Xingchu Gong,Junlin Guo,Jingjing Pan,Zhenfeng Wu

1 Key Laboratory of Modern Preparation of TCM,Ministry of Education,Jiangxi University of Traditional Chinese Medicine,Nanchang 330004,China

2 Tianjin University of Traditional Chinese Medicine,Graduate College,Tianjin 300193,China

3 Pharmaceutical Informatics Institute,College of Pharmaceutical Science,Zhejiang University,Hangzhou 310058,China

Keywords:Material quality standard Critical process parameters Control strategy Loganin Formulated granules

ABSTRACT The quality standards for Fructus Corni have been established based on the effects of the manufacturing processes.Three critical process parameters(CPPs)of extraction,filtration,and concentration to prepare Fructus Corni concentrate were identified by Plackett–Burman design with a single batch of Fructus Corni,which were heating medium temperature,extraction time,and water addition.Morroniside yield,loganin yield,and dry matter yield were process critical quality attributes(CQAs).CPPs arranged with a Box–Behnken design were applied to treat different batches of Fructus Corni.After constructing a model that included CPPs,material properties,and process CQAs,loganin content was found to be the critical material attribute(CMA).The design space was calculated with a probability method.According to the limits of process CQAs,the minimum content of loganin in Fructus Corni was calculated with an error propagation method,which was 6.92 mg·g?1.When the content of loganin in Fructus Corni reaches up to 6.92 mg·g?1,the material is considered high-quality and is most suitable for the process.High-quality material can be used for production of Fructus Corni concentrate.This method can also be used to set material quality standards for other Chinese medicines.

1.Introduction

The“Quality by design”(hereafter abbreviated as QbD)concept has been accepted worldwide for the control of drug quality[1,2].Dr.Yu described the objectives,elements,and tools of the QbD concept in detail[3,4].However,more knowledge of the drug product and drug manufacturing process is required for the implementation of QbD.Rational control strategy should be designed and applied based on scientific knowledge [5].Risk management should be carried out on the whole life cycle of a drug.Successful application of QbD concept will improve the batch-to-batch consistency of a drug,and facilitate quality control level from 2–3 sigma to 6 sigma[6,7].

In recent years,the application of the QbD concept to traditional Chinese medicines has been increasing.The critical process parameters(CPPs)of many processes,such as extraction[8],precipitation[9],chromatography[10],and decoloration[11],were optimized with the design space approach.Critical material attributes (CMAs) of the materials used in manufacturing processes,such as density and water content of concentrate of ethanol precipitation process[12,13],were also considered.However,there are few reports of previous research on the development of control strategies for herbal material quality for us to draw upon.

Variation of the quality of herbal material is usually considered to be the main source of variation of herbal drug product quality.Zhong et al.verified that index component contents in herbal extracts were mainly affected by herbal material quality[14,15].Therefore,the quality control of herbal materials is essential for the batch-to-batch consistency of herbal products.

Properly setting material quality standards is a key step for herbal material quality control.There are many herbal material quality standards listed in the Chinese Pharmacopeia (First Part,2015 Edition).However,these standards were set without the consideration of the actual manufacturing processes used to produce Chinese medicines.Good manufacturing processes may reduce the quality variances between different batches of herbal materials.Therefore,setting herbal material quality standards should take into account the influences of manufacturing processes.

Recently,Wang et al.investigated the effects of Panax notoginseng materials and process parameters simultaneously with a central composite circumscribed design[16].They used the score values of principal components obtained from near infrared spectroscopy(NIR)to characterize Panax Notoginseng[16].Though NIR can reflect the information of Panax Notoginseng comprehensively,setting material quality standards was still difficult because the real significance of principal component scores is not clear[16].In previous work,process parameters were arranged with a definitive screening design to treat different batches of Panax Notoginseng[17].By analyzing the correlation of material property and process critical quality attributes (CQAs),CMA of Panax Notoginseng was obtained[17].After building models of CMAs,CPPs,and process CQAs,the quality standards of Panax Notoginseng were set[17].

