Drug adulteration analysis based on complexation with cyclodextrin and metal ions using ion mobility spectrometry

Abstract

Drug adulteration and contamination are serious threats to human health therefore, their accurate monitoring is very important. Allopurinol (Alp) and theophylline (Thp) are commonly used drugs for the treatment of gout and bronchitis, while their isomers hypoxanthine (Hyt) and theobromine (Thm) have no effect and affect the efficacy of the drug. In this work, the drug isomers of Alp/Hyt and Thp/Thm are simply mixed with α-, β-, γ-cyclodextrin (CD) and metal ions and separated using trapped ion mobility spectrometry-mass spectrometry (TIMS-MS). TIMS-MS results showed that Alp/Hyt and Thp/Thm isomers could interact with CD and metal ions and form corresponding binary or ternary complexes to achieve their TIMS separation. Different metal ions and CDs showed different separation effect for the isomers, among which Alp and Hyt could be successfully distinguished from the complexes of [Alp/Hyt+γ-CD + Cu–H]⁺ with separation resolution (R_P–P) of 1.51; whereas Thp and Thm could be baseline separated by [Thp/Thm+γ-CD + Ca–H]⁺ with R_P–P of 1.96. Besides, chemical calculations revealed that the complexes were in the inclusion forms, and microscopic interactions were somewhat different, making their mobility separation. Moreover, relative and absolute quantification was investigated with an internal standard to determine the precise isomers content, and good linearity (R² > 0.99) was obtained. Finally, the method was applied for the adulteration detection where different drugs and urine were analyzed. In addition, due to the advantages of fast speed, simple operation, high sensitivity, and no chromatographic separation required, the proposed method provides an effective strategy for the drug adulteration detection of isomers.

Graphical abstract

Highlights

• •Alp/Thp and Hyt/Thm isomers were analyzed by ion mobility.
• •Alp/Hyt were distinguished by [Alp/Hyt+γ-CD + Cu–H]+ with Rp-p of 1.51.
• •Thp/Thm was baseline separated by [Thp/Thm+γ-CD + Ca–H]⁺ with R_p-p of 1.96.
• •Quantification for the isomers was measured with internal standard.
• •Alp/Hyt and Thp/Thm drugs were adulteration analysis in drugs and urine.

1. Introduction

In recent years, an increasing attention has been paid to drug quality and safety, which is significantly related to human health [1]. However, incidents of drug adulteration and contamination are emerging in an endless stream, which seriously threatens people's health [[2], [3], [4]]. Many unscrupulous traders try to mix similar or ineffective ingredients to replace the main active substance, which undoubtedly deceives patients and may bring potential threats to their health. Therefore, it is necessary to detect the adulteration of drug to legitimate health and interests of patients.

As we all know, allopurinol (Alp) has been used clinically for the treatment of hyperuricemia and its associated complications since 1962, such as gout [5], the prevalence and incidence of which has increased globally, especially among elderly populations [6]. Theophylline (Thp) has considerable pharmacological activity on bronchial and cardiac asthma, and is also used in pharmaceutical intermediates [7]. Alp and hypoxanthine (Hyt), Thp and theobromine (Thm) (Fig. 1), which belong to isomers of each other, are prone to drug adulteration or contamination in the process of drug production [8]. Therefore, it is desirable to develop effective and simple methods to analyze isomers of Alp and Thp drugs to address drug adulteration and contamination issues.

Fig. 1.

Structures of (A) allopurinol (Alp), (B) hypoxanthine (Hyt), (C) theophylline (Thp) and (D) theobromine (Thm).

