ACPA decreases non-small cell lung cancer line growth through Akt/PI3K and JNK pathways in vitro

Therapeutic agents used for non-small cell lung cancer (NSCLC) have limited curative efficacy and may trigger serious adverse effects. Cannabinoid ligands exert antiproliferative effect and induce apoptosis on numerous epithelial cancers. We confirmed that CB1 receptor (CB1R) is expressed in NSCLC cells in this study. Arachidonoylcyclopropylamide (ACPA) as a synthetic, CB1R-specific ligand decreased proliferation rate in NSCLC cells by WST-1 analysis and real-time proliferation assay (RTCA). The half-maximal inhibitory concentration (IC50) dose of ACPA was calculated as 1.39 × 10−12 M. CB1 antagonist AM281 inhibited the antiproliferative effect of ACPA. Flow cytometry and ultrastructural analyzes revealed significant early and late apoptosis with diminished cell viability. Nano-immunoassay and metabolomics data on activation status of CB1R-mediated pro-apoptotic pathways found that ACPA inhibited Akt/PI3K pathway, glycolysis, TCA cycle, amino acid biosynthesis, and urea cycle and activated JNK pathway. ACPA lost its chemical stability after 24 hours tested by liquid chromatography-mass spectrometry (LC–MS/MS) assay. A novel ACPA-PCL nanoparticle system was developed by nanoprecipitation method and characterized. Sustained release of ACPA-PCL nanoparticles also reduced proliferation of NSCLC cells. Our results demonstrated that low dose ACPA and ACPA-PCL nanoparticle system harbor opportunities to be developed as a novel therapy in NSCLC patients that require further in vivo studies beforehand to validate its anticancer effect.

In this study we hypothesized that ACPA may exert a specific CB1R mediated reduction in proliferation and induction in apoptosis of NSCLC cells in vitro. If so, a novel biocompatible polymer-based nanoparticle system with low biodegradability for long-term controlled release of ACPA can be established for potential anticancer therapy. Primary objective of current study is to assess dose-and time-dependent antiproliferative and apoptotic effect and the mechanism of action of ACPA on CB1R expressing A549, H1299, H358, and H838 NSCLC cells by Water Soluble Tetrazolium-1 (WST-1), real time impedance-based proliferation (RTCA), flow cytometry (FCM), transmission electron microscopy (TEM), gas chromatography-mass spectrometry (GC/MS)-based metabolomics and Simple Western methods. Once the half-maximal inhibitory concentration (IC50) dose is set, the second objective is to design and optimize a novel biocompatible ACPA-loaded polycaprolactone (PCL) nanoparticulate delivery system to improve the stability and prolong the action of ACPA as a potential chemotherapeutic drug.

Study design
A randomized in vitro observative study was designed including control-experiment groups as independent, proliferation-apoptosis measurements as dependent variables. Biological replicates were determined with power analysis (G-Power v3.1).
Quantitative real-time polymerase chain reaction (qRT-PCR) CB1R and CB2R gene expression levels were documented in NSCLC lines 18,36 . Total RNA was isolated and cDNA synthesis was accomplished with QuantiTect ® Reverse Transcription Kit (Qiagen). qRT-PCR was done on a LightCycler ® 480 (Roche) instrument according to producers' recommendations. Relative mRNA expression was assessed using PowerUp SYBR-Green Master Mix (Thermo Scientific) fluorescent dye. CB1R and CB2R levels were normalized to house-keeping gene (β-actin). Sequences of primers used are indicated in Supplementary Table 1.

Immunocytochemistry
Indirect immune peroxidase labeling was carried out for CB1 (cat#C2866, Sigma-Aldrich) and CB2 (cat#HPA028718, Sigma-Aldrich) as previously done 33,39 . Percentage of labeled to total cell number on 25 areas at ×400 magnification was evaluated on automated microscope attached digital camera by image analysis program (Leica DMB6B, DFC7000T, LASV3 Wetzlar, Germany).
Cells kept for 24 hours with ACPA and normal medium were fixed in 2% glutaraldehyde and processed as previously performed 40 for ultrastructural analysis by TEM (Jeol JEM-1400, Japan) and attached digital camera (Gatan, Germany).

