Study on Quality Control of Tenofovir Disoproxil Fumarate Enantiomers by High-Performance Liquid Chromatography–Mass Spectrometry

This study aims at developing a high-performance liquid chromatography–mass spectrometry (LC–MS) method to analyze tenofovir disoproxil fumarate (TDF) and its pharmaceutical preparations. Several cyclodextrin mobile-phase additives were applied to reversed-phase and normal-phase chromatography, and the effects of three chiral stationary phases on the TDF separation were investigated in this study. The R-type and S-type of TDF tablets were quantitatively analyzed in the single ion monitoring (SIM) scanning mode with a Unichiral CMD column. This method has been successfully applied to the separation and quantification of TDF and its isomers. The linear ranges of (R)-TDF and (S)-enantiomer were 1–20 and 0.2–16 μg/mL, respectively. The limit of detection for (R)-TDF and (S)-enantiomer was 0.0015 and 0.0012 μg/mL, respectively. (S)-enantiomer was not detected in the formulas from all the seven manufacturers, and the drug content of each took more than 98.5% of the labeled amount, which complies with the regulations. The method shows its advantages on high sensitivity, low detection limit, good practicability, and repeatability. The proposed method may provide a novel platform for separation of TDF enantiomers and quality control of TDF raw materials and preparations.


Introduction
Tenofovir disoproxil fumarate (TDF), a water-soluble tenofovir disoproxil (C 19 H 30 N 5 O 10 P, M:519.44), is a prodrug of nucleotide reverse transcriptase inhibitor. It can be rapidly absorbed and hydrolyzed into active tenofovir after oral administration, and then phosphorylated and converted into active tenofovir diphosphate in cells. Tenofovir diphosphate competitively binds itself to the natural substrate 1 3 5ʹ-deoxyadenosine triphosphate and enters the deoxyribonucleic acid (DNA) to inhibit the activity of the reverse transcriptase of human immunodeficiency virus (HIV) and hepatitis B virus (HBV). Thus, the lengthening of DNA strand and viral replication will be stopped, providing treatment effect against HIV and HBV [1][2][3][4][5]. The hepatoprotective mechanism of TDF is shown in Fig. S1  The structures of (R) and (S) TDF enantiomers are shown in Fig. 1.
The isomers mentioned have almost the same physical and chemical properties. Hence, (R)-TDF and (S)-enantiomer are difficult to isolate and purify. However, pharmacological studies indicate difference in their pharmacological properties, pharmacokinetic properties, and metabolic rates [6]. No pharmacological activity is found from the corresponding (S)-enantiomer [7]. The existence of (S)enantiomer as impurity can remarkably affect the therapeutic properties of the drug. The proportion of (S)-enantiomer in the raw material should not exceed 1.0% [8]. At present, the studies focusing on the chiral separation of TDF are limited. Therefore, efficient separation methods for (R) and (S) TDF enantiomers and chiral quality control of the active pharmaceutical ingredient and preparation products should be developed.
HPLC is a preferred method for efficient chiral separation. Chromatographic chiral separation can be classified into chiral stationary phase (CSP) method, chiral mobilephase additive (CMPA) method and chiral derivative method. Considering that the chiral derivative method has stringent requirement on derivatization reagent and requires a complex derivation process, CSP and CMPA method were commonly used on enantiomer separation [9,10]. Among the existing chromatographic chiral separation methods, CMPA, an effective one to separate direct enantiomer from chiral compounds, which consists of mobile phase with chiral selector dissolved inside and non-chiral stationary phase. CMPA is characterized for its economic efficiency and flexibility.
Commonly used chiral mobile-phase additives, including chiral coordination additives and cyclodextrin molecular recognition additives, have interactions with the enantiomer leading to transient diastereomer complexes [11][12][13][14]. The formation constant and distribution of these complexes differ between the (non-chiral) stationary and mobile phase, thus achieving enantiomeric separation [15]. Seshachalam U et al. [16] performed the enantiomeric separation of tenofovir on a non-chiral HPLC column using L-phenylalanine as chiral mobile-phase additive and copper sulfate as complexing agent. However, Chiral additives must be able to provide effective groups and positions to form diastereomeric complexes with solute enantiomers, and the specificity and separation efficiency of CMPA method are not as good as that of CSP method. Therefore, the CSP method is generally used for the separation of chiral substances through HPLC. Gao LY et al. [11] established a CSP HPLC method for the resolution of the enantiomer of TDF. The retention time (t R ) of (R)-TDF and (S)-enantiomer were 148 min and 162 min, respectively, indicating that the separation was time-consuming.
HPLC is commonly used for quality control of chiral drugs, but its ability on qualitative analysis lacks specificity. Mass spectrometry is characterized by its good specificity and high sensitivity and can provide accurate molecular weight and structural information of compounds [17]. LC-MS can take the advantage of both in separation performance and selectivity, making LC-MS feasible to use for the separation and quantification of TDF enantiomers. Select ion monitoring (SIM) is a detection technique of LC-MS, where the selected ions are scanned by jump scanning, and then the chromatograms of the selected ions are obtained. The background noise is very low in SIM mode, because only characteristic ions are selected, which can greatly improve the sensitivity and selectivity.
The present paper investigated the effects of cyclodextrin (CD) chiral additives, such as β-cyclodextrin (β-CD), CM-β-CD, HP-β-CD, HP-γ-CD, Me-β-CD, and SBEβ-CD, on the separation of (R) and (S) TDF enantiomers on a reversed-phase C18 column. The CMPA using CDs and its derivatives as additives cannot achieve the separation of TDF enantiomers. The effects of chromatographic column filled with cellulose and amylose and the composition of different mobile phases on the separation of TDF enantiomers were investigated. The method for the determination of TDF tablets was established as well. Combined with the high separation ability of HPLC and the high detection ability of mass spectrometry, SIM mode was selected for quantitative analysis of samples, aiming at enhancing the sensitivity and selectivity. The proposed method may provide a novel platform for separation of (R) and (S) TDF enantiomers and chiral quality control of preparation products.
The drug information of TDF tablets from seven manufacturers is summarized in the Supplemental Materials (Table 1). These tablets were randomly obtained from pharmacies.