For the determination and quantification of isomers, the traditional methods are mainly based on spectroscopy and chromatographic separation techniques. Among them, high-performance liquid chromatography (HPLC) [[9], [10], [11], [12], [13]], supercritical fluid chromatography [14,15], and capillary electrophoresis [16,17] play important roles in drug adulteration analysis. For Alp/Hyt and Thp/Thm analysis, the most reported method is based on HPLC. For example, Tada et al. [10] used HPLC to simultaneously determine Alp and oxypurinol in human serum in a concentration of 0.5–5.0 μg/mL; Reinders et al. [11] used reverse phased HPLC with ultraviolet for Alp and oxipurinol quantification in human serum; high-throughput LC-mass spectrometry (MS) was used for the determination of Thp by Camurri et al. [12]; Safranow et al. [13] used HPLC method to analyze six purines including Hyt and Alp in urinary calculi. The method can be used for reference in clinical laboratory and also for the study on pathogenesis of urolithiasis, a disorder of purine metabolism. Despite their widespread use, they still suffer from long analysis time, complex sample preparation, or unsatisfactory detection sensitivity, and the simultaneous separation of their isomers is rarely reported.

Meanwhile, MS is gradually gaining more attention and popularity due to its high sensitivity, specificity, speed, and impurity tolerance, and the advantage of detecting analytes in solvent-free environment [[18], [19], [20]]. Initially, conventional MS was considered incapable of detecting isomers because of their same mass-to-charge ratio (m/z). Therefore, to overcome these challenges, LC or gas chromatography is usually applied to separate or quantify isomers prior to MS [21]. Obviously, the above approaches can be relatively time-consuming or require chemical derivatization; therefore, alternative methods are still desired for fast and precise identification.

Recently, ion mobility spectrometry-mass spectrometry (IMS-MS) has been used in the analysis of complex mixtures, and it has potential applications for isomer analysis on a millisecond time scale [[22], [23], [24]]. In IMS, gas-phase ions can be separated based on their mobility (K₀) differences [25]. Depending on the device geometry and electric field components, several types of IMS are used to address different types of target analytes, which is summarized in Table S1 [[26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]]. In trapped IMS (TIMS), ions are spatially trapped and released by decreasing the field strength in the separation tube with a constant gas, and then separated based on their collisional cross section (CCS) and charge state values [38]. Although the TIMS-MS technology is widely applied, it is also difficult to separate and analyze molecules with small structural differences.

The concept of host–guest noncovalent complexes is commonly used in small molecule compounds, which can convert isomers with fewer conformational differences into their complex with more conformational differences [39]. In our previous work using TIMS-MS, cis/trans and chiral proline and its derivatives were well separated and identified by interaction with natamycin (as the host) and metal ion [[35], [36], [37]]; bi-2-naphthol and its phosphate derivatives with small conformational differences were converted into their complexes with large conformational differences after interacting with cyclodextrin (CD) and metal ions, which allowed their mobility separation to achieve [34]. An effective method for the determination of chiral amino acid (AA) enantiomers was developed, in which all 19 pairs of natural AA enantiomers could be easily distinguished with the addition of CD as the host [35].

In the present work, an effective method was developed by using electrospray ionization-TIMS-MS (ESI-TIMS-MS) to distinguish and quantify two isomer drug pairs of Alp/Hyt and Thp/Thm. First, the isomer pairs of Alp/Hyt and Thp/Thm were simply mixed with metal ions and CD, and the mixture was analyzed by TIMS-MS. The isomers were separated according to the mobility difference in the formed binary and ternary complexes, and their separation effect was studied and compared systematically. In addition, an absolute quantitation method for determining the total content of the isomers was established with 2-amino-6-chloropurine as an internal standard based on MS; a relative quantitative method for determining the isomer ratio was also developed. Furthermore, the method was applied to separation and quality monitoring of Alp and Thp in actual drugs, providing a novel detection method for drug adulteration.

2. Materials and methods

2.1. Materials

Alp (98%, MW:136.11), Hyt (99%, MW:136.11), 2-amino-6-chloropurine (98%, MW:169.57), and Thp (≥99%, MW:180.16) were purchased from Macklin Biochemical (Shanghai, China). α-CD (98%, MW: 972.84), β-CD (98%, MW: 1134.98), λ-CD (98%, MW: 1297.12), and Thm (99%, MW:180.16) were purchased from Aladdin Co., Ltd. (Shanghai, China). All inorganic salts are of analytical grade which are described in “1. Inorganic metal salt” in Supplementary data. Methanol (LC-MS grade) was purchased from Fisher Scientific Inc. (Pittsburgh, PA, USA). Deionized water was prepared using a Milli-Q water purification system (Bedford, MA, USA). The ESI Low Concentration Tuning Mix was supplied by Agilent Technologies (Santa Clara, CA, USA).

