The Development, Validation, and Application of a UHPLC-HESI-MS Method for the Determination of 17 Cannabinoids in Cannabis sativa L. var. sativa Plant Material

2.2. Optimization of the Cannabinoid Analysis Process Using UHPLC-HESI-MS

Several authors have performed chromatographic separation using mass spectrometry [12,16,17]. Optimization of the mass spectrometer operation (voltage on the capillary, gas flow) was realized based on the analysis of variants of the values of these parameters (unpublished data), and the results of this experiment are presented in Section 4.5. Using a high-resolution mass spectrometer made it possible to achieve better sensitivity of the method and obtain very low quantification limits. The sample’s high dilution, described in Section 4.4, indicates that no secondary fragmentation of compounds is required during the analysis. Most of the published work has analyzed fewer cannabinoids than in our work [12,18]. There are first reports of analysis of 17 cannabinoids or more [16,17]. In the studies presented here, the separation of analytes was evaluated using variants presented in the above works. The authors used, among others, water and/or ACN with 0.1% formic acid as mobile phases [12,16,17]. The analyses conducted by the above authors were operated in a mobile phase gradient. In our work, analyses were carried out using the isocratic elution method. Given that several cannabinoids were characterized by the isomerism of these compounds and, as a result, identical molecular weights, chromatographic separation of the tested substances was an important issue. Identification of the compounds in the tested samples was carried out by comparing their retention times and mass spectra with analytical standards for these compounds. Experimentally, it was shown that the most effective mobile phase for the separation of analytes mixture of ACN: 0.02% HCOOH_aq and 5 mM HCO₂NH_{4 aq} (75:25, v/v), which enabled separation in the shortest possible time (10 min). This is one of the first methods to allow analysis of cannabinoids in such a short time. The chromatogram for all 17 cannabinoids analyzed in this study is shown in Figure 3.

The retention time and precursor ions for the compound are shown in Table 1.

2.3. Method Validation

Due to the lack of cannabinoid-free matrix availability, adding these compounds or their internal standards directly to the matrix before extraction is not practiced. Internal standards are most often added to diluted extracts at the final stage of sample preparation [16]. An important element of method validation in the determination of cannabinoids in samples is the determination of the matrix effect. The absolute matrix effect was calculated by comparing the slope of the matched-matrix standard curve with the slope of the standard calibration curve. The matrix effect values obtained for all matrices (inflorescences, leaves, and dried ground hemp) ranged from 97 to 106% (Table 2), and it can be concluded that there is no matrix effect on the final assay result.

To prepare calibration curves of the test substances, solutions of these compounds were prepared. The linearity of each of the analyzed compounds was evaluated based on eight-point calibration curves determined by analyzing standard solutions of different concentration ranges. These ranges were chosen to consider the different levels of cannabinoid content in the plant. The concentration-response relationship of the present method indicated a linear relationship between the concentration and peak area with R² values of >0.995 for all 17 cannabinoids. The LOD (limit of detection) and LOQ (limit of quantification) were evaluated by measuring the response at a signal-to-noise ratio (S/N) of ≥3 for LOD and ≥10 for LOQ, respectively, in all types of samples spiked with the standard solution. The determined LOD and LOQ values (Table 2) were at an appropriate level, which enabled the identification and quantification of compounds with low concentrations in inflorescences, leaves, and dried hemp material. The LOQ tended to be much lower than the range of the calibration curves since the analyzed compounds did not occur at such low levels; nevertheless, the low LOQ values allow for future analysis of these compounds at much lower levels. The range of calibration curves, coefficients of determination, and LOD and LOQ for each substance are shown in Table 2. According to the AOAC [19], the concentration of added analyte should be no less than the initial concentration, and the fortified test sample’s response must not exceed the calibration curve’s highest point [19]. The general assumptions of AOAC [19] and ICH [20] for the validation of analytical methods are that the recovery values of the analytes tested should be in the range of 80 to 120%, while the RSD for recovery must not exceed 15%. Recovery analyses were performed for all materials used: fresh inflorescences, leaves, and dried ground hemp (CRM). The recovery and repeatability results of the method obtained in our case meet these criteria. In the case of inflorescences, leaves, and CRM, the recovery values ranged from 97 to 100%, 94% to 101%, and 94 to 103% (depending on the type of compound), respectively, while the RSD value was not greater than 10% (Supplementary Materials Table S1). Most authors report LOQ values in µg kg⁻¹, but in our work, this parameter is expressed in μg mL⁻¹ due to the wide range of dilutions used, which affect the final LOQ for a given compound. The LOQ values obtained by Brighenti et al. [12] ranged from 1.8 to 2.5 μg mL⁻¹, depending on the substance. A lower LOQ (1 μg mL⁻¹) was characterized by the method developed by Zivovinovic et al. [6]. De Becker et al. [8] developed a determination method whose LOQs for the analyzed compounds were in the range of 0.06–0.25 μg mL⁻¹. These values are much higher than those obtained in our work, which allows us to conclude that the present one allows the determination of cannabinoids at much lower concentrations. Differences in the LOQs obtained for the different methods depend on the technique used for cannabinoid analysis.

