Skip to main content
NCBI home page
RSC Medicinal Chemistry logoLink to RSC Medicinal Chemistry
. 2024 Jun 11;15(7):2372–2379. doi: 10.1039/d4md00244j

A method for estimation of plasma protein binding using diffusion ordered NMR spectroscopy (DOSY)†

Rachel Taylor a, Thomas Swift b, David Wilkinson c, Kamyar Afarinkia a,d,✉
PMCID: PMC11253862  PMID: 39026647

Abstract

The plasma protein binding (PPB) of a drug plays a key role in both its pharmacokinetic and pharmacodynamic properties. During lead optimisation, medium and high throughput methods for the early determination of PPB can provide important information about potential PKPD profile within a chemotype or between different chemotype series. Diffusion ordered spectroscopy (DOSY) is an NMR spectroscopic technique that measures the diffusion of a molecule through the magnetic field gradient, according to its molecular size/weight. Here, we describe the use of DOSY for a rapid and straightforward method to evaluate the PPB of drug molecules, using their binding to bovine serum albumin (BSA) as a model.

We describes a fast and simple method for quantitative determination of plasma protein binding of drug molecules using diffusion ordered spectroscopy (DOSY).graphic file with name d4md00244j-ga.jpg

Introduction

Human blood contains about 6–8% of soluble proteins which are essential for its function.1,2 In particular, these proteins (commonly referred to as plasma proteins) bind and transport many biologically important molecules such as lipids and hormones. Administered drugs also bind plasma proteins and are transported via blood to various tissues and organs. Albumin accounts for just over half of all human blood proteins.1–4 This is followed by various globulins which account for just under 40%.1,2,5 The main component of the remainder is fibrinogen, which is involved in the coagulation of blood in the case of injury.

The plasma protein binding (PPB) of a drug plays a key role in both its pharmacokinetic (PK) and pharmacodynamic (PD) properties.6–9 Without energy-dependent processes, the free drug concentration in plasma correlates to that found in tissues, and it is this tissue drug concentration that promotes binding to the target, which in turn elicits a pharmacological effect.8 The level of unbound drug in the plasma therefore correlates to the concentration of available drug in the tissues and its pharmacological action. Although low plasma protein binding translates to higher fraction of unbound drug, fu, and lower volume of distribution, it can also promote faster clearance or metabolism. So, an understanding of the PPB enables a better appreciation of the factors that can contribute to the in vivo efficacy of a drug or its safety profile. Therefore, methods that enable determination of PPB are highly desirable.9

There are a number of existing methods for determining PPB of drug molecules, each with their own advantages and drawbacks. Many of these protocols use binding to purified albumin in place of plasma proteins, partly because it is cheaper and more easily accessible but also because its quality does not change from one batch to another, reducing variability in experimental results and enabling a more reliable comparison between the results.10–17

The gold standard for measuring the extent of PPB is equilibrium dialysis (ED).15 ED involves partitioning of a drug between two compartments, separated by a semi-permeable membrane that functions as a molecular filter based on differences in molecular size and weight.18 Only molecules below a particular molecular weight threshold have the ability to permeate through the membrane. So, whilst drug molecules permeate through the membrane, heavier proteins or drug–protein complexes do not. Equilibration times are typically long (12–48 hours),18 so one drawback of ED is the slowness of the method. Rapid equilibrium dialysis (RED) is a high-throughput, 96-well plate version which enhances the efficiency of the ED method. RED allows for equilibration times ranging from 1.5 to 4 hours, depending on the rate of agitation.

However, the adsorption of some compounds at surfaces and membranes of ED devices can have an impact on the measurement, particularly since this ‘parasitic’ binding sometimes exceeds 50% of the total concentration.17,18 The Gibbs–Donnan effect, whereby charged particles in close proximity to a semi-permeable membrane do not evenly distribute on either side of the membrane, is also problematic in ED methods and results in an uneven distribution of charged moieties across the membrane.8,16 This effect can however be prevented through the use of a sufficiently strong ionic buffer, or the dilution of whole plasma prior to dialysis. Other drawbacks for ED are the limitations for non-specific, high binding compounds, as well as issues of plasma instability, and the inability of some compounds to diffuse through the dialysis membrane.18

