In the pharmaceutical industry, the development of a drug candidate comes after the drug has been discovered during the research and discovery process and tested in animals during the preclinical studies for in vitro and in vivo safety and efficacy assessments. Those assessments predict the dose or dose range for human clinical trials within the therapeutic window of safety and efficacy of the drug. Drug development is the stage of research and development (R&D) that aims to produce a marketable form of the drug candidate and gain commercial approval supported by data from human clinical trials based on the requirements of worldwide regulatory agencies. The drug production is scaled up in the manufacturing process to provide high-quality material for human clinical trials and commercial manufacturing. This support is divided into two main categories for the drug candidate: drug substance and drug product. Drug substance is the chemical compound or drug candidate, also named as active pharmaceutical ingredient (API), and the drug product is the drug substance in the formulation form when mixed with the excipients or inactive ingredients. The drug product is the final form of the drug that patients will receive during the clinical trials and when the drug becomes commercial.

Nuclear magnetic resonance (NMR) is an analytical technique widely used during the development process of drug substances and drug products. In the case of the drug substances, NMR provides support for the structure elucidation of the drugs and other chemicals involved in their production, quantitation, impurity characterization from in-process development and stability studies, reaction monitoring of the chemical processes to synthesize the drug substances, and solid characterization of the drug substances. For the drug products, NMR supports structure elucidation of impurities caused by the interaction of the drugs with components of the formulation during stability studies, solid characterization of the drugs in the formulation, and quantitation of the drug substances in the solid forms. In this chapter, we discuss the contributions of NMR in the pharmaceutical development with examples to emphasize its significant role supporting the development of drug substances and drug products required by worldwide regulatory agencies to conduct human clinical trials and support their commercialization.

NMR is widely applied during the development stages of drug substances, which focuses on the scale-up of the manufacturing production to provide a sufficient amount to support human clinical trials and commercial manufacturing. Structure elucidation is the most important application of NMR for the structural analysis of starting materials, intermediates, and drug substances. The characterization of the in-process impurities has a critical role for process understanding during scale-up manufacturing and to improve the chemistry with the objective of producing high-quality drug substances materials. NMR is also applied for the analysis of the structures of impurities occurring during the stability studies of the drug substances to determine if their structures are related to toxicity including genotoxicity. Quantitation by NMR (qNMR) is becoming an important assay to determine the purity of materials, drug substances, and reaction yields. Recently, new NMR methods have been applied towards the analysis of the relative configuration of drug substances with more than two stereocenters. Reaction monitoring by NMR is becoming a trend for kinetic analysis and speciation for detailed process understanding of reactions to improve the synthetic production of drug substances. Finally, NMR is also applied for the solid characterization of drug substances and their forms by solid-state NMR (SSNMR). We discuss those topics with recent examples from the literature.

NMR is the standard technique to determine the structure characterization of organic compounds and is well applied to the pharmaceutical industry ranging from starting materials, intermediates, impurities, and drug substances. During the scale-up production of the drug substances, every batch produced must meet certain specifications criteria to provide the appropriate quality of the material. With regards to the structure, NMR and mass spectrometry (MS) are the main techniques employed to prove the structures of the organic materials for the synthesis of the drug substances. Typical NMR experiments used to determine the structures of organic compounds, with drug substances as example, in solution are focused on proton and carbon information from proton (¹H) and carbon (¹³C) chemical shifts and proton scalar J-coupling constants (from 1D ¹H NMR and 1D ¹³C NMR experiments), proton–proton correlations through bond and space (e.g., 2D ¹H,¹H-COSY, 1D ¹H TOCSY, 2D ¹H,¹H-TOCSY, 1D ¹H NOE, 2D ¹H,¹H-NOESY, 1D ¹H ROE, 2D ¹H,¹H-ROESY experiments), and proton–carbon bond correlations (e.g., 2D ¹H,¹³C-HSQC and 2D ¹H,¹³C-HMBC, ¹H,¹³C-1,1-ADEQUATE, ¹H,¹³C-HSQC-TOCSY, ¹H,¹³C-HMBC-TOCSY experiments). If the drug substances have fluorine atoms in their structure, 1D ¹⁹F NMR provides the chemical shift information of the fluorine atoms, but 1D ¹H NMR and 1D ¹³C NMR experiments can determine the location of the fluorine atoms in the molecule indirectly through the scalar J-coupling constants of ¹H–¹⁹F and ¹³C–¹⁹F. 2D ¹H–¹⁹F HMBC can also help determining the position of the fluorine atom in molecules with protons near the location of the fluorine in the molecule. Nitrogen is another element commonly found in drug substances; therefore, ¹H–¹⁵N HSQC and ¹H–¹⁵N HMBC experiments can provide the location and the chemical shifts of nitrogen atoms in molecules with protons near the nitrogen atoms. Extensive literature and books have been written over the years detailing those experiments, which is not the purpose of this chapter, but some examples are given in the reference list.^1–4 

The complete structural analysis of drug substances is normally not disclosed to the public because their molecular structures and their chemical and physical characteristics may be protected by patents from the originator pharmaceutical companies or institutions under intellectual property laws around the world. Occasionally, the structures of potential drug candidates are in the public domain as part of the medicinal chemistry studies, where some of those candidates may become commercial drugs if they succeed in reaching the targeted end points of safety and efficacy during human clinical trial studies and are approved by regulatory institutions.

In our laboratories, we recently tested a new magnet technology towards the application of structure elucidation of typical drug substance structures. The purpose of the study was to demonstrate the capability of a 9.4 T (400 MHz for the ¹H observation frequency) high-temperature superconducting (HTS) power-driven magnet NMR system with standard electronics using a Bruker AVANCE III HD Nanobay console and a 5 mm Bruker BBFO probe, installed in the chemistry laboratory because the magnet does not require liquid cryogens and the related infrastructure. Three drug substances were tested: cinacalcet HCl (Figure 1.1) and two investigational drug candidates.^5 ,6 

The results were compared to a commercial 9.4 T NMR with a standard cryogen magnet providing satisfactory results for the structure elucidation by 1D ¹H, ¹³C, and ¹⁹F NMR and 2D (¹H, ¹H) homonuclear and heteronuclear (¹H, ¹³C) NMR experiments. Figures 1.2 and 1.3 illustrate the differences of the 1D ¹H NMR spectra of cinacalcet HCl for the aliphatic and aromatic regions when acquired at 400 MHz in a conventional LTS magnet with BBFO Smart probe and BOSS3 shims compared to our HST magnet with BBFO probe and BOSS3 shims and also with QNP probe and BOSS1 shims. The QNP probe with BOSS1 shims seem to better match the coil length and design of the HTS magnet than the BBFO probe with BOSS3 shims.^5 ,6 

In the regulatory environment, a batch of the drug substance with high purity (99.5% or greater) and quality (e.g., low-level impurities, no genotoxic impurities) is produced to determine the specifications of the material for identity, potency, quality, and purity. This batch is considered the drug substance primary reference standard and is used to compare with other drug substance manufacturing batches to determine if they meet the specifications for clinical use. Complete characterization of the primary reference standard by NMR and other analytical techniques is essential for authentic confirmation of the structure of the drug substance. The assays used to determine the specifications must be performed in the same conditions for every batch produced. In the case of NMR, the deuterated solvent and temperature of the experiments must be the same to avoid changes in chemical shifts due to different experimental conditions when comparing the drug substance reference standard with other batches for reliability of the analysis. When a drug substance is commercial, pharmacopeia reference standards may be available to compare with commercial manufacturing batches of the drug. However, when the pharmacopeia reference is not available, the primary reference standard of the drug substance is used to validate all the assays for qualification of other batches. Merkulov et al.⁷  developed an NMR validation method based on ¹³C spectroscopy to confirm the authenticity of buserelin acetate (Figure 1.4) as a pharmacopoeia analysis to be used for drug substance pharmaceutical business import because the drug substance was not produced in Russia based on the author’s information. Buserelin acetate, a synthetic analogue of the natural gonadotropin-releasing hormone, is a nonapeptide used to treat many conditions (e.g., infertility, endometriosis). The study was performed in two different deuterated solvents (D₂O and DMSO-d₆) for 1D ¹H and ¹³C NMR with the purpose to determine the best conditions for best separation of the resonances of the drug substance in solution. The results of the study indicated that D₂O was the preferred solvent for better resolution of the resonances (¹H and ¹³C) of buserelin acetate, and that condition was selected to validate the NMR method for the pharmacopoeia analysis.

Small-molecule drug substances may come from the isolation of natural product sources or from organic synthesis. In the case of synthetic drugs, the characterization of the starting materials and intermediates is necessary towards the production of the drug substances. The NMR characterization of the materials is the same as for the drug substances, described in Section 1.2.1 of this chapter.

The synthetic process to prepare drug substance molecules may be complex depending on the structural complexity of the molecules. When the drug candidate or potential drug substance is synthesized during medicinal chemistry for in vitro and in vivo animal testing, only small quantities are produced, and the synthetic process is not optimized. When the drug candidate becomes a drug substance to be tested in human clinical trials, the synthetic process requires optimization to increase its production yield, make the whole process environmentally friendly, and lower the total production cost. All the synthetic steps of the drug substance production from medicinal chemistry have to be re-evaluated during the drug development process to scale up the production to support human clinical trials in preparation for commercial manufacturing synthesizing high-quality drug substance material. During phase I and II of human clinical trials, process chemists and chemical engineers investigate every step of the synthesis to determine the best synthetic processes for the production of the drug substance. By the time the drug substance reaches phase III human clinical trials, the objective is that the whole synthetic process must be established for manufacturing, including commercial manufacturing, and the process is locked down. This is a requirement from the worldwide regulatory agencies to make sure high-control processes are implemented for the best quality of the drug substance production. The use of lower-cost materials, one-pot reactions, and control of regioisomers are some of the many examples to improve the scale-up synthesis. We will show a few cases to illustrate some of these points and how NMR has impacted the chemical processes.