2.2. Sample preparation

Stock solutions were prepared at 1 mmol/L using deionized water or methanol according to their solubility and then diluted to 10 μmol/L in MeOH/water solvent (1:1, V/V) for analysis. The samples of the complexes were prepared by mixing two or three stock solutions of the CDs, isomers, and metal ions to obtain the final solution.

2.3. TIMS-MS and data analysis

All measurements were performed using a Bruker TIMS-time-of-flight (TIMS-TOF) instrument (Bremen, Germany). Details of the experiment parameters and data analysis were reported in “2. Trapped ion mobility spectrometry-time-of-flight mass spectrometer (TIMS-TOF MS)” in Supplementary data [34,40] and Table S2.

2.4. Standard curve

The establishment of standard curves consists of relative quantification and absolute quantification. For relative quantification, samples were prepared in five different molar ratios, from 1:5 to 15:1 and 1:5 to 20:1. For absolute quantification, the purine drugs stock solutions were diluted to a series of concentrations with MeOH/water solvent (1:1, V/V), ranging from 0.00735 to 1.84 μmol/L for Alp/Hyp and 0.00556–1.39 μmol/L for Alp/Hyp containing 0.294 μmol/L of 2-amino-6-chloropurine as internal standard. The former was based on the ratio of mobility peak intensities, while the latter was based on the ratio of mass spectrum peak intensities.

2.5. Analyte extraction from drugs and urine

Four kinds of drugs consisted of Alp tablets and capsules, Thp sustained release tablets and capsules, which were all purchased from a local pharmacy in Ningbo (Zhejiang, China) and stored at 4 °C for detection. To extract the analytes, the drugs were first removed from the thin film coating and ground to a uniform powder using an agate mortar and pestle. A portion of the powder was dissolved in deionized water and sonicated for 30 s. The extract was then filtered; the filtrate was diluted a thousand-fold with deionized water as a stock solution and stored at 4 °C.

Urine samples were collected from healthy volunteers and stored at 4 °C. With 1.0 mL stored in a 10 mL centrifuge tube, the urine sample was diluted 5-fold with 50% (V/V) methanol solution, and proteins were precipitated by centrifugation at 6,000 rpm for 10 min. 1.0 mL of the centrifuged supernatant was taken with another 10 mL centrifuge tube, diluted 3-fold with 50% (V/V) methanol solution, and filtered with a 0.22 μm membrane filter for TIMS-MS analysis. For spiked samples, four isomer standards were spiked at a concentration of 0.1 μmol/L before sample pretreatment; After the above pre-treatment, γ-CD was configured to 0.4 μmol/L, and CuCl₂ and CaCl₂ were configured to 0.16 μmol/L, respectively, and filtered with a 0.22 μm membrane filter after mixing for TIMS-MS detection.

3. Results and discussion

3.1. TIMS-MS analysis for the isomers and their CD-complexes

The mass spectrometry analysis of Alp/Hyt and Thp/Thm isomers, as well as their mixtures with three different CDs, was performed in positive ion mode. As shown in Figs. 2A–D, the protonated isomers of [Alp/Hyt + H]⁺ (m/z 137.05) (Fig. 2A) and [Thp/Thm + H]⁺ (m/z 181.07) (Fig. 2B) can be found in the same m/z value. And with the addition of α-CD, the m/z 1109.36 and 1153.39 can be found, which attributed to the complexes of [Alp/Hyt+α-CD + H]⁺ (Fig. 2C) and [Thp/Thm+α-CD + H]⁺ (Fig. 2D), revealing that the corresponding binary complexes can be found. Also when mixed with β-CD (Figs. 2E and F) and γ-CD (Figs. 2G and H), the corresponding binary complexes of [Alp/Hyt + β/γ-CD + H]⁺ and [Thp/Thm + β/γ-CD + H]⁺ were also formed. Hence, the isomers of Alp/Hyt and Thp/Thm could mix with the three different CDs, and non-covalently interact to form their corresponding binary complexes.

