In Silico Prediction of N-Nitrosamine Formation Pathways of Pharmaceutical Products

Abstract

The recent discovery of N-nitrosodimethylamine (NDMA), a mutagenic N-nitrosamine, in pharmaceuticals has adversely impacted the global supply of relevant pharmaceutical products. Contamination by N-nitrosamines diverts resources and time from research and development or pharmaceutical production, representing a bottleneck in drug development. Therefore, predicting the risk of N-nitrosamine contamination is an important step in preventing pharmaceutical contamination by DNA-reactive impurities for the production of high-quality pharmaceuticals. In this study, we first predicted the degradation pathways and impurities of model pharmaceuticals, namely gliclazide and indapamide, in silico using an expert-knowledge software. Second, we verified the prediction results with a demonstration test, which confirmed that N-nitrosamines formed from the degradation of gliclazide and indapamide in the presence of hydrogen peroxide, especially under alkaline conditions. Furthermore, the pathways by which degradation products formed were determined using ranitidine, a compound previously demonstrated to generate NDMA. The prediction indicated that a ranitidine-related compound served as a potential source of nitroso groups for NDMA formation. In silico software is expected to be useful for developing methods to assess the risk of N-nitrosamine formation from pharmaceuticals.

Introduction

N-Nitrosamines are a class of organic compounds with the chemical structure O=N-NR₂ (where R is alkyl, aryl, and other groups) and are generally formed through reactions between a secondary amine and nitrous acid.¹⁾ Certain N-nitrosamines exhibit carcinogenic and mutagenic properties through metabolic activation in vivo. In recent years, pharmaceutical companies have voluntarily recalled pharmaceuticals found to be contaminated with N-nitrosodimethylamine (NDMA) and other N-nitrosamines, and this global recall and long-term suspension of supply has adversely impacted clinical practice. In response to this crisis, various regulatory authorities, including the U.S. Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the Ministry of Health, Labour and Welfare, now require manufacturers and distributors of pharmaceuticals to confirm that the concentration of N-nitrosamines are below stipulated limits.^2–4) According to ICH M7 guidelines, which focus on the evaluation and control of mutagenic impurities, N-nitrosamines are classified as a “cohort of concern”⁵⁾ composed of highly mutagenic carcinogens. N-Nitrosamines have recently been detected in some pharmaceuticals, e.g., NDMA in sartan, ranitidine, and metformin (Fig. 1). Contamination of these pharmaceuticals by NDMA is attributed to three main factors: (1) manufacturing process of the pharmaceutical (e.g., sartan),¹⁾ (2) degradation of the pharmaceutical (e.g., ranitidine),⁶⁾ and (3) chemical reactions between the pharmaceutical or its degradation products and ingredients in the container or packaging (e.g., metformin).^7,8) In the case of sartans, NDMA is believed to form through the reaction between nitrite and dimethylamine (DMA), which is generated during the decomposition of dimethylformamide (DMF), a solvent used in the manufacturing process.⁹⁾ Although the source of ranitidine contamination has not been completely elucidated, NDMA can form even from ranitidine alone, leading to the hypothesis that NDMA arises from the dimethyl and nitro groups in the ranitidine molecule.^10,11) Therefore, information about known impurities specified in individual monographs of Pharmacopoeias could be useful to predict NDMA self-production. NDMA may form through reactions between the dimethylamino group or degradation products of metformin and nitrocellulose of press-through-package (PTP) sheets, which serve as the container packaging.

Fig. 1. Pharmaceuticals Recalled Owing to NDMA Contamination

Owing to the high reactivity of N-nitrosamines, such as NDMA, with DNA, the concentrations of N-nitrosamines must be three or more orders of magnitude lower than those of typical mutagenic impurities (tens to hundreds of ppb). When N-nitrosamine contamination arises during the research, development, or production phases of pharmaceuticals, it becomes crucial to investigate the cause of the contamination and devise effective prevention measures. Apart from evaluating the manufacturing processes of the relevant pharmaceutical product, it is also imperative to assess alterations in storage procedures and validate the resulting changes. In severe scenarios, the contamination can significantly hinder the advancement of the pharmaceutical. As a result, there exists a compelling necessity for techniques capable of forecasting the likelihood of N-nitrosamine contamination. While certain studies have predicted the formation of degradation products and stability of select pharmaceuticals, no studies to date have predicted or demonstrated the formation of N-nitrosamines. In this context, developing methods to predict the formation of high-risk impurities, including N-nitrosamines, during the manufacturing and storage of pharmaceuticals is of utmost importance. Considering the increasing number of pharmaceuticals identified as potential N-nitrosamine precursors,¹²⁾ investigating each in detail would be both costly and labor-intensive. Hence, there is a need for streamlined and cost-effective approaches to assess and address the risk of pharmaceutical contamination by N-nitrosamines.