Analysis for stability of ACPA
ACPA was measured in all cells and cell supernatants on days 1, 2, and 3 by liquid chromatography-tandem mass spectrometry (LC-MS/MS) after its supplementation at 10 −6 M (LCMS-8030, Shimadzu, Japan) (n = 3). Chromatographic separation was accomplished by using a C18 column (Hypersill-ODS4, 50 × 3.0 mm, 2.1 µm) with a mobile phase of acetonitrile and water (both containing 0.1% formic acid) with 0.3 ml/min flow rate. Daily calibration curve of ACPA was prepared at 6 different concentrations (1000-1.0 ng/ml) and constructed with the peak area ratio of ACPA to anandamide as IS versus concentration. Proteins in cell media with acetonitrile were centrifugated at 15000 rpm for 10 min and reconstituted using mobile phase including IS. Sample preparation for supernatant was done the same as in metabolomics except derivatization step.
Samples with ACPA at 10 µg/ml were prepared from stock ACPA solution, analyzed at 24 and 48 hours after storing at 37°C and 4°C (n = 3), and compared with freshly prepared ACPA solution at the same concentration.
Preparation and characterization of ACPA-PCL nanoparticles PCL MW:80,000 Da (cat#440744-250 G, Sigma-Aldrich) was used to prepare ACPA-loaded nanoparticles via nanoprecipitation method 44 . Acetonitrile:ethanol (9:1; v/v) including 0.1% (w/v) PCL and 0.01% (w/v) ACPA formed the organic phase was added to ultra-pure water (1:2 v/v) containing Pluronic F68 0.05% (w/v) dropwise at RT and subsequent colloidal mixture was magnetically stirred at 550 rpm for 30 min. Organic solvent was removed under vacuum to obtain final nanoparticle aqueous dispersion. in vitro characterization was performed by measuring particle size distribution, polydispersity index (PDI), zeta potential, ACPA encapsulation efficiency, and release as previously conducted by Zeta Sizer Nano ZS (United Kingdom) 45 (n = 3). Encapsulation efficiency was determined after unbound ACPA was removed by centrifugation at 3500 rpm at RT. The supernatant was lyophilized and dissolved in dichloromethane which was then removed under nitrogen atmosphere to quantitatively analyze for ACPA content with LC-MS/MS. For in vitro release study, nanoparticle dispersion was kept in dialysis membrane immersed in release medium and stirred at 37°C. Samples were analyzed with LC-MS/MS at 0, 1, and 4 hours and 1, 2, 3, 5, and 7 days. Release profile graph was obtained as %ACPA cumulative release.

Cell viability assay for ACPA-PCL nanoparticles
Same protocol of RTCA was used. ACPA-PCL nanoparticles releasing low dose window were applied to cells once whereas ACPA was solely applied every day. Blank PCL nanoparticles were diluted to same volume as ACPA-PCL nanoparticles. Stable and prolonged effect of ACPA was observed (n = 3).

Statistical analysis
Whole data exhibited normal distribution by Shapiro-Wilk test. One-way analysis of variance (ANOVA) and post-hoc Duncan's test were used for comparison of parametric results. Student's t-test was conducted for metabolomics and Simple Western. Descriptive data were presented as mean ± SEM. Spearman's test was performed for correlation analysis. All data were evaluated within 95% confidence interval.

NSCLC cells express high CB1R and low CB2R
A549, H1299, H358 and H838 cells presented high CB1 and low CB2 mRNA expression and immune reactivity by qRT-PCR and immunocytochemistry respectively (Fig. 1a,  b). No CB1 relative mRNA expression was noted for H1975 and SW-1573 cells (Fig. 1a). All cell lines presented significantly higher CB1 immune labeling when compared to CB2 (Fig. 1b) and exhibited mitotic figures, increased nucleus-to-cytoplasm ratio and cell pleomorphism under phase-contrast microscope (Fig. 1c). Labeling percentages presented strong positive correlation with CB1 mRNA expression levels (Fig. 1d) and a weak positive correlation with CB2 (Fig. 1e).
Early (Fig. 3a, b, Supplementary Fig. 1) and late (Fig. 3a, c, Supplementary Fig. 1) apoptotic cell numbers by FCM were higher with IC50 dose of ACPA application at 24 to 72 hours in A549, H1299, H838, and H358 cells compared to untreated controls. Apoptotic effect of ACPA decreased significantly over time in A549 cells but remained constant from 24 to 72 hours in H1299, H358, and H838 cells (Fig. 3b). A549 (Fig. 3e-h), H1299 (Fig. 3i-l) and H358 (Fig. 3m-p) cancer cells exhibited cytoplasmic organelle composition with obvious nuclear and nucleolar chromatin material and the lack of tubular systems by TEM. ACPA treated A549 (Fig. 3g, h), H1299 (Fig. 3k, l) and H358 (Fig. 3o, p) cells had giant mitochondria with degenerated cristae adjacent to varying amount of autophagic vacuoles. Ultrastructural morphology of ACPA-treated groups qualitatively supported the evidence of apoptosis by FCM. Cell viability percentages by RTCA presented positive correlation with total apoptosis rates by FCM in A549, H1299, and H838 cells and strong negative correlation in H358 cells (Fig. 3d).

ACPA became unstable after 24 hours
ACPA lost its chemical stability at 37°C when compared to 4°C on day 1 and 2 ( Fig. 6a). Amount of ACPA in cell-free and cell-containing culture media decreased from day 1 to 3 (Fig. 6b). Intracellular amount of ACPA was measured between 0.4 and 1.2 nM on day 1 (Fig. 6c).

ACPA-PCL nanoparticles were characterized
Mean particle size, PDI, and zeta potential of APCA-PCL nanoparticles were revealed as 162.2 ± 2.3 nm, 0.251 ± 0.008, and −29.4 ± 0.6 mV, respectively. A total of 39.9 ± 14.7% ACPA was loaded into the PCL nanoparticles and its sustained cumulative release from the nanoparticles has been noted as 63.9% for a period of 7 days (Fig. 6d).

Discussion
In this study A549, H1299, H358, and H838 NSCLC cells presented high CB1R mRNA expression and exhibited diffuse CB1 immune labeling that quantitatively correlated with qRT-PCR results. Our data is in line with previous studies investigating CB1 mRNA expression in A549 18,34,35 , H1299 34,36,37 , H358 35,36 , and H838 36,37 cells also supporting that CB1Rs are expressed higher than CB2 in bronchi in lungs 22 . Cannabinoid trafficking occurs within a dynamic milieu and cannabinoid receptors cycle constitutively between plasma membrane and cytoplasm [46][47][48] which may also reflect lower CB2 immune labeling in NSCLC cells.
Here we report that 10 −9 -10 −12 M doses of ACPA significantly reduced proliferation of A549, H1299, H358, and H838 cells in a dose-and time-dependent manner. Antiproliferative IC50 dose of ACPA was determined as 1.39×10 −12 M by RTCA which also exerted antiproliferative effect by decreasing viability, increasing early and late apoptosis mainly on day 1 by FCM. Specific CB1 antagonist AM281 reversed the antiproliferative effect of ACPA on A549, H1299, and H838 cells on day 1. Antiproliferation and apoptosis data significantly correlated with CB1R mRNA expression. Previous reports 49,50 regarding the impact of variant CB1 antagonists are coherent with our findings. Our research group previously determined IC50 dose of ACPA as 9.3 × 10 −6 M on CB1R expressing Ishikawa endometrial cancer cells at 46 hours in vitro which induced 8.9% early and 62.8% late apoptosis by FCM 33 55,56 . Taken together, ACPA may act as an efficient synthetic anticancer drug candidate functioning purely via CB1R to induce apoptosis in NSCLC cells at a low dose window comparing to other cannabis ligands.
ACPA induces cell death on NSCLC cells via Akt/PI3K, glycolysis, pentose phosphate pathways; amino acid biosynthesis, urea, and TCA cycles in our study. Degradation in various metabolites can be associated with p-Akt degradation and p-JNK activation in ACPA-treated A549 cells comparing to control on day 1 by Simple Western. Exogenous cannabinoids have been reported to mediate antitumor activity via Akt/PI3K inhibition 13,17,18,34,[57][58][59][60][61] and JNK activation 18,27,61-63 through CB1R and/or CB2R in various cancer cells including NSCLC. ACPA induced autophagy by activating AMPK, inhibiting glycolysis 30 and proliferation in Panc1 cells by blocking pyruvate kinase-2 which is found in lung and pancreatic islets 32 . Our data regarding p-Akt degradation and p-JNK activation is coherent with previous findings reporting antiproliferation nature of Akt and JNK function. Lung tumors can be characterized with glycine/serine/threonine upregulation 64,65 , pentose phosphate and glycolysis pathways, and TCA cycle 66 regulating cell proliferation and reduced glutathione as an oxidative stress marker. Those findings could improve our understanding of metabolomic results that ACPA can stimulate apoptotic pathways by reducing critical metabolites in NSCLC cells.