Multistep reactions require more controls and provide lower yield of the desired product overall. Understanding the reactions to reduce them as one single or one-pot reaction where all the reactants are added facilitates the control and better yield of the product. One-pot reactions are more successful to be scaled up than several reactions to obtain the same final product yield. Lu et al.⁸  developed the chemistry of synthesizing substituted pyrimidinones from a three-step process to one-pot reaction. Their interest in pyrimidinones was due to the fact that aryl/pyridyl pyrimidinones are inhibitors of p38 MAP kinase (p38α) and proposed for the treatment of inflammatory disorders. Low yield was observed when the synthesis was carried out in three steps with the requirement to crystallize one of the intermediates to improve the yield of the corresponding pyrimidinone. The authors described the one-pot synthesis of 5-aryl-3-methyl-2-methylsulfanyl-6-pyridin-4-yl-3H-pyrimidin-4-one 3 from arylacetic acid ethyl ester 1 (Figure 1.5). All the structures synthesized with this one-pot method were determined by NMR, but the most significant contribution by NMR was the determination of the double bond configuration of the isolated intermediate 2 to be the (Z)-2 instead of the (E)-2 based on NOEs between the aromatic protons of the aryl and pyridine rings, indicating evidence of stacking of the aromatic rings and supporting the proposed mechanism of the reaction.⁸ 

The control of the formation of regioisomers is critical for the synthesis of intermediates towards the formation of corresponding drug substances. Obtaining the wrong regioisomer produces the wrong drug substance; therefore, using the appropriate analytical technique to determine which regioisomer has been synthesized is essential. NMR is in a unique position to determine the correct regioisomer compared to other structural analytical techniques such as MS where the molar mass is the same for all regioisomers with different chemical arrangements. Farrell et al.⁹  developed a regioselective amination of unsymmetrical 3,5-disubstituted pyridine N-oxides using inexpensive saccharin as an ammonium surrogate to lower the cost of the manufacturing route used for a clinical candidate (Figure 1.6). Saccharin provided high conversion and regioselectivity that was easily removed in acidic conditions as a one-pot reaction.

The authors prepared an extensive variety of derivatives to determine the best conditions of the reaction with all the products analyzed by NMR to determine which regioisomer was synthesized in each case. NOE was the main key NMR experiment to determine the correct structures of the regioisomers synthesized.⁹  Figure 1.7 illustrates an example of three potential saccharin intermediates for the amination of 3-bromo-5-methoxypyridine.

Figure 1.8 shows that 1D ¹H NOE can determine the correct synthesized regioisomer. In this case, irradiating the methoxy proton at 3.87 ppm (H-21) elicited a strong NOE signal with the most upfield shifted aromatic proton H-16 of the 2,3,5-trisubstituted pyridine, indicating that the nitrogen N-2 of the 1,2-benzothiazolone ring is attached to the aromatic quaternary carbon C-12 of the 2,3,5-trisubstituted pyridine as shown in structure I, 2-(5-bromo-3-methoxypyridin-2-yl)-1,2-benzothiazol-3(2H)-one, 1,1-dioxide.⁹ 

The goal of scale-up manufacturing goes beyond the demand of the drug substance to support human clinical trials. It is also important to forecast the commercial manufacturing processes in preparation of the approval of the drug substance by the worldwide regulatory agencies to benefit patients in need of the drug. Providing a steady supply of the drug requires a commercial chemical process that fulfils the regulatory expectation of meeting the specifications of high quality and purity of the drug for every batch produced. The process becomes more complex when the drug substance contains several stereocenters, which requires the use of enantioselective resolution and diastereoselective reactions. Caille et al.^10 ,11  developed a commercial chemical process to manufacture AMG 232, a molecule with four stereocenters and an oncology clinical candidate as an inhibitor of the p53-MDM2 protein–protein interaction. The production of AMG 232 for first-in-humans (FIH) of phase-I clinical trials produced a yield of 32%, but with several changes on the reactions including a better rejection of impurities, the authors improved the yield of the drug substance to 49.8% (Figure 1.9). NMR played a significant role determining the structures of all the intermediates and critical in-process impurities in the chemical process for process understanding, which facilitated changes in the process to reject those impurities and to have a better purity and higher yield of the drug substance.^10 ,11 

Environmental sustainability is a major goal when scaling up production of drug substances in preparation to support clinical trials and future commercial manufacturing. Route scouting efforts are pursued for each synthetic step to convert the chemical process of producing the APIs from medicinal chemistry to be scalable for large manufacturing production with high yields and product quality. Challenges become apparent especially for catalytic reactions due to the high costs of catalysts, and in particular if the reaction cannot be performed at high concentrations. This is the case of ring-closing metathesis (RCM) reactions that are carried out under diluted conditions with the need of using expensive Grubbs catalysts. RCM are well known for creating carbocyclic and heterocyclic rings, which are important building blocks to create cyclic structures. St.-Pierre et al.¹²  optimized the RCM reaction to form a 16-membered ring for macrocycle PG, which is an intermediate of a potential oncology drug substance (Figure 1.10). The authors combined high-throughput experimentation and computational analysis using density functional theory to identify the best route (PG = 4-BrBz vs. H in Figure 1.10) with volume reduction from 800 to 50 L kg⁻¹ and over 20% yield increased compared to the discovery route from medicinal chemistry. NMR was critical, supporting the chemical processes of all the RCM reactions tested by characterizing the structures from starting materials to products to determine if the reactions were producing the desired products with the trans double bond in the 16-membered macrocycle.¹²  Completed characterization of all the materials was carried out by 1D and 2D homonuclear and heteronuclear NMR experiments.¹² 

Regulatory agencies around the world, such as the United States Food and Drug Administration (US FDA), the European Directorate for the Quality of Medicines (EDQM), the Therapeutic Goods Administration of the World Health Organization (WHO), and other health agencies, require high-quality materials especially when used for human consumption. The purity of those materials together with their impurity profiles are part of the specifications of the synthetic processes to produce drug substances and require prior approval to commercialize them. Impurities may come from the crystallization process, the synthetic process, and during storage. Specific analytical techniques are deployed to determine the presence of impurities on drug substances, such as high-performance liquid chromatography (HPLC), gas chromatography (GC), MS, and NMR. These impurities can be organic and inorganic, with the organic coming as residual solvents or from the synthetic process as by-products or residual unreactive substrates, including those from previous steps that may be carried over during the synthesis. These impurities are considered minor impurities; however, if they reach the threshold of 0.1%, regulatory agencies require the structural identification for toxicology analysis to determine any potential toxic effects to patients. Control of these impurities during the manufacturing of the drug substances is critical to the chemical processes. Batches with individual impurities above 0.1% and total impurities above 0.5% are out of specification (OOS), and they are rejected for further use. An investigation is required to provide the corrective action preventive action (CAPA) that identifies the root cause of the problem and the corrective actions to prevent recurrence in future batches. Expected impurities in the drug substances from chemical processes must be isolated or synthesized for characterization and qualified as impurity standards to be used as markers to determine if they are present in any of the manufactured drug substance batches. NMR becomes an essential technique for the structural characterization and confirmation of these impurities.^13 ,14  In this section, we describe the role of NMR in cases of impurity profiling, structures of unknown impurities detected on large batches of the drug substance, the concern of mutagenesis on impurities, impurities from intermediates, and NMR developments for the identification of impurities in mixtures.

Unusual impurities may occur, presenting challenges to determine their structures and explain their origin. For example, during the analytical impurity profiling of the drug ulipristal acetate (ULIPA) (Figure 1.11), the drug substance of ESMYA® for pre-operative treatment of uterine fibroids, a minor unknown chromatographic peak was detected by HPLC.¹⁴  The peak was detected during the HPLC method development for the impurity profiling of the drug substance. Based on LC-HRMS data, the mass of the new impurity was a single unit higher than ULIPA, suggesting a mono-isotopologue of ULIPA. The initial quantity of the impurity was not sufficient for NMR studies. Enrichment and isolation of the impurity provided sufficient quantity in the microgram range to perform selective excitation and presaturation experiments to propose the structure as D-ULIPA (Figure 1.11) where a deuterium atom was located in the methyl group of the dimethylamino moiety. Further confirmation of the structure was provided by synthesis to compare the NMR and MS data with the enriched isolated material.¹⁴ 

The safety and efficacy of a drug substance can be compromised when impurities are present even in small amounts, impacting its quality and compromising patients’ health. Determining the presence of these impurities prior to their structural analysis requires the development of HPLC methods that can detect them. The official US Pharmacopeia (USP) monograph describes two UV(ultraviolet)-HPLC methods to examine the impurity profile of the synthetic drug substance levothyroxine sodium (Figure 1.12) depending on the synthetic process used. Levothyroxine sodium is the sodium salt of the l-isomer of thryroxine, an active physiological hormone found in the thyroid gland. The hormone is necessary for cell differentiation, cellular metabolism, and balanced neurological and physiological functions. Patients with hypothyroidism require supplements of the hormone. Determining the structures of the impurities can provide options to reject them during the synthetic process, thereby increasing the quality of the hormone for the better safety and efficacy of the drug. Ruggenthaler et al.¹⁵  examined a second USP method of impurity profiling for levothyroxine sodium to determine unknown impurities not investigated. Through high-resolution (HR) MS (HRMS), high-resolution tandem MS/MS (HRMS/MS) including online H/D exchange analysis, and 2D NMR, in particular heteronuclear HSQC and HMBC experiments, 24 unknown impurities were characterized. Five of the impurities were isolated and synthesized for further confirmation of their structures. With this study, a second USP method for impurity profiling of levothyroxine sodium was well established.¹⁵ 

In preparation for human clinical trials, the impurity profile of the drug substance provides information to interpret its toxicity and safety. The International Conference on Harmonization (ICH) guideline Q3A(R2)¹⁶  provides recommendations on controlling impurities (organic, inorganic, and residual solvents) in the drug substances. Organic impurities are of major concern due to potential toxicity and are produced during the synthetic process. Knowing the structures of the impurities is essential to establish critical control points during the synthetic process of the drug substances to ensure quality and safety for patients. HPLC-UV is the standard technique to determine the impurity profiles of the drugs and to meet their specifications in weight percentages. Based on the ICH Q3A(R2) guidelines¹⁶  for commercial products, if an impurity is above 0.1%, structural elucidation and toxicology studies on that impurity must be performed. Techniques such as liquid chromatography–MS (LC–MS) and NMR are employed to determine the structures of unknown impurities. During the scale-up manufacturing of the first kilogram-scale batch of the investigational acute and chronic pain drug AMG 517¹⁷  (Figure 1.13), the HPLC-UV impurity profile method revealed the presence of two late-eluting impurities that were not detected in smaller production batches during the medicinal chemistry synthetic scheme. Several good manufacturing practice (GMP) batches prepared to support first-in-human (FIH) clinical trials prompted the investigation of the characterization of those impurities because they were present at the 0.1% level and above. Initial MS analysis indicated that one impurity was a homodimer of the drug substance, and the other was a dimer with 32 extra mass units. NMR analysis of an enriched fraction containing both impurities revealed that both were homodimers with some differences in the chemical shifts of the aromatic protons of the benzothiazole ring, the location where the monomers were connected. HRMS of the enriched fraction revealed that the dimer with an addition of 32 units contained a sulfur atom as a thioether link between the monomers. The structure of the impurities could not be explained during their formation from the synthesis of the drug substance. To understand the origin of the dimer impurities, extensive investigation was pursued on the starting materials by HPLC-UV and LC–MS. It was discovered that the N-acetyl benzothiazole was of lower purity than what the vendor certification of analysis (COA) claimed. In addition, several minor impurities were detected as dimers, one as simple homodimer and another as homodimer with a thioether linkage. The presence of those impurities could be carried over during the synthesis of the drug substance forming the impurities detected in the GMP batches. Communication with the vendor to provide higher quality of the material provided subsequent batches with an absence of those impurities.¹⁷  An important lesson from this study indicates that the vendors’ COA may not reflect the quality described, and better control of their materials may be needed including testing their batches prior using them, beside demanding more accurate description of their materials.

