Monitoring quality consistency of Ixeris sonchifolia (Bunge) Hance injection by integrating UV spectroscopic fingerprints, a multi-wavelength fusion fingerprint method, antioxidant activities and UHPLC/Q-TOF-MS

Abstract

Traditional Chinese medicines/herbal medicines (TCM/HM) are too complicated mixture systems to investigate their quality consistency effectively using a single detection technique. Hence, finding an integrated, rapid and effective quality control method is of great importance. In the present study, a novel strategy integrating ultra violet (UV) spectroscopic fingerprints and multi-wavelength fusion fingerprints has been proposed for quality consistency of the popular patent medicine Ixeris sonchifolia (Bge.) Hance Injection (ISHI). The quality grades of 28 batches of the ISHI samples were evaluated by the systematic quantitative fingerprint method (SQFM) from qualitative and quantitative perspectives. As a result, UV spectroscopic fingerprints agree well with multi-wavelength fusion fingerprints. In order to investigate the chemical components of the ISHI sample, ultra-high performance liquid chromatography coupled with quadrupole-time-of-flight mass spectrometry (UHPLC/Q-TOF-MS) has been exploited in the negative ion mode, providing an important chemical structure foundation for further bioactivity and quality control studies. In addition, the orthogonal projection to latent structures (OPLS) was established to investigate the fingerprint-efficacy relationship between high performance liquid chromatography (HPLC) fingerprints and antioxidant activities. These results indicate that integrating UV spectroscopic fingerprints and multi-wavelength fusion fingerprints coupled with antioxidant activities and UHPLC/Q-TOF-MS provide a rapid and effective approach to monitor the quality consistency of TCM/HM.

---

1. Introduction

TCM/HM usually contain a great number of components, which contribute to the therapeutic effects all together with multiple targets.¹ Obviously, it is inadequate to control the quality of the TCM/HM by quantifying only one or a few marker substances. Fingerprint techniques, especially chromatographic fingerprints, as a rational and powerful method, systematically characterize the complicated TCM/HM, and have been widely accepted and adopted for quality control of TCM/HM in recent years.^2,3 However, single wavelength fingerprints cannot reflect the sample chemical information systematically. In this situation, a multi-wavelength fusion fingerprints method^4,5 has been adopted to make up for this limitation. By combining the macroscopical fingerprint characteristics and computer technology, the fusion fingerprints technique could realize the overall evaluation of the samples.⁶

Although chromatographic fingerprints performed on HPLC,³ UHPLC,⁷ gas chromatograph (GC),⁸ capillary electrophoresis (CE),⁹ etc. are frequently used for quality control of TCM/HM, cannot offer information about all the chemical components in the TCM/HM system, which limit the overall quality control of multi-constituents synergistically. In this situation, spectroscopic techniques, such as UV, near-infrared (NIR) and mid-infrared (MIR) have been proposed;^10,11 and UV spectroscopic fingerprint is preferred using flow injection analysis (FIA)¹² to monitor the quality consistency of TCM/HM, owing to its apparent advantages, such as short measurement time, simple sample preparation and low experimental expenses. As we all know, UV spectra can reflect the chemical bond information of π–π*, n–π* and n–σ*.¹³ Since UV spectra reflects the features of unsaturated bonds of components qualitatively and quantitatively, it seems to be a straightforward and effective method for TCM/HM quality control.

TCM/HM possess complex components and variety of therapeutic efficacies, thus, it is necessary to carry out the chemical profiling of TCM/HM for further bioactivity and quality control studies. In recent years, UHPLC/Q-TOF-MS has been proved to be an efficient analytical tool for the chemical components of TCM/HM.^14,15 Compared with the conventional HPLC, UHPLC provides lower consumption of samples and higher sensitivity in a shorter analysis time.⁷ Meanwhile, Q-TOF-MS offers both valuable fragment ions and reliable formulas, which is benefited by the high resolution for structural characterization.¹⁶ Accordingly, UHPLC/Q-TOF-MS supplies a rapid and powerful method to investigate TCM/HM components.

Antioxidants can scavenge free radicals to reduce the risk of numerous diseases, which is closely related to the harmful effects, such as lipid peroxidation, protein peroxidation, and DNA damage of free radicals.¹⁷ Accordingly, research on natural antioxidants and their antioxidant properties in foods, plant materials and TCM/HM is attracting increasing attention.^2,18 2,2-Diphenyl-1-picrylhydrazyl (DPPH) is a stable and commercially available radical, and the DPPH radical scavenging assay providing simple, rapid, sensitive and reproducible procedures, has been employed to evaluate the antioxidant activity of TCM/HM.^3,19,20 These information encourage us to investigate the antioxidant activities of TCM/HM and correlate them with chromatographic fingerprints.