To address the abovementioned problem, we aimed to develop a method to assess the risk of N-nitrosamine formation by referring to information about impurities specified in the purity tests of the United States Pharmacopeia (USP) and European Pharmacopoeia (EP). We predicted the pathway of N-nitrosamine formation by investigating the degradation reactions of model pharmaceuticals, namely gliclazide and indapamide, using Zeneth^®,^13,14) a commercially available knowledge-based software designed for predicting the degradation pathways of organic compounds. This software uses information about structural transformations during forced degradation studies to predict the degradants of organic compounds. The knowledge base of the software covers various degradation reactions, such as hydrolysis, oxidation, and photoreaction, and each predicted degradation reaction is assigned score (values from 0 to 1000) indicating the likelihood of the reaction occurring (Supplementary Fig. S1). The accuracy of predictions has been improved through the accumulation and refinement of the knowledge base. Zeneth^® has been used to predict the degradation products of pharmaceuticals under forced degradation conditions,^15–17) although the dependency of the N-nitrosation transformation step scores on the conditions under which nitrosamine degradants are experimentally observed has not been demonstrated in the literature. In the present study, we validated the reaction conditions and associated scores of the predicted degradation pathways by analyzing actual degradation products using chemical ionization mass spectrometry (APCI–MS).

Results and Discussion

In Silico Prediction of the Degradation Pathways of Pharmaceuticals

In this study, we selected gliclazide and indapamide as model pharmaceuticals (Fig. 2) for the following reasons: 1) Although the formation of N-nitrosamines in pharmaceuticals is generally known to occur through the reaction of secondary amines with nitroso sources, such as nitrous acid, the formation of N-nitrosamine through the oxidation of hydrazine is less known.^18–20) 2) Inherent impurities of these pharmaceuticals are well defined, and thus N-nitrosamine formation can be assessed through the degradation reactions of these pharmaceuticals and analogues. Additionally, we predicted the degradation product formation pathways using ranitidine, a pharmaceutical known to produce NDMA.

Fig. 2. Chemical Structures of (A) Gliclazide and Impurities (Imps.) A, B, C, D, E, F, and G and (B) Indapamide and Imps. A, B, and C

Initially, we conducted in silico prediction of the degradation pathway of gliclazide, using Zeneth 9 (Lhasa Limited, KB 2022.1.2). Gliclazide is a sulfonylurea-based oral hypoglycemic drug used in the treatment of type 2 diabetes, and one of its inherent impurities is gliclazide impurity B (EP), a compound with an N-nitrosamine structure. Prediction of the degradation pathway of gliclazide revealed two routes: hydrolysis of the urea structure to produce D1, and photolysis to produce D2 (Fig. 3). D1 is a hydrazine structure, which upon oxidation produced the N-nitrosamine structure known as gliclazide impurity B. In contrast, D2 did not undergo further degradation reactions. However, if hydrolysis of the urea moiety occurred, D2 transformed into D1, and subsequent oxidation of D1 generated gliclazide impurity B. Interestingly, the step score for the degradation pathway from D1 to gliclazide impurity B was different under different pH conditions. The step scores were 100, 500, and 700 at pH 1, 7, and 14, respectively. This indicated that the reaction likelihood depended on the pH.

Fig. 3. Structures of Gliclazide and Predicted Degradation Products

Decomposition pathways predicted by Zeneth are indicated by solid lines. Pathways not predicted by Zeneth are indicated by dashed lines. Detected compounds are enclosed in squares. Values in curly brackets are Zeneth-step likelihood scores.

We also carried out in silico prediction of the degradation pathway of indapamide, a thiazide-like diuretic drug used for hypertension. Indapamide has analogues containing the N-nitrosamine structure, specifically indapamide impurity A (EP). According to the predicted degradation pathway, hydrolysis followed by oxidation of indapamide produced indapamide impurity A, similar to the degradation of gliclazide. Additionally, the software predicted the formation of D3 and D5 structures with dehydrogenated indoline backbones (Fig. 4).

Fig. 4. Structures of Indapamide and Predicted Degradation Products

Decomposition pathways predicted by Zeneth are indicated by solid lines. Pathways not predicted by Zeneth are indicated by dashed lines. Detected compounds are enclosed in by squares. Values in parentheses are Zeneth-multiplied step pathway likelihood scores.

Detection of Degradants from Gliclazide and Indapamide

To validate the predicted degradation pathway, APCI–MS in combination with standards was used to analyze the actual degradation products as well as each pharmaceutical and related compounds (Supplementary Figs. S2 and S3). First, we checked whether the oxidation of D1, which comprised a hydrazine structure, produced gliclazide impurity B. The experimental results showed that in the presence of hydrogen peroxide (H₂O₂) serving as the oxidant, D1 oxidized to form gliclazide impurity B with the N-nitrosamine structure (Fig. 5A). Because in silico prediction showed that the step score for each degradation pathway from D1 to gliclazide impurity B was different under different pH conditions, model reactions were evaluated under different pH conditions (pH 1, 7, or 14). The results confirmed that gliclazide impurity B formed under all pH conditions. Notably, the formation of gliclazide impurity B was significantly pronounced at pH 14 (Supplementary Fig. S4). Interestingly, the results suggested a correlation between the in silico assigned step scores of predicted degradation pathways and the reactivity of actual degradation reactions. This is probably due to both the higher rate of decomposition of hydrogen peroxide under alkaline pH conditions compared to acidic pH and the lower ratio of protonation at the amino group of hydrazine. Prior studies have also reported that amine oxidation is easier under alkaline conditions than acidic conditions.^16,21) We proceeded to verify that gliclazide degraded to form D1 and subsequently gliclazide impurity B at pH 14. The results confirmed that gliclazide impurity B indeed formed when gliclazide was treated with H₂O₂ at pH 14 (Fig. 5B). This finding strongly supports the predicted degradation pathway in which hydrolysis of the urea structure and subsequent oxidation of the hydrazine moiety lead to the N-nitrosamine structure. According to these results, one effective approach to prevent N-nitrosamine contamination of this drug is to avoid alkaline conditions and excipients containing H₂O₂.

Fig. 5. APCI–MS Spectra of Degradation Products

(A) Gliclazide Imp. B standard, D1 standard, and D1 in the presence of H₂O₂ at pH 14. (B) Gliclazide standard and gliclazide in the absence or presence of 3% H₂O₂ at pH 14.

The degradation products of indapamide were investigated using a model reaction under the same condition as that for gliclazide (Fig. 6A). The outcomes revealed that indapamide impurity A, an N-nitrosamine derivative, also formed through hydrolysis followed by oxidation, which released hydrazines. Moreover, the model reaction leading to the formation of indapamide impurity C (EP), which was expected to occur through the hydrolysis of indapamide, also led to the formation of indapamide impurity A (Fig. 6B). Additionally, D3, a compound with the dehydrogenated indoline structure, was also detected in the reaction. These results support the in silico prediction in which N-nitrosamine forms through a specific degradation pathway.

Fig. 6. APCI–MS Spectra of Degradation Products

(A) Indapamide standard and indapamide in the presence of 3% H₂O₂ at pH 14. (B) Indapamide Imp. C standard and indapamide Imp. C in the presence of 3% H₂O₂ at pH 14.

Structure–Activity Relationship (SAR) for N-Nitrosamines

The mutagenic potential of the identified N-nitrosamines was evaluated through SAR analysis employing Derek Nexus (Version 6.2.1, KB 2022 2.0) (Lhasa Limited). Derek uses expert derived structural alerts, defined by the relationship between compound structures and toxicological effects, to predict potential hazard. This SAR approach was employed to evaluate the mutagenicity of gliclazide impurity B and indapamide impurity A, which were both rated PLAUSIBLE. In fact, the skeletal structure of gliclazide impurity B has been classified as exhibiting a higher TD₅₀ (the dose required to induce 50% tumor incidence in a lifetime exposure) than that of indapamide impurity A.²²⁾

In Silico Prediction of Degradation Pathway of Ranitidine

Finally, in silico prediction was performed to predict the degradation pathway of ranitidine, a pharmaceutical known to produce NDMA (Fig. 7). While no direct degradation pathway leading to NDMA formation was predicted, the software did predict the formation of DMA, a key precursor necessary for NDMA generation. Using the software, the compatibility between DMA and nitrous acid, a nitroso source in excipients, was assessed and a formation of NDMA was predicted (Fig. 8).