4.2. Chemicals and Reagents

The certified reference materials of cannabidiol (CBD), cannabidiolic acid (CBDA), cannabigerol (CBG), cannabichromene (CBC), cannabinol (CBN), cannabidiolic acid (CBNA), cannabidivarinic acid (CBDVA), cannabicyclol (CBL), and cannabicyclic acid (CBLA) were provided at 1.0 mg mL⁻¹ solutions in methanol (MeOH) or acetonitrile (ACN) from Restek GmbH (Bad Homburg, Germany). Cannabigerolic acid (CBGA), cannabichromenic acid (CBCA), Δ⁹-tetrahydrocannabinol (Δ⁹-THC), Δ⁸-tetrahydrocannabinol (Δ⁸-THC), Δ⁹-tetrahydrocannabinolic acid (Δ⁹-THCA-A), Δ⁹-tetrahydrocannabivarinic acid (Δ⁹-THCVA), and cannabidivarin (CBDV) were purchased from LGC Standards. Δ⁹-tetrahydrocannabivarin (Δ⁹-THCV) was provided at 1.0 mg mL⁻¹ solutions in MeOH or ACN from Cerilliant Corporation (Round Rock, TX, USA). The certified purity value for all the CRMs was >98.00%. Liquid chromatography–mass spectrometry (LC-MS)-grade water, ACN, MeOH, ethanol (EtOH), and n-hexane were purchased from Witko (Lodz, Poland). Formic acid and ammonium formate (LC-MS grade) were obtained from Sigma-Aldrich (St. Louis, MO, USA).

4.3. Preparation of Standard Solutions and Calibration Curves

Standard 100 μg mL⁻¹ solutions for all 17 cannabinoids were prepared by dissolving 1.0 mL of the compound reference standard in ACN or MeOH using 10 mL volumetric flasks separately. This step was repeated as it was necessary to prepare higher dilutions for most compounds except CBD and CBDA. All solutions were stored at <−80 °C. Eight-point curves were prepared for 17 cannabinoids in different ranges, which were generated using Thermo TraceFinderTM software, version 5.1 (Thermo Fisher Scientific, Pleasanton, CA, USA). Curves in matrix were prepared by adding specific amounts of the standard (as in the curve in solvent) to a previously prepared and appropriately diluted extract (fresh inflorescence, fresh leaves, CRM).

4.4. Preparation of Test Samples

Before starting cannabinoid extraction, all samples previously frozen (−60 °C) were ground to a fine powder using a Grindomix GM200 grinder (Retsch, Haan, Germany).

Fresh material: 0.1 g was weighed into 50 mL Falcon vial samples of inflorescences or leaves and extracted with MeOH in a volume of 5 mL.

Dried material (ground dried hemp—CRM and hemp tea): 0.1 g to 50 mL Falcon vial samples were weighed and extracted twice with 10 mL of MeOH.

Samples were homogenized (2 min. 5000 rpm) using a Unidrive X 1000 homogenizer, Cat SCIENTIFIC (CAT Scientific Inc., Paso Robles, CA, USA). The prepared samples were then centrifuged (2 min. 10,000 rpm) using an MPW-380R centrifuge from MPW Med. Instruments (Warsaw, Poland). The obtained extracts of the dried material after extraction were mixed. For analyses, 1 mL of extract filtered through a 0.22 µm 13 mm syringe filter was used (LLG Labware, Meckenheim, Germany). For compounds for which concentrations outside the upper range defined by the calibration curve were observed, the extracts were diluted accordingly with the extraction liquid. Dry weight content was determined using a WPS 30S balance dryer (Radwag, Radom, Poland) to express cannabinoid levels relative to the dry mass. The level of the studied cannabinoids was calculated according to the following formula (Equation (2)):

where: C—the content of the studied compound (in µg kg⁻¹) calculated per dry weight; a—determined content of the studied compound (in µg kg⁻¹); b—dry weight content (in % of the fresh weight content).

4.5. UHPLC-HESI-MS

Cannabinoids were analyzed using ultra-high-performance liquid chromatography-Q-Exactive Orbitrap mass spectrometry setup operating with a heated electrospray interface (UHPLC-HESI-MS) (Thermo Fisher Scientific, Waltham, MA, USA). Analyses were performed on a 2.1 × 100 mm, C18 Cortecs, 1.6 µm chromatography column (Waters, Milford, MA, USA). The mobile phase consisted of a mixture of ACN: 0.02% HCOOH_aq and 5 mM HCO₂NH_{4 aq} (75:25, v/v). Elution of analytes was carried out in isocratic mode. A constant flow rate of 0.3 mL min⁻¹ (10 min) and a constant injection volume of 2 µL were used for all analyses. The column thermostat temperature was 20 °C. During ionization in the mass spectrometer, the heated electrospray ionization (HESI) mode of operation was used in both positive and negative polarizations with a scan range of 100–1000 m/z. The parameters of the ionization source were as follows: spray gas flow rate (nitrogen): 50 L h⁻¹; auxiliary gas flow rate (nitrogen): 1 L h⁻¹; spray voltage: +2300 V and −2000 V; capillary temperature: 305 °C.

4.6. Method Validation

The method validation was based on AOAC International guidelines [19] and by the Validation of Analytical Procedures: Text and Methodology Q2 to meet the requirements of the International Council for Harmonization [20]. The proposed analytical method was validated for linearity, limits of detection (LOD) and quantification (LOQ), recovery (R), and precision (RSD). Matrix effects were measured by comparing the ratio of the slope coefficient of the calibration curve prepared from the analysis of standards dissolved in a solvent to the slope of the same calibration curve resulting from the study of these standards in the all-tested matrix. A value of 100% indicates that there is no matrix effect. There is signal amplification if the values are >100% and signal attenuation if the values are <100%. The extraction recovery was measured by comparing the peak area of all analyzed matrix spiked with standards before and after extraction.