Other commonly used methods for the evaluation of PPB include ultrafiltration and ultracentrifugation. Ultrafiltration is one of the simplest and fastest methods for fu determination, making it a useful tool in drug monitoring studies.8 It is similar to ED except in that it utilises the application of pressure to increase the speed of analysis by forcing the solution through the membrane.18 Ultrafiltration still suffers from the same drawbacks as ED, including the Gibbs–Donnan effect, protein leakage through the membrane, and nonspecific binding of compounds to the filter membrane. Due to the high pressure enforced for rapid analysis times, the volumes of ultrafiltration cannot exceed 10% of the total sample volume in order to maintain equilibrium.8,16 Ultracentrifugation is also related to ED, but instead of separation via a membrane, centrifugal force is used to separate the protein from the free drug.18 The advantage of this method is that it eliminates membrane effects, such as the Gibbs–Donnan effect and adsorption, but ultracentrifugation also presents its own issues. A lipid layer can form as high-density proteins like albumin will sink, but lower-density lipoproteins will float, leaving the protein-free drug layer in the middle. This can cause difficulties in the sampling for this method. Moreover, the method requires expensive equipment, is low-throughput, and can give falsely high results for binding due to free-drug sedimentation, viscosity, or back-diffusion.8,18

Each of the above methods involves physical separation of constituent molecules, but various spectroscopic techniques17–20 particularly nuclear magnetic resonance (NMR) spectroscopy17,21–23 which do not involve separation, can also be employed in the determination of PPB. NMR methods are used to determine PPB by analysing information on changes that occur in the spectrum of an unbound drug (or unbound protein) following protein-drug-binding particularly at high-affinity binding sites. For example, one dimensional 1H NMR studies measure the degree of line perturbation of small molecule signals following plasma protein interaction.

One drawback of using NMR methods is that due to the timescale of experiments, signals are often averaged of the bound and unbound molecules. Nevertheless, this can allow for the derivation of linear relationships, providing quantitative information on the bound fraction of the molecule.17 However, using NMR to determine PPB binding does benefit from a number of important features compared to its counterparts. To start with, the sample preparation is minimal and fast, and the NMR measurement data acquisition speed is comparatively fast compared to other techniques. In addition, running of samples, and the materials required, are inexpensive, and the role of each of the protein components of plasma can be separately assessed in binding via NMR method.17,21–23 NMR spectrometers are already ubiquitously used in most medicinal chemistry laboratories. Perhaps most importantly though, NMR based techniques can relatively easily be adapted for automation which allows rapid collection of data on a series of chemotypes. However, one-dimensional NMR methods also have numerous drawbacks. In particular, spectra of most proteins are complex and it is often difficult to distinguish signals for individual groups of protons within the spectra of a mixture of protein, unbound drug and drug–protein adduct.18 To overcome this many groups favour high power, high field instruments to improve resolution of complex overlapping protein peaks.24 Whilst this does not affect per sample complexity, cost or run time, it does require a vast increase in capital expenditure to both access and house these instruments.

On the other hand, two dimensional NMR spectroscopy such as DOSY, enable us to discern signals from mixtures of compounds. DOSY measures diffusion constants of molecules by measuring the rate by which they are displaced within a magnetic field gradient. The diffusion constant for a molecule depends on a number of factors in particular viscosity of the medium, and its size (molecular volume or weight).

Here, we describe a quick, straightforward and cost effective DOSY-based method to rank binding affinity of different molecules, including drug molecules, to bovine serum albumin (BSA) and show that this data is in agreement with information from the literature on PPB binding of these molecules.