When new impurities are detected during the preparation of commercial pharmaceutical products, the ICH Q3B(R2) guidance¹⁶  requires to identify and characterize them if they are present above 0.05%. During the oxidation step of isoleuco acid to obtain isosulfanblue (Figure 1.14), a new unknown impurity below 0.10% was observed and detected by HPLC.¹⁸  Isosulfanblue is a dye used as a contrast agent for cancer diagnosis and to guide surgical removal of cancerous tissues. The impurity was isolated and characterized by LC–MS and NMR. LC–MS analysis indicated an addition of 16 amu compared to isosulfanblue, suggesting the addition of a hydroxyl group to the molecule. The ¹³C NMR spectrum showed the benzylic carbon for the impurity at 64 ppm compared to isosulfanblue at 126.9 ppm (sp² carbon in isosulfanblue), indicating loss of conjugation of that carbon and supporting the bonding to a hydroxyl group. The authors proposed a mechanism of formation of the impurity by the nucleophilic attack of a hydroxyl group on that benzylic carbon during the oxidation due to the presence of water in the reaction conditions.¹⁸ 

Commercial drug substances that have been approved for a period of time may be listed in the USP monograph with information of their purity analyses, normally by thin layer chromatography (TLC). If a new impurity is detected in new batches, the drug substance becomes out of specification (OOS) and requires further investigation and a mitigation plan to reject the impurity in the final product. This was the case of clorsulon (Figure 1.15), a drug substance used as a suspension for the treatment and control of adult liver flukes in cattle.¹⁹  Based on the Harmonization of Technical Requirements for Registration of Veterinary Medical Products (VICH) guidelines,²⁰  any unknown impurity at the threshold level of 0.5% must be identified and qualified as impurity standard. In contrast, USP specifies a lower level to characterization of new unknown impurities as 0.15%. In this case, the impurity was over 0.5% in several batches of the drug substance. An investigation was conducted to determine the structure of the impurity for further toxicology studies. HPLC was used instead of TLC to determine the presence of the impurity in several batches and to isolate it for its structural characterization. After LC–MS analysis based on the molecular ion and molecular formula, NMR analysis indicated that the impurity was methylated as N-methyl of the primary amino group attached directly to the aromatic ring of clorsulon.¹⁹ 

Constant testing of batches of commercial drug substances is necessary to control the quality of the material and to detect any potential issues during the manufacturing process that can affect the quality of the drug substances to be out of specifications. Repaglinide (Figure 1.16) is an antidiabetic drug for the treatment of type-2 non-insulin-dependent diabetes mellitus, and seven unknown impurities were observed from the UPLC analysis of several manufacturing batches during the process optimization to produce the drug.²¹  The impurities were in the range of 0.05–0.10%, which reduced the safety and efficacy of the drug substance. Using semi preparative HPLC, the impurities were isolated for their characterization. Extensive work on NMR and MS determined the structure of seven impurities. Only one impurity required crystallization to solve its structure by X-ray crystallography. Once the structures were determined, they were synthesized for further confirmation of their structures and also to create material as impurity reference standards to be used as markers for spiking studies to analyze subsequent batches and consider them part of the impurity profile of the drug substance. To understand the root cause of the impurities for better control of the manufacturing process of the drug substance, an investigation was pursued to determine if the synthetic process could be the source of the impurities. The results of the investigation indicated that several impurities were formed during step 1 for the condensation reaction to form the amide bond, and others were formed during step 2 of the ester hydrolysis. The information regarding the impurities and their source was used to implement some modifications to reduce their formation and to improve the quality and purity of the drug substance.²¹ 

Impurities are more difficult to control for large pharmaceutical doses of the drug substances. The chemical processes to manufacture the drugs on the large scale require careful optimization to minimize the presence of impurities. Ceftolozane (Figure 1.17), a cephalosporin antibiotic with broad-spectrum gram-negative activity, requires doses of 1 g per day. With that large amount of the dose, the quality of the material becomes challenging to control impurities.²²  During the development of the chemistry for the second generation of the process, an unknown new impurity was observed in the range of 0.03–0.05% in the final drug substance. The structure of the impurity was determined as a dimer through the LC-HRMS, HRMS/MS, H/D exchange, and 2D NMR studies. Once the structure was determined, the next step was to discover in which step of the preparation of the drug substance was the dimer impurity formed. Based on the dimeric structure of the impurity, the drying process of the drug substance was under investigation. The level of the impurity was sensitive to the solvent composition of the wet cake, which is the final drug substance after filtration but with the presence of some residual solvents from the last chemical reaction. A kinetic study was pursued to understand the formation of the dimer impurity depending on the ratios of the residual solvents. It was discovered that the impurity formation was dependent on the amount of water present in the cake, which points towards the beginning of the drying process where more water was present in the cake. Other solvents like acetonitrile (ACN) and acetone were less critical for the impurity formation. The knowledge acquired during the investigation helped to make the appropriate changes to mitigate the formation of the dimer impurity in subsequent batches.²² 

Impurities formed during the reaction of intermediates may also be important to characterize for process understanding and to reject potential impurities that could be carried over to the final chemical step of the drug substances. The knowledge of key impurities in intermediate steps helps to improve the yield of those intermediates and provide cost savings to the processes. During the formation of a tetrazole ring as part of the synthesis of an investigational drug for diabetes, two drug-related impurities were observed (Figure 1.18).²³  Those two impurities had the same mass-to-charge ratio, but 15 mass units larger than the expected product of the reaction. To further understand how to improve the formation of the intermediate by eliminating or rejecting the impurities, isolation of both impurities from the mother liquor was pursued. Initial isolation using acidic conditions provided only one of the impurities with the loss of the BOC (tert-butoxycarbonyl) functional group. The structure was determined by 1D and 2D NMR experiments where the molecule incorporated an NH group between one of the aromatic rings and the dimethylcarbamoyl moiety. 1D NOE provided spatial correlations between the aromatic protons and the methyl groups of the dimethylcarbamoyl group. The connectivity was also supported by ¹H,¹³C-HMBC and ¹H,¹⁵N-HMBC experiments. The explanation of the addition of the NH group was proposed as a “Schmidt-like” reaction mechanism forming a urea-type in-process impurity due to the excess of sodium azide in the reaction. Isolation of the two impurities without the removal of the BOC was achieved in neutral conditions. Both impurities had an NH incorporated between one of the aromatic rings and the corresponding dimethylcarbamoyl group but proposed to be different isomer for each impurity. One of the impurities was more unstable but it provided the same structural information as the more stable impurity (6) (Figure 1.18) with some minor differences in the proton and carbon chemical shifts and different retention times by HPLC, indicating that they were different impurities with similar structural arrangements. The structural knowledge was sufficient to modify the process to reject them. Changes of the crystallization process were sufficient to minimize their presence in the formation of the desired intermediate.²³ 

Impurities that have the potential for being mutagenic require high control to provide patient safety. The ICH M7 guidance²⁴  provides recommendations for safety and quality risk management for the levels of mutagenic impurities that are not supposed to possess carcinogenic risk for patients. Atovaquone (Figure 1.19) is a drug substance approved in the United States for the treatment and prophylaxis for Pneumocystis carinii infection and for the treatment and prevention for Plasmodium falciparum for malaria.²⁵  Atovaquone may be mutagenic based on its quinone substructure, but it was negative during the standard GLP genotoxic tests (e.g., Ames Salmonella mutagenicity assay). In the regulatory submission and marketing application, the confirmation of purging potential mutagenic impurities has to be specified. GSK developed a second-generation chemical route to produce the drug substance that was not registered for commercial purpose. However, it was used for a control strategy of mutagenic impurities and not implemented because the process involved the formation of potential mutagenic impurities. In addition, this process was generated prior to the publication of the ICH M7 guidance.²⁴  Impurities generated in the second-generation route were analyzed to determine their structures and in silico analysis by quantitative structure–activity relationship (QSAR) screening to predict if any of those impurities could be mutagenic. Furthermore, the compounds were tested for the Ames test (a test using bacteria, such as Salmonella typhimurium, to assess if chemicals can have potential carcinogenic effects by causing mutations in the DNA of the bacteria), with many of those giving negative results. The study also revealed that the concentration of the impurities was mutagenic and thus chemical processes were employed to purge the impurities. The role of NMR was critical to determine the structure of those impurities for in silico mutagenic analysis and to provide risk control of the process. ¹H NMR spectroscopy with selective excitation was also used to determine the presence of the potential impurities in the trace level sensitivity down to 1 μg g⁻¹ (a dose by weight of 1 μg of impurity per gram of drug in this case) for a particular impurity, which at that level was observed with good signal-to-noise ratio and with minimum distortions in the spectrum.²⁵ 

The best option to determine the structures of impurities is by isolating them for NMR structural analysis. As a requirement from the worldwide regulatory agencies and followed by the ICH Q3A(R2) and ICH Q3B(R2)¹⁶  recommendation guidelines, the structures of impurities present in the drug substances and intermediates that are above 0.1% must be known, and toxicology studies should be performed. However, when the impurities are in low quantity even above 0.1%, isolation of impurities at those low levels is not trivial. Mixture analysis become critical to obtain structural information on NMR signals for the smaller components, but when they are structurally related to the major components of the mixture, many of the signals from the smaller components overlap, making it difficult to determine their partial structures. The development of selective excitation NMR experiments provides an opportunity to obtain partial information of the structure of the minor impurities in the samples. The development of a 2D selective-TOCSY HMBC proved useful when analyzing regioisomers as minor impurities of a drug substance without the need for isolation.²⁶  The authors presented several cases to demonstrate the utility of the experiment. One drawback of a HMBC type of experiment is the relatively low sensitivity of the experiment which required acquiring data overnight to obtain sufficient structural information for impurities at a level of 0.4% or above. The method was tested with examples of impurities ranging from 0.4% to 8.2% in three case studies to be particularly useful for final drug substance materials when isolation becomes a burden.²⁶ 

The hyphenation of analytical separation techniques with NMR to determine the structures of components in complex mixtures analysis has been in development for decades and applied to fields beyond the pharmaceutical environment.^27–29  The conventional analytical separation technique used to hyphenate with NMR has been HPLC; however, one drawback is that only a percentage of the chromatographic peak is seen by NMR due to the relatively small volume of the NMR flow cell (10, 30, 60, or 120 μL depending on the flow cell used) compared to the volume of chromatographic peaks. The replacement of HPLC by ultra-performance liquid chromatography (UPLC) has brought the advantage that UPLC enables the concentration of the chromatographic peak to a similar volume as that of a 30 μL NMR flow cell thereby increasing the sensitivity or signal-to-noise ratio up to threefold compared to standard LC-NMR.³⁰  Testing was conducted to prove this technology as part of a methodology for impurity structural identification during the development of pharmaceutical products.