Fig. 2.

Mass spectra of (A) Alp/Hyt and (B) Thp/Thm, and their mixtures with (C, D) α-CD, (E, F) β-CD, and (G, H) γ-CD, respectively. Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

Moreover, the mobility separation for the protonated isomers and their binary complexes was investigated. As shown in Figs. 3A and B, the same extraction ion mobility (EIM) with the R_p–p of 0 was displayed for [Alp + H]⁺ and [Hyt + H]⁺, indicating that isomers of Alp and Hyt could not be directly separated by TIMS; while a slight EIM differences were observed for [Thp + H]⁺ and [Thm + H]⁺ with an R_p–p of 0.41, but it was insufficient for efficient separation and analysis due to their partial overlap of EIMs. In order to achieve baseline separation for the isomers, their CD-related complexes were also investigated (Figs. 3C and D). It can be observed that obvious mobility difference can be found with the addition of CDs. An R_p–p of 0.38 was measured for [Alp/Hyt + α-CD + H]⁺, and the same R_p–p of 0.17 was measured for [Alp/Hyt + β-CD + H]⁺ and [Alp/Hyt + γ-CD + H]⁺. Meanwhile, different CDs had different separation effects on Thp and Thm. For example, the R_p–p values for the formed complexes of [Thp/Thm + α-CD + H]⁺, [Thp/Thm + β-CD + H]⁺ and [Thp/Thm + γ-CD + H]⁺ were different, which were 0, 0.57 and 1.08, respectively. The results revealed that as the radial size of the CD increased, the separation resolution of its complex increased for the separation of Thp and Thm. However, baseline separation was not achieved for the isomers of Alp/Hyt and Thp/Thm, but the addition of γ-CD generally had a positive effect on the separation of drugs isomers. Thus, further mobility separation studies for Alp/Hyt and Thp/Thm were all based on interaction with γ-CD.

Fig. 3.

Ion mobilograms of the protonated adduct ions. (A) [Alp/Hyt + H]⁺, (B) [Thp/Thm + H]⁺, (C) CD-related complex ions [Alp/Hyt + α/β/γ-CD + H]⁺ and (D) [Thp/Thm + α/β/γ-CD + H]⁺. The corresponding mobility and R_P–P values are also given. Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

3.2. Ion mobilograms of metal ion coordinated γ-CD-related complexes

According to our previous studies [[34], [35], [36], [37]], the addition of metal ions can effectively regulate the separation effect for the isomers. Therefore, with γ-CD as the separation selector, different metal ions were added and studied to improve the separation effect for Alp/Hyt and Thp/Thm. Herein, 16 kinds of readily available metal ions were added separately (see Supplementary data), but the results of experiment revealed that only 11 kinds of metal ions could interact between Alp/Hyt and γ-CD (Fig. 4A), and 12 kinds of metal ions could interact between Thp/Thm and γ-CD, and formed their corresponding ternary complexes (Fig. 4B). Interestingly, it can be noted that the analysis of EIMs of the ternary complexes allowed different degrees of mobility differences for the isomers, revealing that the mobility separation difference was strongly dependent on the metal ion.

Fig. 4.