Fig. 7. Chemical Structures of Ranitidine Hydrochloride and Impurities (Imps.) A, B, C, D, E, F, G, H, I, J, and K

Fig. 8. Structures of Ranitidine and Predicted Degradation Products

Decomposition pathways predicted by Zeneth are indicated by solid lines. Pathways not predicted by Zeneth are indicated by dashed lines. Values in parentheses are Zeneth-multiplied step pathway likelihood scores.

Recently, several pharmaceuticals have been recalled owing to contamination by N-nitrosamines derived from the structures of active pharmaceutical ingredients (APIs), referred to as nitrosamine drug substance related impurities (NDSRIs).¹²⁾ The presence of these NDSRIs may result from the reaction of a secondary or tertiary amine structure of the pharmaceutical with a nitroso source. These reactions may be predicted by considering the coexistence of APIs and excipients (e.g., nitrous acid), a capability offered by Zeneth^®.

Additionally, the formation of ranitidine-related compound H (USP) (Imp. H) was predicted. Previous studies have shown that ranitidine Imp. H serves as a nitroso source and reacts with DMA, leading to the production of NDMA.¹⁰⁾ These results indicate that even when the direct formation of NDMA is not predicted in the absence of an obvious nitroso source, the risk of N-nitrosamine formation should be assessed by considering the chemical reactivity of the degradation products and exercising appropriate judgment.

Conclusion

In this study, we used Zeneth^® to predict the degradation pathways of model pharmaceuticals, namely gliclazide and indapamide, and validated the predictions by analyzing actual degradation products using APCI–MS. In the case of gliclazide, we confirmed the formation of gliclazide impurity B, an N-nitrosamine, in the presence of H₂O₂, especially under alkaline conditions, which validated the Zeneth-step likelihood score. According to the pathway predicted using the expert-knowledge system, the N-nitrosamine form was generated through hydrolysis of the urea moiety followed by oxidation of the hydrazine structure. Moreover, we confirmed that the corresponding N-nitrosamine of indapamide was generated along a similar degradation pathway. The prediction indicated that the degradation of ranitidine, a drug known to produce NDMA, did not directly lead to the formation of NDMA unless an obvious nitroso source, such as nitrous acid, was present. However, the software did predict the formation of another degradation product, ranitidine-related compound H (USP), which served as a potential nitroso source for the generation of N-nitrosamine.

It is essential to acknowledge that not all predicted degradation products may actually form or can be detected. However, for the risk management of highly mutagenic substances, ‘sensitivity’ (i.e., the number of truly observed degradants that are predicted by the system) holds significant importance.¹⁴⁾ It is crucial to recognize that the actual formation of degradation products depends on various factors, such as temperature, pH, and packaging conditions, among others. These factors influence the likelihood of specific degradation pathways and whether degradants predicted by the software are experimentally observed. Despite possible variations, focusing on ‘sensitivity’ would help in effectively managing the risks associated with highly mutagenic substances.

The global concern over N-nitrosamine contamination of pharmaceuticals persists, necessitating ongoing risk assessments to prevent contamination. The potential carcinogenicity of the newly identified N-nitrosamines necessitates comprehensive examination, taking into account the carcinogenic potency categorization approach (CPCA) recently introduced by the EMA for assessing the carcinogenic risk posed by nitrosamines found in pharmaceutical products.²³⁾ To this end, evaluation techniques, such as (Q)SAR, should be considered.

While the ultimate evaluation of in silico prediction results requires expert judgment, the continuous accumulation of knowledge from future forced degradation studies is anticipated to continually enhance the accuracy of predictions. Through in silico prediction of degradation products, we can effectively assess the possibility of N-nitrosamine formation from pharmaceuticals and their impurities. This information can be useful in implementing contamination prevention measures and developing risk assessment methods. Moreover, this approach aligns with the goal of medicinal chemistry, which is to expedite the development of novel and safe pharmaceuticals for patients. With the aid of predictive tools, researchers can swiftly bring new medications to patients with ensured safety. This is expected to enable the prompt and steady supply of high-quality pharmaceutical products to patients and the market.

Experimental

Chemicals and Reagents

Gliclazide (>98.5%, Part# G0381) was purchased from Tokyo Chemical Industry (Tokyo, Japan), Indapamide (98.0 + %, Part# 091-04831) was purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Hexahydrocyclopenta[c]pyrrol-2(1H)-amine hydrochloride (97.00%, Part# A165954) and 2-methylindolin-1-amine hydrochloride (97.00%, Part# A1152602) were purchased from Ambeed (Arlington Hts, IL, U.S.A.). Gliclazide impurity B, Indapamide impurity A, and Indapamide impurity C were prepared according to the reported method.^24–26)

In Silico Prediction of Degradation Pathway of Compounds

Degradation products from gliclazide and indapamide were predicted by Zeneth^® (Version 9.0, KB 2022.1.2) (Lhasa Limited, Leeds, U.K.). In silico prediction was performed With the default parameters, including all the transformations (condensations and additions, eliminations and fragmentations, hydrolysis, isomerisations and rearrangements, oxidations and photochemical reactions) with the exception that pathway likelihood was selected as multiplied step likelihood (conditions; pH 1, 7, or 14 at 80 °C, maximum number of steps in a pathway: 4, maximum number of degradants was 1000, minimum score in multiplied step pathways: 0). Zeneth scores, a measure of the likelihood of the reaction occurring are as follows; 0 : impossible, 1–199 : very unlikely, 200–399 : unlikely, 400–599 : equivocal, 600–799 : likely, 800–999 : very likely, 1000 : certain.

Sample Preparation for Decomposition Reaction

To a solution of compounds in MeOH (2 mg/mL) was added an equal volume of aqueous solution [pH 1 (adjusted with HCl aq.), 7 or 14 (adjusted with NaOH aq.)], then added one-tenth volume of 30% H₂O₂ (final concentration : approximately 3%). The resulting mixture (approximately 1 mg/mL) in glass vials with hermetic caps were heated at 80 °C in an oil bath for 1 h.

Detection of Degradation Product

The detection of degradation products in the degradation reaction was performed using Expression CMS^®-S (Advion Interchim Scientific, Ithaca, NY, U.S.A.). Mass calibration was carried out using APCI/APPI Tuning Mix (Part# G2432A, Agilent Technologies, Santa Clara, CA, U.S.A.). After cooling to room temperature, the reaction mixture was directly introduced into the APCI-MS probe using a melting point capillary (ASAP, positive ion mode).

Structure–Activity Relationship (SAR) Assessment

SAR assessments were performed by Derek Nexus (Version 6.2.1, KB 2022 2.0) (Lhasa Limited, Leeds, U.K.). Mutagenicity in vitro of gliclazide impurity B and indapamide impurity A were predicted with the default setting and the results were both PLAUSIBLE.

Acknowledgments

The authors would like to express their deepest appreciation to Dr. Mariko Matsumoto and Dr. Yasumasa Murata (National Institute of Health Sciences) for their technical support in the SAR analysis. This study was supported in part by grants from AMED under Grant numbers JP23mk0101220 (to Y.D.) and JP21mk0101208 (to Y.D., E.Y., N.U., S.M.).

