The stability of CBD was quantified and compared with two additional CBD analogs: an isolated cannabidiol quinone denoted as CBD-Q and a synthesized cannabidiol-diacetate (CBD-DA) (SI S1 Synthesis Procedures). To prevent sample degradation, CBD and its analogs were stored under an 99% argon atmosphere at -20 °C for up to one month prior to cell culture experiments. Upon retrieval, they were dissolved in DMSO or ETOH and used immediately.²⁸ For investigating the long-term stability of CBD in solution, CBD in DMSO and CBD in EtOH solutions were also stored in the dark for one month at 4 ºC, and then characterized using mass spectroscopy (SI S2. Mass Spectrometry). After one month, mass spectros-copy revealed that both the DMSO and EtOH samples displayed a decrease in the relative abundance of CBD. While the relative abundance of CBD in EtOH decreased with 20% remaining and 80% converted to CBD-Q, CBD in DMSO solution completely degraded with no trace of CBD after one month with 100% converted to CBD-Q (Figure 1a).

The 100% loss of CBD in DMSO, and the relative abundance of CBD-Q in both samples indicate that oxidation of CBD to CBD-Q is greater in DMSO. Cyclic voltammetry was performed to assess the oxidative susceptibility of CBD and CBD-DA (SI S3. Cyclic voltammetry procedures). The cyclic volt-ammogram (Fig. 1b) shows a distinct oxidation potential for CBD at approximately 1600 mV, indicating its high oxidative susceptibility. This supports the conclusion that storing CBD in an oxygen-rich environment leads to oxidation products like CBD-Q. In contrast, no significant oxidation potential was observed for CBD-DA analog, suggesting that the hydroxyl groups on CBD are likely involved in the formation of CBD-Q (Fig. 1c). Together with prior research demonstrating that DMSO exhibits a lower oxygen solubility compared with EtOH, factors such as storage temperature and light may also be contributing to the production of oxidation products.²⁹ Knowing how CBD may degrade into CBD-Q in solution over time, we used any dissolved CBD analog samples for cell culture studies immediately upon retrieval form storage in -20 °C argon.

For cell toxicity assays, we followed previously described protocols.^26,27 We diluted the analog-vehicle solutions into EGM-2 (Lonza, CC-3162) cell culture medium and treated HUVECs to working concentrations of 1 and 10 µM for 24 hrs. Afterward, cells were stained with Calcein AM (Invitrogen, C3099) and imaged at 10x using a fluorescent microscope. Images were then analyzed using FIJI with pre-determined size-exclusion thresholds used for cell counting, followed by a student t-test to determine statistically significant differences between conditions.³⁰ Detailed methods can be found in SI S4. Cell cutlure and S5. Quantitative analysis of cytotoxicity and S6. Statistical analysis.

At the lowest 1 µM concentration of CBD, CBD-Q, and CBD-DA, all cannabinoid analogs yielded a slight decrease in the average live cell count compared to the solution control (Fig 2a and 2b). Still, this decrease was only statistically significant in the EtOH control compared with the 1 µM CBD-DA (Fig 2b).

Additionally, the endothelial cell morphologies confluency appears slightly different in several conditions compared to the controls, but cell morphology differences are likely related to confluency in each image.

Because the toxicity of the oxidized metabolites may not directly correlate to the oxidation potentials, it is essential to establish which cannabinoids yield products posing the most significant risk for adverse health effects.

At the highest concentration, 10 µM, as hypothesized, all conditions displayed a more significant decrease in the average live cell count than for 1 µM conditions, with the CBD and CBD-Q in DMSO and EtOH exhibiting the largest cytotoxicities. However, for the CBD, CBD-Q, and CBD-DA in DMSO, a student t-test analysis indicated all of the 10 µM DMSO conditions possessed an averaged live cell count that was significantly lower than the control (Fig 3a), with 10 µM CBD-DA exhibiting the highest cell survivability, with cell counts at 75 ± 2 per field of view. In contrast, CBD and CBD-Q measured 60 ± 10 and 30 ± 9 cell counts per field of view, respectively. Moreover, 10 µM CBD, CBD-Q, and CBD-DA treatment groups were not statistically different, demonstrating that at 10 µM, all CBD analog conditions displayed significant cytotoxicity.

The analogs delivered in the EtOH vehicle exhibited similar trends, with CBD and CBD-Q in EtOH treatments significantly decreasing cell survivability relative to the EtOH control. However, the average live cell count of CBD-DA in EtOH was only slightly lower than the EtOH control and this decrease was not statistically significant. When comparing the CBD, DBD-Q, and CBD-DA only (not against the control), the cells treated with CBD-DA in EtOH possessed a significantly greater average live cell count per field of view and considerably higher survivability than CBD and CBD-Q in EtOH. The decrease in cell viability in CBD and CBD-Q in both DMSO and EtOH is easily observed visually (Fig. 3 c-j), where a complete eradication of the cell monolayer and overall reduction of cell attachment is evident while the CBD-DA micrographs from both DMSO and EtOH solution treatments (Fig. 3f and Fig. 3j) show that significantly more live cells remain. Still, some differences in cell morphology can be seen when comparing control samples and CBD-DA micrographs. The calcein stain appeared more continuous and localized around the center of the cell body in the controls while in the CBD-DA micrographs, one could observe a slight speckling of the fluorescent signal, suggesting a reduction of fluorescent calcein production, which could result from cell apoptosis. Regardless, the presence of the cell within the micrograph indicates that the cells remained attached after washing with PBS and before imaging after 24 hours. We also observed some edge-effect-dependent toxicity (not shown), where cells at the well edges experienced a greater cell death compared with those at the center of the well. This greater cell death might have been due to the lower cell density at the well-edges, suggesting toxicity also to be cell density-dependent. To mitigate this, all images presented and analyzed were of images taken at the approximate center of the well.

Taking these considerations into account, our results demonstrate that CBD in DMSO or EtOH is unstable and has the potential to degrade into CBD-Q, a significantly more cytotoxic analog oxidatively. Even though EtOH exhibits a higher oxygen solubility, cytotoxic cell assays reveal similar results in the effects of CBD analogs, regardless of vehicle, on HUVEC monolayers.

Based on our observations from cyclic voltammetry and cytotoxic assays, we conclude that CBD-Q is a potential culprit for decreased cell survivability. Within any given cell assay, solvents are exposed to some degree of light, oxygen, and/or heat either during the preparation, storage, or experimental process. Since CBD is sensitive to oxygen, degradability can cause experimental results to be unclear as to whether CBD is toxic or whether byproducts, like CBD-Q, are the true source of toxicity. Since most cytotoxic assays are done within a period of 24–72 hours in an open-air environment, it is recommended to reduce any sample exposure to an oxygen-rich environment and preserve sample purity for future cell experiments.

We recommend a few sample storage methods to minimize the oxidation of solvents. In our case, CBD was kept as a crystalline powder, where the powder was pumped and purged to remove air and moisture and then kept in a dark, inert argon environment. Additionally, we used prepared analog solvents immediately, as preliminary results had revealed an increased cytotoxicity depending on the length of analog-solvent storage. This suggests that future cytotoxicity studies on CBD must distinguish whether CBD’s toxicity arises from byproducts produced by CBD or due to the compound’s inherent toxicity. The sensitivity of our samples also highlights the importance of detailing the exact storage conditions and durations of their analog-solvent solutions.

We conclude by emphasizing how new CBD analogs, such as the newly synthesized CBD-DA, can be designed with improved oxidation resistance and reduced cytotoxicity