Rapid and sensitive liquid chromatography-tandem mass spectrometric method for the quantitative determination of potentially harmful substance 5,5'-oxydimethylenebis (2-furfural) in traditional Chinese medicine injections.

With the rapid development and wide application of traditional Chinese medicine injection (TCMI), a number of adverse events of some TCMIs have incessantly been reported and have drawn broad attention in recent years. Establishing effective and practical analytical methods for safety evaluation and quality control of TCMI can help to improve the safety of TCMIs in clinical applications. In this study, a sensitive and rapid high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) method has been developed and validated for the quantitative determination of potentially harmful substance 5,5'-oxydimethylenebis (2-furfural, OMBF) in TCMI samples. Chromatographic separation was performed on a C18 reversed-phase column (150 mm × 2.1 mm, 5 µm) by gradient elution, using methanol-water containing 0.1% formic acid as mobile phase at the flow rate of 0.3 mL/min. MS/MS detection was performed on a triple quadrupole mass spectrometer with positive electrospray ionization in the multiple reaction-monitoring mode. The method was sensitive with a limit of quantification of 0.3 ng/mL and linear over the range of 0.3-30 ng/mL (r=0.9998). Intra- and inter-day precision for analyte was <9.52% RSD with recoveries in the range 88.0-109.67% at three concentration levels. The validated method was successfully applied to quantitatively determine the compound OMBF in TCMIs and glucose injections. Our study indicates that this method is simple, sensitive, practicable and reliable, and could be applied for safety evaluation and quality control of TCMIs and glucose injections.

Introduction

As a new dosage form of traditional Chinese medicine (TCM), traditional Chinese medicine injection (TCMI) is considered to be a great achievement of modernization of TCM. TCMIs have been extensively used in China to treat a variety of diseases, including bacterial and viral infections, musculoskeletal disorders, cancer, cardiovascular and cere-brovascular dysfunction1, 2, 3. However, many serious adverse drug reactions (ADRs) of TCMI in some patients, including anaphylactic shock and fatal anaphylaxis, have been reported in recent years4, 5, 6. Because TCMI is a complex concoction made from extracts derived from a single herb or a group of herbs in a composite formula, and chemical ingredients in herb medicine, which vary greatly with the geographical origin of the species, time of harvest, cultivation practice, methods of processing, and storage condition, further contribute to the complexity and instability of TCMI7, 8, 9, 10, 11. Moreover, the quality control of TCMI is still unresolved for its complex composition. Particularly, current methods for monitoring the potentially harmful components in TCMI produced in the procedure of preparation, transportation and storage, are inadequate. Thus, it is imperative and urgent to develop practicable and reliable analytical methods for the purposes of improving the safety and quality of TCMI.

5-Hydroxymethyl furfural (5-HMF, C6H6O3, Fig. 1A), a common product of the Maillard reaction, is generated by acid-catalysed thermal dehydration from fructose, saccharose and to a lesser degree from glucose12, 13, 14. Thus, it can be easily found in many processed sugar- or starch-rich foods and heat-sterilised glucose/fructose solutions for pharmaceutical preparations15, 16, 17. Excessive 5-HMF can cause skin irritation, damage to striated muscles liver cancer, or induce aberrant crypt foci in the colon. Therefore, the content of 5-HMF in dextrose injection was strictly limited in Chinese Pharmacopoeia and United States Pharmacopoeia21, 22. In our previous study, we have found that there were significant differences in the content of 5-HMF in TCMI samples produced by different manufacturers or even different batches from the same manufacturer.

Figure 1

The chemical structures of (A) 5-hydroxymethyl furfural and (B) 5,5′-oxydimethylenebis (2-furfural).

Due to the unstable character of 5-HMF, it is readily hydrolyzes to levulinic acid and formic acid under acidic aqueous conditions24, 25. In addition, 5-HMF can also participate in hydrogenation, esterification and polymerization reactions. 5,5′-Oxydimethylenebis (2-furfural, OMBF, C12H10O5, Fig. 1B), a coloured polymer of 5-HMF, is a by-product of the Maillard reaction, and is generally produced from acid-catalyzed dehydration reaction of 5-HMF28, 29. We occasionally found this compound in TCMI samples through imitating the high-temperature/high-pressure sterilization procedures of glucose injection production. Our further study revealed that OMBF has immunosensitizing potential by acting as a neo-antigen or neo-epitope to elicit a mixed type-1 and type-2 immune response, and exposure to OMBF may represent a safety concern for humans. However, no studies have been reported on screening this potentially harmful substance in TCMIs and glucose injections. We herein developed a rapid, simple and sensitive HPLC–MS/MS method for quantitative determination of OMBF in TCMIs and glucose injections. The developed method in this study could rapidly and sensitively determine OMBF with short analysis time, low limits of detection and quantification, and could also contribute to improve the safety and quality of TCMI and glucose injections.

Material and methods

Chemicals and reagents

OMBF was prepared using the procedure outlined by Larousse et al. A purity of >98% was detected by HPLC. Formic acid was purchased from Sigma—Aldrich (St. Louis, MO, USA). HPLC grade methanol was purchased from Merck (Muskegon, MI, USA). Pure water was obtained from the Wahaha Group Co., Ltd. (Hangzhou, China). Other chemicals were of analytical grade.

TCMI samples (Shuxuening Injection, Qingkailing Injection, Chaihu Injection, Huangqi Injection, Xuesaitong Injection, Shuanghuanglian Injection, Shenmai Injection, Mailuoning Injection, Zhiyinhuang Injection, Tianmasu Injection, Gegensu Injection, Chuanhuning Injection, Guanxinning Injection, Dengzhanxixin Injection, Shengmai Injection, Ciwujia Injection, Chuanxinlian Injection, Danxiangguanxin Injection, Xiyanping Injection) and glucose injection samples produced by different Chinese pharmaceutical factories were purchased at local pharmacies.

Sample preparation

Accurately weighed OMBF was dissolved in methanol to prepare a 0.335 mg/mL stock solution. Standard working solutions of OMBF at concentrations of 0.3, 0.6, 1.2, 2.4, 5, 15 and 30 ng/mL were prepared by serial dilutions of the stock solution with methanol. TCMI samples and glucose injection samples were filtered through a 0.22 µm Nylon membrane and kept at 4 °C before use.

HPLC–MS/MS conditions

Samples analyses were performed on an Agilent 1200 series rapid resolution liquid chromatography system (Agilent technologies, Waldbronn, Germany) equipped with a binary gradient pump, autosampler, column oven and diode array detector. A Zorbax SB C18 column (150 mm × 2.1 mm, 5 µm; Agilent, USA) was used for separation. The mobile phase consisted of water with 0.1% (v/v) formic acid (A) and methanol (B) with a gradient elution. The gradient conditions were as follows: initial 5% B maintained for 10 min to balance the column, 0−5 min, maintained at 5% B; 5−18 min, increased to 100% B in 18 min; and 18−30 min, 100% solvent B. The flow rate was 0.3 mL/min, and the column temperature was kept at 35 °C throughout the analysis. The injected sample volume was 10 µL.

Mass spectrometry experiments were analyzed on an API 4000 triple quadrupole mass spectrometer (Applied Biosystems Sciex, Ontario, Canada), with an electrospray ion (ESI) source in the positive ion mode. The optimized mass spectrometric conditions for the compound OMBF included the following conditions: source temperature, 450 °C; ESI source voltage, 5 kV; nebulizer gas (Gas1), 60 psi; turbo gas (Gas2), 55 psi; curtain gas (CUR), 30 psi; entrance potential (EP), 10 V; declustering potential (DP), 55 V; and collision energy (CE), 18 eV. Nitrogen gas was used for both nebulizing and drying. The dwell time was 50 ms, with a 5 ms pause between scans. The MS/MS detection was operated at unit resolution in the multiple reaction monitoring (MRM) mode. The mass transitions of the protonated precursor/product ion pairs that were used to record the selected ion mass chromatograms of OMBF were m/z 235.1→109.0. Data acquisition and processing were performed using Analyst 1.5.1 supplied by AB SCIEX (Foster City, CA, USA).

Method validation procedure

The quantitative HPLC–MS/MS method was validated by assessing linearity, limit of detection (LOD) and quantification (LOQ), precision, accuracy, stability and recovery, according to the US Food and Drug Administration (FDA) and China Food and Drug Administration (CFDA) guidelines for the validation of analytical methods.

The LOD and LOQ of the analyte were calculated by analyzing a series of dilute standard solutions of known concentration at signal-to-noise ratios (S/N) of 3:1 and 10:1, respectively. The LOQ served as the lowest standard on the calibration curve in this analytical method. Standard seven point calibration curves, covering the range 0.3–30 ng/mL, were used for determination of linearity. A weighted (1/x) least squares regression analysis was used to determine the intercepts, slopes and correlation coefficients (r). Linearity was considered to be acceptable when correlation coefficients were higher than 0.99 and calibrators had accuracies of 85%–115% and precisions within ±15% RSD.

The intra-day precision and accuracy of the analytical method were assessed by processing and analyzing five replicates of the OMBF standard solutions at three concentration levels. The inter-day precision and accuracy were evaluated over 3 days by analyzing 15 samples (n=5 for each concentration level) each day. Precision of determination was expressed as the percentage relative standard deviation (% RSD) and accuracy was expressed as the percentage of nominal values. The acceptance limits were <15% RSD for precision and 85%–115% for accuracy.

The recovery was also evaluated by adding OMBF standard solution to a TCMI sample. In this experiment, Mailuoning injection (Lot No. 20140544) containing 75.0 ng/mL of OMBF was used. The Mailuoning Injection was diluted 10 times with water. 1 mL of diluted Mailuoning Injection was added to a 2 mL volumetric flask. Then 0.4, 0.5 and 0.6 mL of OMBF standard solution (15 ng/mL) were added to the flasks, respectively, and the volumes were adjusted to 2 mL by adding water. To determine recovery at LOQ, the Mailuoning Injection was diluted to 0.6 ng/mL. A 0.2 mL of OMBF standard solution (15 ng/mL) was spiked into 1 mL of diluted Mailuoning Injection, and then the volumes were adjusted to 2 mL by adding water. These samples were prepared in triplicate for each concentration level and were disposed as described above, and analyzed with the procedure. The average recovery was estimated by the Eq. (1):

Stability studies were carried out as part of the method validation. OMBF stability in terms of short-term stability and long-term stability were assessed by analyzing replicates (n=3) of standard solution samples at concentrations of 0.6, 12 and 24 ng/mL, and a TCMI sample (Mailuoning Injection, Lot No. 20140544). The standard solutions of OMBF were prepared in methanol. The short-term stability was evaluated after exposure of the standard solution samples of OMBF and the TCMI sample Mailuoning Injection to room temperature for 0, 2, 4, 8, 12 and 24 h, and the long-term stability was examined at room temperature over 0, 2, 3, 7 and 15 days.

The absolute matrix effect was also determined by comparing the chromatographic peak areas of OMBF spiked into blank solvent with peak areas obtained from the same concentration of OMBF in the glucose injection. The matrix effect was calculated using the Eq. (2):

The experiment was also evaluated at three concentrations of the analyzed compound (n=6 for each concentration level).

Residual action was also evaluated in this experiment by analyzing blank samples after injection high concentration of OMBF (30 ng/mL, n=6). The acceptance limit was <20% LOQ.

Results and discussion

Optimization of HPLC–MS/MS conditions

To optimize the chromatographic separation, a serial of preliminary experiments were performed, testing different mobile phases consisting of methanol, acetonitrile or mixture of acetonitrile and methanol as an organic phase and water with different mobile phase additives, such as formic acid and acetic acid at various concentrations. The addition of aqueous formic acid and acetic acid reduced peak tailing and improved the response of OMBF in positive ESI mode. Finally, methanol/0.1% aqueous formic acid (v/v) was selected as the best mobile phase for the chromatographic separation.

Full scan and MS/MS mass spectra were obtained from infusion of 10 ng/mL standard solution of OMBF at a flow rate of 0.1 mL/min. The protonated molecular ion [M+H]+ and the sodium adduct ion [M+Na]+ of OMBF were observed at m/z 235.1 and 257.0, respectively, in the positive ion mode mass spectrum (Fig. 2). The MS/MS spectrum of the precursor ion at m/z 235.1 displayed characteristic product ions at m/z 207.1, 177.1, 109.0 and 81.0. The product ion at m/z 109.0, which had the highest relative intensity, was chosen as the quantitative ion.

Figure 2

ESI-MS spectrum (A) of 5,5′-oxydimethylenebis (2-furfural), and product ion scan (MS/MS) spectrum (B) of the precursor ion [M+H]+ in the positive ion mode.

The optimization of ESI-MS/MS parameters was performed for analyte in infusion experiments: 2 ng/mL standard solution of OMBF was infused at a constant flow-rate of 5 µL/min into the mass spectrometer using a syringe pump. The following parameters were tested: nebulizer gas, turbo gas, curtain gas, spray voltage, capillary temperature, entrance potential, declustering potential and collision energy. The optimal conditions are given in the experimental section.

Method validation

Sensitivity and linearity

The calibration curve for OMBF was y=62169x+814.94 (x, concentration of reference substance; y, peak area). Good results were achieved in the range of 0.3–30 ng/mL for OMBF, with an excellent correlation coefficient (r=0.9998). The LOD (S/N=3) and LOQ (S/N=10) for OMBF were 0.1 and 0.3 ng/mL, respectively. The LOQ was determined as the lowest concentration point on the calibration curve that could be quantitated with an accuracy within ±15% bias of nominal concentration and precision not exceeding 15% coefficient of variation.

Precision and accuracy

Precision and accuracy were determined by analyzing high, medium and low standard concentrations of OMBF samples on the same day (intra-day, n=5) and continuously for 3 days (inter-day, n=15). Intra- and inter-day precision and accuracy, as shown in Table 1, were in the range 1.50%–9.52%, and 99.08%–113.33%, respectively. The results revealed that all values were in the acceptable ranges, indicating the method was reliable and reproducible for the determination of OMBF in traditional Chinese medicine injections.

Table 1

Intra- and inter-day assay precisions and accuracies of the developed HPLC–MS/MS method for the determination of OMBF.

Nominal (ng/mL)	Intra-day (n=5)			Inter-day (n=15)
	Determined (mean±SD)	Precisiona (%)	Accuracyb (%)	Determined (mean±SD)	Precisiona (%)	Accuracyb (%)
0.6	0.69±0.01	7.42	113.33	0.68±0.07	9.52	113.33
12	12.64±0.23	1.82	105.33	12.50±0.21	1.66	104.17
24	23.78±0.36	1.50	99.08	23.95±0.40	1.65	99.79

Expressed as RSD(%) = (SD/mean) × 100.

Calculated as (mean determined concentration/nominal concentration) × 100.

Recovery and stability

For the recovery test, known amounts (low, medium, and high) of the OMBF were spiked into samples and then prepared as test solutions. The determination was performed in triplicate, and the average recoveries and RSD were calculated and summarized in Table 2. The developed method had good accuracy with the overall recovery of 88.0%–109.67%, with the RSD of 7.27%. Thus, the recoveries of OMBF were consistent, precise and reproducible within the acceptance criteria.

Table 2

Recoveries of OMBF in traditional Chinese medicine injection.

No.	Original amount (ng/mL)	Spiked (ng/mL)	Measured (ng/mL)	Recovery (%)
1	0.3	1.5	1.62	88.0
2	0.3	1.5	1.58	88.33
3	0.3	1.5	1.65	90.0
4	3.75	3.0	6.76	100.33
5	3.75	3.0	6.74	99.67
6	3.75	3.0	7.04	109.67
7	3.75	3.75	7.81	108.27
8	3.75	3.75	7.44	98.40
9	3.75	3.75	7.36	96.27
10	3.75	4.5	7.89	92.00
11	3.75	4.5	8.02	94.89
12	3.75	4.5	8.21	99.11

The results of short-term stability and long-term stability of OMBF in TCMI are shown in Table 3. All the results indicated that the analyte was stable at room temperature for 24 h, and also unaffected by storage at room temperature for 15 days.

Table 3

Results of stability tests carried out on standard solutions of OMBF at three concentrations, and a TCMI sample (Mailuoning Injection) under different storage conditions.

Theoretical concentration (ng/mL)	Short-term stability								Long-term stability
	Determined concentration (ng/mL)						Precision (RSD%)	Accuracy (%)	Determined concentration (ng/mL)					Precision (RSD%)	Accuracy (%)
	0 h	2 h	4 h	8 h	12 h	24 h			0 d	2 d	3 d	7 d	15 d
0.6	0.76	0.68	0.65	0.74	0.62	0.67	7.7	114.7	0.69	0.74	0.60	0.81	0.62	12.32	115.2
12	12.0	12.4	12.7	12.8	13.0	12.5	2.8	104.7	12.84	12.54	12.32	12.7	12.8	1.69	105.3
24	23.2	24.2	25.1	25.2	25.3	24.1	3.4	102.1	24.2	24.26	23.8	24.1	24.3	8.30	100.6
75.0	74.8	74.6	79.9	81.2	78.9	76.2	3.6	103.5	76.2	74.8	79.2	79.2	79.4	2.73	103.7

Matrix effect and residual action

The matrix effects of the analysis were within the range of 95%–105%, indicating that no significant ion suppression or enhancement of glucose solution was observed using the current method. After injection high concentration of OMBF sample, the residue in blank sample was 5.8% of LOQ. The residual action in the present method thus meet the acceptance criteria.

Based on all of these validation results, the present method was considered to be suitable for the quantitative analysis of OMBF in TCMI samples.

Quantitative determination of OMBF in TCMI and glucose injection samples

59 TCMIs and glucose injection samples from different batches or different manufacturers were analyzed using the established HPLC–MS/MS method. A representative MRM chromatogram of TCMI samples is shown in Fig. 3A. Peak identity was confirmed by both retention time compared to that of the reference analyte (Fig. 3B) and by the characteristic ion pairs. The results of quantitative determination of OMBF in TCMI and glucose injections are shown in Table 4. OMBF was detected in 6 samples, and the content of OMBF was in the range 0.37–127 ng/mL. Obviously, the concentrations of OBMF were significantly different between different batches from one manufacturer as well as between different manufacturers. Such significant differences in the content of OMBF would likely to be associated with the variability of sugar content in the raw material and the pH value of TCMIs. In our previous study, we found Mailuoning Injection (Lot No. 20120451) contained a high concentration of 5-HMF (1420 μg/mL). In this study, two batches of Mailuoning Injection (Lot Nos. 20140928 and 20140544) produced by the same manufacturer, were also found to contain high concentration of OMBF. These products may increase the risk to cause adverse drug reaction. In addition, it is worth noting that the OMBF was also detected in a glucose injection sample. Therefore, further toxicity studies on compound OMBF should be conducted, and it is strongly recommended that limit criteria for the content of OMBF in TCMI and glucose injection should be established to ensure the safety in clinical application.

Figure 3

Representative MRM chromatograms of (A) a TCMI samples (Mailuoning Injection, Lot No. 20140544), (B) standard solution of OMBF at 2.4 ng/mL analyzed by HPLC–MS/MS method. The retention time for OMBF was 14.8 min.

Table 4

Results of OMBF quantification in 59 samples of TCMI and glucose injection using the developed HPLC–MS/MS method.

No.	Sample name	Lot No.	Content (ng/mL)
1	Shuxuening injection#1	1041411041	n.d.
2	Shuxuening injection*1#1	1041410242	n.d.
3	Shuxuening injection#1	140726B1	n.d.
4	Shuxuening injection*1#1	140725B2	n.d.
5	Qingkailing injection#2	14052902	n.d.
6	Qingkailing injection#2	14052305	n.d.
7	Qingkailing injection*2#2	14011208	n.d.
8	Qingkailing injection*2	5140830B2	n.d.
9	Qingkailing injection#2	14120951	n.d.
10	Qingkailing injection#2	14122752	n.d.
11	Qingkailing injection#2	14121862	n.d.
12	Qingkailing injection*2#2	14122663	n.d.
13	Chaihu injection*3	1501092	n.d.
14	Chaihu injection*3	14121622	n.d.
15	Chaihu injection*3	311405092	n.d.
16	Chaihu injection*3	141022D1	n.d.
17	Huangqi injection#4	141106A3	n.d.
18	Huangqi injection#4	141126A3	n.d.
19	Huangqi injection*4#4	141127A2	n.d.
20	Huangqi injection*4	A20141201	n.d.
21	Huangqi injection*4	1410273	n.d.
22	Xuesaitong injection#5	B20140416	n.d.
23	Xuesaitong injection#5	B20140710	n.d.
24	Xuesaitong injection#5	B20140402	n.d.
25	Xuesaitong injection#5	B20140306	n.d.
26	Xuesaitong injection#5	B20140709	n.d.
27	Xuesaitong injection*5#5	B20140708	n.d.
28	Xuesaitong injection#5	ZGA1457	n.d.
29	Xuesaitong injection*5#5	ZGA1446	n.d.
30	Xuesaitong injection#5	14HJ205-11	n.d.
31	Xuesaitong injection#5	14GJ206-11	n.d.
32	Xuesaitong injection#5	14GJ205-21	n.d.
33	Xuesaitong injection*5#5	14GJ205-22	n.d.
34	Shuanghuanglian injection#6	14094012	n.d.
35	Shuanghuanglian injection#6	14094013	n.d.
36	Shuanghuanglian injection#6	14094112	n.d.
37	Shuanghuang injection*6#6	14094111	n.d.
38	Shenmai injection#7	13071321	n.d.
39	Shenmai injection*7#7	140427F1	n.d.
40	Mailuoning injection#8	20140544	75.0
41	Mailuoning injection*8#8	20140928	127
42	Zhiyinhuang injection	140913B1	n.d.
43	Tianmasu injection#9	14070303-2	n.d.
44	Tianmasu injection*9#9	1411150111	n.d.
45	Gegensu injection#10	1411037	n.d.
46	Gegensu injection*10#10	141004	n.d.
47	Chuanhuning injection#11	141015	n.d.
48	Chuanhuning injection#11	140704	n.d.
49	Chuanhuning injection*11#11	140718	n.d.
50	Guanxinning injection	10214091627	n.d.
51	Dengzhanxixin injection	20150148	Det
52	Shengmai injection*12	13120711	2.44
53	Shengmai injection#12	14060602	n.d.
54	Shengmai injection*12#12	14050704	n.d.
55	Ciwujia injection	2014090521	0.37
56	Chuanxinlian injection	14062201A	n.d.
57	Danxiangguanxin injection	251150106	Det
58	Xiyanping injection	2014121403	1.18
59	Glucose injection	D14070210	0.76

n.d., not detected--peak not observed or concentration is lower than the LOD.

Det, detected--peak observed, concentration is lower than the LOQ but higher than the LOD.

*The test samples with the same name from different manufacturers.

#The test samples with the same name but different lot No. from the same manufacturer.

Conclusions

In this study, a rapid and sensitive HPLC–MS/MS method has been developed and validated for the quantitative determination of OMBF in TCMIs and glucose injections for the first time. It was successfully applied to large-scale screening of OMBF in commercially available TCMIs and glucose injections. The developed HPLC–MS/MS method has been elucidated to be a simple, sensitive, practicable and reliable quality control procedure for TCMI and glucose injection. In addition, this method can be further adapted for the analysis of OBMF in other herbal medicines or preparations containing this compound.