Ion mobilograms of the different metal ion bound adduct ions with γ-CD for isomers (A) Alp/Hyt and (B) Thp/Thm. Heatmap obtained from (C) a mixture containing Alp, Hyt, γ-CD, and CuCl₂, and (D) a mixture containing Thp, Thm, γ-CD, and CaCl₂. Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

The EIMs for the ternary complexes are presented in Fig. 4A, and detailed data are listed in Table S3. Taking the R_p–p of [Alp/Hyt + γ-CD + H]⁺ (0.17) as a reference value, the addition of different ions produced different effects on the R_p–p values. For example, the addition of Ag⁺, Ca²⁺, and Sr²⁺ reduced the R_p–p values for the Alp/Hyt separation to 0.11, 0.14, and 0.09. However, the introduction of Na⁺, K⁺, Rb⁺, Cs⁺, Mg²⁺, Mn²⁺, Co²⁺, and Cu²⁺ could increase the separation degree for the Alp/Hyt isomers to 0.61, 0.46, 0.43, 0.45, 0.23, 0.46, 0.41, and 1.51 (Fig. 4A). Satisfactorily, an obvious baseline separation was achieved for the isomers of Alp and Hyt, in which the R_p–p was 1.51 by the complex of [Alp/Hyt + γ-CD–Cu–H]⁺, and the detailed mass spectrum is shown in Fig. S1A. Besides, a more intuitive heatmap obtained from isomers mixture is shown in Fig. 4C, where the ion [(Alp + Hyt) + γ-CD + Cu–H]⁺ has two obvious different EIMs. Besides, to ensure the reliability of the method, it was repeated for three consecutive days, accepted relative standard deviation (RSD%) between 0.31% and 0.42% were measured for R_p–p, which supported the good repeatability of the method.

Meanwhile, the improvement of the mobility separation for Thp and Thm was also investigated by the addition of different metal ions, that is, the TIMS measurement for their ternary complexes was performed systematically. The EIMs for the ternary complexes are shown in Fig. 4B, and their detailed data are listed in Table S4. The R_p–p values ranged from 0.18 to 1.96 for Thp and Thm isomer separation by the complexes of [Thp/Thm + γ-CD + M]⁺, and the RSD% for the reliability in three consecutive days was between 0.32% and 0.45% for the measured R_p–p. Herein, taking the R_p–p of [Thp/Thm + γ-CD + H]⁺ (R_p–p = 1.08) as a reference value, only the R_p–p of [Thp/Thm + γ-CD + Ca–H]⁺ (R_p–p = 1.96) was bigger than that of the reference value (the detailed mass spectrum is shown in Fig. S1B). And in referrence to their heatmap in Fig. 4D obtained from the isomer mixture, an intuitive baseline separation was displayed for Thp and Thm isomer. Besides, the separation of the four isomers by HPLC was also measured, which could not get baseline separation by a C₁₈ column (Fig. S2). Overall, the TIMS separation results revealed that different metal ions had different effects on the isomer separation, and that metal ions as ligands could enhance the R_p–p for isomers, but might also reduce their separation efficiency.

3.3. Conformational simulations

Computational modeling was performed to generate the theoretical structures of [Alp/Hyt + γ-CD + Cu–H]⁺ and [Thp/Thm + γ-CD + Ca–H]⁺ ions, delineate their microscopic interactions, and explain the separation mechanisms. The process of chemical calculations is explained in Supplementary data “3. Chemical calculations” [37,41]. As shown in Fig. 5, four isomers and metal ion all inserted into the cavity of γ-CD, but the spatial location of their interactions was somewhat different. In detail, Alp tended to lie diagonally inside the γ-CD, while Hyt was diagonally inserted into the γ-CD cavity. Meanwhile, for [Thp + γ-CD + Ca–H]⁺ and [Thm + γ-CD + Ca–H]⁺, the angle of interaction between Thp and Thm and γ-CD was different, where the two –CH₃ functional groups of Thp exposed in the large cavity of the γ-CD, while the one –CH₃ functional group of Thm stuck out directly outside the small port of γ-CD. Besides, the distance between metal ion and the analyte was different, and the specific number is marked the top view of the conformation (Fig. S3). Accordingly, CCS of the favored structure of the complexes was measured and compared in Table 1. The results revealed that the theoretical calculations agreed with the TIMS experimental values, and the relative errors were ≤11.79%. Overall, TIMS measurements and theoretical calculations all revealed that the method of simply mixing with CD and metal ions can make their conformation different and realize the simple and fast mobility separation.

Fig. 5.