In the pharmaceutical industry, stress testing or forced degradation studies are an essential part of the development process of drug substances in preparation for regulatory filing and future commercialization. Those tests are intended to determine the stability of drugs in different conditions and stabilize the shelf-life of the drugs in both forms, as drug substances and drug products. ICH guidelines Q1A(R2) and Q1B¹⁶  provide the recommendations on the appropriate experiments to determine the stability of the drug substances and drug products, together with the thresholds for impurities based on GMP to assess the risk management on drugs and maintain the patient safety. Typically, stress testing studies are conducted in chambers where the drugs remain for a certain time (several months but taking monthly aliquots for analysis) under specific conditions of humidity and temperature. Other stability testing requires the drugs to be exposed to acid, base, oxidation, and light. Those stability testing protocols are required when drug substances are intending to be filed for phase III human clinical trials and not necessarily required for early clinical trials, but regulatory authorities may require them. In fact, it is becoming standard to provide initial stability data obtained when the drugs are stressed for several months (3–6) when filing for early phases (I and II). The guidelines for the acceptable thresholds of impurities present in the drugs when exposed to forced degradation studies and to determine their structures and their toxicological effect are the same as for impurities present in drug substances and drug products. Those guidelines are the ICH Q3A(R2) for drug substances and the Q3B(R2) for drug products.¹⁶ 

NMR has a significant impact for the structure elucidation of impurities as we have seen in the previous section. Therefore, NMR becomes the main structure elucidation technique for those impurities. The procedure is to analyze the samples during the degradation studies by HPLC to determine the growth of the impurities on those conditions. LC–MS provides an initial information on the structures of those impurities. When impurities become above 0.1% of the sample, they are isolated for a more thorough study to determine their structures by NMR. Few cases are presented here to illustrate the role of NMR in the study of the structures of impurities during stability testing of drugs.

Amlodipine maleate is a commercial calcium channel blocker to treat high blood pressure and chest pain. During the stability studies of amlodipine maleate (Figure 1.20) in accelerated conditions at 40 °C and 75% of relative humidity (RH), a new unknown impurity was observed by LC–MS. Isolation of the impurity was carried out to elucidate its structure by MS, NMR, and IR (infrared spectroscopy) (Figure 1.20).³¹  The isolation was performed after the drug was exposed under stress conditions at 105 °C for 3 days to accelerate its growth and produce a sufficient amount for identification. Based on 1D and 2D NMR experiments and supplemented with fragmentation pattern by MS and IR information, amlodipine maleate salt lost the maleate salt and the amino ethoxy group and underwent cyclization to form a six-membered ring.³¹ 

Asthma is a heterogeneous disease that affects the inflammation of the airway smooth muscle in the lungs and can be a chronic disease for many patients including children. Bronchodilators help relax the airway smooth muscle, providing a short-term treatment effect for patients especially when they have asthma attacks to minimize hospitalizations. Medication used over time can be detrimental if impurities can grow with time. Impurity profile and forced degradation studies in stress conditions determine the safety and stability of the drugs for the shelf-life expiration time. A series of bronchodilators used for asthma were investigated for impurity profiles and stability studies.³²  Those drugs are short acting β2-agonists (salbutamol, arformoterol, bambuterol, clenbuterol, formoterol, carmoterol, indacaterol, vilanterol, PF-610355), muscarinic antagonists short acting (ipratropium, tiotropium, glycopyrrolate), and xanthines (caffeine, doxophylline, CH-13584). The drugs were exposed to acid, base, heat, humidity, oxidation, and light. The impurities were detected by HPLC for quantitation and their structures, when above 0.10% following the ICH guidelines,¹⁶  were determined in preference by NMR spectroscopy together with MS. Those impurities in larger amounts were tested for toxicology.³² 

Perampanel (Figure 1.21) is an antiepileptic drug administered as a cocktail combined with other drugs for the treatment of partial and generalize seizures. It is also considered as a Schedule III controlled substance, or moderate to low potential abuse. Being FDA approved in 2012, Saida et al.³³  indicated that few HPLC methods for the analysis of perampanel in pharmaceutical formulation were reported, but no UPLC methods were found for the drug for stability, degradation, isolation, and complete characterization of perampanel impurity profile. The authors conducted stability studies following the Q1A(R2) ICH guidelines¹⁶  to determine the impurity profile under the force degradation conditions and the structure of the impurities. The degradation studies were conducted under acidic conditions for 12 h at 60 °C, basic conditions for 16 h, under oxidation conditions with peroxide, heat, and light. All the studies were monitored by UPLC-MS. The studies under thermal and light conditions indicated that the drug did not degrade. In acid conditions an unknown impurity at the level of 17.02% was found; degradation was negligible in basic conditions, but another unknown impurity was present in oxidation conditions at 11.68%. Both impurities were isolated, and ¹H and ¹³C NMR experiments were acquired to elucidate their structures. The cyano group became an amide group in the impurity isolated from the acid stress conditions, and in the case of the oxidation conditions, the impurity was an N-oxide in the nitrogen of the pyridine ring. The UPLC method was validated for the impurity profile of the stability studies for perampanel.³³ 

Levosimendan (Figure 1.22), an inodilator used for short-term treatment of decompensated heart failure, was granted Fast Track status by the FDA. It is known that the drug deteriorates with variation of temperature and humidity conditions; therefore, drug stability studies become critical for its quality and safety attributes. Accelerated stability studies were conducted instead of time-consuming real-time stability studies. No stability studies were previously reported; therefore, Mehta et al.³⁴  conducted accelerated stability studies of the drug under 40 °C and 75% RH in a stability chamber for 6 months following the ICH guidelines.¹⁶  Stress degradation studies were also conducted under basic conditions and reflux for 9 h at 60 °C, then neutralized prior to analysis. Oxidative degradation studies were conducted by exposing the drug to hydrogen peroxide at 60 °C for 6 h. Two major degradation impurities, DP1 and DP1, were formed during the studies at 40 °C/75% RH, reaching a threshold of 0.09% and 0.13% and targeted for identification through isolation. DP1 was also present in the oxidation studies. Extensive NMR and MS studies were conducted to determine their structures where DP1 was having one of the cyano groups of the carbonohydrazonoyl dicyanide moiety converted to a primary amide (CONH₂) and DP2 lost the carbonohydrazonoyl dicyanide moiety of levosimendan. The structures were synthesized for further confirmation, and a mechanism of formation for both degradation products was proposed based on an oxidative pathway. The accelerated stability studies provided information to predict the storage conditions of the drug during process development. Cool and dry conditions were required to avoid drug degradation to ensure quality during the manufacturing of the drug.³⁴ 

From the foregoing equation, %TI is the percentage of total impurities by HPLC, %TS is the percentage of total solvents or volatiles by GC, and %M is the percentage of moisture by Karl Fisher titration. The measurement by HPLC requires the molecule of interest to have an UV response factor, which makes it less accurate for molecules with poor or lack of chromophores. In addition, impurities present in the sample must be separated by the HPLC method used for accurate measurement and using the same UV response factor as for the main component, which may deviate for the impurities depending on their structures.

In the foregoing equation, %P(x) is the percentage purity of the analyte, I(x) and I(std) are the integrals of the signals selected for the analyte and the internal standard, N(x) and N(std) are the number of nuclides (e.g., protons for 1D ¹H NMR) of the selected signals for the analyte and the internal standard, M(x) and M(std) are the molar masses of the analyte and the internal standard, m(x) and m(std) are the weights of the analyte and the internal standard, and P(std) is the purity of the internal standard. The methodology has been extensively tested and validated.³⁵  This methodology has several advantages: NMR does not depend on response factors, a single experiment provides all the information to calculate the purity of the analyte of interest, and any nuclide with nuclear spin can be used to do the quantitation with the main consideration of acquiring the data with sufficient relaxation delay to account for T₁ relaxation.