In our research,the quality standard for Fructus Corni was set based on our analysis of the influences of material properties and the processes for preparation of Fructus Corni concentrate.Fructus Corni concentrate is an important intermediate product of Fructus Corni formulated granules[18].An excipient,such as maltodextrin,will be added to Fructus Corni concentrate before drying and granulation.Fructus Corni concentrate was obtained through water extraction,filtration,and concentration.CPPs were determined by Plackett-Burman experimental designed experiments using a single batch of material.Different batches of Fructus Corni were then used in Box–Behnken designed experiments.After modeling with material attributes,CPPs,and process CQAs,the CMA of Fructus Corni was identified.The design space based on the CMA was calculated with a Monte Carlo method.The standard for Fructus Corni was developed,as seen in Fig.1.Verification experiments were then performed.

2.Materials and Methods

2.1.Materials and chemicals

Fructus Corni was collected from three provinces,Zhejiang,Shaanxi,and Henan (China),as seen in Table S1 Standard substances of morroniside and loganin were purchased from Shanghai Ronghe Pharmaceutical Technology Co.,Ltd.(Anhui,China).Methanol (HPLC grade) and acetonitrile (HPLC grade) was purchased from Merck(Darmstadt,Germany).The formic acid(HPLC grade)was purchased from ROE (Darmstadt,Germany).Deionized water was producing using a Milli-Q academic water purification system(Milford,MA,USA).

Fig.1.Schematic diagram of the procedures for setting material standard.

2.2.The preparation of Fructus Corni concentrate

The Fructus Corni concentrates were prepared according to the“Technical Requirements for Quality Control and Standard Formulation of Traditional Chinese Medicine Formula Granules (Draft for Comments)”[19].150 g of Fructus Corni was soaked in water for a set amount of time and the reflux was extracted twice with a measured amount of water.The amount of water added for the first extraction was 2 ml·g?1higher than the set value.Considering Fructus Corni is a tonic herbal material,the extraction time for the two extractions was the same.The two extracts were filtered and combined.The combined extracts were concentrated to obtain concentrates.All the reflux extraction experiments were performed in a jacket extraction tank made from glass.The heating medium was a mixture of water and glycerol.

2.3.Experimental design

To identify the CPPs,some process parameters were investigated,including soaking time(X1),water addition in the extraction process(X2),heating medium temperature(X3),extraction time(X4),screen mesh number(X5),concentration temperature(X6),and concentrate density at end point (X7).The coded and uncoded values of each parameter are shown in Table 1.Plackett-Burman design was utilized to analyze the effects of these seven parameters on morroniside yield(Y1),loganin yield(Y2),and dry matter yield(Y3),as shown in Table 2.Batch SZY-09 of Fructus Corni was used in Plackett-Burman designed experiments.

After Plackett–Burman designed experiments,some parameters were fixed as follows.The soaking time was 30 min.The filter used for both filtrations was 100 mesh.The concentration temperature was 70°C,and the concentrate density at the end point was 1.04 g·ml?1.The Box–Behnken design was used to study the quantitative relationships among the CPPs,potential CMAs,and process CQAs,as shown in Table 3.After the material standard was developed,verification experiments with the conditions listed in Table 4 were carried out.

2.4.Analytical methods

2.4.1.Dry matter contents

The dry matter contents in Fructus Corni concentrate were determined gravimetrically.Samples were heated in an oven (DHG -9146A,Shanghai Jing Hong Laboratory Instrument Co.,Ltd.)at 105°C for 3 h,and then weighed precisely.

2.4.2.Content of active components

To determine the content of morroniside and loganin in Fructus Corni,the preparation of test samples was carried out as described in the Chinese Pharmacopeia (First Part,2015th Edition).Fructus Corni was pulverized and screened with a 50-mesh sieve.200 mg of powder was accurately weighed and extracted with 80 ml of 80%(v/v)methanol solution for 60 min.The extract solution was analyzed with the HPLC method described in Chinese Pharmacopeia(First Part,2015 Edition).An XDB-C18 column(250 mm×4.6 mm,5 μm)was used.

Morroniside content and loganin content in Fructus Corni concentrate were determined using a HPLC method developed in priorresearch[18].A HPLC(1260,Agilent Technologies,USA)system with a UV detector was used.Analyses were performed on a XDB-C18 column(250 mm×4.6 mm,5 μm)with the column temperature controlled at 35 °C.The solvent flow rate was maintained at 1 ml·min?1.The sample injection volume was set at 20 μl.0.05%(v/v)formic acid was added at both the water phase and the acetonitrile phase to form mobile phases A and B,respectively.The solvent gradients were as follows:0–10 min,1%–2%B;10–20 min,2%–10%B;20–23 min,10%–11% B;23–33 min,11%B;33–55 min,11%–40%B;55–60 min,40%–50% B;60–61 min,50%–100% B.The detection wavelength was fixed at 240 nm.Concentrate samples were diluted with water and then centrifuged at 12000 rpm for 10 min,after which the supernatant was collected and analyzed.A typical chromatogram of Fructus Corni concentrate is shown in Fig.2.

Table 1 Parameters and their levels for Plackett-Burman design

Table 2 Plackett-Burman designed experiments and results

Table 3 Box–Behnken design and results

Table 4 Verification of experimental conditions and results

Fig.2.HPLC chromatogram of a Fructus Corni concentrate sample.Data Analysis(1)Morroniside;(2)Loganin.

2.5.Data analysis

The calculation formulas for morroniside yield,loganin yield,and dry matter yield were as follows:

where Y is the yield,C is the concentration in the concentrate,W is the weight;subscript M,L,C,F,and DM refer to morroniside,loganin,concentrate,Fructus Corni,and dry matter,respectively.

Multiple linear regression analysis was used to determine the CPPs according to Eq.(4).

where Y is the actual value of morroniside yield,loganin yield,or dry matter yield;a0is a constant;biis a partial regression coefficient;and Xiis potential CPPs.A process parameter with a P value smaller than 0.05 was considered to be an acceptable CPP.

By applying multivariate regression analysis,quadratic models were built to obtain the quantitative models between CPPs,potential CMAs,and process CQAs according to Eq.(5)[17].

In this equation n is the number of CPPs;m is the number of potential CMAs;bii,bij,and ckare regression coefficients;Zkis a material property.Stepwise regression was used to simplify the models.Insignificant variables were removed by stepwise regression.The significance levels for adding terms and removing terms were both set to 0.1.The material property left in the model was considered to be the CMA.Data analysis was performed by Design Expert(version 8.0.6,Stat-Ease Inc.,USA).

Based on the specific goals of CQAs,a Monte-Carlo method was performed using a MATLAB program (R2016a,Version 9.0,The Math Works Inc.,USA)to calculate the design space.The detailed calculation processes were described in previous work[20].A brief description is given as follows.The experimental results were assumed to follow a normal distribution.The mean value of the normal distribution was assumed to be the measured response value.The relative standard deviation(RSD)of the normal distribution was assumed to be the same as that of the center points in the Box–Behnken designed experiments.Random response values were then obtained and modeled by Eq.(5)with every simulation.The prediction values of process CQAs were obtained using models built with random response values.The probability of meeting all the process CQA criteria was then calculated based on the model prediction results.The simulation was repeated 2000 times.The calculation step length used for loganin content,added water,heating medium temperature,and extraction time were 0.1,0.1,0.7 and 1.5 respectively.The design space was defined with a probability higher than 0.90.

3.Results

3.1.Material properties

Most of the morroniside content and loganin content of different batches of Fructus Corni were taken from previously published work[18],as shown in Table S1.The morroniside content was between 8.995 and 13.10 mg·g?1,and the loganin content was between 6.184 and 7.937 mg·g?1.Morroniside content was always higher than loganin content for the same batch of Fructus Corni in this round of experiments.

3.2.CPP identification

The results of the Plackett-Burman designed experiments are shown in Table 2.The yield of morroniside was between 10.43 and 13.49 mg·g?1,the yield of loganin was between 5.993 and 7.045 mg·g?1,and the yield of dry matter was between 510.5 mg·g?1and 613.9 mg·g?1.The ANOVA results of multiple linear regression analysis of each response using Eq.(4)are shown in Table 5.According to theP values,the water addition in the extraction process and the extraction time significantly affected the yield of morroniside.Loganin yield was remarkably affected by heating medium temperature and extraction time.Therefore,the water addition in the extraction process,the extraction time,and the heating medium temperature were selected as CPPs.

Table 5 Regression coefficient values,determination coefficients and ANOVA results

Table 6 ANOVA results for multiple regression models

Fig.3.Contour plot of morroniside yield(Loganin content=7.0 mg·g?1).Water addition=6 ml·g?1.

3.3.Effects of CPPs and CMAs on process CQAs

The results of the Box–Behnken experiments are shown in Table 3.The yield of morroniside was between 11.29 and 14.23 mg·g?1,the yield of loganin was between 5.508 and 6.904 mg·g?1,and dry matter yield was between 497.3 mg·g?1and 583.0 mg·g?1.The partial regression coefficients and variance analysis results of the models are shown in Table 6.The determination coefficient(R2)of each model was higher than 0.74,indicating that the models can explain most of the variation of experimental data.

In Table 6,the term of loganin content in Fructus Corni(Z2)was in both the models of morroniside yield and loganin yield.It means that loganin content in Fructus Corni was a CMA.Because the regression coefficients were positive,higher loganin content led to higher morroniside yield and loganin yield.

Figs.3-4 shows the contour plots of CPPs on process evaluation indices.Higher heating medium temperature,longer extraction time,and more water addition in the extraction process all resulted in higher yields of morroniside,loganin,and dry matter.

3.4.Design space development and verification

To obtain optimum extraction results,the design spaces were calculated based on the specific goals of each CQA.The acceptable range of the CQA and the probability requirements for compliance are shown in Table 7.The design space was calculated and is shown in Fig.5.The combinations of the CMA and the CPPs are listed in Table S2.

In order to verify the accuracy of the design space,verification experiments were performed within and outside the design space,as seen in Fig.5(b).Results of verification are shown in Table 4.All the predicted values are close to the experimental values,indicating that the models had good predictive performance.

Fig.4.Contour plots of dry matter yield(Loganin content=7.0 mg·g?1).(a)Extraction time=30 min,(b)Water addition=6 ml·g?1.

Fig.5.Three-dimensional design space and verification points(color bar refers to the probability of attaining the process CQA criteria;○and △,verification points).(a)Material:SZY-09,(b)Material:SZY-09;heating medium temperature:125°C.

3.5.Quality control strategy of Fructus Corni

In most cases,the manufacturing of herbal extracts will be performed at fixed process parameter values commonly used in theindustry.In our research,CPPs were fixed at 6.0 ml·g?1of water addition in the extraction process,125°C of heating medium temperature,and 30 min of extraction time.Models shown in Table 6 can be simplified,as seen in Eq.(6) and (7).It seems that the yields of morroniside and loganin were affected only by the loganin content of Fructus Corni.

Table 7 Upper and lower limits of process CQAs and probability requirements for compliance

An error propagation method was used to estimate the material quality standard.This method was also used in the estimation of Danshen extract quality standards[21].Taking Eq.(6)as an example,the residuals were assumed to follow a normal distribution with a mean value of zero.The standard deviation(SD)of the normal distribution of Eq.(6) was assumed to be the same as that of the residuals,which was 0.3364.If the yield of morroniside is expected to be greater than 12.40 with a probability of 90%,then Eq.(8)was used to calculate the minimum value of the content of loganin in Fructus Corni.

where U0.1is the boundary value of the normal distribution,which is 1.28.The minimum content of loganin in the Fructus Corni should be 6.92 mg·g?1.

In the same manner,the standard deviation of the normal distribution of Eq.(7) was assumed to be the same as the residual,which is 0.1584.If the probability that the loganin yield is greater than 5.50 with a probability of 90%,Eq.(9)was used to calculate the minimum value of the content of loganin in Fructus Corni.

From Eq.(9),a minimum value of loganin in Fructus Corni could be calculated,which was 6.13 mg·g?1.

Therefore,when the content of loganin in Fructus Corni satisfies Z2≥6.92 mg·g?1,the material can be considered high-quality.When the content of loganin was lower than 6.92 mg·g?1,the Fructus Corni is low-quality material and should not be released for production of Fructus Corni concentrate without any other pretreatments.

4.Discussions

4.1.Morroniside yield

We found that some of morroniside yield values are higher than its content in Fructus Corni by comparing Table A1(Supplementary Material)and Table 3.This means that some other components in Fructus Corni degraded and formed morroniside in the extraction process.Zhou et al.observed a similar phenomenon with morroniside yield higher than morroniside content in Fructus Corni when heating medium temperature was higher than 80°C[22].The morroniside yield could be about 145% of its original content in Fructus Corni [23].7β-Omethylmorroniside in Fructus Corni was found to degrade in aqueous solution and form morroniside [18].Therefore,the high yield of morroniside in this study can be explained.

4.2.The usage of low-quality Fructus Corni

In our work,Fructus Corni were divided into two classes according to the loganin content.A batch of low-quality Fructus Corni,cannot be used with the fixed process parameters mentioned in Section 3.5.However,if the low-quality Fructus Corni is still qualified material according to Chinese Pharmacopeia (1st Section,2015 Edition),it could be used in a mixture with other qualified batches of Fructus Corni to form a new batch.Lau et al.researched and described the mixing method for single-herb extraction and multi-herb extraction[24,25].Mixing different batches of herbal materials is also an effective method to improve batch-to-batch consistency of herbal drug products [14].The mixing method was also encouraged in the“Technical Requirements for Quality Control and Standard Formulation of Traditional Chinese Medicine Formula Granules(Draft for Comments)”issued by Chinese Pharmacopeia Commission[19].

5.Conclusions

In this work,process CQAs for the preparation of Fructus Corni concentrate were morroniside yield,loganin yield,and dry matter yield.After Plackett-Burman designed experiments,CPPs were found to be heating medium temperature,extraction time,and water addition.The influences of CPPs and different material properties were studied simultaneously.Loganin content of Fructus Corni was found to be the CMA.Quantitative models using loganin content of CMA,CPPs,and process CQAs were established.Design space was obtained with a Monte Carlo method simulating the experimental errors.In the verification experiments,experimental results agreed well with prediction results.After fixing the values of CPPs,the quality standard of Fructus Corni was developed.Fructus Corni with loganin content higher than 6.92 mg·g?1was the high-quality material.Because the batches of Fructus Corni used in this work are still limited,the standard for highquality material may be revised with more data collected.If more properties of Fructus Corni can be measured,it is probable to find more CMAs when simplifying Eq.(5)with the stepwise regression method.Accordingly,high-quality Fructus Corni would be defined with more standards.In general,this method for setting material quality standards can also be applied to the manufacturing of other Chinese medicines.

Acknowledgements

This Study was Supported by the Open Fund of Key Laboratory of Modern Chinese Medicine Preparations,Ministry of Education,Jiangxi University of Traditional Chinese Medicine and First-class Discipline Construction Project of Jiangxi Province (JXSYLXK-ZHYAO009,JXSYLXK-ZHYAO010).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.04.006.