were placed in different wells of the Tetherchip containing 100,000 cells per well and allowed to tether for 45 min. All experiments were performed in triplicate.

IF stainings of CK (1:70 in PBS/1% FBS, 1 h, RT) (Amgen, Thousand Oaks, CA, USA) (RRID: AB_11218704)/PD-L1 (1:100 in PBS/1% FBS, 1 h, RT) (Novus Biologicals, NB300-903, Littleton, CA, USA) (RRID: AB_10986627)/CD45 (1:100 in PBS/1% FBS, 1 h, RT) (Santa Cruz Biotechnology, sc-1178 AF647, Dallas, TX, USA) (RRID: AB_627074), CK/VIM (1:100 in PBS/1% FBS, 1 h, RT) (Santa Cruz Biotechnology, sc-6260) (RRID: AB_628437)/GLU (1:100 in PBS/1% FBS, 1 h, RT) (Abcam, ab48389, Cambridge, MA, USA) (RRID: AB_869990), and CK/M30 (1:100 in PBS/1% FBS, 1 h, RT) (Roche Diagnostics GmbH, 12 140 322 001, Mannheim, Germany) (RRID: AB_1613872)/CD45 were performed according to protocols described in [37], while the CK/CXCR4 (1:100 in PBS/1% FBS, 1 h, RT) (Abcam, ab124824) (RRID: AB_10975635)/JUNB (1:100 in PBS/1% FBS, 1 h, RT) (Santa Cruz Biotechnology, sc-46) (RRID: AB_2130002) staining was conducted as detailed in [17].

IF staining of Wheat germ agglutinin (WGA) (1:100 in PBS/1% FBS, 1 h, RT) (Thermo Fisher Scientific, W11261)/CD45 was also performed to visualize CTCs’ McTNs as in [37]. The current study focused on evaluating AS’s effect on CTC viability and expression of prognostic biomarkers [JUNB (Santa Cruz Biotechnology, sc-46), CXCR4 (Abcam, ab124824), VIM (Santa Cruz Biotechnology, sc-6260), GLU (Abcam, ab48389)].

Regarding all IF staining panels, fixation was carried out using 3.7% formaldehyde in PBS for 10 min, followed by permeabilization with 0.1% Triton-X 100 in PBS for 10 min at RT. Non-specific binding was blocked by incubating the cells in 5% FBS in PBS for 1 h at RT. For the observation of McTNs, WGA conjugated to Alexa Fluor 488 (1:100 dilution in PBS/1% FBS, Thermo Fisher Scientific, W11261, 45 min, RT) was used, as WGA binds to mammalian cell membranes, highlighting the presence of McTNs [38]. After three washes with PBS, CD45 conjugated to Alexa Fluor 647 (sc-1178, Santa Cruz Biotechnology) was added to verify whether cells with McTNs were CTCs. For nuclear staining, Hoechst nuclear stain (Cell Signaling Technology, 4082S, Danvers, MA, USA) (1:2500 dilution in PBS/1% FBS, 15 min, RT) was used. Visualization of the stained cells was performed using the VyCAP Puncher system version 1.5.1 (VyCAP B.V.).

All results are expressed as mean ± SEM unless otherwise specified. For experiments with n = 2, data are presented as the mean of two independent experiments, and error bars represent the range. Cell viability was assessed and reported as viability (%) or decrease in viability (%), normalized to control conditions. Depending on the experiment, the control was either 0.1% DMSO-treated cells (for concentration-response experiments) or untreated cells (for single-dose comparisons). Statistical analyses were performed using GraphPad Prism (version 8, La Jolla, CA, USA). Concentration-response relationships were fitted using nonlinear regression models. IC₅₀ values were converted to pIC₅₀, where values are expressed on a scale that better approximates normality. Paired samples non-parametric tests or paired samples t-tests were applied where appropriate. A p-value of <0.05 was considered statistically significant. For n = 2, due to limited sample size, statistical significance was not calculated; however, the range indicates the consistency of the observed effect, and results were interpreted descriptively.

To examine the effect of AS on the viability of various cell lines from different tissues, cells were treated with a range of concentrations of AS for 48 h in serum-free conditions. Results on viability (%) were normalized against carrier-treated cells (0.1% DMSO corresponding to 100 μM AS; control). IC₅₀ (concentration at which 50% inhibition was observed) and pIC₅₀ (−logIC₅₀) values for the different cell lines were calculated. The MTT assay was used to determine viable cells. AS was found to reduce cell viability in a concentration-dependent way in all cancer cell lines, irrespective of their tissue origin (Fig. 1, Fig. 2 and Fig. 3; closed circles).