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1) Beard J. C., Seager T. M., J. Org. Chem., 86, 2037–2057 (2021).
• 2) U.S. Food and Drug Administration, “Information about Nitrosamine Impurities in Medications.”: ‹https://www.fda.gov/drugs/drug-safety-and-availability/information-about-nitrosamine-impurities-medications›, cited 3 August, 2023.
• 3) European Medicines Agency, “Nitrosamine impurities.”: ‹https://www.ema.europa.eu/en/human-regulatory/post-authorisation/referral-procedures/nitrosamine-impurities›, cited 3 August, 2023.
• 4) Ministry of Health Labour and Welfare of Japan, “Self-inspection on Risks of Contamination with Nitrosamines in Drugs.”: ‹https://www.pmda.go.jp/files/000243438.pdf›, cited 31 March, 2020.
• 5) European Medicines Agency, “ICH guideline M7(R1) on assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk.”: ‹https://www.ema.europa.eu/en/documents/scientific-guideline/ich-guideline-m7r1-assessment-control-dna-reactive-mutagenic-impurities-pharmaceuticals-limit_en.pdf›, cited 3 August, 2023.
• 6) King F. J., Searle A. D., Urquhart M. W., Org. Process Res. Dev., 24, 2915–2926 (2020).
• 7) Zmysłowski A., Książek I., Szterk A., Molecules, 25, 5304 (2020).
• 8) Jireš J., Kalášek S., Gibala P., Rudovský J., Douša M., Kubelka T., Hrubý J., Řezanka P., J. Pharm. Biomed. Anal., 195, 113877 (2021).
• 9) Masada S., Tsuji G., Arai R., Uchiyama N., Demizu Y., Tsutsumi T., Abe Y., Akiyama H., Hakamatsuka T., Izutsu K., Goda Y., Okuda H., Sci. Rep, 9, 11852 (2019).
• 10) Yokoo H., Yamamoto E., Masada S., Uchiyama N., Tsuji G., Hakamatsuka T., Demizu Y., Izutsu K., Goda Y., Chem. Pharm. Bull., 69, 872–876 (2021).
• 11) Harmon P., J. Pharm. Sci., 112, 1220–1224 (2023).
• 12) Schlingemann J., Burns M. J., Ponting D. J., Avila C. M., Romero N. E., Jaywant M. A., Smith G. F., Ashworth I. W., Simon S., Saal C., Wilk A., J. Pharm. Sci., 112, 1287–1304 (2023).
• 13) Parenty A. D. C., Button W. G., Ott M. A., Mol. Pharm., 10, 2962–2974 (2013).
• 14) Kleinman M. H., Baertschi S. W., Alsante K. M., Reid D. L., Mowery M. D., Shimanovich R., Foti C., Smith W. K., Reynolds D. W., Nefliu M., Ott M. A., Mol. Pharm., 11, 4179–4188 (2014).
• 15) Radić I., Runje M., Babić S., J. Pharm. Biomed. Anal., 201, 114091 (2021).
• 16) Golla V. M., Kushwah B. S., Dhiman V., Velip L., Samanthula G., J. Pharm. Biomed. Anal., 227, 115280 (2023).
• 17) Dhiman V., Chavan B. B., Ramarapu N., Samanthula G., Eur. J. Mass Spectrom., 29, 123–131 (2023).
• 18) Lunn G., Sansone E. D., Andrews A. W., Zeiger E., Environ. Mol. Mutagen., 17, 59–62 (1991).
• 19) López-Rodríguez R., McManus J. A., Murphy N. S., Ott M. A., Burns M. J., Org. Process Res. Dev., 24, 1558–1585 (2020).
• 20) Wichitnithad W., Nantaphol S., Noppakhunsomboon K., Rojsitthisak P., Saudi Pharm. J., 31, 1558–1585 (2023).
• 21) Lunn G., Sansone E. B., Chemosphere, 29, 1577–1590 (1994).
• 22) Dobo K. L., Kenyon M. O., Dirat O., Engel M., Fleetwood A., Martin M., Mattano S., Musso A., McWilliams J. C., Papanikolaou A., Parris P., Whritenour J., Yu S., Kalgutkar A. S., Chem. Res. Toxicol., 35, 475–489 (2022).
• 23) European Medicines Agency, “Questions and answers for marketing authorisation holders/applicants on the CHMP Opinion for the Article 5(3) of Regulation (EC) No 726/2004 referral on nitrosamine impurities in human medicinal products.”: ‹https://www.ema.europa.eu/en/documents/referral/nitrosamines-emea-h-a53-1490-questions-answers-marketing-authorisation-holders/applicants-chmp-opinion-article-53-regulation-ec-no-726/2004-referral-nitrosamine-impurities-human-medicinal-products_en.pdf›, cited 9 August, 2023.
• 24) Jiang A., CN106518747 A (2012).
• 25) Guo W., Cai L., Xie Z., Mei W., Liu G., Deng L., Zhuo X., Zhong Y., Zou X., Zheng L., Fan X., Org. Chem. Front, 8, 3838–3846 (2021).
• 26) Sączewskf F., Wasilewska A., Hudson A. L., Ferdousi M., Rybczyńska A., Boblewski K., Lehmann A., Acta Pol. Pharm., 72, 277–287 (2015).