Human chronic myeloid leukemia K562, KU812, and MOLM-6 cell lines were obtained from DSMZ-German Collection of Microorganisms and Cell Cultures GmbH (DSMZ, Braunschweig, Germany) and were maintained in RPMI-1640 medium (Euroclone Ltd., Devon, United Kingdom) supplemented with 10% heat-inactivated fetal calf serum (Euroclone), 2 mM L-glutamine, 100 IU/ml of penicillin, and 100 μg/ml of streptomycin.

The buffy coat, no longer used for transfusion, kindly provided by the Transfusion Center of Macerata Hospital after authorization of the hospital management (Direzione Medica Presidio Ospedaliero Unico AV3, Medical Director Dr. Carlo Di Falco), was only used in vitro to obtain an enrichment of normal myeloid cells using a RosetteSep^TM HLA Myeloid Cell Enrichment Kit (STEMCELL Technologies, Cambridge, United Kingdom).

Capsazepine (CPZ), capsaicin (CPS), imatinib mesylate, ionomycin, propidium iodide (PI), and 2′,7′-dichlorofluorescin diacetate (DCFDA) were purchased from Sigma-Aldrich (Milan, Italy). The Fluo-3 AM calcium indicator and JC-1 were purchased from Thermo Fisher Scientific (Waltham, MA, United States). A784168 was purchased from Bio-Techne S.R.L (Milan, Italy). N-oleoyl-dopamine (OLDA) was purchased from Tocris (Bristol, United Kingdom). CPZ, CPS, and OLDA were dissolved in DMSO and were used as vehicle (maximum percentage 0.05, considered non-toxic) (de Abreu Costa et al., 2017).

Antibodies (Abs) were used according to manufacturer’s instructions: anti-phospho-histone γH2AX (Ser139) (γH2AX), anti-caspase 3, anti-binding immunoglobulin protein (BiP), anti-phospho-ubiquitin (pSer65), anti-microtubule-associated protein-1 light chain 3 (LC3), anti-activating transcription factor 4 (ATF4), anti-autophagy protein 12 (ATG5–ATG12), and anti-GAPDH were purchased from Cell Signaling Technology (1:1000, Danvers, MA, United States). Anti-transient receptor potential channel vanilloid 1 (TRPV1) was purchased from Invitrogen (1:1000, Waltham, Massachusetts, United States). Secondary Abs used were: HRP-conjugated anti-rabbit IgG (1:5000, Jackson ImmunoResearch Europe Ltd., Ely, United Kingdom); HRP-conjugated anti-mouse IgG (1:2000, Cell Signaling Technology); and PE-conjugated goat anti-rabbit Ab (1:40, BD Biosciences, Milan, Italy). The OxyBlot Protein Oxidation Detection Kit was purchased from Merck Life Science (Milan, Italy).

A density of 2 × 10⁵ CML cells/mL was plated in 12-well plates. The next day, cells were exposed to different concentrations of OLDA (0.5–100 μM), CPS (10–300 μM), or the respective vehicle for 24 h. Then, cells were stained with trypan blue and counted using the TC20 automated cell counter according to the instrument instruction (Bio-Rad, Milan, Italy). This automated cell counter uses multifocal plane analysis to assess cell viability. Three replicates were carried out for each treatment. IC₅₀ (half-maximal inhibitory concentration) values, shown as the mean ± standard deviation (SD), were calculated using GraphPad Prism^® 9.1 (GraphPad Software, San Diego, CA, United States). In some experiments, cell counting was performed on siTRPV1 or siGLO CML cells treated with OLDA (IC₅₀) or vehicle. OLDA was used in combination with imatinib mesylate for 24 h. Synergistic activity of OLDA–imatinib combination was determined by isobologram analysis and combination index (CI) methods (CompuSyn Software, ComboSyn, Inc.). The CI was used to express synergism (CI < 1), additivity (CI = 1), or antagonism (CI > 1).

TRPV1 (siTRPV1) and siGLO non-targeting siRNA (used as the control), FlexiTube siRNA, were purchased from Qiagen (Milan, Italy). The day before transfection, CML cells were diluted at the density of 6 × 10⁵ cells/mL in the culture medium. After 24 h, cells were collected, counted, and plated at the density of 4 × 10⁵ cells/mL, and siTRPV1 or siGLO (50 nM) was added according to the HiPerFect Transfection Reagent protocol (Qiagen). Cells were then harvested at 48 h post transfection. Silencing efficiency was evaluated by qRT-PCR and Western blotting. No differences in TRPV1 expression and cell viability were observed in siGLO-transfected cells compared with siGLO-untransfected cells.