Regarding the lung cancer cell line H1299, IC₅₀ (pIC₅₀) values of 9.30 ± 1.74 μM (5.11 ± 0.10, n = 10) were calculated (Fig. 1A; closed circles). In breast cancer cell lines, IC₅₀ (pIC₅₀) values of 19.3 ± 2.34 μM (4.75 ± 0.064, n = 10), 13.6 ± 6.00 μM (5.00 ± 0.20, n = 4), and 34.6 ± 5.00 μM (4.47 ± 0.062, n = 4) were calculated for MDA-MB-231 (Fig. 2A), MDA-MB-436 (Fig. 2B), and MCF-7 (Fig. 2C), respectively. MCF-7 had statistically higher IC₅₀ compared to MDA-MB-231 and MDA-MB-436 (p < 0.01 and p < 0.05, respectively) (Fig. 2A–C; closed circles).

Finally, in colon cancer cell lines, IC₅₀ (pIC₅₀) values of 39.0 ± 17.4 μM (4.58 ± 0.24, n = 4) and 45.9 ± 21.3 μM (4.35 ± 0.21, n = 2) were calculated for HT-29 (Fig. 3A) and SW-620 (Fig. 3B) cells, respectively (Fig. 3; closed circles).

Based on the origin of the cells, lung cancer cells appeared to have lower IC₅₀, followed by breast and colon cancer cells, suggesting an increased sensitivity of these cells, respectively. AS was also found to dose-dependently reduce the viability of non-tumorigenic, normal human bronchial epithelial cells, BEAS-2B (Fig. 4A; closed circles). BEAS-2B cells, however, were less sensitive to AS with an IC₅₀ value of 102 ± 37.8 μM (pIC₅₀ = 4.11 ± 0.16, n = 5) at 48 h.

Selectivity indices were calculated by dividing IC₅₀ values for cytotoxicity against normal BEAS-2B cells with the IC₅₀ values for cytotoxicity against the lung cancer cell line H1299, showing 11-fold lesser toxicity of AS in normal cells compared to H1299 cells.

Overall, irrespective of tissue origin, cells can be ranked in terms of their sensitivity to AS (based on the IC₅₀ values) with H1299 being more sensitive, followed by MDA-MB-436, MDA-MB-231, HT-29, MCF-7, SW-620, and BEAS-2B, as the least sensitive cells to AS. Interestingly, this sensitivity order correlates with the aggressiveness of cancer cell lines, with higher sensitivity observed in more aggressive models such as H1299, MDA-MB-436 and MDA-MB-231, and lower sensitivity in less aggressive, like MCF-7, HT-29, and BEAS-2B. This implies that AS could be more effective in aggressive tumor types.

To examine the effect of AS on the viability of the same cancer cell lines after preventing their adherence, cells were grown in polyHEMA-coated wells and treated with the same range of concentrations for 48 h in serum-free conditions. This setup was chosen to resemble CTCs’ non-adherent conditions in the bloodstream. AS reduced cell viability in a concentration-dependent manner in all cells (Fig. 1, Fig. 2 and Fig. 3; open circles). A rightward shift of the curve was observed in lung H1299 (Fig. 1A) and breast cancer cell lines (Fig. 2A–C), indicative of an increase in IC₅₀ values in the non-adherent set-up.

Regarding the H1299 lung cancer cell line, IC₅₀ (pIC₅₀) values of 27.0 ± 7.86 μM (4.72 ± 0.12, n = 10) were calculated in the non-adherent condition vs. 9.30 ± 1.74 in adherent cells (Fig. 1A). In breast cancer cell lines, IC₅₀ (pIC₅₀) values of 40.0 ± 8.00 μM (4.46 ± 0.074, n = 10), 22.7 ± 19.1 μM (4.69 ± 0.39, n = 2), and 45.3 ± 10.6 μM (4.39 ± 0.12, n = 4) were calculated for MDA-MB-231 (Fig. 2A), MDA-MB-436 (Fig. 2B), and MCF-7 (Fig. 2C), respectively, while the corresponding IC₅₀s as previously described in adherent conditions were 19.3 ± 2.34, 13.6 ± 6.00 and 34.6 ± 5.00, respectively. MCF-7 again had a higher IC₅₀ compared to MDA-MB-231 and MDA-MB-436, but not statistically significant. Especially for H1299 and MDA-MB-231, the increase of IC₅₀ values in the non-adherent set-up was statistically significant (p < 0.05) compared to the values of the adherent cells.