Chromatographic Conditions
Unichiral CMD (4.6 × 250 mm, 5 μm) column was selected for chromatographic separation. The mobile phase was anhydrous ethanol, eluted with isocratic elution. The mobile phase was anhydrous ethanol, eluted with isocratic elution. The flow rate was controlled at 1 mL/min, the detection wavelength was 260 nm, the column temperature was 30 °C, and the injection volume was 20 μL.

MS Conditions
Mass spectrometry was conducted using a Shimadzu LC-MS/MS-2020 single quadrupole mass spectrometer equipped with an electrospray ionization interface. SIM in the positive mode was adopted to detect TDF enantiomers. The molecular ion peak [M + Na] + of tenofovir plus sodium was selected as the monitoring ion for (R)-TDF or (S)-enantiomer with m/z of 542. Other parameters were as follows: dry gas flow rate of 3 L/min, atomizer flow of 1.5 L/min, interface temperature of 350 °C, DL temperature of 250 °C, and heating block temperature of 200 °C. The analyses were performed in scan mode within the range of 100-600 m/z.

Standard Solution Preparation
7.50 mg enantiomeric standard of (R)-TDF and 5.00 mg enantiomeric standard of (S)-enantiomer were precisely weighed and dissolved separately in methanol and fixed into 25

Sample Preparation
Ten drug tablets were taken from each of the different manufacturers and weighed accurately. The tablets were ground, and then a certain amount of powder was weighed and dissolved with methanol in the 100 mL capacity bottle. After filtered, 1 mL of the filtrate was pipetted into a 100 ml volumetric flask and diluted to the scale with methanol. Then, the 10 μg/mL sample was prepared and ready to be tested. The exact weighing of each tablet is shown in Table 1.

