Optimization of Processing Technology for Tiebangchui with Zanba Based on CRITIC Combined with Box-Behnken Response Surface Method

Summary

The present protocol describes an efficient and standard detoxification processing method for Zanba-stir-fried Tiebangchui using CRITIC combined with the Box-Behnken response surface method.

Abstract

The dried root of Aconitum pendulum Busch., called Tiebangchui (TBC) in Chinese, is one of the most famous Tibetan medicines. It is a widely used herb in northwest China. However, many cases of poisoning have occurred because of TBC's intense toxicity and because its therapeutic and toxic doses are similar. Therefore, finding a safe and effective method to reduce its toxicity is an urgent task. A search through the Tibetan medicine classics shows that the processing method of TBC stir-fried with Zanba was recorded in the "Processing specification of Tibetan medicine of Qinghai Province (2010)". However, the specific processing parameters are not yet clear. Thus, this study aims to optimize and standardize the processing technology of Zanba-stir-fried TBC.

First, a single-factor experiment was conducted on four factors: the slice thickness of TBC, amount of Zanba, processing temperature, and time. With monoester and diester alkaloid contents in Zanba-stir-fried TBC as indexes, CRITIC combined with the Box-Behnken response surface method was used to optimize the processing technology of Zanba-stir-fried TBC. The optimized processing conditions of Zanba-stir-fried TBC were a TBC slice thickness of 2 cm, three times more Zanba than TBC, a processing temperature of 125 °C, and 60 min of stir-frying. This study determined the optimized and standard processing conditions for the usage of Zanba-stir-fried TBC, thus providing an experimental basis for the safe clinical use and industrial production of Zanba-stir-fried TBC.

Introduction

The dried root of Aconitum pendulum Busch and A. flavum Hand.-Mazz., one of the most famous Tibetan medicines, is called Tiebangchui (TBC) in Chinese^1,2. The dried roots of TBC are helpful in dispelling cold and wind, reducing pain, and calming shock. It was recorded in the first volume of "Drug Standards (Tibetan Medicine) of the Ministry of Health of the People's Republic of China," which states that the dried roots of TBC are commonly used to treat rheumatoid arthritis, bruises, and other cold diseases³. However, the clinical therapeutic dose of TBC is similar to its toxic dose, and incidents of poisoning or death have been frequently reported due to improper use⁴. Therefore, reducing the toxicity and preserving the efficacy of TBC has become a research hot spot over the years.

In Tibetan medicine, processing is one of the most effective methods to attenuate the toxicity of TBC. According to "Processing specification of Tibetan medicine of Qinghai Province (2010)", the original herbs (TBC) should be placed in an iron pot and stir-fried with Zanba until the Zanba turns yellow, after which Zanba is removed and the herbs are dried in air^5,6. However, no specific process parameters have been documented, which makes controlling the processing technology and the quality of Zanba-stir-fried TBC difficult. The CRITIC method is an objective weight method that can avoid fuzzification and subjectivity, and enhance the objectivity of weighing⁷. The Box-Behnken response surface method can directly reflect the interaction between each factor through polynomial fitting⁸. The combination of the Box-Behnken response surface and CRITIC method is commonly used to optimize processing technology to acquire the optimized processing protocol^9,10. In this paper, a monoester-diterpenoid alkaloid (MDA) (benzoylaconitine) and two diester-diterpenoid alkaloids (DDAs) (aconitine, 3-deoxyaconitine) were used as evaluation indexes. CRITIC combined with the Box-Behnken response surface method was applied to optimize the processing technology of Zanba-stir-fried TBC and establish a standard processing method for clinical safe use.

Protocol

The Zanba-stir-fried TBC processing method was optimized and standardized by CRITIC combined with the Box-Behnken response surface method. Benzoylaconitine, aconitine, and 3-deoxyaconitine were used as evaluation indexes during this procedure.