Method

Caffeine (C0750), diclofenac (SML3086), propranolol (P8688), acetyl salicylic acid (S5922), theophylline (T1633), d-glucosamine hydrochloride (G1514), d-ribose (R7500), cimetidine (C4522), esomeprazole (E7906), DMSO (472301), bovine serum albumin (A7030) and deuterium oxide (151882), were all purchased from sigma. The relevant quantity of each drug was dissolved in D2O (>99.5% isotope purity, 5 mL), to afford a 10 mM solution (solution A). BSA (332.4 mg) was dissolved in D2O (5 mL), yielding a 1 mM solution (solution B). The required solutions of BSA, drug and BSA plus drug were prepared by combining aliquots of solutions A, B and D2O to a total of 600 μL volume in a 5 mm diameter NMR tube. For example, five 600 μL solutions were prepared by adding 100 μL of solution B (BSA), x μL of solution A (drug) and (500–x) μL of D2O, where x is 0, 100, 200, 300, 400 μL. In this example, the final concentration of BSA is 167 μM and the final concentrations of drug, which depends on quantity of x, are zero, 10 fold (1.67 mM), 20 fold (3.34 mM), and 40 fold (6.68 mM). Concentration of solution A can be adjusted to afford a wide range of drug : BSA ratios. The DOSY spectra were collected using a Bruker Spectrospin 400 Ultrashield NMR spectrophotometer operating at 400 MHz, with samples maintained at 298.2 K via a condensed gas feed (400 lph). 1H DOSY NMR measurements were set up with 16 gradient scans with 16 N repeat samples across each gradient. All measurements were initiated using Bruker TopSpin (version 3.7) with IconNMR automation. The pulse sequence ledbpgp2s was used (see ESI† for more details). Calculation of the diffusion coefficient were done using a Bruker Dynamic Centre (version 2.4.11). All experiments were carried out in triplicate. All fits were carried out using a single exponential diffusion decay with an R2 > 99%. GraphPad Prism 8 was used for the production of graphs and quantification of the results. Multiple exemplar analysis of raw data are provided in ESI† file.

Results

Investigation of binding of caffeine to BSA

We first set out to test the binding of caffeine to BSA. Caffeine has a relatively low-medium binding affinity to plasma proteins25,26 and is often used as a standard/control in the existing methods for PPB determination. DOSY was used to determine diffusion coefficients (D) on samples containing different concentration of caffeine only, BSA only and caffeine/BSA mixtures which correspond to caffeine : BSA ratios ranging from 2 : 1 to 40 : 1 (Fig. 1A). For both caffeine and BSA, diffusion coefficients are constant over the concentration range (Fig. 1A, blue crosses for caffeine and green crosses for BSA). As expected however (see Discussions), there is a steady rise in the observed, averaged diffusion coefficient for caffeine in the presence of BSA which eventually reaches a plateau at just below the diffusion coefficients of pure caffeine (Fig. 1A black crosses). An observed Dmax value can be calculated by fitting a hyperbolic curve on these datapoints. Repeating experiment (n = 3) afforded very similar observed Dmax values (standard deviation = 1.67 × 10−11 m2 s−1 or 2.5% of the mean) confirming the reproducibility of the experiments (Fig. 1B). Although caffeine is quite soluble in (deuterated) water, many investigative molecules are not. In these instants, it is commonplace to dissolve test articles in dimethyl sulfoxide (DMSO), prior to addition to aqueous media to assist with their solubilisation. Therefore, we investigated the effect of adding deuterated DMSO (up to 3% v/v). Again, the experiment afforded very similar observed Dmax value (Fig. 1B) confirming small quantities of DMSO do not affect the values for observed Dmax. Interestingly, although there is a steady increase in the observed Dmax value from these experiments, the values were still just below the theoretical Dmax (that for caffeine alone). To show that the observed D values eventually converge to the D value for caffeine alone, we repeated the experiment at 100 : 1 ratio of caffeine to BSA. As expected, the observed Dmax value raised to nearer the value for theoretical Dmax for caffeine alone (Fig. 1C).

Fig. 1. (A) Changes in diffusion coefficient, D, of caffeine over a concentration range in the absence of BSA (blue crosses) and in the presence of BSA (black crosses). (B) Comparison between the diffusion coefficients of caffein/BSA mixtures in the presence or absence of 3% v/v DMSO (C) the maximum value for diffusion coefficient, Dmax, of a mixture of caffeine and BSA steadily rises towards the value for caffeine alone as the ratio of caffeine : BSA increases.