In the pharmaceutical industry, the measurement of purity of compounds by NMR is becoming a trend to determine the purity of drug substances, but also for their intermediates, starting materials, and impurities. In addition, qNMR is used during the chemical processes of producing the drugs with the purpose to calculate the correct quantity of substrates to be charged in reactions to aim for the highest yield of the desired products and to determine the yield of reactions. The method of adding internal standards is well utilized in the industry. A general quantitative method by NMR was developed to calibrate atropine sulfate (Figure 1.23), an anticholinergic drug, as a reference standard. In solution, several conformations of the tropane ring have been observed, and they are sensitive to temperature. The quantitative method using maleic acid as internal standard and D₂O as solvent was validated, showing good linearity, range, limit of quantitation, stability, and precision, showing good results from mass balance calculations.³⁶ 

A synthetic molecule named 6-oxo (Figure 1.24), 2-methyl-5,6,7,8-tetrahydro-2H-chromen-4(3H)-one, is a new AI-1 quorum sensing inhibitor of the bacteria Vibrio harveyi.³⁷  During its early development, no certified reference material was available in the market to develop an appropriate chromatographic method to determine its purity. qNMR was tested using two internal standards, maleic acid and benzoic acid, in CDCl₃ and DMSO-d₆ with decoupling of the C-13 satellites in the 1D ¹H NMR spectra. Considerations were taken for the acquisition parameters to allow sufficient time for the T₁ relaxation to avoid affecting the results. The assessment of both internal standards provided information on the selection of the appropriate internal standard for the purity analysis of 6-oxo. The presence of an impurity was observed only when using benzoic acid, indicating some chemical reactivity occurred with the analyte 6-oxo. Maleic acid as internal standard provided reproducible good results. In fact, this methodology was also used to determine the mass fraction of crude samples of 6-oxo during its synthesis.³⁷ 

The analysis of qNMR is normally performed using high-field NMR instruments especially when using ¹H NMR for greater chemical shift dispersion and sensitivity. However, Jaki et al.³⁸  recently reported the pioneering work of George M. Hanna at the US FDA between 1984 and 2006, where Hanna used a 90 MHz NMR instrument to perform qNMR analysis of small molecules of biomedical importance. Hanna’s work is relevant to the current benchtop NMR instruments available today, making NMR more accessible to non-NMR users and their laboratories avoiding special facility requirements for installation and use. Lee et al.³⁹  investigated the use of external standards for qNMR on benchtop NMR spectrometers to determine the major source of errors for the analysis. The authors indicated that the main source of errors for benchtop NMR instruments applying qNMR methodology are coming from the fixed tuning and matching that is less predominant in dilute samples used in high-field NMR, the phase and baseline corrections, and broad signals typical in low-field benchtop NMR systems. However, using automated integration with quantitative global spectral deconvolution (qGSD) instead of manual integration produces more reliable results.³⁹ 

Following the pioneering work or George M. Hanna at FDA, scientists from industry, academia, nonprofit institutions, and governmental bodies have collaborated for over a decade to standardize the practices of qNMR. The qNMR Summit 5.0⁴⁰  was attended by scientists from the United States, Canada, Japan, Korea, and Europe with the purpose of discussing the status of qNMR methodology focused on the pharmaceutical industry, quantitation algorithms, and other developments.⁴⁰ 

Many pharmaceutical drug substances have one or more stereocenters in their molecular structures, which bring another level of complexity to the development of drug substances. In general, the potency of the drug is related to one enantiomer, with the other being inactive or toxic; therefore, associating the potency to the right enantiomer is critical for the safety of patients. Chemically it is not trivial to produce only one enantiomer because chemical synthesis is not as selective as nature. When possible, enantioselective catalytical reactions and chiral chromatography are the typical methods to attempt the selectivity and separation of enantiomers. The golden standard analytical technique to determine the absolute stereochemistry of molecules is X-ray crystallography; however, not all the molecules can easily form single crystals, which is required for the X-ray analysis. NMR cannot determine the difference between enantiomers because they are mirror images, and they have the same bond and spatial correlations. However, NMR can determine the relative configuration in many cases when more than one stereocenter are present in the molecules of interest. The typical NMR experiments in solution used for spatial correlations through protons are 1D NOE/ROE and 2D NOESY/ROESY. Those experiments can provide information on the relative configuration of stereoisomers when their stereocenters are in spatial proximity of less than 5 Å, but not when they are farther apart. Recently, a new approach has emerged that is independent of the location of the stereocenters to obtain the relative configurations of molecules with at least two stereocenters, distinguishing their diastereomeric structures. The technique requires to extract values on C–H residual dipolar couplings (RDCs) from 2D pure-shift ¹H,¹³C-HSQC, or its J-resolved version when measuring the sample in isotropic and anisotropic media because RDCs contain information of the orientation of the molecule in the gel (anisotropic media) based on its configuration and more stable conformations in solution. The method requires comparing the experimental RDCs with all possible density functional theory (DFT) calculated structures to determine which DFT molecules fit best the experimental NMR data. RDCs have been applied to some cases in the pharmaceutical industry. A few examples from the pharmaceutical industry are discussed subsequently.

When the sample is in the gel, a percentage of molecules will be aligned to the magnetic field depending on how much the gel is compressed or stretched. Weak alignment provides small RDCs that are prompt to inaccurate measurements. However, strong alignment degrades the quality of the spectra due to extensive anisotropic effects from the gel. The best option is to be in between weak and strong alignment. RDCs must be in the range of 10–30 Hz for the measurement to be reliable, but it is possible to measure them at higher level of alignment. Strychnine and retrorsine (Figure 1.25) were tested to measure RDCs at a range of –300 to +300 Hz using the PBLG polymer, poly-γ-(benzyl-l-glutamate), with the standard 2D heteronuclear J-resolved experiment.⁴¹  Strychnine was used as a test molecule to compare the results obtained with PBLG to the data published using PMMA gel, poly(methyl-methacrylate), providing good confidence on this methodology.⁴¹ 

Complementary NMR information combined with RDCs can be used as a strong case to prove the relative configuration and information of the more stable conformations of molecules in solution. The cyclic hexapeptide cyclo-[Leu-d-Leu-Leu-Ley-d-Pro-Tyr] required RDCs, ¹J_NH-Hα scalar couplings, and NMR temperature coefficients (T_c) in two deuterated solvents, CDCl₃ and DMSO-d₆.⁴²  The cyclic peptide was rigid enough to provide sufficient RDCs to determine the preferred conformations of the molecule, and the two solvents were used to understand its molecular permeability and bioavailability. The authors proved that this approach was superior to the traditional NOE studies.

Another type of data that can provide more information on the relative orientation of the ¹H–¹³C-dipole bonds of molecules in the magnetic field is through the measurement of residual chemical shift anisotropy (RCSA) because it is related to the orientation of the carbon chemical shift shielding tensor when the sample is measured in the gel at different compressing strengths. This technique does not depend on carbons attached to protons, which is a benefit when HSQC cannot provide enough RDCs for molecules lacking protons in their structures. Molecules with different masses were tested with this method (estrone with 270 Da, retrorsine with 351 Da, and cryptospirolepine with 303 Da) (Figures 1.25 and 1.26) and with two polymers, PMMA compatible with CDCl₃ and poly(20-hydroxyethylmethacrylate) compatible with DMSO-d₆.⁴³ 

Another experiment that can be used when molecules are deficient in protons is the J-modulated ADEQUATE that shows ¹³C–¹³C correlations. Because of the low sensitivity and natural abundance of ¹³C, the experiment may require several days of acquisition depending on the amount of sample. This approach was tested to distinguish the Z- and E-isomers of thiazolidinedione (Figure 1.27) with PBLG polymer to confirm their structures when compared the experimental data with the DFT calculated structures.⁴⁴ 

Recently, the stereochemical analysis via measurement of anisotropic NMR data through RDCs has been evaluated for the first time with a prototype 400 MHz cryogen-free high temperature superconducting (HTS) NMR spectrometer, installed in the fumehood of one of our chemistry laboratories.^5 ,6 ,45  Three commercial natural products, artemisinin, artemether, and dihydroartemisinin (Figure 1.28), were chosen for this assessment using poly-HEMA gel in DMSO-d₆ as anisotropic media.⁴⁵ 

Figure 1.29 illustrates how to extract the measurement of RDCs from the J-scaled BIRD HSQC using RESET F2 homonuclear decoupling (pure-shift) for artemether as example. Two J-scale HSQC spectra with scaling factor 4 are overlapped, the isotropic in DMSO-d₆ in grey and the anisotropic in poly-HEMA gel/DMSO-d₆ in blue, for the region of the CHs signals at positions 1, 7, and 10 in the molecule. The spectra overlap but are slightly shifted for better visual comparison. The RDC measurement or ¹D_CH (Hz) is the difference in value from the total coupling in anisotropic media or ¹T_CH (Hz) and the J-coupling from the isotropic media or ¹J_CH (Hz), as ¹D_CH = ¹T_CH–¹J_CH (Hz). The table illustrated in Figure 1.29 shows the experimental values measured from these spectra. The accuracy of measurement of the anisotropic NMR data with the HTS magnet spectrometer was evaluated through the computer-assisted 3D structure elucidation (CASE-3D) fitting protocol implemented in Mestrenova­StereoFitter software program. Artemisinin and artemether provided the expected results.⁴⁵  The case of dihydroartemisinin, as β-epimer in the commercial solid material used for this testing, was more challenging because it epimerized in solution as a mixture of α- and β-epimers, with the α-epimer kinetically favored in solution. The combination of ¹³C chemical shifts, RDCs, and the two J-coupling values around the OH functionality provided sufficient discrimination to determine unambiguously the structures of both epimers.⁴⁵ 

Low-field benchtop NMR instruments have not been applied yet to measure RDCs and RCSAs due to its narrower chemical shift dispersion especially in ¹H and its lower sensitivity. However, testing and evaluation would be needed to determine if benchtop NMR instruments can be used for some cases in anisotropic NMR.

When a drug candidate coming from medicinal chemistry is selected to conduct human clinical trials, its synthetic process is revised to improve lowering the cost of production, supporting the scale-up manufacturing, obtaining high yields and quality of synthetic materials, and making the synthesis environmentally friendly by reducing chemical waste. Control of reactions for better quality of products is part of the development process of the drug substances. Several analytical techniques are used to monitor reactions to provide kinetic and speciation information of reactions. Some analytical techniques require taking aliquots of the reaction overtime, quenching them, and performing preliminary treatment prior to off-line analysis (e.g., HPLC, UPLC, tube NMR). This method provides some information of the process, but it is not timely, and the treatment may give inaccurate information affecting the quality of the results. The best method to provide timely information of the reaction without the need of any treatment of samples is the online reaction monitoring by in situ or online analytical techniques connected to the reactor. Infrared and Raman probes are used for in situ reaction monitoring, providing some structural information of the components of the reaction. Recently, NMR has emerged, providing a more comprehensive structural information of the species in the reaction and kinetic information due to its intrinsic quantitative properties. In addition, NMR has been applied in the low- and high-magnetic-field instrumentation for reaction monitoring, and a few examples from the pharmaceutical industry are given subsequently.