ACPA can penetrate the cell membrane to bind its receptor in any intracellular location and intracellular/ extracellular ACPA levels matter for its antiproliferative effect. Therefore, we initially demonstrated that amount of ACPA in-cell free and cell-containing culture media diminished after 24 hours at 37°C which accounts for the instability of ACPA upon storage. We prepared ACPA-PCL nanoparticles according to IC50 dose of ACPA to mimic the time-dependent release within the cellular milieu and noted approximate ACPA loading efficiency as 40% with a cumulative release of 64% for 7 day-period.  Cannabinoids are lipophilic molecules with low solubility 67 and drug delivery systems have recently been developed with cannabinoids due to shorter half-lives 11,[68][69][70] . Various cannabinoids were loaded with 60-100% efficiency into lipid 71,72 and polymer-based 69,[73][74][75][76] nanoparticles showing about 60-100% cumulative release after 7 hours-30 days of loading. PCL is a biopolymer approved by the FDA for therapeutic use in drug delivery systems due to its biocompatibility and hydrophobicity facilitating encapsulation of poorly soluble molecules such as ACPA 77 . Although our loading efficiency data was lower than expected, the releasing profile of ACPA was evaluated within the range stated in the literature which thus provided a basis for our novel ACPA-PCL nanoparticle system. As ACPA has poor stability in handling and storage conditions, PCL provided an optimum matrix structure in which ACPA could be encapsulated based on weak electrostatic interactions, mostly within the polymer matrix but also adsorbed on the nanoparticle surface. As PCL nanoparticles do not have a lipid core, loading efficiency was lower than lipidic nanocarriers 71,72 . However, PCL compensates the lower loading data and suggests a promising nanomedicine to develop ACPA as a therapeutic product. Moreover, hydrophobicity is the main parameter controlling the release rate of ACPA from PCL nanoparticles which was observed as a slowrelease up to 7 days. Complete release is not yet achieved in this 7-day period suggesting a longer release time. Initial release may be a result of the burst effect of ACPA adsorbed on nanoparticle surface followed by gradual release of encapsulated ACPA governed by diffusion, partition coefficient of ACPA, and matrix erosion of the nanoparticle liberating the encapsulated drug.
PCL nanoparticles releasing ACPA also reduced proliferation of all cells at different time intervals. PCL nanoparticles releasing exogenous cannabinoids inhibited proliferation of basophilic leukemia 76 , colon 76,78 , breast 69,79 , and NSCLC 78 cancer cells. Here we provided ACPA-PCL nanoparticles releasing low dose window inducing a significant antiproliferation on NSCLC cells in vitro.
Low doses of ACPA and ACPA-PCL nanoparticles have antiproliferative and/or apoptotic effects on NSCLC cells via Akt/PI3K and JNK pathways. Since this is a preliminary study, results were limited to in vitro conditions which requires to be completed with further functional in vitro and in vivo studies. This limitation, however, does not obstruct further in vivo and clinical researches since statistical accuracy and parametric distribution were validated. Moreover, RTCA has been reported as one of the most reliable in vitro methods permitting the assessment of potential personalized therapeutics with high sensitivity and specificity before clinic 80 . Lack of evaluation of CB1R and CB2R expressions in primary lung epithelium in our study may be a limitation. Still, CB1R mRNA expression results contribute to the literature. Lack of assessing p-Akt/p-JNK expressions in ACPAtreated A549 cells in the presence of Akt/JNK inhibitors is a limitation which needs to be analyzed in further studies. Since ACPA has a lipophilic characteristic and a short half-life, maintaining its stability was difficult throughout in vitro studies. To overcome this limitation, we first developed ACPA-PCL nanoparticles. In conclusion, anticancer effect of ACPA and novel ACPA-PCL nanoparticle system through CB1R agonism could represent a promising nanomedicine candidate for in vivo studies and further clinical trials for the eventual reduction of potential adverse effects of systemic EGFR or EML4/ALK targeting chemotherapeutics that are used for lung cancer in the clinic.