Fig. 1

Open in a new tab

Determination of binding of diclofenac and propranolol to BSA

In contrast to caffeine which is reported to have a low-medium PPB binding, diclofenac is reported to have a high PPB binding.25,27 So we set out to investigate the D value for diclofenac : BSA mixtures and compare it with that of caffeine : BSA mixtures over the same range (2 : 1 to 40 : 1). Again, DOSY was used to determine diffusion coefficients of diclofenac over the concentration range to show that these values remain constant (Fig. 2A). As expected, there is a steady rise in the observed, averaged diffusion coefficient for diclofenac with an increase in the ratio of diclofenac : BSA. However, in contrast to the observation with caffeine, the fitted curve is shallower and shows a significantly slower rise. Because of this, and to show that the D values have not yet reached a plateau at 40 : 1 ratio, we increased the ratio first to 100 : 1 and then to 600 : 1. As expected, the observed D value raises steadily towards the Dmax value for diclofenac (Fig. 2B). This observation is wholly consistent with a higher PPB binding for diclofenac than for caffeine. Because of the higher affinity of diclofenac, the proportion of free drug to bound drug is low, meaning that contribution of the free drug to the observed D remains small unless a much larger ratio of diclofenac is used.

Fig. 2. (A) Changes in diffusion coefficient, D, of propranolol (blue) and diclofenac (green) over a concentration range in the presence of BSA (blue crosses). Control values for absence of BSA are also shown. (B) The maximum value for diffusion coefficient, Dmax, of a mixture of caffeine and BSA steadily rises towards the value for diclofenac alone as the ratio of diclofenac : BSA increases.

Fig. 2

Open in a new tab

Propranolol (295.80 g mol−1) has a similar molecular weight to diclofenac (296.15 g mol−1), however it is reported to have a weaker binding to BSA.25,28 To demonstrate that the differences in observed D values for diclofenac and caffeine where not merely a consequence of the two molecules' different molecular weights, we measured the D value for propranolol : BSA mixtures and compare it with that of diclofenac : BSA mixtures over the 2 : 1 to 40 : 1 concentration range (Fig. 2A). Averaged diffusion coefficient for propranolol increases with the ratio of drug : BSA. However, the fitted curve is steeper and shows a significantly faster rise than observed for diclofenac. This suggests that propranolol has a weaker binding to BSA and in deed the reported value for PPB of propranolol25,28 is lower than that for diclofenac.25,27

Determination of binding of theophylline, acetylsalicylic acid and glucosamine to BSA

As was the case for propranolol and diclofenac, theophylline (180.16 g mol−1), glucosamine (179.17 g mol−1) and acetylsalicylic acid (180.16 g mol−1) have similar molecular weight but very different physical and chemical properties which means we expect each of them to bind very differently to BSA. Theophylline is however structurally very similar to caffeine which means we expect its binding to BSA to be very similar to that of caffeine.

DOSY was used to determine diffusion coefficients of theophylline, acetylsalicylic acid and glucosamine over the concentration range (2 : 1 to 40 : 1). As can be seen below (Fig. 3A), theophylline and caffeine not only have similar observed Dmax values, but their respective fitted curves are also similar in shape. In contrast, acetylsalicylic acid has a very different observed Dmax and a much shallower fitted curve suggesting it binds to BSA more strongly than theophylline (Fig. 3B).

Fig. 3. (A) Similarity in changes in diffusion coefficient, D, of theophylline (blue) and caffeine (black) over a concentration range in the presence of BSA. (B) Changes in diffusion coefficient, D, of theophylline (blue), acetylsalicylic acid (black) and glucosamine over a concentration range in the presence of BSA. Control values for absence of BSA are not shown for clarity.

Fig. 3

Open in a new tab

Quantitative comparison of binding to BSA

In addition to the drugs and chemical compounds mentioned above, we also investigated the binding of ribose, cimetidine and esomeprazole (as magnesium sulfate salt) to BSA (Fig. 4). Examination of the curves confirms that esomeprazole is a high binding molecule whilst cimetidine is a low binding drug. Ribose, which appears as medium binding in comparison with the other two molecules has not been previously investigated for PPB. Interestingly, in contrast to the similarity between theophylline and caffeine, the curves for ribose and glucosamine are quite different, suggesting very different binding affinities to BSA, even though the molecules are both carbohydrate sugars.

Fig. 4. Differences in changes to the averaged diffusion coefficient of different molecules in the presence of BSA: (A) diffusion coefficient, D, of ribose over a concentration range in the presence of BSA. (B) Diffusion coefficient, D, of cimetidine over a concentration range in the presence of BSA. (C) Diffusion coefficient, D, of esomeprazole over a concentration range in the presence of BSA. (D) Changes in diffusion coefficient, D, of ribose, cimetidine and esomeprazole over a concentration range in the presence of BSA. Control values for absence of BSA are not shown for clarity.