The development of NMR technology during the last decade has reduced the footprint of low-field NMR instruments as benchtop systems and improved their performance to become standard instruments in many laboratories. A 45 MHz picoSpin-45 (currently under Thermo Fisher Scientific) benchtop NMR instrument was used successfully to monitor several reactions: a Fischer esterification for proof of concept, a Suzuki coupling to evaluate the complexity of the NMR spectra in the low field with the dynamic range of solvents and species in solution, and an oxime reaction as a two-phase reaction with lack of chromophores that was part of a drug substance synthesis.⁴⁶  The study proved to be successful, demonstrating the capability of low-field NMR for reaction monitoring. A 42.5 MHz for ¹H frequency benchtop NMR was used to understand the stability of an immobilized lipase enzyme to study the Prins cyclization reaction to obtain the desired chiral pyranone acyl adduct from the racemic mixture of pyranal cis alcohols. NMR was chosen because of the lack of chromophores in the pyranal alcohols and the importance of obtaining real-time data independent of response factors, which are needed for UV-HPLC analysis. The reaction was also carried out off-line at high field to demonstrate the value of the low-field benchtop NMR instrument.⁴⁷  Maiwald et al.⁴⁸  integrated a 43 MHz compact NMR spectrometer installed in a certified pressurized housing to be explosion-proof as an online NMR sensor in a pharmaceutical pilot plant to monitor a metal–organic reaction. Absolute quantification of the reaction was successfully achieved through one-point calibration from the known raw material concentration. This example demonstrates the approach of implementing low-field benchtop NMR into a commercial-scale pilot plant successfully under automation production conditions.⁴⁸ 

When high-field NMR instruments are used to support scale-up reactions in the manufacturing setting, those reactions are carried out at smaller scale to provide sufficient information to improve the conditions of the reaction. High-field NMR instruments are nonstandard instruments to be installed in the manufacturing floor because of the required infrastructure and safety concerns due to the need of liquid cryogens. However, they provide valuable information to reactions even when they are performed at smaller scale than the manufacturing scale. During the scale-up production of the cholesterol ester transfer protein (CETP) inhibitor evacetrapib drug substance for its GMP manufacturing, NMR played a critical role mimicking the batch reaction process to study the reductive amination reaction between a secondary amine and a trans-aldehyde catalyzed by Pt/C to produce a tertiary amino that needed improvement.⁴⁹  The NMR data provided information of the mechanism of reaction, kinetics, and the key species involved in the reaction including detecting dissolved hydrogen. The kinetic knowledge through technology transfer facilitated the control in the manufacture facility under GMP conditions to synthesize the drug for clinical use.⁴⁹  One of the problems of reactions where gases are formed is to determine the rate of the gas–liquid mass transfer; however, in situ NMR has been applied to quantitate the amount of gasses dissolved in solvents to monitor homogeneous and heterogeneous catalytic reactions under pressure.⁵⁰  This method was developed as hydrogenation reactions are commonly used in the pharmaceutical industry.

Benchtop NMR instruments are ideal to be sensors of reactions in chemistry laboratories and manufacturing plants. However, they have limitations compared to high-field instruments due to their narrower chemical shift dispersion especially for ¹H NMR, lower sensitivity, limited range of temperature control especially when reactions run at different temperatures than the temperature of the permanent magnet, and limited probes available for nuclides to measure. Those factors may suggest the use of benchtop NMR instruments for simple reactions. Commercial high-field LTS magnets in NMR spectrometers require special facility infrastructure to support their liquid cryogen needs, meaning they take up a lot of space in chemistry laboratories. However, the recent developments in magnet technology using ceramic materials to build HTS magnets, which do not require liquid cryogens, have provided a solution to have high-field NMR instruments installed and integrated in chemistry laboratories with no additional infrastructure needs for the HTS magnets. As described in the structure elucidation section (1.2.1), we have installed in our chemistry laboratory a 4.7 T (400 MHz for the ¹H observation frequency) HTS magnet NMR system that was tested successfully for standard NMR experiments and compared to a commercial LTS magnet NMR system.^5 ,6  As this instrument is liquid-cryogen free, high field at 400 MHz, with a broadband probe tuneable for a broad range of nuclides, and capable to handle a large range of temperatures depending on the probe (e.g., −20 °C to 130 °C for our BBFO probe), we tested if for the application of reaction monitoring with protonated solvents, understanding that the magnetic field drift of the HTS magnet is different than for LTS magnets. We selected the ring-closing metathesis (RCM) reaction of diallyl malonate with Grubbs second-generation catalyst in protonated methylene chloride to form the product with a 5-membered ring structure (Figure 1.30). The reactor was connected by tubing to the Bruker InsightMR flow cell that was inserted inside the magnet to monitor the reaction online.⁵¹  Rate constants for the reactant and the product were obtained. The results of the study demonstrated the utility of high-field HTS magnets for reaction monitoring carried out in the chemistry laboratory, which could also be carried out in a manufacturing plant.⁵¹  After this demonstration, we have been able to utilize this prototype 400 MHz cryogen-free HTS magnet NMR system to study the mechanism of several reactions (radical and photochemistry benzylic bromination and nitration reactions) to support scale-up cases in the manufacturing plant. The cases are described subsequently.

The radical benzylic bromination reaction of 2-fluoro-3-nitrotoluene with N-bromosuccinimide (NBS) decreased the yield around 15% at the manufacturing run compared to the laboratory scale demo.⁵²  The reaction was monitored using the 400 MHz cryogen-free HTS magnet NMR system but in the NMR tube due to corrosiveness. ¹⁹F NMR was used to monitor the fluorinated species of the fluoro-nitrotoluene substrate and products and ¹H NMR for NBS and succinimide. Figure 1.31 illustrates the reaction and the kinetic graph of the different species in the reaction. Reaction progress kinetic analysis (RPKA) and variable time normalization analysis (VTNA) were used to set the design of experiments (DoE) and determine the rate constants and reaction orders of the species in the reaction. The findings indicated that molecular bromine (Br₂) was the brominated species of the reaction instead of NBS; therefore, the addition over time of NBS in portions was changed to a continuous slurry addition to maintain low solution phase of Br₂ minimizing the losses of reagent from the solution phase. The main improvements observed were reductions on the reaction time from 21 h to 8 h, a decreased amount of equivalent of NBS needed from 1.68 to 1.40, a decrease in impurity formation above 50%, and an increase in the yield in the manufacturing run of 15% to become 80% of total yield. Those changes provided a more sustainable process and high cost savings overall beyond the lower amount needed for the NBS reagent, producing better product quality.⁵² 

Photochemistry reactions require the use of a special tube or flow cell with a light source to monitor the reaction by NMR. Ji et al.⁵³  demonstrated the direct measurement in situ of accurate quantum determination of several chemical actinometers using an LED-based NMR illumination device. In our laboratories, we used the same LED-based NMR illumination device to develop a photochemical bromination/alkylation sequence as part of a continuous process for the synthesis of an intermediate by replacing the radical bromination reaction of 2-fluoro-3-nitrotoluene discussed previously for a photochemical reaction.⁵⁴  The photochemical reaction was monitored using the 400 MHz cryogen-free HTS magnet NMR system to determine its best reaction conditions.⁵⁴  Figure 1.32 illustrate the reaction and the kinetic graph of the fluorinated species obtained from ¹⁹F NMR. It should be noted that the time scale of the bromination reaction went from hours in radical conditions (Figure 1.31)⁵²  to minutes with light (Figure 1.32).⁵⁴  Running several reactions in continuous processes provide advantages over batch processes by reducing reaction times, eliminating work ups which reduces waste, and allowing direct crystallization of products from reaction mixtures. NMR plays an important role to determine the optimal conditions of individual reaction steps in preparation of the continuous process of combined reactions in one step.

Nitration reactions have been well known for centuries but are difficult to control at the large scale. The nitration of the (2-fluoro-3-methylphenyl)boronic acid to the corresponding 2-fluoro-1-methyl-3-nitrobenzene was carried out with NH₄NO₃/TMS-Cl.⁵⁵  The drawbacks for the scale-up of the reaction are the high cost of TMS-Cl and the explosiveness of NH₄NO₃. During the process of optimizing the reaction for large-scale production, it was observed that adding some equivalents of water improved the yield from 86% without water to 92% with water and reduced 25% the formation of an undesirable by-product. The proposed intermediate of the reaction, TMS-O-NO₂, was questionable to exist in the presence of some water. To understand the mechanism of the reaction and the nature of the nitrating agent, the reaction was monitored in a high-pressure NMR tube due its exothermicity using the 400 MHz cryogen-free HTS magnet NMR system. ¹⁹F NMR was used to monitor the fluorinated species of the fluoro-toluene boronic acid and its nitrated product, and ¹H NMR for TMS-Cl and its product TMS-OH. RPKA and VTNA were used to set the DoE and determine the rate constants and reaction orders of the species in the reaction. The findings indicated that in the absence of water, boronic acid is in equilibrium with its boroxine trimer, making the reaction slower with an induction period of 5 h prior product formation. However, in the presence of water, TMS-Cl hydrolyses, first decreasing the induction period to 1 h and forming the product eliminating the equilibrium of the boronic acid with its boroxine trimer and with dimerization not occurring due to condensation. It was postulated that NO₂ radical (nitrogen dioxide radical) was the mediated nitration species of the reaction, and it was proven by running the reaction with fuming nitric acid directly.⁵⁵  However, nitric acid is a highly corrosive and toxic material that requires extreme care especially for scale-up reactions. NMR was used to compare with calorimetry to provide information on safety considerations for the nitration reaction with nitric acid.⁵⁶  Both techniques provided the same information. This study demonstrated that the role of NMR is beyond kinetics studies, and it is also a valuable tool to determine the safety conditions of reactions.⁵⁶ 

SSNMR is widely applied in the pharmaceutical industry during the development of drug substances as one of the techniques for the solid characterization of materials. In the solid state, drug substances can appear in different forms from amorphous to a variety of crystal forms. Only one form will be used for the final drug substances prior to being part of the drug products, and that form has the specific characteristics to support clinical trials. Many drug substances require an improvement in their solubility to meet the desired absorption and bioavailability criteria as part of their characteristics and specifications. Solubility of drug substances can be improved by forming cocrystals, solid phase dispersions for amorphous materials, and nanocrystals as examples. In addition, SSNMR has been applied to determine and quantify the different forms of drug substances. Examples in those areas are discussed subsequently. Extensive literature has been published on SSNMR, which is not the purpose of this chapter, but some examples are given in the reference list.^57–60 