Total RNA was extracted using the RNeasy Mini Kit (Qiagen), and cDNA was synthesized using the iScript Advanced cDNA Synthesis Kit (Bio-Rad) according to manufacturers’ protocol. qRT-PCR was performed using QuantiTect Primer Assays for human TRPV1 (QT00046109) and GAPDH (QT00079247), as a reference gene (Qiagen), using the iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad). The PCR parameters were in accordance with the primer datasheet. All samples were assayed in triplicate. Gene expression analysis was performed using iQ5 software.

Intracellular Ca²⁺ influx was measured using Fluo-3 AM and FACS analysis. Briefly, 1.5 × 10⁶ CML cells/mL were first washed in calcium- and magnesium-free PBS supplemented with 4.5 g/L of glucose and then incubated in calcium- and magnesium-free PBS/glucose medium supplemented with 7 μmol/L Fluo-3 AM for 30 min in the dark at 37°C and 5% CO₂. After washing, cells were resuspended in calcium- and magnesium-free PBS/glucose medium containing 2 mmol/L Ca²⁺ and were stimulated with OLDA (IC₅₀ dose) or with vehicle up to 3 min. In some experiments, CML cells, loaded as described previously, were treated with OLDA in combination with CPZ (10 µM) or A784168 (1 µM). Ionomycin (5 μg/ml) treatment was used as a positive control for measuring calcium influx. Fluo-3 AM fluorescence was measured using the BD Accuri C6 Plus flow cytometer and its software (Beckton Dickinson, San Jose, CA, United States).

Lysate from CML cells, treated or not treated with OLDA at the IC₅₀ dose, was extracted using lysis buffer (10 mM Tris; 100 mM NaCl; 1 mM EDTA; 1 mM EGTA; 1 mM NaF; 20 mM Na₄P₂O₇; 2 mM Na₃VO₄; 1% Triton X-100; 10% glycerol; 0.1% SDS; 0.5% deoxycholate; and 1 mM PMSF) containing protease-inhibitor cocktail (Euroclone). Proteins were separated on 8%–14% SDS-polyacrylamide gels and transferred using Bio-Rad systems. Non-specific binding sites were blocked with 5% low-fat dry milk or 5% BSA in PBS containing 0.1% Tween 20 for 1 h at room temperature. Membranes were incubated overnight at 4°C with anti-LC3, anti-ATG12, anti-ATF4, anti-BiP, anti-γH2AX, anti-caspase 3, anti-phospho-ubiquitin (pSer65), or anti-GAPDH Abs, followed by corresponding HRP-conjugated secondary Abs. In some experiments, Western blot analysis was performed on siTRPV1 and siGLO (control) CML cells treated with OLDA or vehicle to assess BiP expression levels. Moreover, lysates from CML cells were treated with OLDA at the IC₅₀ dose and imatinib mesylate (0.5 µM), either alone or in combination, for 24 h were incubated with anti-γH2AX and anti-caspase 3.

In addition, the protein oxidation products were identified in CML cells, which were treated as described previously, by scanning carbonyl groups using the OxyBlot^TM Protein Oxidation Detection Kit according to the manufacturer’s instructions. In brief, dinitrophenylhydrazine was added to the crude total proteins (20 μg) to derive the carbonyl groups from the protein side chains. Carbonylated proteins were resolved by SDS-polyacrylamide gel electrophoresis, and Western blot analysis was performed using the provided anti-DNP antibody (1:150).

The detection was performed using LiteAblot PLUS kit, ChemiDoc, and Quantity One software (Bio-Rad). GAPDH was used as the loading control. SHARPMASS VI–VII (Euroclone) and SeeBlue Plus2 (Invitrogen) were used as pre-stained protein markers.

CML cells, treated with OLDA or vehicle at the IC₅₀ dose for 24 h, were incubated with 2 μg/mL PI for 30 min at 37°C. After washing, the fluorescence intensity was analyzed using BD Accuri C6 Plus software. The fluorescent probe DCFDA was used to assess oxidative stress levels in siTRPV1 and siGLO CML cells after treatment with OLDA (IC₅₀). Cells were incubated with 20 μM DCFDA for 20 min prior to the harvest time point. After washing, the fluorescence was assayed using the BD Accuri C6 Plus flow cytometer and its software. ∆Ψm was evaluated by JC-1 staining according to the manufacturer’s protocol in CML cells, treated with OLDA or vehicle at the IC₅₀ dose for 24 h. Samples were then analyzed using the BD Accuri C6 Plus flow cytometer and its software.