ISHI has been widely used for its anti-inflammatory and haemostatic effects, its influence on improving blood circulation, and its potential protection against ischemia brain injury with relatively low side effects.²¹ The compositions of ISHI mainly including organic acids, flavonoids, sesquiterpene lactones, nucleosides and other compounds.²² Previous studies on chemical components of Ixeris sonchifolia and its injection (i.e., ISHI) have been reported. For example, Liu et al. characterized a total of 33 constituents from Ixeris sonchifolia and its injection, including 16 organic acids, 7 flavonoids, 7 sesquiterpene lactones, and 3 nucleosides;²³ A total of 6 sesquiterpene lactones, 6 phenolic acids and 7 flavonoids were identified or tentatively characterized in Ixeris sonchifolia by Shi et al.;²⁴ Wang et al. characterized a total of 25 phenolic acids in ISHI;²⁵ A total of 14 chlorogenic acids and 16 flavonoids have been screened and identified in ISHI by Wang et al.²⁶ After a thorough literature search, 7 compounds including 2 organic acids (vanillic acid and (E)-2,5-dihydroxy cinnamic acid), 3 sesquiterpene lactones (sonchifolactone E, 3-hydroxydehydroleucodin and 11,13-dihydroixerinoside) and 2 phenols (3-(4′-hydroxy-3′-methoxyphenyl)-1-propanol and vanillin) were first identified in ISHI in the present study. In addition, published studies have shown that the antioxidant components, especially organic acids and flavonoids in Ixeris sonchifolia and its injection, can provide neuro-protective effects against ischemia-induced cellular injury.^21,27 Accordingly, the antioxidant activity of the ISHI was performed by the DPPH assay in this study, and the predictive model was established using the OPLS method.²⁸

In the present study, UV spectroscopic fingerprints and a multi-wavelength fusion fingerprints method were developed for evaluating the quality consistency of 28 batches of the ISHI samples, where the fusion fingerprint is based on the three wavelengths that displayed strong UV absorption bands. In fingerprint assessments, SQFM was established for quality analysis of the ISHI from qualitative and quantitative perspectives. In addition, the fingerprint-efficacy relationship between HPLC fingerprints and antioxidant activities was established utilizing OPLS. UHPLC/Q-TOF-MS has been exploited to systematically investigate the chemical components in ISHI in the negative ion mode, which provided information on chemical substances for bioactivity and quality control studies. It has been demonstrated that integrating UV spectroscopic fingerprints and multi-wavelength fusion fingerprints method coupled with antioxidant activities and UHPLC/Q-TOF-MS offers an effective and powerful method for quality control of ISHI.

2. Experimental

2.1 Chemicals and reagents

Reference standards, caffeic acid, 5-caffeoylquinic acid and ferulic acid were acquired from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China); chicoric acid and luteolin-7-O-β-D-glucuronide were supplied by Chengdu Puri France Science and Technology Development Co., Ltd. (Chengdu, China); luteolin-7-O-β-D-glucoside and luteolin were provided by Shanghai Winherb Medical Technology Co., Ltd. (Shanghai, China); uridine, guanosine and DPPH were purchased from Sigma Chemical Co. (St. Louis, MO, USA). A total of 28 batches of the ISHI samples (numbered S1–S28) were obtained from Shenyang Shuangding Pharmaceutical Co., Ltd. (Shenyang, China; Manufacturer A, including S1–S23) and Tonghua Huaxia Pharmaceutical Co., Ltd. (Jilin, China; Manufacturer B, including S24–S28), respectively. In the UHPLC/Q-TOF-MS analysis, acetonitrile (MS grade) was purchased from Merck (Darmstadt, Germany); formic acid (HPLC grade, purity >95%) was obtained from Sigma (Saint Louis, MO, USA). In the HPLC and UV spectroscopic analysis, methanol (HPLC grade), acetonitrile (HPLC grade) and glacial acetic acid (HPLC grade) were purchased from Yuwang Industry Co., Ltd. (Shandong, China). High-purity deionized water was prepared by a Milli-Q system (Bedford, MA, USA). Other reagents were all of analytical grade.

2.2 Preparation of standard and sample solutions

The reference standards of caffeic acid, 5-caffeoylquinic acid, ferulic acid, chicoric acid, luteolin-7-β-D-glucuronide, luteolin-7-glucoside, luteolin, uridine and guanosine were accurately weighed separately and dissolved in methanol, and stored at 4 °C prior to use.

Each of the tested ISHI samples was filtered through a 0.45 μm Millipore filter (Beijing Sunrise T&D Company, China) prior to use.