Fig. 4

Open in a new tab

Clearly, examination of the curves provides a qualitative estimation of the binding affinities of different drug molecules to BSA. As the concentration of drug relative to a fixed concentration of BSA is increased the observed averaged diffusion coefficient for the molecule also increases approaching the value for the diffusion coefficient for the molecule in solution on its own. For molecules with low binding to BSA, in this example cimetidine, the rise to its corresponding maximum is sharper, whereas for molecules with high binding to BSA, in this example esomeprazole, the rise to its corresponding maximum is shallow.

Although the observation of fitted curves allows a qualitative measure of binding of different molecules to BSA, we wanted to quantitatively estimate the fraction of bound drug. One way to do this is to measure the relative observed diffusion coefficient (Dobs) to the Dmax value at a given ratio of drug : BSA for each drug/molecule.

We argued that the Dobs at any given drug concentration is an averaged value from all bound and unbound molecules, or Dobs = (Db × fu) + (Du × fb) where Db would be the diffusion coefficient of entirely bound drug, Du would be the diffusion coefficient of entirely unbound drug, and fb and fu would be the fraction of bound and unbound drug respectively. Since fb + fu =1, then fb = (Du − Dobs)/(Du − Db). The value of Du is the same as the diffusion coefficient of the drug in the absence of BSA (Dmax) which can be experimentally determined (Fig. 5). The value of Db however, can only be estimated. Value of Db corresponds to when there is a large excess of BSA so that even if the drug molecule has little affinity, majority of it would be bound to BSA. However, as the ratio of drug : BSA is reduced, it would be more difficult to discern the peaks due to the small molecule drug from protein, and hence not possible to determine diffusion coefficient. We found that for all molecules in this study, determination of diffusion coefficient was only possible at drug : BSA ratios >2.

Fig. 5. Estimation of the relative values of fb at a given drug : BSA ratio.

Fig. 5

Open in a new tab

Table 1 shows the estimated values of relative BSA binding (fb) at 5 : 1, 10 : 1 and 20 : 1 ratio of drug : BSA for molecules in this study, as well as previously reported values for the whole human plasma binding of the drugs from the literature. Obviously, this qualitative comparison carries a number of caveats. There may be differences between human and bovine serum albumin. Furthermore, albumin is only one, albeit the major component of blood proteins, and therefore, binding to other protein components of plasma can significantly change values for PPB than the ones estimated from binding to BSA. Nevertheless, as can be seen the ranking of BSA binding from our study closely mirrors the PPB binding values reported for these drugs.

Comparison of the PPB values reported in the literature to the values for fraction binding to BSA (fb) in this study.

Molecule PPB reported in literature (%) Fraction bound (fb) to BSA (%) from this study
At 5 : 1 (%) At 10 : 1 (%) At 20 : 1 (%)
Diclofenac 99.5 (ref. 25 and 27) 100 94 90
Esomeprazole 97 (ref. 28) 100 98 96
Propranolol 88,25,29 86–88 (ref. 30) 79 59 51
Acetylsalicylic acid 55;25 <50 (ref. 31) 90 84 52
Theophylline 56,25 40 (ref. 32) 71 41 23
Caffeine 37.5 (ref. 26) 73 33 29
Cimetidine 20,33 22.5 (ref. 34) 58 34 26
d-Glucosamine N/A 74 54 29
Ribose N/A 89 75 64
Open in a new tab

Discussion

Poor pharmacokinetic/pharmacodynamic (PKPD) profile is one of the main reasons for the lack of in vivo efficacy and failure of clinical progression of drugs.35–37 PPB values have a strong association with PK/PD parameters which is the reason their evaluation is an integral part of any drug development programme. It has been recognised that during lead identification and lead optimisation steps, tools that can inform the likely levels of PPB are useful in guiding the direction of a drug discovery programme.7,8 However, the scale and speed of lead identification and optimisation programmes dictates that methods for estimating PPB must be rapid, simple and cost-effective, in order to be able to properly inform the progression of potential drugs through preclinical stages. As the demand for evaluation of PPB data for drug development programmes has grown, so too has an interest in the development of newer methods that meet these criteria.