The favored structure obtained by chemical calculation. (A) [Alp + γ-CD + Cu–H]⁺, (B) [Hyt + γ-CD + Cu–H]⁺, (C) [Thp + γ-CD + Ca–H]⁺, and (D) [Thm + γ-CD + Ca–H]⁺. Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

Table 1.

The collisional cross section (CCS) obtained by experiment and theoretical calculation.

Complex	m/z	Experiment CCS (Å2)	Theoretical CCS (Å2)	Relative error (%)
[Alp + γ-CD + Cu–H]+	1494.39	342.3	370.6	8.27
[Hyt + γ-CD + Cu–H]+		347.5	373.8	7.57
[Thp + γ-CD + Ca–H]+	1515.44	337.5	377.3	11.79
[Thm + γ-CD + Ca–H]+		345.5	384.2	11.20

Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

3.4. Quantitative analysis

Before the quantitative analysis, the effects of solvents, salts and ionization voltages are discussed in the Supplementary data “4. Optimize of the experimental condition” (Fig. S4). The optimized conditions were that the solvent was 50 vol% methanol solution, the molar ratio of the metal salt to analyte was 0.8:1, and the ionization voltage was 3000 V. In the field of analytical chemistry, quantitative analysis plays an important role in obtaining the relative and absolute levels of adulterated components to determine whether they are within the safe limits of use.

Relative quantification enables the determination of the ratio between Alp and Hyt, Thp and Thm. In this case, various molar ratios of isomers were mixed with γ-CD and Cu²⁺/Ca²⁺ for TIMS analysis, i.e., 1:5, 1:1, 5:1, 10:1, 15:1 for Alp to Hyt, and 1:5, 1:1, 5:1, 10:1, 20:1 for Thp to Thm. Herein, the content ratios were quantified by the peak area intensity ratio of their mobility peaks, and the accuracy (RSD%) of the method was calculated by comparing the slopes of the standard curves (n = 5). As displayed in Table 2, a linear regression equation was obtained as the ratio of molar concentration (Alp to Hyt) versus their peak area intensities of the mobility peaks, yielding an R² higher than 0.99 and an RSD of 2.12%. Meanwhile, good linear quantification (R² = 0.9926) was also plotted as molar concentration ratio (Thp to Thm) versus the relative intensity, with RSD being 2.36%. The limit of detection (LOD) was calculated on the basis of signal-to-noise ratio (S/N) of 3 measured for the molarity ratios, which ranged from 1:20.8 to 36.3:1 for Alp/Hyt, and 1:32.5 to 30.7:1 for Thp/Thm. In addition, the calibration curves of relative quantification, also established in Fig. S5, demonstrated the accuracy and feasibility of the method for the relative quantitative determination of the isomer drugs. Meanwhile, MS/MS experiments were also performed for the isomers (Fig. S6), and the similar fragment ions were observed for isomers of [Alp + H]⁺ and [Hyp + H]⁺ (Fig. S6A). However, different fragment ions were obtained for [Thp + H]⁺ and [Thm + H]⁺, such as 124.07 for [Thp + H]⁺, 138.05 and 163.05 for [Thm + H]⁺. In this case, relative quantitation for Thp/Thm was measured by the relative intensity of the different fragment ions of 124.07 and 138.05 (Fig. S6B). The relative quantification measurement by the ion mobility and the MS/MS method was compared; taking the molar ratio of Thp/Thm of 3:1 and 1:3 as examples, the back calculated ratios were 3.00 ± 0.12:1.00 ± 0.09 and 1.00 ± 0.11:3.00 ± 0.13, and 3.00 ± 0.10:1.00 ± 0.07 and 1.00 ± 0.09:3.00 ± 0.10 by MS/MS method and ion mobility method, respectively. Overall, the results calculated by the two methods were basically the same, revealing the reliability of the proposed method.

Table 2.

Relative and absolute quantification of Alp/Hyt and Thp/Thm.