CD Chiral Additive Method
Among all chiral separation techniques, HPLC is the most simple, reproducible, selective, sensitive, and widely used method. CMPA was employed at the beginning of the experiment because of its simplicity and flexibility [15].
CDs usually contain 6-12 D-glucopyranose units. Each glucose unit forms a slightly conical ring with 1,4-glycoside bond, has a hydrophilic shell and a hydrophobic cavity structure [12,13]. CDs can form an inclusion complex with the enantiomer for the separation and recognition of chiral compounds [13]. The cyclodextrin containing 7 glucose units called β -CD is the focus of the study [13,14]. In the experiment, six kinds of CDs, including 15 mmol/mL β-CD, 10 mmol/mL CM-β-CD, 25 mmol/mL HP-β-CD, 20 mmol/ mL HP-γ-CD, 20 mmol/mL Me-β-CD, and 5% SBE-β-CD were dissolved in aqueous phase as chiral additives. On C18, these chiral additives and methanol or acetonitrile formed a chiral mobile phase to separate TDF enantiomers. However, the results showed that no separation occurred within 60 min. Then, the TDF enantiomers were separated on a normal-phase ChromCore TM CN column. First, 5 mmol/ mL Me-β-CD was used as chiral additive. The mobile phases were 5 mmol/mL and 0.5 mmol/mL Me-β-CD with ethanol:n-hexane (35:65), respectively. The composition of the mobile phase is 5 mmol/mL Me-β-CD with ethanol: n-hexane (35:65) was too sticky for the experiment. The TDF isomer was injected at the concentration of 0.5 mmol/ mL Me-β-CD with absolute ethanol: n-hexane (35:65) as the mobile phase, and the separation was not realized.

Crown Ether Chiral Additive
Crown ethers are macrocyclic polyether compounds that can bind to some cations, such as basic ions or protonated primary amines, and are potential chiral additives with high selectivity and affinity [18]. Using 18-crown ether-6 as a chiral additive in ethanol and n-hexane as mobile phase, resolution experiments were carried out on a normal-phase ChromCore ™ CN column. The results showed that the enantiomers of TDF were not separated.
In conclusion, the CMPA using CDs and its derivatives or crown ethers as additives cannot achieve the separation of TDF enantiomers.

Optimization of Chromatographic Conditions
CSPs were a common approach to chiral chromatographic separation, and polysaccharide-based CSPs derived from cellulose and amylose are the most commonly used CSP in HPLC [19][20][21][22]. To separate the two enantiomers of TDF, we investigated the effects of three chromatographic columns and different mobile phases on the separation of (R)tenofovir and (S)-enantiomer. The chromatographic columns included CHIRALPAK IB N-5 [cellulose-tris(3,5-dimethylphenylcarbamate) immobilized to silica gel], CHIRALCEL OJ-RH [Cellulose tris(4-methylbenzoate) coated on silica gel], and Unichiral CMD [Amylose-tris(3,5-dimethylphenylcarbamate)]. The results showed that the separation effects differed from different surface functional groups and polysaccharides.
CHIRALPAK IB N-5 is a normal-phase column. N-hexane and ethanol were chosen to compose the mobile phase, and the separation effect of different proportions of n-hexane and ethanol on (R)-TDF and (S)-enantiomer was determined.
When the mobile phase was n-hexane:anhydrous ethanol (80:20) at the flow rate of 0.6 ml/min, both (R)-tenofovir and (S)-enantiomer were eluted simultaneously at 33.88 min, which means the separation was not achieved under this condition. At ethanol ratios of 100%, 95%, 90%, and 85%, the retention time of mixed analytes was shortened to 8.53-8.89 min, and no peak separation was observed. Under above conditions, no resolution was observed with CHIRALPAK IB N-5 as stationary phase. Therefore, the column is considered as not suitable for the separation of (R)-TDF and (S)-enantiomer.
Then, CHIRALCEL OJ-RH column and Unichiral CMD column were chosen for the study forward. The CHIRAL-CEL OJ-RH chiral reverse phase column was coated with cellulose-tris(4-methylbenzoate) on the surface of silica gel. The results showed that when the mobile phase was water:acetonitrile (87:13) at the flow rate of 0.5 mL/min, (R)-tenofovir and (S)-enantiomer could be separated. The t R of (R)-TDF and (S)-enantiomer were 217.29 min and 236.81 min, the peak width at half height were 5.95 min and 7.37 min, and the numbers of theoretical plates were 7355 and 6804. The degree of separation was 1.81, but the analysis was time-consuming. As shown in Fig. 2a, the separation effect was not ideal. To shorten the separation time and improve the separation degree, 0.1% tetrahydrofuran was added to the mobile phase, but the results showed that the separation effect was not good. When the mobile phase was water:methanol (50:50) at the flow rate of 0.8 ml/min, the results showed that the separation effect of (R)-tenofovir and (S)-enantiomer was the best, and the separation degree was 2.17. As shown in Fig. 2b, the retention times of (R)tenofovir and (S)-enantiomer were 85.51 min and 69.48 min, shorter than the previous values, with half-peak widths of 3.99 min and 4.92 min, respectively.
Amylose-tris(3, 5-dimethyl-phenylcarbamate) is a polysaccharide-based chiral selector. The separation is thought to be achieved by the formation of hydrogen bonds between the -NH or -CO groups on the carbamate structure and the analyte, as well as by π-π and van der Waals interactions with the benzene ring. The carbamate bond is around the cavity of the main chain of polysaccharide, providing more flexibility to the benzene ring, thus the cavity can be adjusted for different analytes, and the π-π interaction can be maximized [23]. The (R)-TDF and (S)-enantiomer contain C = O and -NH functional groups, which can form hydrogen bonds with carbamates in amylose-tris(3,5-dimethylphenylcarbamate). (R)-tenofovir and (S)-enantiomer have different structures, and the hydrogen bond forces between them and stationary phase are different. Therefore, a high degree of separation can be obtained on UniChiral CMD column.

Optimization of MS Conditions
The mixed solution of TDF was analyzed by mass spectrometry, and Fig. 3 shows the mass spectra of 50 μg/mL (R)-TDF and 10 μg/mL (S)-enantiomer with retention times of 7.787 min and 16.164 min, respectively. The ion peak of m/z 520.2 was tenofovir dipivoxil quasi-molecular ion [M + H] + and that of m/z 542.2 was [M + Na] + , the two major fragment ion peaks were m/z 300.2 and 270.1. Figure 4 shows cleavage diagram of tenofovir disoproxil.
In comparison with the full scanning mode, SIM enables the mass spectrometer to spend more time in detecting the ion current of the selected ions, thereby improving the sensitivity for quantitative analysis. Based on the results of full scan (Fig. 3), the signal intensity of ion peak [M + Na] + was the highest. Therefore, [M + Na] + was selected as the quantitative measuring ion of (R)-TDF and (S)-enantiomer. According to the retention time of chromatographic peak, mass spectrum data, and literature [8], the two peaks at 7.633 min and 16.100 min were attributed to (R)-tenofovir and (S)-enantiomer, respectively.

Calibration Curve, Precision, Repeatability, and Stability Test
The calibration curve was obtained by triple injection of a series of working solutions and constructed by plotting the peak areas (A) versus reference concentration (C, μg/ mL). Table 2 shows that the compounds have a good linear relationship in the corresponding linear range, with the corresponding correlation coefficient (r 2 ), linear range (μg/mL), LOD, and LOQ. The limit of detection (LOD) and limit of quantitation (LOQ) of (R)-TDF and (S)-enantiomer were calculated at S/N values of 3 and 10.
Mixed standard solutions of (R)-TDF and (S)-TDF at concentrations of 5 μg/mL, respectively, were prepared as QC samples. And during the LC-MS experiment, a QC sample was injected after every 10 injections of sample analysis to verify the stability of the instrument. The results showed that the retention times of all QC samples were stable and the RSD of signal response was less than 5.0%, indicating good instrument stability. For precision, the mixed reference solution of 5 μg/ml (R)-tenofovir and 1 μg/ml (S)-enantiomer was selected. The chromatographic peak areas of (R)-tenofovir and (S)enantiomer were determined (n = 6). The RSD values of (R)tenofovir and (S)-enantiomer were 1.14% and 1.42%, respectively, indicating that the precision of the method was good.
For repeatability, the mixed reference solution of 5 μg/ ml (R)-tenofovir and 1 μg/ml (S)-enantiomer was selected, and its response for 3 continuous days was measured by MS. The RSD values of (R)-tenofovir and (S)-enantiomers were determined to be 1.34% and 1.68% (n = 6), respectively, indicating that the method had good repeatability.
For stability, TDF solution (production batch no: a867401, Aspen Port Elizabeth Pty. Ltd.) was prepared by the method under item ("Sample Preparation" section) and placed under room temperature for 0, 1, 2, 4, 8, and 12 h. Then, 20 μL of the solution was injected for LC-MS analysis at each time point. The RSD values of chromatographic peak area of (R)-tenofovir and (S)-enantiomers were determined to be 1.83% and 2.26% (n = 6), respectively, indicating that the stability of the test sample was good within 12 h.