BloodSpot, Stemformatics, and GEO are open-access downloaded bio-database that provide visualization and are analyzing tools for large-scale genomics datasets. In particular, BloodSpot (https://www.bloodspot.eu) provides gene expression profiles of healthy and malignant hematopoiesis in humans or mice, encompassing a total of more than 5,000 samples analyzed using a oligonucleotide microarray chip and by RNA-seq assay (Bagger et al., 2016). Stemformatics (https://www.stemformatics.org/) is an established gene expression data portal containing over 420 public gene expression datasets derived from microarray, RNA sequencing, and single-cell profiling technologies. Its major focus is on pluripotency, tissue stem cells, and staged differentiation (Choi et al., 2019). The Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo) is an open functional genomics database of a high-throughput resource (Barrett, 2004).

Analysis of data from database IDs 6326 and 6610 from Stemformatics (accessed on 4 November 2022) and GSE24759 and GSE13159 from GEO was performed in silico (accessed on 13 December 2022). The hierarchical tree was analyzed in the BloodSpot online database (accessed on 13 December 2022).

The statistical significance was determined by Student’s t-test and by ANOVA with Dunnett’s post hoc test. No statistically significant differences were found between siGLO-untransfected and -transfected CML cells, treated or not treated with a vehicle at different times (data not shown). Given the absence of differences, for simplicity, as the control, in time-course analysis, cells treated with vehicle for 24 h were shown.

We recently demonstrated the expression of TRPV1 on K562, KU812, and MOLM-6 CML cell lines, common myeloid progenitors, normal myeloid cell enrichment, and PBMCs, respectively (Maggi et al., 2022). To enhance our analysis of TRPV1 expression, we have also performed in silico analyses with datasets 6326/6610 from the Stemformatics database and GSE13159 from the GEO repository, demonstrating its expression in all types of leukemia cells (Supplementary Figure S1A). Interestingly, TRPV1 expression in common myeloid CML progenitors is modulated during the different phases of the disease with higher levels during the chronic phase (Supplementary Figure S1B).

It is well known that TRPV1 is the receptor for CPS; however, several sources of evidence demonstrated that the endogenous compound OLDA is a more potent and selective TRPV1 agonist (Chu et al., 2003; Seebohm and Schreiber, 2021). Thus, CML cells were treated with CPS (10–300 μM) and OLDA (0.5–100 μM) for 24 h and analyzed by cell viability assay. OLDA induced a stronger decrease in cell viability than CPS, as shown by the IC₅₀ values: 14.1 μM vs 140.1 μM in K562, 1.4 μM vs 129.8 μM in KU812, and 8.4 μM vs 173.9 μM in MOLM-6 (Figures 1A, B). For this reason, we selected OLDA as the TRPV1 agonist, and doses in the IC₅₀ range (10 µM for K562 and MOLM-6; 1 µM for KU812) are considered for the subsequent experiments.

To assess the involvement of TRPV1, cell viability assay was evaluated in CML cells silenced for TRPV1 expression (siTRPV1) (Supplementary Figures S2A, B). TRPV1 silencing significantly reduced the OLDA effects compared to siGLO control cells, confirming that OLDA decreases cell viability by activating this channel (Figure 1C). We also assessed cell viability in normal myeloid cells, enriched from the blood of healthy donors, treated or not treated with OLDA at 1 µM and 10 µM (Supplementary Figure S3A). Results showed that in normal mature myeloid cells, OLDA is much less effective at reducing growth. This result is in line with our in silico analysis, in which the downregulation of TRPV1 expression occurs along the differentiation process (Supplementary Figure S3B).

The opening of the TRPV1 ligand-gated ion channel promotes transmembrane Ca²⁺ entry, which affects the fine balance between death and survival signaling pathways (Zhai et al., 2020). To assess the signaling pathway induced by OLDA, we first performed calcium mobilization assay for up to 3 min (data not shown). The increase in [Ca²⁺]_i was found in CML cells treated with OLDA for 1 min after the stimulation compared to the vehicle (Figure 2A). This effect was inhibited by both the TRPV1 antagonists CPZ and A784168, validating the TRPV1 involvement in OLDA-induced effects (Figure 2B). Moreover, the [Ca²⁺]_i overload was associated with a clear rise in ROS production induced by OLDA in a TRPV1-dependent manner as shown by the evident increase in DCFDA fluorescence in siGLO but not in siTRPV1 CML cells (Figure 2C; Supplementary Figure S4A). Mitochondria dysfunction is often the consequence of [Ca²⁺]_i influx and ROS accumulation (Brookes et al., 2004). Thus, by performing JC-1 staining and FACS analysis, we showed that the treatment of CML cells with OLDA for 24 h induces a marked reduction in JC-1 red fluorescence, demonstrating mitochondrial depolarization (Figure 2D; Supplementary Figure S4B). Given that the interplay between [Ca²⁺]_i and ROS, leading to mitochondrial impairment, is responsible for changes in macromolecules with consequent alterations of the cellular redox state (Feno et al., 2019), Western blot analysis was performed to better elucidate the oxidation levels. Our results showed that OLDA treatment induces a robust enhancement in the oxidation of total proteins in all CML cell lines (Figure 2E). In addition, we also found that phospho-ubiquitin (Ser65) levels, correlated with injured mitochondria (Hepowit et al., 2022), are strongly enhanced in OLDA-treated CML cells (Figure 2F), supporting that OLDA, by triggering TRPV1, induces oxidative stress and mitochondrial dysfunction.