Diffusion based NMR experiments were first proposed as a broad technique for measuring ligand-protein binding in 2004.38 Since then, several studies have employed DOSY to study molecule-protein interactions. For example, changes in molecular diffusion have been used as evidence of chemical interaction between proteins and two small molecules.39 DOSY has been shown as an excellent quantitative indicator of binding constants between small molecules,39,40 and between small molecules and synthetic polymers.41

In 2015 Aroulmoji et al. demonstrated that DOSY studies on hyaluronate – methotrexate conjugates with bovine serum albumin produced dissociation constants of an equivalent order of magnitude to that from classical fluorescence titrations,42 and shortly afterwards Denis-Quanquin et al. demonstrated that the diffusion constants between paramagnetic inorganic complexes and proteins could be easily extrapolated to monitor protein binding to tris-dipicolinate lanthanide complexes.43 Although these are some examples where DOSY has been employed to provide binding information, NMR method traditionally used to ascertain this information is primarily via1H chemical shift changes.43

The work described here, builds on these earlier studies to use of DOSY to determine binding affinity of drug molecules to bovine serum albumin, as a model for human plasma proteins. Therefore, it can quickly and easily provide information that is critical to understanding PK profile of potential drugs. The methodology is straight forward, experimentally simple to perform, and reasonably quick. As we have shown here, the method is reproducible and using binding to BSA as a model, gives quantitative estimation of plasma protein binding. The issue with solubility of test articles can be avoided since up to 3% v/v deuterated DMSO is tolerated in the experiments.

We should note here that the exact values of diffusion coefficients can be affected by temperature gradient within the sample which lead to convection44 and sample viscosity.45,46 There are pulse sequences that can further compensate for convection within a sample,47 however strict control of the rate of gas flow used for temperature regulation, is required to ensure repeatability in diffusion coefficient measurements. The method provided maintains a constant concentration of protein, reducing issues with variation of viscosity between samples, which is a known concern when sampling chemical diffusion, particularly if mixed solvents are involved. In this study we have observed no significant between diffusion coefficient values were seen in 3% v/v deuterated DMSO in D2O compared with D2O alone. Despite these issues, methodology we have described has significant versatility and a wide scope. For example, whilst we have focused on binding to BSA as a model system representing blood plasma proteins, the methodology can also be applied, in principle to other individual proteins in plasma, to human serum albumin as a better measure of PPB binding in humans, or even to whole human plasma. Therefore, we plan to further expand the use of DOSY for estimating PPB binding of drugs and will report our results in due course.

Data availability

The data generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Author contributions

Experimental work was carried out by R. T. and T. S. Experiments were designed by R. T., D. W., and K. A. Manuscript was written by K. A. and all authors have reviewed the manuscript.

Conflicts of interest

The authors declare no competing interests.

Supplementary Material

MD-015-D4MD00244J-s001

Acknowledgments

We thank University of Bradford for financial support.

†

Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4md00244j

References

  1. Leeman M. Choi J. Hansson S. Storm M. U. Nilsson L. Anal. Bioanal. Chem. 2018;410(20):4867–4873. doi: 10.1007/s00216-018-1127-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Schenk S. Schoenhals G. J. de Souza G. Mann M. BMC Med. Genomics. 2008;15(1):41. doi: 10.1186/1755-8794-1-41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Sethi P. K. White C. A. Cummings B. S. Hines R. N. Muralidhara S. Bruckner J. V. Pediatr. Res. 2016;79(3):409–415. doi: 10.1038/pr.2015.237. [DOI] [PubMed] [Google Scholar]
  4. Vincent J. L. Best Pract. Res., Clin. Anaesthesiol. 2009;23(2):183–191. doi: 10.1016/j.bpa.2008.11.004. [DOI] [PubMed] [Google Scholar]
  5. Hovaida M. Torrazza-Perez E. McCarthy D. Dig. Dis. Sci. 2018;63(1):32–35. doi: 10.1007/s10620-017-4849-7. [DOI] [PubMed] [Google Scholar]
  6. Liu X. Chen C. Hop C. E. Curr. Top. Med. Chem. 2011;11(4):450–466. doi: 10.2174/156802611794480918. [DOI] [PubMed] [Google Scholar]
  7. Howard M. L. Hill J. J. Galluppi G. R. McLean M. A. Comb. Chem. High Throughput Screening. 2010;13(2):170–187. doi: 10.2174/138620710790596745. [DOI] [PubMed] [Google Scholar]
  8. Bohnert T. Gan L.-S. J. Pharm. Sci. 2013;102(9):2953–2994. doi: 10.1002/jps.23614. [DOI] [PubMed] [Google Scholar]
  9. Trainor G. L. Expert Opin. Drug Discovery. 2007;2(1):51–64. doi: 10.1517/17460441.2.1.51. [DOI] [PubMed] [Google Scholar]
  10. Suresh P. K. Divya N. Nidhi S. Rajasekaran R. Spectrochim. Acta, Part A. 2018;193:523–527. doi: 10.1016/j.saa.2017.12.069. [DOI] [PubMed] [Google Scholar]
  11. Alam M. A. Awal M. A. Subhan N. Mostofa M. J. Health Popul. Nutr. 2009;27(1):20–30. doi: 10.3329/jhpn.v27i1.3315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. SreedharanNair S. Unni K. K. Sasidharanpillai S. Kumar S. Aravindakumar C. T. Aravind U. K. Eur. J. Pharm. Sci. 2022;172:106141. doi: 10.1016/j.ejps.2022.106141. [DOI] [PubMed] [Google Scholar]
  13. Li T. Cheng Z. Cao L. Jiang X. Food Chem. 2016;194:740–748. doi: 10.1016/j.foodchem.2015.08.067. [DOI] [PubMed] [Google Scholar]
  14. Wani T. A. Bakheit A. H. Abounassif M. A. Zargar S. Front. Chem. 2018;6:47. doi: 10.3389/fchem.2018.00047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lambrinidis G. Vallianatou T. Tsantili-Kakoulidou A. Adv. Drug Delivery Rev. 2015;86:27–45. doi: 10.1016/j.addr.2015.03.011. [DOI] [PubMed] [Google Scholar]
  16. Bowman C. M. Benet L. Z. Eur. J. Pharm. Sci. 2018;123:502–514. doi: 10.1016/j.ejps.2018.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gallo M. Matteucci S. Alaimo N. Pitti E. Orsale M. V. Summa V. Cicero D. O. Monteagudo E. J. Pharm. Biomed. Anal. 2019;167:21–29. doi: 10.1016/j.jpba.2019.01.049. [DOI] [PubMed] [Google Scholar]
  18. Vuignier K. Schappler J. Veuthey J. L. Carrupt P. A. Martel S. Anal. Bioanal. Chem. 2010;398(1):53–66. doi: 10.1007/s00216-010-3737-1. [DOI] [PubMed] [Google Scholar]
  19. Pacholarz K. J. Garlish R. A. Taylor R. J. Barran P. E. Chem. Soc. Rev. 2012;41(11):4335–4355. doi: 10.1039/C2CS35035A. [DOI] [PubMed] [Google Scholar]
  20. Tatlidil D. Ucuncu M. Akdogan Y. Phys. Chem. Chem. Phys. 2015;17(35):22678–22685. doi: 10.1039/C5CP03583J. [DOI] [PubMed] [Google Scholar]
  21. Maciążek-Jurczyk M. Sułkowska A. Bojko B. Równicka-Zubik J. Sułkowski W. W. Spectrochim. Acta, Part A. 2011;82(1):181–190. doi: 10.1016/j.saa.2011.07.030. [DOI] [PubMed] [Google Scholar]
  22. Fielding L. Rutherford S. Fletcher D. Magn. Reson. Chem. 2005;43(6):463–470. doi: 10.1002/mrc.1574. [DOI] [PubMed] [Google Scholar]