1. Sample solution preparation

1. Prepare the reference substance stock solution. Weigh precisely 9.94 mg of benzoylaconitine, 8.49 mg of aconitine, and 6.25 mg of 3-deoxyaconitine (Table of Materials) on an electronic analytical balance and place them in a 10 mL volumetric flask. Then, add 0.05% hydrochloric acid methanol solution to dissolve the solids and make up the volume to 10 mL. Finally, shake the mixture well to obtain the reference substance stock solution with mass concentrations of 0.9940 mg/mL benzoylaconitine, 0.8490 mg/mL aconitine, and 0.6250 mg/mL of 3-deoxyaconitine.
   CAUTION: Hydrochloric acid is a highly corrosive material¹¹. Use proper protection, such as gloves, a lab coat, goggles, and a fume hood.
2. Prepare the test sample solution.
   1. Weigh 2 g of Zanba-stir-fried TBC powder in a conical flask.
      1. Prepare Zanba-stir-fried TBC by weighing 30 g of TBC (2 cm) and 90 g of Zanba and adding them into the preheated stir-fry machine. Set the time and temperature of the stir-fry machine to 40 min and 140 °C, respectively. Set the machine to complete processing.
      2. Use a high-speed smashing machine to grind the Zanba-stir-fried TBC separately into powder samples that can pass through a 50 mesh (0.355 mm) sieve.
   2. Add 3 mL of ammonia solution and 50 mL of a mixed solution of isopropyl alcohol and ethyl acetate (a ratio of 1:1 v/v) into the above conical flask, based on previous studies^12,13.
      NOTE: To prepare the ammonia solution, add 40 mL of concentrated ammonia solution into a 100 mL volumetric flask and fill with purified water to the measuring line. Take appropriate protective measures when using concentrated ammonia solution as it has a strong smell.
   3. Weigh the above sample and conical flask and record the weight. Ultrasonicate for 30 min (voltage: 220 V, frequency: 40 kHz).
      NOTE: Aconitine alkaloids are easily decomposed by heat. Thus, the temperature of ultrasonic extraction must be below 25 °C.
   4. Weigh the sample and conical flask after ultrasonic extraction.
   5. Make up for the lost weight by adding a mixture of isopropyl alcohol and ethyl acetate (ratio of 1:1 v/v).
   6. Filter the sample solution. Evaporate 25 mL of the filtrate to dryness using a rotary evaporator at 40 °C.
   7. Dissolve the residue by adding 5 mL of 0.05% hydrochloric acid methanol solution, filter the solution through a 0.2 µm syringe filter, and analyze it by performing high-performance liquid chromatography (HPLC).
3. Prepare a mixed reference solution that contains 0.1988 mg/mL benzoylaconitine, 0.0509 mg/mL aconitine, and 0.0938 mg/mL 3-deoxyaconitine.
   NOTE: Each standard (0.9940 mg of benzoylaconitine, 0.2545 mg of aconitine, and 0.4690 mg of 3-deoxyaconitine) is dissolved in a 5 mL volumetric flask in 0.05% hydrochloric acid methanol as the dissolution medium.
4. Prepare 0.04 M ammonium acetate buffer by dissolving 6.16 g of ammonium acetate (Table of Materials) in 2 L of ultrapure water (mobile phase A). Adjust the pH to 8.50 using ammonia.
   CAUTION: Ammonia is a hazardous material. Use proper protection, such as gloves, a lab coat, goggles, and a fume hood.
5. Filter 2 L of ultrapure 100% acetonitrile (mobile phase B) and degas it.
   CAUTION: Acetonitrile is a hazardous material¹³. Use proper protection, such as gloves, a lab coat, goggles, and a fume hood.