About 70% of small-molecule drug substances form different polymorphs that need to be characterized. X-ray diffraction is the gold standard technique but can only be applied when single crystals are available. Alternative techniques are X-ray powder diffraction (XRPD) and SSNMR to determine the crystalline, hydrated, dehydrated, and salt forms of drug substances. Atorvastatin calcium (Figure 1.33), a statin drug prescribed to lower blood cholesterol, exists in a variety of polymorphic forms, 41 crystalline forms and 2 amorphous forms. Form I is the most stable crystalline form that has been studied by ¹³C, ¹⁹F, and ¹⁵N SSNMR magic angle spinning (MAS) with complete ¹³C and ¹⁹F chemical shift assignments based on ¹³C,¹H-heteronuclear correlation spectroscopy (HETCOR) and ¹³C,¹⁹F-HETCOR data.⁶¹  The NMR analysis indicated asymmetry in the unit cell with two atorvastatin molecules. The authors used the density functional theory and the experimental NMR data to propose the structure of Form I for the drug.⁶¹ 

Several solid-state characterization techniques are commonly used to determine the crystalline polymorphs in drug substances. A demonstration case is the combination of SSNMR experiments using the classic ¹³C cross-polarization magic angle spinning (CP/MAS) and electron diffraction (ED) techniques that were used to determine the crystal forms of several l-histidine samples as l-histidine HCl H₂O, orthorhombic, l-histidine, and monoclinic l-histidine.⁶² 

Cocrystals are widely employed to improve the solubility and bioavailability of drug substances. In cocrystals, the drug substance is mixed with another molecule where both components interact in the crystal structure through hydrogen bonds and charge transfers depending on their stoichiometric ratio. The drug metaxalone (Figure 1.34), a muscle relaxant and pain reliever, was studied forming cocrystals as metaxalone–nicotinamide (1 : 1), metaxalone–isonicotinamide (1 : 1), metaxalone–salicylamide (1 : 1), metaxalone–3-hydroxybenzoic acid (1 : 1), and metaxalone–4-hydroxybenzoic acid (1 : 1) hydrate by ¹³C SSNMR CP/MAS providing detailed structural information together with molecular conformation and hydrogen bonding in the cocrystals.⁶³ 

Tenapanor (Figure 1.35), a drug in development to treat irritable bowel syndrome, has been studied in two different solid forms, a crystalline anhydrous free base and an amorphous dihydrochloride salt.⁶⁴  SSNMR and dynamic nuclear polarization (DNP)-SSNMR were used for the study. DNP-SSNMR ¹³C–¹³C incredible natural abundance double quantum transfer experiment (INADEQUATE) was selected to study the free base because of the enhancement of the signal for low sensitivity nuclides. The salt was not compatible with the DNP radicals, and its structure was analyzed by SSNMR. The study showed the formation of microcrystalline structures in the amorphous salt indicating that both forms were in equilibrium, but it was pH-dependent during the crystallization process requiring control of the pH to avoid the formation of the undesired salt.⁶⁴ 

Polymorphs of drug substances may interconvert depending on temperature and humidity conditions and also when influenced by the composition of the excipients in the formulation. SSNMR is a technique that can determine changes in polymorphs and quantify them. Su et al.⁶⁵  developed a method for the crystalline and amorphous forms of γ-polymorph of indomethacin (Figure 1.36), a nonsteroidal anti-inflammatory drug, to determine their presence for further batch analysis. The protocol included the assigning of the ¹H, ¹³C, and ¹⁵N peaks for both forms using 1D and 2D homo-and heteronuclear correlation experiments. The authors used ultrafast spinning at 60 kHz to increase resolution. In addition, 1D ¹³C-edited spectra and 2D ¹³C-detected heteronuclear correlations were important to identify the spin systems and assign overlapping ¹³C signal.⁶⁵ 

Drug products are the final form of the drugs given to patients and are denominated as the formulated drug substances. The most common formulation is in the form of tablets or capsules for typical oral drugs, commonly used for small-molecule drugs. Other formulation types depend on factors such as patients’ age (e.g., children, adults, and geriatric adults), organs targeted, and others to have the appropriate route of administration, such as syrups for oral, creams for topical, and inhalation, injectables, and intravenous for parenteral administration. Biologics are formulated in appropriate buffer solutions for a parenteral route of administration, but they are not covered in this chapter. The formulated drugs are a mixture of the drug substances with the excipients, which are inactive ingredients that provide other enhancing properties to the drug substances such as appropriate absorption and bioavailability. NMR is employed in two main states: liquid and solid. In the liquid state, the focus is the study of the structure of impurities isolated from drug products during stability studies. In the solid state, the objective is to determine the forms of the drugs, any degradation due to interactions with the excipients, and quantitation of the drugs in the formulated solid forms. Several examples in those areas are discussed in the following sections.

Formulated drug products are exposed to forced degradation conditions to determine their stability profile and the shelf-life of the drug products. The studies are conducted under the ICH Q3B guidelines for drug products.¹⁶  Following are several examples of stability studies demonstrating the role of NMR for the structure elucidation of the impurities formed during those studies.

In the formulated dosage form, the drug substance is in physical contact with the excipients. The excipients are nominally inactive components that are supposed not to interact with the drug substance. Compatibility studies between excipients and drug substances are conducted to minimize undesirable interactions. However, under the conditions of heat and moisture, some unwanted interactions may occur. Drug substances containing primary and secondary amine moieties may react with carbohydrates containing the aldehyde functionality in the open-ring form as in glucose, maltose, and lactose. However, starch, a polysaccharide, is not typically part of compatibility studies as carbohydrates. A study carried out by LC–MS and NMR to understand the compatibility of desloratadine (Figure 1.37) and starch showed that degradation occurred indicating incompatibility of both components.⁶⁶  The formulated desloratadine was stressed under different forced degradation conditions. Under acid conditions, a new impurity was detected, and enough material was produced and isolated for its structural analysis. The impurity was an adduct of desloratadine with an isomer of acetylformoin and maltose (Figure 1.37). The authors proposed the mechanism of formation of the novel impurity.⁶⁶ 

Bisoprolol (Figure 1.38), a beta blocker used to treat hypertension and other related symptoms, is formulated as film-coated tablets. Accelerated and long-term degradation studies with heat and humidity were carried out to determine the stability of the drug product. The test in accelerated conditions produced an impurity at levels up to 0.47% by HPLC area percent, which based on ICH guidelines¹⁶  (ICH, Q3B(R2)), required complete structural characterization because it was at a higher level than the threshold of 0.2%.⁶⁷  The impurity increased in proportion when bisoprolol was exposed to temperature and calcium hydrogen phosphate. The impurity was isolated for full characterization by NMR and MS (MS and MS²). A combination of 1D ¹H NMR and ¹³C NMR, with 2D ¹H,¹³C-HSQC, and 1D ³¹P NMR experiment, provided the structure of the impurity as phosphomonoester of bisoprolol (Figure 1.38), which was agreeable with the MS data. In silico prediction of genotoxicity was conducted and resulted in no evidence.⁶⁷ 

Following the ICH guidelines,¹⁶  stability studies were performed on dolutegravir (Figure 1.39) as drug substance and in its formulated tablet dose. Dolutegravir is an antiretroviral approved for the treatment of HIV-1 infections. The degradation products were isolated for structural characterization by NMR and MS. The structures of two degradation products were determine and isolated from acid stress and oxidative stress, respectively.⁶⁸ 

During the accelerated stability study at 40 °C/75% RH of pregabalin (Figure 1.40),⁶⁹  an approved drug for anxiety disorders, two impurities were observed above the threshold of the Q3B ICH guidelines threshold.¹⁶  One of them was known from previous degradation conditions, pregabalin lactam (Figure 1.40), but the other was of unknown structure and never observed. The new impurity grew up to 0.33% in the stability samples stored for 3 months in accelerated stability conditions and also when the pregabalin extended-release tablets were in storage at 30 °C/65% RH. The novel impurity was isolated for its structural elucidation by NMR and MS. The proposed structure for the impurity was a pregabalin lactam methylene dimer (Figure 1.40) and it was synthesized for final structural proof. A proposed mechanism of formation was based on its structure to explain the dimerization and lactamization by degradation. In addition, kinetic studies were performed to determine the rate of formation of the dimer impurity and found to be as zero order.⁶⁹ 

Two important roles of solid formulations are to increase the solubility and the bioavailability of insoluble drug substances through the formation of nanocrystals, amorphous solid dispersions (ASD), and amorphous spray dried dispersions (SDD). In addition, polymorphs may interconvert in the solid formulation during the compression process to manufacture the final form of the drug products, the one delivered to patients. SSNMR plays an important role as a key analytical technique to study those processes to determine the characterization of the drug substances in solid formulations. Examples in those areas are discussed subsequently.

When drug candidates have poor solubility, forming small size crystals may be an option for consideration. With the recent development of manufacturing technologies, nanocrystals can be formulated into thin films, appropriate for continuous manufacturing. MAS SSNMR has been employed to study the kinetic process of nucleation and the structures of nanocrystals inside polymers with the examples of glycine, acetaminophen, and ibuprofen nanocrystals in porous cellulose membranes and cellulose acetate membranes.⁷⁰  In this study, T₁ (¹H) relaxation times were measured to determine any potential perturbation that might have occurred in the individual nanocrystal forms of the compounds in the membranes.⁷⁰ 

When crystalline solids have poor solubility and poor stability to form salts, ASD is selected as the drug delivery system with the objective to increase in vivo bioavailability. However, amorphous materials are thermodynamically unstable and can interconvert to crystalline forms undergoing chemical degradation. Polymers are used to minimize the instability of the drug substances and prevent the crystallization. The miscibility of the polymer–drug depends on how the drug is molecularly dispersed in the polymer matrix. AMG 517 (structure in Figure 1.13), an investigational drug to treat acute and chronic pain, has limited solubility with poor absorption and low bioavailability. In vivo tests indicated that formulation as ASD showed better exposure than as cocrystal. Therefore, the drug substance was formulated as ASD with the polymer hydroxypropyl methylcellulose-acetate succinate (HPMC-AS). SSNMR through the measurement of T₁(¹H) and T_1ρ(¹H) relaxation times indicated that the drug and the polymer were in close proximity over 10–20 nm in length, suggesting good miscibility of the drug in the polymer during the spray drying process.⁷¹  The same polymer, HPMC-AS, was used to form ASD with the GENE-A molecule (Figure 1.41) developed for rheumatoid arthritis. Relaxation times through SSNMR provided kinetic information to determine the miscibility of the drug in the polymer. The drug was sensitive to humidity; therefore, stability studies were conducted at various temperatures and humidity conditions (60 °C and 40 °C/75% RH) for 6 months. Based on T_1ρ(¹H) and power X-ray diffractometry (PXRD), 20% of the ASD sample remained amorphous during the stress time.⁷² 