  23. Maity S. Gundampati R. K. Kumar T. K. S. Nat. Prod. Commun. 2019;14(5):1–17. doi: 10.1177/1934578X19849296. [DOI] [Google Scholar]
  24. Skinner A. L. Laurence J. S. J. Pharm. Sci. 2008;97(11):4670–4695. doi: 10.1002/jps.21378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Zhang F. Xue J. Shao J. Jia L. Drug Discovery Today. 2012;17(9–10):475–485. doi: 10.1016/j.drudis.2011.12.018. [DOI] [PubMed] [Google Scholar]
  26. Blanchard J. J. Pharm. Sci. 1982;71(12):1415–1418. doi: 10.1002/jps.2600711229. [DOI] [PubMed] [Google Scholar]
  27. Chan K. K. Vyas K. H. Brandt K. D. J. Pharm. Sci. 1987;76(2):105–108. doi: 10.1002/jps.2600760204. [DOI] [PubMed] [Google Scholar]
  28. https://go.drugbank.com/drugs/DB00736, (Accessed June 2024)
  29. Ahmad S. Baker D. Murnane D. Spooner N. Gerhard U. Bioanalysis. 2021;13(14):1101–1111. doi: 10.4155/bio-2021-0109. [DOI] [PubMed] [Google Scholar]
  30. Kelly J. G. McDevitt D. G. Br. J. Clin. Pharmacol. 1978;6(2):123–127. doi: 10.1111/j.1365-2125.1978.tb00836.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lee S. Johnson D. Klein J. Eppler J. Vet. Hum. Toxicol. 1995;37(3):224–225. [PubMed] [Google Scholar]
  32. https://go.drugbank.com/drugs/DB00277, (Accessed June 2024)
  33. Somogyi A. Gugler R. Clin. Pharmacokinet. 1983;8(6):463–495. doi: 10.2165/00003088-198308060-00001. [DOI] [PubMed] [Google Scholar]
  34. https://go.drugbank.com/drugs/DB00501, (Accessed June 2024)
  35. Sun D. Gao W. Hu H. Zhou S. Acta Pharm. Sin. B. 2022;12(7):3049–3062. doi: 10.1016/j.apsb.2022.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Harrison R. K. Nat. Rev. Drug Discovery. 2016;15(12):817–818. doi: 10.1038/nrd.2016.184. [DOI] [PubMed] [Google Scholar]
  37. Cook D. Brown D. Alexander R. March R. Morgan P. Satterthwaite G. Pangalos M. N. Nat. Rev. Drug Discovery. 2014;13(6):419–431. doi: 10.1038/nrd4309. [DOI] [PubMed] [Google Scholar]
  38. Lucas L. H. Larive C. K. Concepts Magn. Reson., Part A. 2004;20:24–41. doi: 10.1002/cmr.a.10094. [DOI] [Google Scholar]
  39. Tanoli S. A. K. Tanoli N. U. Usmani S. Ul-haq Z. Ferreira A. G. Cent. Eur. J. Chem. 2014;12:332–340. [Google Scholar]
  40. Di Carluccio C. Forgione M. C. Martini S. Berti F. Molinero A. Marchetti R. Silipo A. Carbohydr. Res. 2021;503:108313. doi: 10.1016/j.carres.2021.108313. [DOI] [PubMed] [Google Scholar]
  41. Doan A. T. N. Doan V. T. H. Katsuki J. Fujii S. Kono H. Sakurai K. ACS Omega. 2022;7(13):10890–10900. doi: 10.1021/acsomega.1c06194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Aroulmoji V. Bergamin M. Marega R. Murano E. Int. J. Adv. Sci. Eng. 2015;1:12–17. [Google Scholar]
  43. Denis-Quanquin S. Riobé F. Delsuc M. A. Maury O. Giraud N. Chem. – Eur. J. 2016;22(50):18123–18131. doi: 10.1002/chem.201603460. [DOI] [PubMed] [Google Scholar]
  44. Swan I. Reid M. Howe P. W. Connell M. A. Nilsson M. Moore M. A. Morris G. A. J. Magn. Reson. 2015;252:120–129. doi: 10.1016/j.jmr.2014.12.006. [DOI] [PubMed] [Google Scholar]
  45. Li D. Kagan G. Hopson R. Williard P. G. J. Am. Chem. Soc. 2009;131(15):5627–5634. doi: 10.1021/ja810154u. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Hrab J. Kaur G. Guilfoyle D. N. J. Med. Phys. 2007;32(1):34–42. doi: 10.4103/0971-6203.31148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Nilsson M. Morris G. A. J. Magn. Reson. 2005;177(2):203–211. doi: 10.1016/j.jmr.2005.07.019. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

MD-015-D4MD00244J-s001
MD-015-D4MD00244J-s001.pdf (243.9KB, pdf)

Data Availability Statement

The data generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Articles from RSC Medicinal Chemistry are provided here courtesy of Royal Society of Chemistry

RESOURCES