2. Chromatographic condition

1. Inject 10 µL of the pretreated sample solutions into an HPLC system with binary pumps. Use an HPLC system employing an ODS-3 column (5 µm x 4.6 mm x 250 mm; working at 30 °C) with mobile phases A and B for the MDA and DDAs separation. Inject each sample three times for technical replication.
2. Program the method as shown in Table 1 for the ODS-3 column. Set a flow rate of 1.0 mL/min and the detection wavelength as 235 nm.
3. Record the peak areas of every target compound.
   NOTE: Details of the instruments can be found in the Table of Materials.

3. System adaptability test

NOTE: Refer to section 2 for the chromatographic conditions to perform steps 3.1-3.5.

1. Investigate the linear relationship between the concentration and peak area.
   1. Prepare various concentrations - 19.88, 39.76, 59.64, 159.04, 198.80, and 497.00 µg/mL - of benzoylaconitine solution.
   2. Prepare various concentrations - 8.49, 16.98, 25.47, 33.96, 50.94, and 169.80 µg/mL - of aconitine solution.
   3. Prepare various concentrations - 1.875, 12.50, 37.50, 62.50, 93.75, and 125.00 µg/mL - of 3-deoxyaconitine solution.
   4. Inject the above reference solutions from low mass concentration to high mass concentration and record the peak areas.
   5. Obtain three linear regression equations from the plot of the reference solution concentration (µg/L) against the peak area.
      NOTE: Ensure that the concentrations of benzoylaconitine, aconitine, and 3-deoxyaconitine fall within the linear range of this standard curve.
2. Perform precision testing by continuously injecting six repeats of 10 µL of the sample solution into the HPLC system and run the samples under the same HPLC conditions described in section 2. Record the peak areas of benzoylaconitine, aconitine, and 3-deoxyaconitine.
3. Perform stability testing experiments by injecting 10 µL of the prepared sample solution and determine the peak areas after 0 h, 2 h, 4 h, 8 h, 12 h, and 24 h.
   NOTE: The peak areas are recorded automatically by the referenced HPLC system. These time points were based on relevant literature^15,16,17.
4. Perform the reproducibility test by taking the same batch of Zanba-stir-fried TBC to prepare six test sample solutions in parallel according to the method in step 1.2. Inject 10 µL of each sample into the HPLC system and run the samples as described in section 2.
   NOTE: Reproducibility was assessed by comparing the concentration differences between the six samples.
5. Perform the recovery experiment by preparing six portions of the same batch of Zanba-stir-fried TBC for the test solution. Then, add ~100% of the reference substance of each index component into six portions of the test solution to calculate the recovery rate. Inject these samples (10 µL) into the HPLC system under the same conditions described in section 2 and calculate the recovery rate using Equation (1):
        (1)
   NOTE: In Eq. (1), A is the amount of the component to be measured in the test solution, B is the amount of reference substance added, and C is the measured value of the solution that contains the reference substance and the Zanba-stir-fried TBC sample.