Parameters	Relative quantification		Absolute quantification (μmol/L)
	Alp/Hyt	Thp/Thm	Alp/Hyt	Thp/Thm
Linearity	1:10–20:1	1:10–20:1	0.00735–1.84	0.00556–1.39
Linear regression	ya = 0.0144 + 0.529xa	ya = 0.0387 + 0.474xa	yb = 0.0974 + 5.458xb	yb = 0.00585 + 3.401xb
R	0.9996	0.9963	0.9998	0.9986
R2	0.9992	0.9926	0.9996	0.9971
RSD% (n = 5)	2.12	2.36	2.66	1.95
LOD	1:20.8–36.3:1	1:32.5–30.7:1	0.0196	0.0042

R: Linear correlation coefficient; R²: determination coefficient; RSD: relative standard deviation; LOD: limit of detection. x_a: the isomer ratios (the molarity of Alp to Hyt); y_a: the peak area intensity ratio of ion mobility peak (Alp to Hyt); x_b: total content containing two isomers (1:1); y_b: the intensity ratio of mass spectral peak (analytes to internal standard). Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; and Thm: theobromine.

Absolute quantification can be used to determine the total content of isomers. Herein, 2-amino-6-chloropurine (MW:169.57) selected as internal standard was measured to improve the quantitative accuracy. The isomers were mixed in a 1:1 ratio and spiked over a concentration range of 0.00735–1.84 μmol/L for Alp/Hyt and 0.00556–1.39 μmol/L for Thp/Thm, then prepared and analyzed at various concentrations of the analytical isomers containing a constant 0.294 μmol/L internal standard. The analytes could be quantified by the MS peak intensity ratio of [Alp/Hyt + H]⁺ (m/z 137.05) or [Thp/Thm + H]⁺ (m/z 181.07) to [2-amino-6-chloropurine+H]⁺ (m/z 170.02). Standard curve equations are shown in Table 2, with an acceptable R² of 0.9996 for Alp and Hyt, and 0.9971 for Thp and Thm. And the LOD (S/N = 3) calculated for the isomers was 0.0196 and 0.0042 μmol/L for Alp/Hyt and Thp/Thm, respectively. Moreover, the repeatability of the linear regression was also measured (n = 5), and the RSD for the measured slope of the linear regressions was ≤2.66%. Finally, for further investigation, a recovery by spiking Alp/Hyt (1:1) at a total concentration of 0.368 μmol/L before sample pre-treatment was investigated as an example. The acceptable recovery value of 102.6% was obtained, demonstrating the feasibility of the developed method.

3.5. Adulteration detection of purine drugs

Based on the study of standard samples, adulteration detection of drugs was carried out using the TIMS-MS by mixing γ-CD and Cu²⁺ for Alp/Hyt, and γ-CD and Ca²⁺ for Thp/Thm. Herein, the tablet and capsule drugs used in this experiment were all diluted 10⁶-fold.

First, the ions of [Alp + γ-CD + Cu–H]⁺ (m/z 1494.39) and [Thp + γ-CD + Ca–H]⁺ (m/z 1515.44) could be detected in both tablets (Figs. 6A and B) and capsules (Fig. S7). And the EIMs for the ions were studied and compared with the standard sample to identify the components in the actual sample. As shown in Fig. 6C, the 1/K₀ values of the standard Alp and Hyt samples were 1.698 V·s/cm² and 1.725 V·s/cm². In particular, the 1/K₀ value of Alptablet was also 1.698 V·s/cm², which was consistent with that of standard Alp. Besides, considering the complex of the actual samples, the MS/MS spectra for the ions were further identified. Upon observation of the illustrations, high degree of consistency of fragment ions between the actual sample and the standard sample was observed. The results indicated that the Alp existed in the pretreated tablets, without ingredient components such as Hyt. Meanwhile, as shown in Fig. 6D, the 1/K₀ values of the standard Thp and Thm samples were 1.668 V·s/cm² and 1.708 V·s/cm², respectively. It is worth noting that a single peak was shown in the Thp tablet, where the 1/K₀ was 1.668 V·s/cm²; the consistency of mobility indicated that this tablet contained the ingredient Thp, but not its isomer Thm.