Autophagy, activated in response to several conditions including oxidative stress and mitochondrial alteration, is an essential pathway aimed at eliminating unwanted intracellular elements, such as unfolded oxidized proteins or damaged organelles, to promote cell survival (García Ruiz et al., 2022). To assess the induction of autophagy by OLDA treatment, we first analyzed the ATG12–ATG5 complex essential for autophagosome formation (Kharaziha and Panaretakis, 2017). We found that in OLDA-treated CML cells, no increase in the expression levels of ATG12–ATG5 complex was evident, supporting that the execution of autophagy is not stimulated (Figure 3A). In addition, the conversion of the soluble form of LC3-I to the lipidated and autophagosome-associated form (LC3-II) was investigated. We found that OLDA treatment was not able to increase the expression of LC3-II in all CML cell lines (Figure 3B). To further strengthen our data, we also investigated the expression of ATF4, a transcription factor involved in stress-induced autophagy gene expression (B’chir et al., 2013). We showed that the treatment with OLDA was not able to upregulate the expression of ATF4 compared to vehicle-treated CML cells, highlighting once again that autophagy is not activated (Figure 3C).

These results prompted us to investigate endoplasmic reticulum stress, found to be stimulated under cellular stress conditions, which is characterized by the accumulation of misfolded proteins detected by the unfolded protein response (UPR). BiP, a chaperone protein abundant in ER, is considered the master stress sensor involved in UPR activation (Kopp et al., 2019). Therefore, to investigate the ER stress induced by OLDA via TRPV1, CML cells silenced for TRPV1 expression were treated with OLDA for 24 h. We found that the treatment with OLDA induces a strong upregulation of BiP expression levels in siGLO CML cells, used as the control, compared to siTRPV1 CML cells, indicating that ER stress is activated in a TRPV1-dependent manner (Figure 3D).

Our next step was to investigate cell death in CML cells treated with the TRPV1 agonist OLDA, given that the UPR pathway may play a dual role in either providing survival benefits or triggering cell death (Mlynarczyk and Fåhraeus, 2014). The cytofluorimetric analysis showed that OLDA enhances the percentage of PI fluorescent cells, indicating the activation of cell death (Figures 4A, B), and, moreover, upregulates the expression of the Ser139-phosphorylated variant of histone 2A (γH2AX), supporting the presence of DNA double-strand breaks (Figure 4C). Finally, to assess the type of cell death, the analysis of caspase-3 cleavage was performed. The presence of cleaved caspase-3 fragments in OLDA-treated CML cells was found, demonstrating that the oxidative stress induced via TRPV1 by OLDA promotes apoptotic cell death in CML cells (Figure 4D).

Presently, the common therapeutic strategy for the treatment of CML is based on TKIs such as imatinib. Therefore, as new strategies and new drug targets are always embraced to increase treatment possibilities, we used a computational approach to analyze the experimental data in order to elucidate the nature of the interaction between OLDA and imatinib. Cells were exposed to different doses of OLDA and imatinib for 24 h, and then the cell viability test was performed. Isobologram analysis demonstrated that several combinations of the two drugs provoke increased levels of cytotoxicity, when compared to single treatments (Figure 5A). In particular, the CI values obtained by combining OLDA at the IC₅₀ dose with all the three doses of imatinib are < 1, indicating synergistic effects (Figure 5B). To support these data, we also performed Western blot analysis to assess DNA damage and apoptosis on CML cells treated with the combination of OLDA (IC₅₀) and imatinib (0.5 µM). Our data demonstrated that the co-administration of both drugs is more effective in inducing upregulation of γH2AX and increasing caspase-3 fragment levels than the single treatment (Figures 5C, D), confirming the synergistic effects.