Cannabidiol was obtained from Tocris Bioscience (Bristol, UK). CBD was initially dissolved in DMSO (stock concentration of 75 mM) with subsequent dilutions in serum-free tissue culture medium for in vitro experiments. Doxorubicin hydrochloride was purchased from Abcam (Cambridge, UK) and the stock dilution in DMSO to 25 mM was diluted in serum-free medium. The following antibodies were purchased from Cell Signaling Technology [New England Biolabs (NEB), Hertfordshire, UK]: PathScan^® Multiplex Western Cocktail I (#5301), anti-Phospho-mTOR (Ser2448) (D9C2) XP^® Rabbit mAb (#5536), β-Actin (8H10D10) Mouse mAb (#3700), anti-Rabbit IgG, HRP-linked Antibody (#7074) and anti-Mouse IgG, HRP-linked Antibody (#7076). Concanavalin A was from Sigma–Aldrich, UK (#C2272) with stock solutions stored at 5 mg/mL in dH₂O.

The human leukaemic cell line Jurkat were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (complete medium defined as nutrient rich conditions) (both from Invitrogen, Paisley, UK). For routine culture, medium was changed every 2–3 days and cell density was maintained between 5 × 10⁵ and 2 × 10⁶ cells/mL. Standard incubator conditions were 21% O₂, 5% CO₂, and 37°C, denoted AtmosO₂. Serum-free medium (defined as nutrient poor conditions) was used in signaling experiments and to initiate the CBD effect in recovery experiments, with complete medium used in the recovery phase. Cells were also cultured in physiological normoxia (12% O₂ and 5% CO₂, denoted PhysO₂) in an H35 Hypoxystation from Don Whitley Scientific, Shipley, UK. Normal cell culture used 5% serum-containing RPMI. For activation experiments, 10⁶ cells/mL from both AtmosO₂ and PhysO₂ were washed once with RPMI and cultured in normal growth medium supplemented with 5 μg/mL Concanavalin A for 48 h.

Exponentially proliferating cells were counted and 10⁵ cells/mL were seeded into 96-well plates in medium containing 10, 5 or 1% serum, as indicated. Cells were incubated for 72 h with or without compounds, as indicated. After the time period, 20 μL of PrestoBlue^® reagent (Invitrogen, Paisley, UK) was added to wells and cells further incubated for 2 h. Changes in fluorescence were measured at 560 and 590 nm. For doxorubicin experiments, cells were incubated for 72 h in 5% serum conditions in both AtmosO₂ and PhysO₂. All experimental values were determined from triplicate or quadruplicate wells. After subtracting the average fluorescence values of the no-cell control wells from all the experimental wells, the data was averaged and depicted as a percentage of untreated control.

Viable cells exclude trypan blue, whereas dead and dying/apoptotic cells are able to take up the dye since their membranes are compromised. Normally cultured viable Jurkat cells are between 10 and 14 microns (Rosenbluth et al., 2006). Using an automated cell counter TC20 (Bio-Rad), small cells were regarded as those within 5–9 microns and only those that excluded trypan blue were counted. Cells greater than 14 microns were also present, but in negligible numbers, that did not change. Giant cells greater than 18 microns were rare in normal culture. Viability experiments were carried out in 6-well plates and starting cell densities were always 10⁶ cells/mL. After treatments and times, as indicated, cells were resuspended and aliquots were counted in 1:1 trypan blue (0.4%) solutions within 1–2 min to avoid toxicity. Whole cell counts were recorded as live cells/mL and percentage viability, with subsequent gating for live small cell counts (S) and live normal size cell counts (N), based on the above parameters. These counts were expressed as a percentage of the total live cell count or as cells/mL as indicated. In addition, cell aliquots stained with 1:1 trypan blue (0.4%) were placed on microscope slides and coverslips were applied. Phase contrast images of the slides were acquired using a VisiCam TC10 tablet (VWR International, Leicestershire, UK) fitted onto a Motic BA210 Upright Microscope (Motic, Hong Kong) with a 20× objective lens.

Proliferating cells from culture were resuspended in serum-free RPMI medium at 10⁶ cells/mL and treated with or without CBD at 10 μM for 24 h. Cells were counted and resuspended in complete medium (10⁶ cells/mL) without CBD for a further 24 h. This was repeated for two more rounds up to 96 h (72 h post-treatment). In other experiments, cells were allowed to recover for 24 h and then CBD treated again for 24 h, with a subsequent 24 h recovery.