Percentage of Labeled Amount Test
The sample solution prepared under the condition of item ("Sample Preparation" section) was filtered with 0.22 μm organic microporous membrane, and the content was determined by LC-MS. The percentage of labeled amount of TDF tablets produced by seven manufacturers was 98.5%-104.7%  (Table 1), and the (S)-enantiomer was not detected. At present, TDF is not included in the 2020 edition of Chinese Pharmacopoeia. In 2020, the Chinese Pharmacopoeia Commission released the draft National Drug Standard Table of tenofovir disoproxil fumarate, which stipulated that the percentage of labeled amount of drug should be at least 98.5% of the labeled amount [24]. The area of each impurity peak shall not be larger than that of the main peak of the control solution, if there is any (impurity limit < 1.0%). If the peak area is smaller than that of the main peak of the sensitivity solution, the peak area in the chromatogram of the test solution is negligible, and the percentage of labeled amount of TDF tablets produced by seven manufacturers meets this standard. According to relevant literature, the percentage of labeled amount of TDF in Chinese National Medical Products Administration import drug registration standard (No: jx20080009) is 90%-105% [25], and that of TDF tablets produced by seven manufacturers also meets this standard. The USP Medicines Compendium requires that the percentage of labeled amount of TDF be between 98.0% and 102.0% [26]. The percentage of labeled amount of TDF tablets produced by five manufacturers meets this standard. The Indian Pharmacopoeia requires that the percentage of labeled amount of TDF be between 90% and 110.0% [27], and the percentage of labeled amount of TDF tablets produced by seven manufacturers meets the standard.
In summary, the drug content of TDF tablets produced by seven manufacturers met the standard of the Draft National Drug Standard for TDF issued by the Chinese

Sample Recovery Rate
TDF solution (production batch no: 210104, Beite pharmaceutical) was prepared by the method under item ("Sample Preparation" section). The solution was added with appropriate mixed reference solution at low, medium, and high levels, and the volume was fixed to 2 mL. Precisely, 20 μL of the solution was injected into LC-MS after filtered through a 0.45 μm microporous membrane. The content of each component was calculated according to the standard curve. The experimental results are shown in Table 3. The recovery of this method in the sample ranged from 95.1% to 105.7%, indicating good recovery. The recovery rate is calculated using the following formula: where: A: measured quantity of samples added with reference materials B: "true value" of the substance in the sample C: standard substance quality added

Comparison Between LC-MS and Other Methods
Limited studies have investigated the enantiomeric separation and quantification of TDF, and the quantitative determination of TDF by SIM has not been reported. In the present paper, the LOD of LC-MS method was lower than that of general method, and the result of qualitative analysis was reliable. The results are summarized in Table 4.

Conclusion
In this paper, the LC-MS method was used for the quantitative analysis of TDF, and the sensitivity was greatly improved in the SIM mode. The values of LOD and LOQ were less than 1/2 of the literature data. The (S)-enantiomers Recovery rate (% ) = 100 × A B + C , were not detected in the TDF tablets produced by seven manufacturers, and the drug content were all above 98.5% of labeled amount. In addition, under the conditions of reversed-and normal-phase chromatography, several CD chiral additives were excluded for they are not feasible in CMPA method for TDF resolution. Finally, the results confirm that the Unichiral CMD column with amylose-tris(3,5dimethylphenyl carbamate) as CSP was suitable for the chiral separation of TDF. The method established in this study showed its advantages on high sensitivity, low detection limit, good practicability, and repeatability. It was successfully applied to the separation and quantification of TDF and its isomers. The method provides a sensitive, simple and rapid approach for the quality control of TDF.
the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.