4. Single-factor experiments

1. Comparison of slice thickness
   1. Prepare five groups for tests, each with 30 g of TBC, where the thickness of the TBC is 0.5, 1, 2, 3, and 4 cm, respectively. Weigh an amount of Zanba that is three times as much as that of TBC (90 g).
      NOTE: TBC is toxic. Use proper protection, such as gloves, a lab coat, goggles, and a fume hood, and be careful during the cutting process. Through the pre-experiment, it was found that three times the amount of Zanba was required for complete contact between TBC and Zanba. Therefore, in the formal experimental design, the study selected three times the amount of Zanba when examining the slice thickness.
   2. Set the temperature and the time of the automatic stir-fry machine to 140 °C and 40 min, respectively.
   3. Add ~30 g of TBC and 90 g of Zanba into the machine after the automatic stir-fry machine has heated up to the set temperature.
   4. Prepare the sample solutions by following step 1.2. Calculate the contents of the MDA and DDAs in different processing products according to the standard curve (Table 2). Calculate the comprehensive score based on the results via the CRITIC method in section 6.
   5. In this way, compare the amounts of Zanba, as well as processing temperatures and times for optimization of the conditions.
2. Comparison of the amount of Zanba
   1. Perform five groups of tests, each with 30 g of TBC (2 cm), where the amount of Zanba is one, two, three, four, and five times as much as TBC, respectively.
   2. Turn on the stir-fry machine for processing. Set the time and the temperature of the stir-fry machine at 40 min and 140 °C.
   3. Prepare the sample solutions by following step 1.2. Calculate the content of the MDA and DDAs in different processing products according to the standard curve (Table 2). Calculate the comprehensive score based on the results via the CRITIC method in section 6.
3. Comparison of processing temperature
   1. Perform five groups of tests, each with 30 g of TBC (2 cm) and 90 g of Zanba.
   2. Turn on the stir-fry machine for processing. Set the processing temperature to 100 °C, 120 °C, 140 °C, 160 °C, and 180 °C. Set the processing time as 40 min.
      NOTE: Through pre-experiments, it was found that the speed of Zanba yellowing is very low when the processing temperature is below 100 °C, and Zanba is easy to burn and turn black if the temperature is too high (above 180 °C). Therefore, 100 °C and 180 °C were set to be the minimum and maximum values of temperature during processing, respectively.
   3. Prepare the sample solutions by following step 1.2. Record the peak areas of the MDA and DDAs. Calculate the content of the MDA and DDAs in different processing products according to the standard curve (Table 2). Calculate the comprehensive score based on the results via the CRITIC method in section 6.
      NOTE: The experiment involves high temperatures of 160 °C and 180 °C. Pay attention to safety during the experiment, according to the safety code of the laboratory.
4. Comparison of processing time
   1. Perform five groups of tests, each with 30 g of TBC (2 cm) and 90 g of Zanba.
   2. Turn on the stir-fry machine for processing. Set the processing time to 20, 40, 60, 80, and 100 min. Set the temperature to 140 °C.
   3. Prepare the sample solutions by following the description in step 1.2. Record the peak areas of the MDA and DDAs. Calculate the quality of the MDA and DDAs in different processing products according to the standard curve (Table 2). Calculate the comprehensive score based on the results via the CRITIC method in section 6.

5. Processing technology optimization of Zanba-stir-fried TBC using response surface methodology (RSM)

1. Box-Behnken response surface design
   1. Determine the range of slice thickness (A, 1-3 cm), the amount of Zanba (B, 2-4x), the processing temperature (C, 100-140 °C), and the processing time (D, 40-80 min) by preliminary experiments using single-factor tests (step 4.1-4.4).
      NOTE: The coded values of four variables and their levels are shown in Table 3. Three levels of each variable were coded as -1, 0, and 1.
2. Use software to generate the matrix and analyze the response surface models.
   NOTE: The screenshots for the software usage are shown in Supplementary File 1.
   1. Use a three-level-four-factor Box-Behnken design composed of 24 experiments (as done in this study), and measure five replicates (run order 1, 9, 14, 16, and 25) to calculate the pure error sum of squares (Table 4). Set the comprehensive score (Y) as the response (steps 1-4, Supplementary File 1).
      1. On the home page, click on New Design (step 1, Supplementary File 1), and in the left panel of the Design page, click on Response Surface | Box-Behnken and set the parameters of the four factors in the table (step 2, Supplementary File 1).
      2. Click on Next (step 2, Supplementary File 1), set the response names, and click on Finish (step 3, Supplementary File 1).
      3. Generate the response surface design through the above operation (step 4, Supplementary File 1).
3. Complete the experiment based on the 29 scenarios designed for the response surface.
4. Prepare the sample solutions by following step 1.2.
5. Record the peak areas of the MDA and DDAs.
   NOTE: The peak areas are recorded automatically by the referenced HPLC system.
6. Calculate the quality of the MDA and DDAs in the different processing products.
7. Calculate the comprehensive score based on the results via the CRITIC method in step 6.
   NOTE: The specific method is illustrated in step 6.
8. Input the obtained comprehensive score of 29 trials into the computer and analyze it using the referenced software (step 5, Supplementary File 1).
9. Perform the statistical validation of the polynomial equations and response surface analyses plotted in 3D model graphs through the software (steps 6-8, Supplementary File 1).
   1. In the left Navigation pane, under Analysis (+), click on Y, and then click on Start Analysis in the Configure window (step 6, Supplementary File 1).
   2. Click on ANOVA in the top menu and observe the table of results displaying variance analysis (step 7, Supplementary File 1).
   3. In the top menu, click on Model Graphs and then 3D Surface to obtain the response surface plots reflecting the effects of processing parameters on the synthetic scores (step 8, Supplementary File 1).
10. Perform the validation of the response surface model in triplicate under the predicted optimum conditions (step 9, Supplementary File 1) to verify the stability of the processing technology. In the left Navigation pane, under Optimization, click on Numerical Then, in the top menu, click on Solutions. Observe the predicted optimal conditions.