Similar to ASD, SDD is defined as a mixture of amorphous drug substance and a polymer such as the polymer hydroxypropylmethylcellulose acetyl succinate M grade (HPMCAS-M) but with no other excipients in the formulated drug product. Two drugs with similar structures, BMS-903452 and BMS-986034 (Figure 1.42), were SDD formulated with HPMCAS-M to study their behavior under stress conditions in preparation for in vivo animal testing. BMS-903452 was the only one showing good stability with no evidence of loss of dissolution performance or crystallization during the studies when formulated at 30% (w/w%) and left at 40 °C/75% RH for 6 months. On the contrary, BMS-986034 was unstable when formulated with HPMCAS-M showing exothermic phase separation over three days at 37 °C/100% RH.⁷³ 

To understand the miscibility of the drug substances with polymers for ASD formulations, studies are conducted using 1D and 2D HETCOR SSNMR experiments to obtain information of the physical stability of amorphous formulations and to understand the interaction point at the molecular level. Posaconazole (POSA) (Figure 1.43), a second-generation triazole antifungal drug, has been formulated with two polymers, HPMCAS and HPMCP (hypromellose phthalate) using melt-quenching or spray-drying techniques. Studies on 1D and 2D SSNMR including ¹⁹F–¹³C, ¹⁵N–¹³C, and ¹⁹F–¹H polarization transfer, spin correlation, and ultrafast MAS with isotopic labelling provided information on the interaction between the polymers and POSA through the triazole and the difluorophenyl rings. The interactions were intramolecular hydrogen bonding (O–H ⋯ O = C and O–H ⋯ F–C), π–π aromatic packing, and electrostatic interactions. POSA was compared to itraconazole, which has a chlorine instead of the fluorine in POSA. The studies suggested that the presence of fluorine in POSA provided better binding interaction to fungal proteins through hydrogen bond donors of polar residues or hydrophobic side chain. This result may be related to POSA exhibiting better efficacy than other azoles, such as fluconazole, voriconazole, and itraconazole, to a few drug-resistant fungal infections.⁷⁴ 

Compounds can change their solid form during the process of manufacturing their drug products. Process-induced phase transformations (PIPTs) are conversions between polymorphs, polymorphic transitions, hydration to dehydration, dehydration to hydration, crystalline to amorphous, and amorphous to crystalline transformations that must be under control to manufacture stable and high-quality drug products. During the process of compression of the drug product to form pharmaceutical tablets, induced-phase transformations, one common PIPT type, may occur. ¹⁹F SSNMR experiments were used to investigate and quantify the amorphous content in POSA compressed tablets.⁷⁵  Two diluents, microcrystalline cellulose (MCC) and dibasic calcium phosphate anhydrous (DCPA), and a lubricant, magnesium stearate (MgSt), were evaluated during the compression of POSA in the tablets form. When the drug loads were low (5% w/w%), MCC and DCPA promoted the amorphization of POSA. When MCC-POSA tablets were blended with MgSt (1% w/w%), the amorphous content of POSA decreased, but was barely observed when using DCPA. The significance of the study was the critical effect on certain excipients that can induce amorphization during the compression process with the crystalline drug substance to form the tablets.⁷⁵  Further developments on understanding the interconversion to other solid forms of the drug substance pointed to more complex SSNMR experiments and the advantage of using ¹⁹F for fluorinated drugs. 3D SSNMR experiments based on ¹H and ¹⁹F as ¹H–¹⁹F–¹H, and ¹⁹F–¹⁹F–¹H correlations were applied to the fluorinated drug aprepitant (Figure 1.44) and its commercial nanoparticle formulation EMEND (Merck & Co., Inc.). The sensitivity was greatly enhanced by using 60 kHz ultrafast MAS for the analysis of samples in the milligram range.⁷⁶ 

Quantitation of the drug substances in the formulated drug products has been done in the solid form. The main purpose of the analysis is to determine the physical and chemical stability of the form because of the possibility of interconversion to other crystalline forms or amorphous. SSNMR has disadvantages due to the complexity of the instrumentation, large size, complex operations, maintenance, and installation in the manufacturing floor where the tablets are produced.

Studies have been conducted with a drug containing ³¹P using ³¹P SSNMR as reference method to develop a quantitative Raman approach using multivariate partial least squares (PLS) regression calibration technique.⁷⁷  Samples with different percentage of amorphous drug substance were prepared to test the methodology. Once Raman provided satisfactory results, it was implemented as an at-line process analytical tool (PAT) in the manufacturing floor. A chemometric model was developed for the measurement, and Raman provided fast and good results with sensitivity as a high-throughput method taking less than 5 min per tablet to quantitate amorphous conversions.⁷⁷ 

In our laboratories, we developed a methodology to measure the average content of fluorinated drugs in fluorine-containing drug products by low-field time-domain (TD) NMR.^78 ,79  The method requires the acquisition of a calibration curve with the drug substance prior to measuring the content of drug substance on tablets or capsules. Three drug products with fluorinated drug substances were tested, SENSIPAR® tables with cinacalcet HCl (structure in Figure 1.1) as the drug substance, PREVACID® and the generic version lansoprazole capsules with lansoprazole (Figure 1.45) as the drug substance, and generic cipro tablets with ciprofloxacin (Figure 1.45) as drug substance.

The methodology requires measuring the T₁ (¹⁹F) relaxation time of the drug substance to adjust the relaxation delay of the experiment for reliable quantitation. The results were comparable with the standard HPLC method used to measure the content uniformity of the drug substance in the formulated drug product. This TD-NMR method has several advantages: (1) it does not require destroying the tablets; (2) it is faster than the standard HPLC method; (3) it does not require developing a different method per drug product like for HPLC; (4) the benchtop NMR is simple to use by non-NMR experts and easy to maintain, which facilitates the use of this methodology in the manufacturing floor. Figure 1.46 illustrates the case of ciprofloxacin drug substance when using the free base and the HCl monohydrate as standards to build the calibration curves. In the case of the standard drug substance free base, the weight percentage of ciprofloxacin in intact and crushed tablets was indicated to be 137%. However, when using HCl monohydrate of the drug substance as standard, the average content of free ciprofloxacin in the intact and crushed tablets was 66%, which is around the expected value of the drug substance calculated as free base in the tablets (65%). The main consideration is to build the calibration curve using the drug substance with the same chemical and physical forms as when formulated in the drug product to account with the same relaxation time and NMR signal decay.

The role of NMR is fundamental during the development of drug substances and drug products in the pharmaceutical industry. Structure elucidation and quantitation of all the materials to produce the drug substances are a requirement for quality control and confirmation of their correct structures. In addition, structure determination of in-process impurities provides chemical process understanding, and structural analysis of impurities from stability studies supports stability data for drug substances. Relative configuration of drugs can be assessed by anisotropic NMR using anisotropic media, which can be used as an orthogonal method during the manufacturing of the drug substances for their production control strategies. NMR also plays an important role in reaction monitoring for process understanding, kinetic analysis, and quantitation of materials for mass balance and reaction yields. SSNMR is key to determine the polymorphism analysis of drug substances towards the stability of the drug. When the drugs are formulated as drug products, impurity structural analysis during stability studies provides information of the shelf-life of the drugs. In addition, SSNMR studies the interactions of the excipients with the drug substances, which are key to determine the stability and compatibility of drug substances in formulated drug products. Another application is the content of drug substances in the drug products in the solid form performed by low- and high-field NMR instruments. All those studies strongly describe the extensive role of NMR in the development of drugs for the pharmaceutical industry as drug substances and their formulated drug products.

ACN

Acetonitrile

ADEQUATE

Adequate double quantum transfer experiment

API

Active pharmaceutical ingredient

ASD

Amorphous solid dispersions

BOC

tert-Butoxycarbonyl

COA

Certification of analysis

COSY

Correlation spectroscopy

CP

Cross polarization

DCPA

Dibasic calcium phosphate anhydrous

DFT

Density functional theory

DNP

Dynamic nuclear polarization transfer experiment

ED

Electron diffraction technique

EDQM

European Directorate for the Quality of Medicines

GC

Gas chromatography

GLP

Good laboratory practice

GMP

Good manufacturing practice

HETCOR

Heteronuclear correlation spectroscopy

HMBC

Heteronuclear multiple bond correlation spectroscopy

HPLC

High-performance liquid chromatography

HPMC-AS

Hydroxypropyl methylcellulose-acetate succinate polymer

HPMCAS-M

Hydroxypropylmethylcellulose acetyl succinate M grade polymer

HPMCP

Hypromellose posaconazole phthalate polymer

HRMS

High-resolution MS

HRMS/MS

High-resolution tandem MS/MS

HSQC

Heteronuclear single quantum coherence spectroscopy

HTS

High-temperature superconducting material in coils for magnets

ICH

International Conference on Harmonization of Technical Requirement for Pharmaceuticals for human use

INADEQUATE

Incredible natural abundance double quantum transfer experiment

IR

Infrared spectroscopy

LC–MS

Liquid chromatography mass spectrometry

MAS

Magic angle spinning

MCC

Microcrystalline cellulose

MgSt

Magnesium stearate

MS

Mass spectrometry

NMR

Nuclear magnetic resonance

NOE

Nuclear Overhauser effect

NOESY

Nuclear Overhauser effect spectroscopy

OOS

Out of specification

PAT

Process analytical tool

PBLG

Poly-γ-(benzyl-l-glutamate) polymer

PIPT

Process-induced phase transformations

PLS

Partial least squares regression technique

PMMA

Poly(methyl methacrylate) polymer

PXRD

Power X-ray diffractometry

qNMR

Quantitation by NMR

QSAR

Quantitative structure activity relationship

RCSA

Residual chemical shift anisotropy

R&D

Research and development

RDC

Residual dipolar couplings

RH

Relative humidity

ROE

Rotating frame Overhauser effect

ROESY

Rotating frame Overhauser effect spectroscopy

SDD

Amorphous spray dried dispersions

SSNMR

Solid-state NMR

TD

Time domain

TLC

Thin layer chromatography

UPLC

Ultra-performance liquid chromatography

US FDA

United States Food and Drug Administration

USP US

Pharmacopeia monograph

UV

Ultraviolet

VICH

Harmonization of Technical Requirements for Registration of Veterinary Medical Products guidelines

WHO

World Health Organization