Fig. 6.

Trapped ion mobility spectrometry-time-of-flight (TIMS-TOF) mass spectra formed by actual samples of (A) Alp tablets and (B) Thp tablets. Extraction ion mobility (EIM) comparison of actual sample (upward) and standard sample (downward) of (C) Alp tablets and (D) Thp tablets. Alp: allopurinol; Thp: theophylline; Hyt: hypoxanthine; Thm: theobromine; and CD: cyclodextrin.

In addition, adulteration detection was also carried out under the same experimental conditions for the actual capsule drugs. The formation of ternary complexes could be observed in the mass spectra containing [Alp + γ-CD + Cu–H]⁺ (m/z 1494.39) and [Thp + γ-CD + Ca–H]⁺ (m/z 1515.44), shown in Fig. S7A and Fig. S7B, respectively. EIMs for the ions were also compared in Fig. S7C and Fig. S7D, in which the mobility peaks and MS/MS peaks of Alp and Thp in actual samples were highly consistent with those of standard samples. The results revealed that the actual capsule medicines were almost qualified without interfering substances Hyt and Thm.

Additionally, the proposed method was applied for the four isomers analysis of Alp/Hyp and Thp/Thm in the urine sample. In detail, three urine samples of healthy volunteer served as analysis samples, and the four isomers could not be detected in the healthy urine (Fig. S8A). Additionally, the four isomers were spiked at 0.1 μmol/L in urine before sample preparation, and then for quantitative analysis. As shown in Fig. S8B, the spiked isomers were detected in a concentration of 0.800 ± 0.042 μmol/L, and the molar ratio detection for Alp:Hyp and Thp:Thm was 1.00 ± 0.67:1.00 ± 0.72 (Fig. S8C), which indicated that the detected results had good consistency with the spiked amount, revealing the studied method can also be used to detect four isomers in urine samples.

4. Conclusion

In this work, we investigated the isomer drug separation of Alp/Hyt and Thp/Thm by TIMS-MS, based on complexation with CD and metal ion to increase the mobility differences. TIMS-MS results showed that Alp/Hyt and Thp/Thm isomers could interact with α-, β-, γ-CD and metal ions and form the corresponding ternary complexes to obtain their mobility separation. Different metal ions and CDs had certain differences in the separation effect of isomers, among which the isomers of Alp and Hyt could be successfully distinguished by the complexes of [Alp/Hyt + γ-CD + Cu–H]⁺ with R_p–p of 1.51, whereas Thp and Thm could be baseline separated by [Thp/Thm + γ-CD + Ca–H]⁺ with R_p–p of 1.96, and with good repeatability of RSD% between 0.32% and 0.45% for the measured R_p–p for the reliability in three consecutive days. Besides, chemical calculations were made for the complex, revealing that the microscopic spatial interaction was different, led to the structure difference, and then realized the mobility separation. In addition, relative and absolute quantitation of the two pairs of isomers was established with a good linear of R² value higher than 0.99 for determination of molar ratios and the total content of the isomers, respectively. Furthermore, adulteration detection of four drugs was carried out successfully. Finally, the results indicated that the proposed method can be a promising approach to isomers distinguishment and quantitation of adulterants, and thus can be applied in the safety monitoring of isomers in drugs.

CRediT author statement

Zhigang Liang: Experimental, Conceptualization, Methodology, Data curation, Writing - Original draft preparation, Reviewing and Editing; Huanhuan Wang: Experimental, Conceptualization, Methodology, Validation, Investigation, Data curation, Writing - Original draft preparation; Fangling Wu: Conceptualization, Methodology, chemical calculations, Data curation, Supervision, Writing - Original draft preparation, Reviewing and Editing, Funding acquisition; Longfei Wang and Chenwei Li: Validation, Data curation; Chuan-Fan Ding: Conceptualization, Resources, Supervision, administration, Writing - Reviewing and Editing, Funding acquisition.

Declaration of competing interest

The authors declare that there are no conflicts of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article: