Determination of Absolute Purities of Hygroscopic Substances by Quantitative NMR Analysis for the Standardization of Quantitative Reagents in the Japanese Pharmacopoeia (Part 2)

Abstract

As a new absolute quantitation method for low-molecular compounds, quantitative NMR (qNMR) has emerged. In the Japanese Pharmacopoeia (JP), 15 compounds evaluated by qNMR are listed as reagents used as the HPLC reference standards in the assay of crude drug section of the JP. In a previous study, we revealed that humidity affects purity values of hygroscopic reagents and that (i) humidity control before and during weighing is important for a reproducible preparation and (ii) indication of the absolute amount (not purity value), which is not affected by water content, is important for hygroscopic products determined by qNMR. In this study, typical and optimal conditions that affect the determination of the purity of ginsenoside Rb₁ (GRB1), saikosaponin a (SSA), and barbaloin (BB) (i.e., hygroscopic reagents) by qNMR were examined. First, the effect of humidity before and during weighing on the purity of commercial GRB1, with a purity value determined by qNMR, was examined. The results showed the importance afore-mentioned. The results of SSA, which is relatively unstable in the dissolved state, suggested that the standardization of humidity control before and during weighing for a specific time provides a practical approach for hygroscopic products. In regard to BB, its humidity control for a specific time, only before weighing, is enough for a reproducible purity determination.

Introduction

Quantitative NMR (qNMR) has emerged as a new absolute quantitation method for low-molecular compounds, including reference standards for pharmaceuticals. The Japanese Pharmacopoeia (JP), supplement II to the 16th edition, introduced qNMR as a method for the determination of the purity of reagents used as HPLC reference standards in the assay of crude drugs and Kampo extract formulations.^1–4) In supplement II to the JP 17th edition, published in June 2019, the following 15 substances were evaluated by qNMR testing and are listed as the reagents for the assay: (E)-cinnamic acid, geniposide, magnolol, magnoflorine, paeonol, rhein, rosmarinic acid, saikosaponin b₂, loganin, [6]-gingerol, [6]-shogaol, (E)-ferulic acid, 10-hydroxy-2-(E)-decenoic acid, evodiamine, and sinomenine. The HPLC quantification of 15 markers in 11 crude drugs and 23 Kampo formulae is performed using the reagents as reference standards.^1–4) In addition, mangiferin evaluated by qNMR is listed as the reagent for the assay of the Byakkokaninjinto extract, newly listed in the JP18.⁵⁾

To establish the HPLC reference standards with purity values determined by qNMR in the crude drug section of the JP, there were several problems to be solved. We previously resolved problems related to: 1) the establishment of the reference standards for qNMR, 2) complications due to signals from impurities in the reference standards for qNMR and targeted marker compounds, and 3) peak unity of signals of targeted marker compounds in HPLC.^6–9) Furthermore, the stability of the reagents is crucial and we showed that the preparation and distribution of reagents as commercial products with the determined purity values were important for JP users.

Purity evaluation of reagents by qNMR requires accurate weighing. However, the weighing of hygroscopic reagents may be affected by the environment.¹⁰⁾ In a previous study, we examined the hygroscopicity of 21 standard products for crude drug tests in the JP by water sorption-desorption analysis using thermal gravimetric analysis (TGA) equipment. The results showed that ginsenosides and saikosaponins are hygroscopic.¹¹⁾ It can be predicted that the weights of hygroscopic reagents easily change depending on humidity. Consequently, their purities change. We demonstrated this using saikosaponin b₂ (SSB2) in the previous report and that (i) humidity control before and during the weighing is important for a reproducible preparation and (ii) indication of the absolute amount (not purity value), which is not affected by water content, is important for hygroscopic products determined by qNMR.¹¹⁾ In this study, typical and optimal conditions that affect the determination of the purity of ginsenoside Rb₁ (GRB1), saikosaponin a (SSA), and barbaloin (BB) (i.e., hygroscopic reagents)¹¹⁾ by qNMR were examined.

Results and Discussion

GRB1

We purchased commercially available reagents of GRB1, the purity of which had been determined by qNMR. The reagents were labeled and re-measured using qNMR to obtain the purity after maintenance under different humidity conditions before and during weighing.

First, quantitation signals of GRB1 were examined. The reagent had no interference signal when the 3rd-position signal (around 3.8 ppm) was used as a signal to be integrated⁸⁾ (Fig. 1). This signal was used as a quantitation signal to calculate the absolute purity. The labeled purity value of the reagent was 90.84%. After the reagent was adjusted under controlled humidity for 3 h, the purity of the reagent was re-determined by qNMR. The purity of GRB1 under controlled humidity of 26% (Lab. C), 43% (Lab. B), 43% (Lab. A), 60% (Lab. D), and 60% (Lab. E) was 92.58, 88.34, 87.19, 86.12, and 86.65%, respectively (Table 1). The humidity during weighing, determined by a digital hygrometer, was 26, 38, 45, 60, and 60%, respectively. Figure 2 shows the purity at each humidity level in Labs A–D. Purity tended to decrease as humidity increased; thus, the purity of the hygroscopic substances determined by qNMR is affected by humidity before and during weighing.

Fig. 1. Quantitative ¹H-NMR Spectrum of Ginsenoside Rb₁ (GRB1) in Methanol-d₄

Table 1. Purity (%) of Ginsenoside Rb₁ (GRB1) Prepared under Different Humidity Conditions across Five Independent Laboratories

Laboratory	C	B	A	D	E
Average (%)	92.58	88.34	87.19	86.12	86.65
S.D. (%)	0.12	0.12	0.15	0.06	0.10

*Purity (%) of ginsenoside Rb₁: 90.84% described in the certificate of analysis determined by qNMR. Standard deviation (S.D.).

Fig. 2. Purity (%) of GRB1 under Different Humidity Conditions across Four Independent Laboratories

The dotted line shows the purity (%) of GRB1 described in the certificate of analysis: 90.84%.

The attached certificate of analysis seemed to be performed in Europe. Our results suggested that the purity of the substances is determined by qNMR after weighing in a place in which the humidity is approximately 30% (Fig. 2). The results also suggested that the absolute purity of hygroscopic reference standards with an indication of purity (or water content) vary if the humidity during the time of preparation of the standard solution for HPLC is different from that at the time of water content determination, as most of the impurity is water.

Next, we examined the weight equilibration time to determine the appropriate duration of humidity control during weight measurement prior to purity analysis (Fig. 3). Under the condition at 25 °C and 60% humidity by TGA equipment, after 3 h (180 min) the rate of weight change per h was less than 0.1% (Fig. 3). Therefore, the humidity control conditions of GRB1 was set as 25 °C and 60% humidity for 3 h or more.

Fig. 3. Examination of Equilibration Time of Weight of GRB1

As mentioned in a previous paper,¹¹⁾ these results indicate that (i) humidity control before and during the weighing of hygroscopic reagents is critical for accurate assessment of purity and (ii) the absolute amount of reagent itself that is not affected by water content should be labeled when the evaluated reagents become commercially available.

SSA

The purity of SSA, which is also hygroscopic,¹¹⁾ was examined. Hosoe et al. reported that the use of the 12th-position signal (around 5.7 ppm), instead of the 11th-position signal (around 5.2 ppm), provide more accurate estimation of purity (quantitative value) without interference peaks for the analysis of SSA⁸⁾ (Fig. 4). Thus, we used the 12th-position signal for quantitation to calculate the absolute purity (quantitative value). First, the purity was determined without humidity control before and during weighing in three facilities (Table 2). The results showed a high variation in purity (89.45 ± 2.79%), which indicated that the substance requires humidity control both before and during weighing. SSA is relatively unstable in the dissolved state (data not shown). We concluded that immediate qNMR measurement is essential for accurate quantitation after weighing.

Fig. 4. Quantitative ¹H-NMR Spectrum of Saikosaponin a (SSA) in Methanol-d₄

Table 2. Purity (%) of Saikosaponin a Prepared under Non-controlled Humidity across Three Independent Laboratories

Laboratory	A	B	C	Average of three labs
Average (%)	86.37	90.18	91.81	89.45
S.D. (%)	0.13	0.20	0.13	2.79

The weight equilibration time of SSA was examined to determine the appropriate duration of humidity control before the weighing of SSA. Under the conditions of 20 °C and 60% humidity by the TGA equipment, the rate of weight change was less than 0.1% per hour after 3 h (180 min) (Fig. 5). Therefore, the humidity control conditions for SSA were set at 3 h or more at 20 °C and 60% humidity.

Fig. 5. Examination of Equilibration Time of Weight of SSA

The subsequent qNMR analysis after the weighing of the reagent with the consideration of the minimum weight (by an ultra-microbalance) under controlled humidity after 4 h of humidity adjustment (at approximately 60% humidity) by a saturated NaBr solution at 22–23 °C in three facilities showed good results with low variation (89.28 ± 0.076%) (Table 3). The humidity around the balance scale, monitored and recorded during weighing with a digital hygrometer in three facilities, was 56–57%.

Table 3. Purity (%) of Saikosaponin a Prepared after Humidity Controlling across Three Independent Laboratories

Laboratory	A	B	E	Average of three labs
Average (%)	89.29	89.36	89.20	89.28
S.D. (%)	0.20	0.10	0.16	0.076

These results suggested that, by stipulating the humidity control method before and during weighing, the reference standard users of reference standards for HPLC quantification are able to use the labeled purity value previously determined by qNMR in the regent company laboratory, even if the reference standard is hygroscopic.

BB

The 6th-position signal (around 7.2 ppm) of BB overlapped with the non-deuterated signal of 1,4-BTMSB-d₄ (i.e., the reference standard for qNMR) as a contaminant signal. No interference signal was observed around the 2nd- and 7th-position signals (around 6.8 ppm) and the 4th- and 5th-position signals (around 6.6 ppm), which were used as signals to be integrated (Fig. 6).

Fig. 6. Quantitative ¹H-NMR Spectrum of Barbaloin (BB) in Methanol-d₄

The weight equilibration time of BB was examined using TGA equipment to determine the appropriate duration of the humidity control before weighing (Fig. 7). The 160 min observation suggested that the time for BB should be set at 2 h or more at 25 °C and 60% humidity. As TGA analysis revealed that BB is less hygroscopic than SSA, we tried to weigh BB with or without humidity control after maintenance for more than 2 h under the 60% humidity by a saturated NaBr solution.

Fig. 7. Examination of Equilibration Time of Weight of BB

The subsequent qNMR analysis after the weighing of the reagent with consideration of the minimum weight with or without humidity control suggested that humidity control during weighing does not affect purity (Table 4). Namely, the purity of the controlled laboratories (A, D, E) was 92.35 ± 0.63(%) and that of non-controlled laboratories was 92.45 ± 0.65(%) (Table 4). As there was almost no difference in quantitative values, equilibration time (more than 2 h) for humidity control of BB was necessary before weighing, although a relatively quick weighing time (less than 30 min) should be considered.

Table 4. Purity (%) of Barbaloin Prepared across Six Independent Laboratories

		Humidity control during weighing			Humidity non-control during weighing
Position	Laboratory	A	D	E	B	C	F
2,7-H	Average	91.73	92.97	92.16	91.73	93.26	92.03
	S.D.	0.15	0.26	0.08	0.11	0.08	0.15
4,5-H	Average	91.85	93.07	92.30	92.18	93.12	92.38
	S.D.	0.39	0.09	0.04	0.09	0.07	0.29
Average (%) of 2,7-H and 4,5-H		91.79	93.02	92.23	91.95	93.19	92.20
Average (%) of three labs		92.35			92.45
S.D. (%)		0.63			0.65

Hygroscopic Substances and Absolute Purities Determined by qNMR

We examined an optimal and reproducible sample preparative approach on hygroscopic compounds, GRB1, SSA, and BB to qNMR. The results showed the importance of humidity control before and/or during weighing. This study used a saturated salt solution for humidity control. Since laboratory users normally do not want to set a saturated salt solution in a balance box, a more convenient method would be preferable. To solve this problem, a constant temperature and humidity box with a built-in balance has been recently developed as a humidity control system. The box can be used in future studies to determine the purity of hygroscopic substances.

Experiment

Facilities

Six investigators from six laboratories (A–F) separately performed an experiment in all operations. The experiment of GRB1 was performed in five laboratories (A–E).

Samples

Reference Standards for qNMR and Solvents

In this study, 1,4-BTMSB-d₄ (1,4-bis(trimethylsilyl)benzene-d₄, MW = 226.50), a certified reference material (NMIJ CRM), was purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan) and used as the calibration standard for qNMR. The material is the same as the certified reference material that has obtained accreditation from the Accreditation System of National Institute of Technology and Evaluation (ASNITE) (Guide34). Methanol-d₄ with the following deuteration rate was used as a solvent for qNMR determination: Lot. TA2142V (99.96 atom % D), EW1886 (99.96 atom % D), and TV1921 (99.96 atom % D) by Isotec Inc. (NY, U.S.A.); Lot. A0355966 (100 atom % D), A0357061 (99.9 atom % D), A0260657 (99.96 atom % D), and A0359314 (99.96 atom % D) by Acros Organics (NJ, U.S.A.); Lot. ECG4123 (99.5 atom % D) and Lot. KPH4454 (99.8 atom % D) by FUJIFILM Wako Pure Chemical Corporation.

Reagents

GRB1 from Sigma-Aldrich (MO, U.S.A.) (Code No. 00170580; Lot. HWI01309, the purity described in the certificate of analysis 90.84%) was used in this study. SSA obtained from Tsumura & Co. (Tokyo, Japan) and FUJIFILM Wako Pure Chemical Corporation (Lot. AWE2493) was used in this study. BB (Lot. DSP2289) commercially available from FUJIFILM Wako Pure Chemical Corporation was used in this study.

Instruments and Equipment

The ultra-micro balance (Mettler Toledo UMX2, XP2U, XP2UV, Sartorius SE2) with a minimum reading of 0.0001 mg was used. JNM-ECA600 (600 MHz), JNM-ECZ600 (600 MHz), JNM-ECA800 (800 MHz), JNM-ECA500 (500 MHz), and JNM-ECZ400 (400 MHz) (manufactured by JEOL Ltd.) were used for NMR. The NMR spectra shown in the Figs. 1, 4, and 6 were all obtained by the same instrument, JNM-ECZ600 (600 MHz). Regarding the chemical shift values, a reference standard for qNMR was also used as the chemical shift reference signal (0 ppm). The δ value was expressed in ppm. Wilmad 528-pp, 535-pp, and 507-pp, Norell S3010 ST500-7, Wako 293-48151, and NMR Test Tube HG-Type were used as NMR sample tubes. A digital temperature-humidity atmospheric pressure gauge (Testo 622) by Testo Co., Ltd. was used as a temperature-humidity meter.

Preparation of Sample Solution

NMR Validation Test

Approximately 5 mg of each reagent and about 1 mg of reference standard for qNMR, which were precisely weighed and placed in the same vial together for each tare, were dissolved in NMR solvent (0.6–1 mL). In the sample solution, this solution (0.6 mL) was sealed in an NMR sample tube.

Humidity Control Conditions

Humidity control conditions for GRB1: Each two GRB1 samples were equilibrated for 3 h in a sealed container with a solution of saturated K₂CO₃ or saturated NaBr salt at 25 °C and 43 and 60% humidity, respectively. The observed laboratory humidity of the day was 26% and one sample was maintained without humidity control for 3 h at 25 °C. Sample weighing was performed under control humidity with or without the corresponding saturated salt solution in a balance box.

Humidity control conditions for SSA: SSA was equilibrated for at least 4 h in a sealed container with a saturated NaBr solution at 24 °C under controlled humidity (60%). Sample weighing was also performed under constant humidity (60%) controlled by a saturated NaBr solution in a balance box at 24 °C.

Humidity control conditions for BB: BB was equilibrated for at least 2 h in a sealed container with a solution of saturated NaBr salt at 24 °C under controlled humidity (60%). The observed humidity in the laboratory was from 25–33%. Sample weighing was performed with or without humidity control by a saturated NaBr solution.

Conditions for qNMR Determination

The observed spectrum width was 20–22 ppm. A digital filter was used. The center of the spectrum was set at 5 ppm. The pulse width was set to the time at which a 90-degree pulse was obtained. Acquisition time, 4 s; digital resolution, 0.25 Hz; and delay time, 60 s. An auto FG shim was used for shim adjustment. The determination temperature was set at room temperature (20–30 °C). ¹³C decoupling with MPF8 was performed. The dummy scan was performed twice. In principle, the measurement was performed three times for each sample in accordance with the internal standard method (AQARI: Accurate quantitative NMR with internal reference substance) to ensure that the S/N of the quantitative signal was 200 or higher. Alice 2 for qNMR, Purity Pro, and Delta ver.4.2.3 manufactured by JEOL Ltd. were used for NMR data processing. The trimethylsilyl peak of the reference standard for qNMR (1,4-BTMSB-d₄) was set at 0 ppm. Phase correction and baseline correction were performed by the manual method. The integration range for each peak was determined using the manual method. All integrated values in this study are expressed in terms of purity (%). The purity of the reagents was calculated using the following formula based on a previous study^12–15):

  

• I = signal area, H = number of protons, W = weight, M = molecular weight, P = purity (%), std, sample = reference standard for qNMR and sample

The following numbers were used for the calculation: number of protons of methyl groups in 1,4-BTMSB-d₄ (reference standard for qNMR), 18; molecular weight of 1,4-BTMSB-d₄, 226.50; molecular weight of GRB1, 1109.29; molecular weight of SSA, 780.98; and molecular weight of BB, 418.39.

TGA Equipment

TGA Q5000SA (TA Instruments) was used as the thermogravimetric analyzer. The conditions for the determination were as follows: purge gas, nitrogen; feed rate in a humidity chamber, 200 mL/min; balance, 10 mL/min; sample weight, 1.6–5.0 mg; sample pan, metalized quartz sample pan; temperature in weight equilibration determination, 20 or 25 °C; relative humidity, 60%.

Acknowledgments

This work was supported by the Research on Regulatory Science of Pharmaceuticals and Medical Devices and Research on Development of New Drugs from the Japan Agency for Medical Research and Development (AMED).

Conflict of Interest

The authors declare no conflict of interest.

References

• 1) The Ministry of Health, Labour and Welfare, Japan. “The Japanese Pharmacopoeia Seventeen Edition (2016).”:‹https://www.mhlw.go.jp/file/06-Seisakujouhou-11120000-Iyakushokuhinkyoku/JP17_REV_1.pdf›, cited 10 March, 2020.
• 2) The Ministry of Health, Labour and Welfare, Japan. “Supplement I to The Japanese Pharmacopoeia Seventeen Edition (2017).”: ‹https://www.mhlw.go.jp/content/11120000/000352508.pdf›, cited 10 March, 2020.
• 3) The Ministry of Health, Labour and Welfare, Japan. “Supplement II to The Japanese Pharmacopoeia Seventeen Edition (2019).”: ‹https://www.mhlw.go.jp/content/11120000/000597173.pdf›, cited 10 March, 2020.
• 4) The Ministry of Health, Labour and Welfare, Japan. “The Japanese Pharmacopoeia Seventeen Edition (2016), General Information, Quantitative Analytical Technique Utilizing Nuclear Magnetic Resonance (NMR) Spectroscopy and Its Application to Reagents in the Japanese Pharmacopoeia.”:‹https://www.pmda.go.jp/files/000217666.pdf›, cited 10 March, 2020.
• 5) Pharmaceuticals and Medical Devices Agency (PMDA), “JP drafts for public comments. Byakkokaninjinto Extract (3 June, 2019).”: ‹https://www.pmda.go.jp/files/000229715.pdf›, ‹https://www.pmda.go.jp/files/000229752.pdf (English)›, ‹https://www.pmda.go.jp/rs-std-jp/standards-development/jp/pub-comments/jp/0161.html›, ‹https://www.pmda.go.jp/english/rs-sb-std/standards-development/jp/pub-comments/jp/0013.html (English)›, cited 10 March, 2020. A request for comments on the original draft of the Japanese Pharmacopoeia (June 2019, part 2).
• 6) Hosoe J., Sugimoto N., Goda Y., Pharmaceutical and Medical Device Regulatory Science, 41, 960–970 (2010).
• 7) Hosoe J., Sugimoto N., Suematsu T., Yamada Y., Hayakawa M., Katsuhara T., Nishimura H., Goda Y., Pharmaceutical and Medical Device Regulatory Science, 43, 182–193 (2012).
• 8) Hosoe J., Sugimoto N., Suematsu T., Yamada Y., Miura T., Hayakawa M., Suzuki H., Katsuhara T., Nishimura H., Kikuchi Y., Yamashita T., Goda Y., Pharmaceutical and Medical Device Regulatory Science, 45, 243–250 (2014).
• 9) Goda Y., Pharmaceutical and Medical Device Regulatory Science, 48, 615–619 (2017).
• 10) Kato H., Bao X., Suzuki A., Ihara T., Bunseki Kagaku, 66, 375–380 (2017).
• 11) Uchiyama N., Hosoe J., Miura T., Sugimoto N., Ishizuki K., Yamada Y., Iwamoto Y., Suematsu T., Komatsu T., Maruyama T., Igarashi Y., Higano T., Shimada N., Goda Y., Yakugaku Zasshi, 140, 1063–1069 (2020).
• 12) Ohtsuki T., Sato K., Abe Y., Sugimoto N., Akiyama H., Talanta, 131, 712–718 (2015).
• 13) Tada A., Takahashi K., Sugimoto N., Suematsu T., Arifuku K., Saito T., Ihara T., Yoshida Y., Ishizuki K., Nishimura T., Yamazaki T., Kawamura Y., Shokuhin Eiseigaku Zasshi, 51, 205–212 (2010).
• 14) Sugimoto N., Tada A., Suematsu T., Arifuku K., Saito T., Ihara T., Yoshida Y., Kubota R., Tahara M., Shimizu K., Ito S., Yamazaki T., Kawamura Y., Nishimura T., Shokuhin Eiseigaku Zasshi, 51, 19–27 (2010).
• 15) Sato-Masumoto N., Nishizaki Y., Saito N., Yamazaki T., Numata M., Ihara T., Sugimoto N., Sato K., Jpn. J. Food Chem. Safety, 24, 75–81 (2017).