Cells from experiments (adjusted to 10⁶ cells) were lysed at the times indicated in 100 μL of Pierce^TM IP Lysis Buffer (#87787) with Halt^TM Protease Inhibitor Cocktail (#78430), both from Thermo Scientific, UK. Lysates were boiled under reducing conditions in 4X Laemmli sample buffer (Bio-Rad, Hertfordshire, UK). Sample proteins were separated using precast 4–16% TGX Stain-Free gels and transferred onto nitrocellulose membrane (Bio-Rad, Hertfordshire, UK) before blocking with 5% bovine serum albumin (BSA) in Tris-buffered saline with 0.1% Tween (TBST) at RT for 1h. Primary antibodies were incubated overnight at 4°C as per manufacturer’s instructions and, after washing membranes in TBST, blots were incubated with the appropriate secondary antibody at RT for 1 h. Blots were exposed using Clarity^TM Western ECL Substrate and imaged with the Bio-Rad ChemiDoc^TM XRS system. Quantitation of blots was derived from the imaging system from which the relative amount of protein to Rab11 was calculated and used for normalization.

Distribution of cells in the cell cycle was determined by propidium iodide (PI) staining and flow cytometry analysis using a FACS Canto II (BD Biosciences, Oxford, UK). Briefly, 10⁶ cells from experiments were fixed in 90% ethanol, washed and resuspended in RNase A (10 μg/mL) and PI (50 μg/mL). Ten thousand events were collected and live cells were gated away from dead cells as well as debris based on the FSC/SSC parameters. This cell population gate is then placed on PE-Width versus PE-Area dot plot to exclude clumped cells from single cells. The mean FSC-H of the single cells was determined as a measure of relative cell size. Data from the flow cytometry measurements were analyzed using the CytExpert software (Beckman Coulter, Buckinghamshire, UK).

Data are expressed as the mean ± SD of n experiments. To determine statistical significance, Student’s t-test was used for comparing a single treatment mean with a control mean, and a one-way ANOVA followed by a Dunnett’s multiple comparisons test was used for analysis of multiple treatment means. P-values < 0.05 were considered significant.

Previous studies have indicated that CBD effects are influenced by the presence of serum, and previous data on Jurkat cells were generated in serum-free conditions (McKallip et al., 2006). To vous findings, Jurkat cells were exposed to cannabidiol (0.01–10 μM) for 72 h in a range of nutrient conditions (Figure 1A). This resulted in a significant inhibitory effect on the viability of cells at concentrations greater than 1 μM in low serum conditions. This effect was dependent on the nutrient conditions since only 10 μM of CBD was able to inhibit cellular respiration in nutrient rich conditions. IC50 values of 6.4 ± 2.9 and 2.5 ± 0.2 μM were obtained for cells kept in 5 and 1% serum, respectively. To check the cell cycle status of treated cells, cells from treatments in serum-free conditions were removed for DNA content analysis using PI (Table 1 and Figure 1B). Cannabidiol delayed the cell cycle when compared to control -, with an increase in G1 cells and a reduction of cells in the S and G2/M phases. The percentage of SubG1 cells (late apoptosis) remained the same.

Cannabidiol has previously been shown to deactivate Protein Kinase B/Akt (PKB) and mammalian Target of Rapamycin (mTOR) in breast cancer cells (Shrivastava et al., 2011), so we performed a dose response over 4 h to confirm this in Jurkat cells (Figure 2A). CBD reduced the phosphorylation levels of PKB and the ribosomal protein S6 only at 10 μM. There did appear to be an increase in the p42/44 MAPKs but this was not significant. We then performed a time course with CBD for up to 8 h (Figure 2B). For the time course, we included the phosphorylation of mTOR to correlate with the S6 response. CBD did deactivate mTOR to some extent, but this was not significant with multiple comparisons. However, the deactivation of S6 was time-dependent, reaching significance by 2 h.

Next, we sought to ascertain whether cells that remain viable after CBD treatment for 24 h in serum-free conditions, could recover their proliferative capacity. Complete medium was replaced daily after the CBD treatment to a density of 10⁶ cells/mL. Using the Trypan Blue exclusion principle, cells were counted at the start and daily throughout the experiment (Figure 3A). Overall, viable cell numbers were reduced by 30% when CBD was applied at 10 μM. This effect persisted into the next day, despite complete medium replacement, with cell numbers returning to their starting level after 72 h recovery. However, this recovery did not reach the same levels as untreated cells, which more than doubled in number. At the same time, we analyzed viability, based on live cells as a percentage of total cell count (Figure 3B). This reflects the cell count analysis; in that cells did not entirely recover basal viability in this time frame. To try and mimic a multiple dosing regime, cells that were allowed to recover for 24 h and then again treated with CBD in serum-free conditions, had a further reduction in viability to 34.2 ± 1.4 %, which did not recover (Figure 3C).

Since we noted the significant reduction in phosphorylation of S6 by CBD and had observed smaller cells in normal cell counts visualized under the microscope (Figure 4B, left panel), we checked the actual cell size change using automated multifocal plane analysis of cells gated for size. Gating was based on previous evidence (Rosenbluth et al., 2006) and we found that in cycling asynchronous cultures, the percentage of so-called small cells increases as cell density increases (Figure 4A). It should be emphasized here that these small cells are not dead or dying cells; they exclude Trypan Blue and are usually counted as viable in total cells counts (which takes into account all viable cells, regardless of size). Treated cells were also imaged (Figure 4B) and analyzed using flow cytometry (Figures 4C,D). Cells were cultured in nutrient rich medium (10% serum, control +), serum-free medium (control -) and with CBD (10 μM) in serum-free medium for 24 h (Figures 4B–D). The images in Figure 4A show the small cells seen by eye. However, the mean FSC-H, which is a measure of relative cell size based on the granularity of cells, was found to be significantly less in cells in serum-free medium compared to normal culture conditions (Figure 4D). More importantly, CBD further reduced the mean FSC-H, seen as a left-shift in the histogram in Figure 4C and quantified in Figure 4D. Performing the recovery experiment (see Figure 3B) and automated multifocal plane analysis, we found that CBD significantly increased the percentage of small cells over that of control (Figure 4E) and that despite medium replenishment, this population ratio persisted.

To try and replicate these results and potentially explore further mechanisms for our results, we sought to improve our cell model to more closely reflect model the in vivo conditions that Jurkat T cells originated from. Cells were transferred and cultured in complete medium in an environment-controlled incubator at 12% O₂ (termed PhysO₂). Whilst T cells can be found in a wide range of oxygen environments in the body, arterial blood is at 12% (Atkuri et al., 2007). Doxorubicin (Dox) is a drug used to treat many types of cancer and is approved for use in ALL. It is an anthracycline antibiotic and has previously been shown to induce apoptosis in Jurkat cells (Finlay et al., 1986; Mendivil-Perez et al., 2015). In our hands, Dox reduced cell viability with an IC50 = 200.8 ± 24.4 nM () in standard culture conditions (AtmosO₂) (Figure 5A). In combination () with a sub-lethal dose of CBD (5 μM) (), there was a significant reduction in the AUC by approximately a third (Figure 5A, inset). However, this was only an additive effect at low concentrations of Dox that did nothing alone (10^-7–10^-8 M), and the level of reduced viability seen in this range was the same as CBD alone (∼40% reduction). Next, performing the dose responses to Dox and CBD in PhysO₂ conditions (Figure 5B), revealed a resistance to CBD alone (), and an increased susceptibility to Dox (IC50 = 46.7 ± 6.7 nM) (). Compare this with an IC50 of 200 nM seen in AtmosO₂ conditions. In addition, there was no additive or synergistic effect of CBD in combination with Dox ().

We found this surprising and sought to verify the lack of response to CBD by performing a dose response up to 300 μM. In 5% serum conditions, Jurkat cells did respond to CBD, but only at concentrations greater than 37.5 μM with an IC50 = 43.4 ± 5.1 μM () (Figure 5C). Since Jurkat cells are T cells and can be activated by mitogen (Daaka et al., 1996), we wondered whether this phenotype would respond to CBD. Activated cells responded to CBD in the same way as non-activated cells in the AtmosO₂ environment (data not shown) but in the PhysO₂ environment, these cells are also resistant to CBD effects up to 37.5 μM (Figure 5C) (). Interestingly, the interaction effect between the activated and non-activated cells are statistically significant, with activated cells more resistant to CBD-induced cell death (IC50 = 63.4 μM). Our earlier result (Figure 1A) showed that 1% serum conditions produced a robust dose-dependent CBD-induced reduction in cell viability over 72 h and the serum-free effects of CBD on Jurkats was also previously shown by McKallip et al. (2006). We then repeated these experiments in 1% serum conditions with both activated and non-activated cells cultured in the PhysO₂ environment (Figure 5D). Under these conditions, there are three significant features. The first is that CBD does reduce cell viability in both cell phenotypes in a dose-dependent manner. Second, the low serum conditions make a difference to the CBD-induced response for both phenotypes and the IC50 for non-activated PhysO₂ cells (3.2 μM) compares favorably with the AtmosO₂ cells in Figure 1A (2.5 μM). Third, the interaction effect is highly significant, in that activated cells in PhysO₂ conditions are significantly more resistant to CBD than non-activated cells (IC50 = 10.9 μM).