6. Model evaluation

NOTE: This step is to be performed after each single-factor experiment or response surface experiment has been completed. After each experiment (e.g., comparison of slice thickness) is completed, the content of the MDA and DDAs in the different samples are measured to obtain five datasets, according to step 1.2 and section 2. The data are shown in Supplementary Table S1.

1. Dimensionless processing of the index
   NOTE: This step transforms the measured value (X_ij) into a dimensionless relative value, so that the value of each index is at the same quantity level. This operation can facilitate comprehensive analysis and comparison of indicators in different units or orders of magnitude¹⁸. For the purposes of illustration, slice thickness values have been used for the calculations shown below (Supplementary Table S1).
   1. Standardize the content of the MDA (obtain y_MDA; MDA refers to benzoylaconitine) by using the formula in Eq. (2).
      NOTE: The index "i" stands for one of four factors, and slice thickness is the first factor investigated. Hence, the value of i is equal to 1. The index "j" stands for each level of factors; thus, when the slice thickness is the first level (0.5 cm), j is equal to 1; When the slice thickness is the fifth level (4 cm), j is equal to 5. The contents of MDA (X_ij) in the processed TBC with thicknesses of 0.5, 1, 2, 3, and 4 cm were 0.9693, 1.0876, 1.3940, 1.4185, and 1.3614 mg/g, respectively. Thus, x_{j, max} is 1.4185 and x_{j, min} is 0.9693.
            (2)
      Thus,
      Here, X_ij is the measured content of the MDA of the experiment in the i-th factor and at the j-th level; x_{j, min} is the minimum content of the MDA in this group of experiments; and x_{j, max} is the maximum content of the MDA in this group of experiments. Thus, i = 1, 2, …, m, and j = 1, 2, …, n.
      NOTE: Thus, the standardized values of the MDA are 0.0000, 0.2634, 0.9455, 1.0000, and 0.8729 using Eq. (2).
   2. Standardize the total content of the DDAs (obtain y_DDAs; DDAs refers to aconitine and 3-deoxyaconitine) by using the formula in Eq. (3).
      NOTE: i is one of four factors, and j is each level of the factors; X_ij is the measured content of the DDAs of the experiment in the i-th factor and at the j-th level; x_{j, min} is the minimum content of the DDAs in this group experiment of data; and x_{j, max} is the maximum content of the DDAs in this group experiment of data. In this way, i = 1, 2, …, m, and j = 1, 2, …, n. The contents of the DDAs (X_ij) in the processed TBC with thicknesses of 0.5, 1, 2, 3, and 4 cm were 0.3492, 0.2692, 0.2962, 0.5354, 0.5124 mg/g, respectively. Thus, x_{j, max} is 0.5354 and x_{j, min} is 0.2692.
           (3)
      
      NOTE: The standardized values are 0.6995, 1.0000, 0.8986, 0.0000, and 0.0864 using Eq. (3).
2. Calculate the corresponding contrast intensity (S_i), conflict (δ_i), information (C_i), and index weight (W_i) according to Eqs. (4) to (7), respectively^19,20.
   NOTE: i = 1, 2, …, m. y_ij is the standardized data of the MDA or DDAs content of the experiment in the i-th factor and at the j-th level.
   1. To estimate the contrast intensity, first calculate the average MDA value.
      
      Where  is the average value of the MDA.
           (4)
      
   2. To calculate the conflict value, first estimate the correlation coefficient γ_ij using the CORREL function in Excel²¹.
           (5)
      
   3. Calculate information values as follows.
           (6)
      
      NOTE: Similarly, C_{1, DDAS} = 0.7210
   4. Calculate the index weight as follows.
           (7)
      
      NOTE: Therefore, the weight coefficients of the MDA and DDAs in comparison of slice thickness were established as 0.4945 and 0.5055, respectively.
3. Calculate the comprehensive scores of slice thicknesses.
   
   
   
   
   
   NOTE: Y₁₃ is the maximum value. Therefore, the best parameter of slicing thicknesses is the third level - 2 cm.

   

Representative Results

In this study, the elution gradient used had a good resolution (Figure 1) for the three index components in Zanba-stir-fried TBC, as determined after repeated debugging. The three index components in Zanba-stir-fried TBC had a good linear relationship within a specific concentration range (Table 2). The precision (Table 5), stability (Table 6), repeatability (Table 7), and sample recovery (Table 8) of Zanba-stir-fried TBC were all within the methodological range specified in the Chinese Pharmacopoeia (Volume 4, 2020)²², indicating that the method was feasible. Therefore, the HPLC method was reliable to conduct the analysis of Zanba-stir-fried TBC.

The effect of each factor on processing technology was elucidated using single-factor tests, the results of which are shown in Figure 2. The trend of the comprehensive score of Zanba-stir-fried TBC under different conditions was visualized. The range of slice thickness (A, 1-3 cm), amount of Zanba (B, 2-4x), processing temperature (C, 100-140 °C), and processing time (D, 40-80 min) were determined using single-factor tests (Figure 2).

The CRITIC method is an objective evaluation method that takes advantage of measured data^19,20. When each index has very different levels, using the original index value directly for analysis results in a larger role of the index with a higher value in the comprehensive analysis and a smaller role of the index with a lower value. Therefore, the original indicator data must be standardized to ensure the reliability of the results, as applied to the experimental values in this study. According to the response surface test results and CRITIC method, the weight coefficients of the MDA and DDAs in the response surface experiment were established as 0.5295 and 0.4705, respectively. The comprehensive score (Y) could be calculated according to Eq. (8).

    (8)

The results of the Box-Behnken experimental design are shown in Table 4, while Table 9 presents the results of the ANOVA and regression coefficients. The polynomial equations of comprehensive scores were also obtained after the software analysis. Values of probability less than 0.05 suggested that model was significant (p < 0.0001)²³; The equation in terms of actual factors was obtained in Eq. (9) (Y: comprehensive score; A: slicing thickness; B: amount of Zanba; C: processing temperature; and D: processing time). The equation indicated that the intensity of the influence on comprehensive scores follows this order: processing time > processing temperature > slice thickness > amount of Zanba for four different factors.

Y = 89.05 + 4.57A + 2.88B + 4.63C - 4.83D + 5.19AB + 4.91AC + 6.97AD + 6.69BC - 7.05BD - 1.17CD - 22.80A² - 21.93B² - 19.58C² - 27.19D²    (9)

The response surfaces and contour plots are shown in Figure 3, demonstrating the changes in synthetic scores as a function of four variables. On the basis of the experimental results, the optimal processing parameters of Zanba-stir-fried TBC were determined to be as follows: slice thickness of 2.117 cm, 3.118 times more Zanba than TBC, processing temperature of 123.106 °C, and processing time of 58.156 min. Depending on the feasibility of the operation, the optimal processing technology of Zanba was adjusted - the optimal slice thickness of TBC was 2 cm, the amount of Zanba was three times, the processing temperature was 125 °C, and the processing time was 60 min. The reliability of the model was proven through three tests that were conducted according to the obtained processing parameters (Table 10).