Finally, in colon cancer cell lines, IC₅₀ (pIC₅₀) values of 5.38 ± 3.17 μM (5.29 ± 0.26, n = 2) and 72.3 ± 77.7 μM (4.22 ± 0.52, n = 2) were calculated for HT-29 (Fig. 3A) and SW-620 (Fig. 3B) cells, respectively.

In the BEAS-2B cell line, prevention of attachment decreased IC₅₀ to 6.19 ± 9.14 μM (pIC₅₀ = 5.38 ± 0.82, n = 2) (Fig. 4A), suggesting that normal cells are very sensitive in free-floating conditions and hence more susceptible to AS.

Overall, irrespective of tissue origin, cells can be ranked in terms of their sensitivity (based on the IC₅₀ values) to AS in non-adherent conditions as follows: BEAS-2B, HT-29, H1299, MDA-MB-436, MDA-MB-231, MCF-7, SW-620, as the least sensitive cells to AS.

Subsequently, the effect of AS on adherent and non-adherent cells was compared to a common chemotherapeutic agent, 5-FU. As a more biologically prominent concentration, 10 μM of AS (approximately the average of IC₅₀s for the more AS-sensitive H1299, MDA-MB-231, and MDA-MB-436) was chosen for the remaining experiments. The same concentration was chosen for 5-FU to maintain experimental consistency and more direct comparisons between the two agents. Interestingly, all the adherent and the majority of non-adherent cell lines were less sensitive to 5-FU than AS.

Analytically, the most pronounced difference in toxicity (decrease of viability), following treatment with 10 μM AS, between adherent and non-adherent cells, was observed in H1299 (35%), followed by MDA-MB-231 (34%), MDA-MB-436 (33%), A549 (25%), DMS 454 (22%) HT-29 (21%), and MCF-7 (16%) cells (Fig. 1B–D, Fig. 2D–F, Fig. 3C and Table S1). This order correlates with the cells’ aggressiveness for the respective types of cancer, implying that less aggressive cell lines are not influenced by the adherent or non-adherent status of the cells in response to AS. Indeed, in terms of the lung, H1299 (more aggressive cells), showed the highest difference in viability between adherent vs. non-adherent conditions compared to the lesser effect observed in A549 and DMS 454. Similarly, MCF-7, the most “benign” of breast cancer cells, showed reduced sensitivity to AS and a reduced difference in toxicity compared to the other breast cancer cell lines such as MDA-MB-231 and MDA-MB-436.

Interestingly, the effect of 10 μM AS on normal lung cancer cells, BEAS-2B, was smaller (22 ± 7% inhibition of viability) compared to all cancer cell lines in adherent conditions (Fig. 4B and Table S1). In addition, as was the case with the IC₅₀ data, the effect of AS on non-adherent normal cells was more pronounced probably due to the vulnerability of normal cells when grown without focal adhesion contacts to the substrate or neighboring cells (53 ± 16% reduction of viability, Fig. 4B and Table S1).

Higher concentration of AS (100 μM, 48 h) was very toxic for most adherent cells (80–100%, except for SW-620, with 60% inhibition of viability) and non-adherent (55%–99%) cell lines (Table S1). 100 μM AS on adherent BEAS-2B decreased viability by approximately 66% and by 69% on non-adherent cells (Table S1), agreeing with the results obtained with 10 μM AS treatment.

5-FU (10 μM, 48 h) had less pronounced effect on the viability of H1299, MDA-MB-231, and SW-620 cells in both adherent and non-adherent conditions (Table S1). 5-FU was more effective in reducing viability of A549, DMS 454, MDA-MB-436, MCF-7 and HT-29 adherent cells (Table S1). In these cancer cells, culturing under non-adherent conditions, followed the pattern observed in AS, leading to a smaller inhibition of viability (not statistically significant).

Time dependence (24 h vs. 48 h) of the effect of AS (10 and 100 μM) and 5-FU for some cells is presented in Fig. S1, whereby longer treatment (48 h) had more effect than 24 h and 5-FU was again less effective than AS in the majority of the cancer cell lines.

To investigate the impact of AS on real metastasis-driven CTCs, we evaluated the effect of AS on CTCs from the first patient-derived colon CTC line, CTC-MCC-41 [24]. CTC-MCC-41 were normally grown in suspension, and AS was found to inhibit their viability in a concentration-dependent way with an IC₅₀ of 54.6 ± 23.9 μM (n = 5; Fig. 4C).

Interestingly, the effect of AS on the CTC line was more prominent than on all the other malignant cell lines, both in adherent and non-adherent status. Namely, treatment for 48 h with 10 μM of AS decreased the viability of adherent cells by 68 ± 3%. This effect was more modest in non-adherent cells with 48 ± 3% inhibition of viability, which was still higher compared to all other cancer cell lines under similar conditions (Fig. 4E and Table S1) and more evident after 48 h of treatment (Fig. S1). 100 μM of AS inhibited the viability of floating colon CTC line by 92 ± 3%.

Subsequently, to confirm previous results, this study assessed the effect of AS on patient-derived CTCs. PBMCs from five SCLC patients were isolated using Ficoll density gradient centrifugation, cultured for 4–5 days (Fig. S2A), and analyzed with TetherChip technology. A representative cell culture photo from an SCLC patient, after five days of culture is provided in Fig. S2A. We have previously demonstrated that this protocol enables the evaluation of drug efficacy on patient CTCs, offering valuable insights into anti-metastatic agents that can limit the dissemination of tumor cells [37].

The initial number of CTCs detected after blood collection (500,000 PBMCs per patient in cytospins) was compared to the number of CTCs after cell cultures and analysis in TetherChips. The detection of CTCs after cell culture was significantly increased by 800% compared to cytospins (p = 0.02; Fig. 5A). Consequently, this protocol provided high enrichment of CTCs per patient, offering adequate CTC numbers to evaluate drug efficacy. The number of isolated CTCs per patient is shown in Table S2.

McTN formation, which facilitates metastatic dissemination, was assessed in CTCs from SCLC patients using WGA staining before and after AS treatment (Fig. 5B,C). Particularly, IF staining (WGA/CD45) was performed on TetherChips to analyze the presence of McTNs in CTCs. CD45 was added to verify that cells with McTNs were indeed CTCs, as CD45 is a hematopoietic marker and it is absent in CTCs.

Distinct phenotypes (CK+/PD-L1/CD45− and CK+/CXCR4/JUNB) related to poor prognosis were also evaluated in CTCs before and after AS treatment. Fig. 5D–F shows representative images of different phenotypes after culture before AS treatment. The intracellular distribution of these different molecules did not change after AS treatment (data not shown); however, the total number of CTCs expressing these molecules was different.

Furthermore, it has been shown that McTNs are supported by GLU and VIM [39]. Therefore, IF staining (CK/VIM/GLU) in CTCs was also performed before and after AS treatment to further investigate its effects on CTC structural dynamics (Fig. 5F).

In all experiments, control samples with 0.1% DMSO were used for normalization.

Particularly, quantitative analysis of the total detectable CTCs after treatment with AS for 24 h revealed a reduction of 64% compared to the total number of CTCs in the control samples (p = 0.001), demonstrating AS’s cytotoxic effect on patient-derived CTCs (Fig. 6A). The sum of the total CTC numbers from the 4 different IF staining experiments per patient (CK/PD-L1/CD45, CK/JUNB/CXCR4, CK/VIM/GLU, and CK/M30/CD45) is shown in Table S3.

To further evaluate the apoptotic effects of AS on patient-derived CTCs, IF staining with M30 (apoptotic marker)/CD45 antibodies was performed (Fig. S2B,C).

The number of (CK+/M30+/CD45–) apoptotic CTCs was significantly higher in AS-treated samples compared to controls (p = 0.021) (Fig. 6B, Table S4), while (CK+/M30–/CD45–) non-apoptotic CTCs were also reduced after AS treatment, but this difference was not statistically significant (Fig. 6B, Table S4). The increase in apoptotic CTCs and the simultaneous reduction in non-apoptotic CTCs following AS treatment indicate that AS could promote apoptotic cell death in SCLC patient-derived CTCs (Fig. 6B).

However, the total number of CTCs harboring McTNs on their surface before and after AS treatment was reduced but not in a statistically significant manner (Table S5, p = 0.250). In addition, quantitative analysis of CTC phenotypes, based on the previously described IF panels, before and after AS incubation, revealed significant reductions of specific CTC-subclones (Fig. 6C, Table S6). Representative images of the CTC phenotypes depicted in Fig. 6C are presented in Figs. S3–S5, regarding the CK/PD-L1/CD45 (Fig. S3), CK/CXCR4/JUNB (Fig. S4), and CK/VIM/GLU (Fig. S5) stainings, respectively. Particularly, CTCs expressing the (CK+/CXCR4+/JUNB–) and (CK+/VIM+/GLU+) CTCs were significantly reduced compared to control samples (Fig. 6C, Table S6), suggesting that AS targets aggressive CTC subpopulations linked to metastatic progression.