(T1130-08-45) Automating Analytical Method Development with LabSolutions MD: Separating Cannabinoid Stereoisomers using Analytical Quality by Design (AQbD) Principles from ICH Q2(R2) and ICH Q14

Purpose: The cannabis market has been rapidly growing with the list of identified cannabinoids constantly increasing. Analytical methods for detecting over 25 cannabinoids by means of HPLC analysis have been explored and currently there is a method to sperate 21 cannabinoids in a single run. Expanding the list of cannabinoids is possible, however stereoisomers challenge the current methods and, therefore, optimization is needed. Shimadzu’s new analytical method development software, LabSolutions MD (Method Development) alleviates the tedious task of testing, analyzing, and comparing all the individual runs. LabSolutions MD uses “Analytical Quality by Design (AQbD)” concepts to determine the optimal method for cannabinoid separation. Experimental design during the entire method development process, identification of the most robust analytical conditions by graphing factor-response relationships, and predictions of chromatograms once the initial screening phase is completed, provide the user with the power to automate analytical methods development. This software can be an asset to many fields and will be used to efficiently separate cannabis stereoisomers.

Methods: Six Daicel chiralpak columns IA-U, IB-U, IC-U, ID-U, IG-U, IH-U: 3.0mmx100mmx1.6µm) with varying stationary phase chemistries were used with Shimadzu’s Method Scouting system and LabSolutions MD to separate several cannabinoids. Mixes of four and twenty-five cannabinoids were prepared at a concentration of 25ppm. The mobile phase was 0.085% phosphoric acid in water and 0.085% phosphoric acid in acetonitrile.

Results: A four and twenty-five cannabinoid mix was created to display the ability of LabSolutions MD to be used with simple and complex sample types. Utilizing the full factorial screening phase<, the six columns were used to run mobile phase concentrations ranging from 40 to 65 organic to determine the optimal column and mobile phase combination. Both mixes were determined to have the best separation with the IG-U (amylose tris(3-chloro-5-methyl-phenylcarbamate) column. All other columns either did not separate all compounds or did not provide ideal baseline resolution.  Using the optimization phase the best mobile phase ratios, the optimal flow rate, and the ideal oven temperature was determined. Separating the four stereoisomers, the optimal conditions were determined to be 50A:50B mobile phase ratio, 0.8mL/min flow rate, and 35℃ oven temperature. When assaying the twenty-five cannabinoid mix, the conditions were different. Using heat and 3D mapping (Figure 1 and 2), the optimal conditions were determined to be 56A:44B mobile phase ratio, 1.3mL/min flow rate, and 38℃ oven temperature. Concentrations above 50 B caused co-eluting peaks and concertation under 40B increased the run time by more than 10 minutes. The optimal conditions differ between the two mixes due to sample complexity and expansion of run time. The robustness of both the four and twenty-five sample mix were tested and confirmed in the robustness phase of LabSolutions MD using ANOVA. 

Conclusion: In conjunction with Method Scouting, LabSolutions MD is powerful tool for method development and optimization. It alleviates the monotonous task of testing, analyzing, and comparing all the individual runs. LabSolutions MD can be used for both simple and complex samples. Heat mapping and 3D modeling assisted in the identification of the most robust analytical condition for both four and twenty-five cannabinoid mixtures. 

Acknowledgments: The authors would like to acknowledge Dr. Weston Umstead from Chiral Technologies, Inc. and Dr. Eberhardt Kuhn Market Manager Cannabis/Hemp from Shimadzu Scientific Instruments.

Figure 1. During the optimization phase, the optimal conditions are selected using statistical analysis. Heat maps for the evaluation value, minimum resolution, and separated peak count are displayed above. The optimal point indicates that a mobile phase ratio of 56A:44B, 1.3mL/min flow rate, and 38℃ oven temperature.



Figure 2. In addition to heat maps, 3D modeling can be utilized to determine the optimal method conditions. At a concentration of 44B, it was determined that the optimal oven temperature was 38℃ and the optimal flow rate was 1.3mL/min.