Figure 1: Chromatograms. The chromatogram of the sample solution (A) and the mixed standard solution (B) (1: benzoylaconitine; 2: aconitine; 3: 3-deoxyaconitine). Please click here to view a larger version of this figure.

Figure 2: The synthetic scores of all single factors. (A) Slice thickness; (B) the amount of Zanba; (C) processing temperature; and (D) processing time. The results showed that the comprehensive scores of Zanba-stir-fried TBC are highest when the slice thickness is 2 cm, the amount of Zanba is three times, the processing temperature is 120 °C, and the processing time is 60 min. So, the results showed the range of slicing thickness (A, 1-3 cm), amount of Zanba (B, 2-4x), processing temperature (C, 100-140 °C), and processing time (D, 40-80 min) to be used to design the next experiment. Please click here to view a larger version of this figure.

Figure 3: Response surface plots (3D) reflecting the effects of processing parameters on comprehensive scores. Please click here to view a larger version of this figure.

Table 1: The HPLC gradient. Please click here to download this Table.

Table 2: The linear relationship of the index components in Zanba-stir-fried TBC. The results suggested that the three index components in Zanba-stir-fried TBC had a good linear relationship within a certain concentration range. Please click here to download this Table.

Table 3: Levels of variables for the experimental design. Please click here to download this Table.

Table 4: The Box-Behnken experimental design with responses. Please click here to download this Table.

Table 5: The results of the precision measurement. The relative standard deviation (RSD) values of the peak areas of benzoylaconitine, aconitine, and 3-deoxyaconitine were 0.42%, 0.71%, and 2.95%, respectively (n = 6). Abbreviation: RSD = relative standard deviation. Please click here to download this Table.

Table 6: The results of the stability test. The RSD values of the peak areas of benzoylaconitine, aconitine, and 3-deoxyaconitine were 1.86%, 0.54%, and 2.81%, respectively (n = 6). Abbreviation: RSD = relative standard deviation. Please click here to download this Table.

Table 7: The results of the reproducibility test. The RSD values of the peak areas of benzoylaconitine, aconitine, and 3-deoxyaconitine were 1.99%, 1.84%, and 2.41%, respectively (n = 6). Abbreviation: RSD = relative standard deviation. Please click here to download this Table.

Table 8: Sample recovery rate measurements. The RSD values of the recovery rate of benzoylaconitine, aconitine, and 3-deoxyaconitine were 2.47%, 1.88%, and 2.33%, respectively. Abbreviation: RSD = relative standard deviation. Please click here to download this Table.

Table 9: Analysis of variance (ANOVA) results of the experiment model. Please click here to download this Table.

Table 10: The results of verifying tests. Please click here to download this Table.

Supplementary File 1: The instructions of the Box-Behnken design software Please click here to download this File.

Supplementary Table S1: The calculation result of slice thickness. Please click here to download this File.

Discussion

TBC is an important Tibetan medicine with the effects of dispelling cold and relieving pain. It has been mostly used to treat traumatic injury and rheumatic arthralgia in China for thousands of years^24,25,26. Diterpenoid alkaloids are both active and toxic ingredients of TBC^27,28,29. The main toxic effects of the aconitum alkaloids of TBC are neurotoxicity, cardiotoxicity, and gastrointestinal toxicity^30,31. TBC is generally processed before oral use to mitigate the risk of toxicity. Various processing methods, such as steaming, decocting, and sand-frying, as well as processing with Hezi decoction, Qingke wine, and Zanba, have been useful in reducing the toxicity of TBC while preserving its efficacy¹. Among them, Zanba-stir-frying is an important processing method. Zanba is produced from highland barley (Hordeum vulgare L. var. nudum Hook. f), which is an important grain for people living in the Qinghai-Tibet Plateau^32,33. However, the precise parameters of formulating Zanba-stir-fried TBC are still unclear, which is why this processing technology needs to be standardized to ensure its quality control and safe application.

The most crucial aspect of the method is that the evaluation index was determined with the CRITIC method. According to recent studies, highly toxic DDAs can be hydrolyzed or pyrolyzed into an MDA with moderate toxicity during the heating process^34,35. Studies have showed that aconitine hydrolysis to benzoylaconine is the typical example³⁶. Therefore, the composition changes in the processing process were taken as the evaluation index in the process technology optimization. The CRITIC method is an objective weight method that mainly considers the variation of indicators and the conflict among indicators, which are expressed by the standard deviation and correlation coefficient, respectively. It has been widely applied in the processing of traditional Chinese medicine^37,38. In this protocol, the weight of the main components of Zanba-stir-fried TBC, including benzoylaconitine, aconitine, and 3-deoxyaconitine, were calculated using the CRITIC weighing method of objective assignment, which was used as the evaluation standard of Zanba-stir-fried TBC.

One of the key experimental procedures is ensuring a constant processing temperature during processing, as the processing temperature greatly affects the decomposition of DDAs. Therefore, the pre-experiment involved the use of many kinds of heating devices, such as an induction cooker, electric ceramic stove, and multifunctional stir-fry machine. The multifunctional stir-fry machine could maintain a constant temperature and stabilize the quality of the processed product.

Although the optimized processing technology could reduce TBC's toxicity effectively, limitations still exist. First, some of the active ingredients in Zanba-stir-fried TBC remain unknown. Therefore, qualitative and quantitative analysis cannot be conducted as the relevant reference product is not available. More attention should be paid to phytochemical investigations to obtain the target quality control components. In addition, the pharmacological comparison of raw and Zanba-stir-fried TBC is unclear. Detoxification and evaluation of the efficacy reservation effects in animal models will be the next objectives.

Traditional Chinese medicine processing culture is mainly passed down from master to apprentice, and the end point of processing is generally judged by people's subjective consciousness, which is not conducive to the establishment of a standardized processing method. In this study, digital process parameters were used to specify the processing endpoint, which can realize the combination of modern technology to a certain extent. In summary, this study standardized Zanba-stir-fried processing technology for the toxic attenuation and efficacy reservation of TBC. This approach can provide useful information and guidance for processing technology of other poisonous ethnic medicines.

Materials

Name	Company	Catalog Number	Comments
3-Deoxyaconitine	Chengdu Desite Biotechnology Co., Ltd.	DST221109-033
Aconitine	Chengdu Desite Biotechnology Co., Ltd.	DSTDW000602
Ammonium acetate	Tianjin Kermel Chemical Reagent Co., Ltd	Chromatographic grade
Benzoylaconitine	Chengdu Desite Biotechnology Co., Ltd.	DSTDB005502
Design-Expert software	Stat-Ease, Inc., Minneapolis, MN, USA	version 13.0
Electronic analytical balance	Shanghai Liangping Instruments Co., Ltd.	FA1004
High performance liquid chromatography	SHIMADZU Co., Ltd.	LC-20A
High-speed smashing machine	Beijing Zhongxing Weiye Instrument Co., Ltd.	FW-100
Millipore filter	Tianjin Jinteng Experimental Equipment Co., Ltd	φ13 0.22 Nylon66
stir-Fry machine	Changzhou Maisi Machinery Co., Ltd	Type 5
Tiebangchui	Gannan Baicao Biotechnology Development Co., Ltd	20211012
Ultra pure water systemic	RephiLe Bioscience, Ltd.	Genie G
Ultrasonic cleansing machine	Ningbo Xinyi Ultrasonic Equipment Co., Ltd	SB2200
Zanba	27 Chuanzang Road, Ganzi County	-