2.3 Experimental conditions

2.3.1 Instrumentation and UHPLC/Q-TOF-MS conditions. Chemical components analysis was performed on a Waters Acquity UHPLC system (Waters, Corp., Milford, MA, USA) coupled with a quaternary pump, an on-line degasser, an auto-sampler and a column oven controller. Chromatographic separation was achieved on a Waters Acquity UHPLC BEH Shield RP C₁₈ column (2.1 mm i.d. × 100 mm, 1.7 μm) at 30 °C. The mobile phase was composed of an aqueous solution containing 0.2% (v/v) formic acid (A) and acetonitrile (B). The gradient elution program was set as follows: 2–5% B at 0–5 min, 5–15% B at 5–10 min, 15–20% B at 10–15 min, 20–30% B at 15–18 min, 30–40% B at 18–20 min, 40–50% B at 20–22 min, 50–90% B at 22–25 min, 90–2% B at 25–30 min. The flow rate and the injection volume were 0.3 mL min⁻¹ and 3 μL, respectively. MS analysis was achieved on a Synapt G2 Q-TOF-MS (Waters, Corp., Milford, MA, USA) equipped with electrospray ionization (ESI) source in the negative ion mode. Ultra-high purity helium (He) and high-purity nitrogen (N₂) were used as the collision gas and the nebulising gas, respectively. Full-scan MS data were collected from 50 to 1200 Da with a collision energy of 20–40 eV. Other MS conditions were listed as follows: capillary voltage 2.5 kV, cone voltage 30 V, extraction cone voltage 5 V, ion source temperature 100 °C, desolvation temperature 300 °C, desolvation gas flow 600 L h⁻¹, and cone gas flow 50 L h⁻¹. Leucine-enkephalin at m/z 554.2615 was utilized to calibrate the mass accuracy.

2.3.2 HPLC chromatographic conditions. HPLC analysis was performed on an Agilent 1100 HPLC system (Agilent, USA) comprised of an online degasser, a low pressure mix quaternary pump, an auto-sampler, and a diode array detector (DAD). The chromatographic separation was carried out on an Arcus EP-C18 column (250 mm × 4.6 mm, 5 μm) (Shanghai, China) at 35 °C. The mobile phase was composed of water–5 mM citric acid and 10 mM sodium dihydrogen phosphate (A) and acetonitrile–1.0% (v/v) glacial acetic acid (B). By the way, the buffer solution of 5 mM citric acid and 10 mM sodium dihydrogen phosphate offered optimal improvement of peak separation and shape, while 1.0% (v/v) glacial acetic acid could prevent the tail of peak effectively. The gradient elution program was as follows: 0–3% B at 0–5 min, 3–7% B at 5–10 min, 7–10% B at 10–16 min, 10–17% B at 16–25 min, 17–19% B at 25–30 min, 19–20% B at 30–40 min, 20–25% B at 40–60 min. The flow rate and the injection volume were 1.0 mL min⁻¹ and 10 μL, respectively. The detection wavelengths were set at 236, 286 and 326 nm.

2.3.3 UV spectroscopic conditions. UV spectra was record by an Agilent 1100 HPLC system (Agilent, USA) equipped with a DAD and a hollow polytetrafluoroethylene (PTFE) pipe (6500 × 0.12 mm), with FIA being used as the analytical principle. Methanol–water (50 : 50, v/v) was adopted as the carrier. The temperature of the PTFE tube was maintained at 35 °C. The flow rate, sample injection volume and wavelength interval was set at 0.5 mL min⁻¹, 1 μL and 1 nm, respectively. Online UV spectra was obtained over a wavelength range of 200–400 nm with a slit width of 1 nm.

2.3.4 Antioxidant activity conditions. The DPPH radical scavenging assay was performed according to Bhandari et al.²⁹ with slight modification. Briefly, the mixture of 2 mL ISHI sample solution and 2 mL 0.127 mM DPPH solution (in methanol) was diluted in methanol to various concentrations. After stand in the dark for 40 minutes, the absorbance was measured at 517 nm against a blank performed on a 722S spectrophotometer (Shanghai Precision Instrument Co., Ltd., Shanghai, China). All tests were performed in triplicates. The radical scavenging capacity is expressed as percent inhibition and calculated using the following equation: % inhibition = [(A_control − A_sample)/A_control] × 100, where A_control is the absorbance of the negative control and A_sample is the absorbance at the presence of the ISHI sample. The antioxidant activity was calculated by plotting the percent inhibition vs. sample concentration, and represented as IC₅₀ value (the concentration of samples required to scavenge 50% of DPPH radicals). The lower IC₅₀ value, the stronger antioxidant activity.

2.4 Theory of SQFM

We assume that the sample fingerprint (SFP) and reference fingerprint (RFP) vector are x⃑ = (x₁, x₂,…, x_n) and y⃑ = (y₁, y₂,…y_n), respectively, where x_i and y_i are the peak area of the co-possessing peak in the SFP and RFP, respectively. Macro qualitative similarity (S_m), a qualitative parameter as shown in eqn (1), can accurately describe the resemblance in terms of the distribution and number of fingerprints between SFP and RFP. Furthermore, projection content similarity (as defined in eqn (2)) and quantitative similarity (P, as shown in eqn (3)) were calculated for quantitative assessment of the fingerprints. Combining the above two quantitative properties yields macro quantitative similarity (P_m, as defined in eqn (4)), which is a measure to monitor the overall content of chemical components in SFP and RFP. Finally, a fingerprint leveling coefficient (α), as defined in eqn (5), is another quantitative parameter that can detect the difference between SFP and RFP.

Accordingly, the quality evaluation method in terms of S_m, P_m and α is named as SQFM,³ by which TCM/HM quality can be classified into 8 grades. The evaluation criteria by SQFM are listed in Table 1. Accordingly, grade 1 belongs to the highest quality and grade 8 to the lowest one, and grades 1–5 are recognized as qualified.

Table 1 The quality grade criteria based on SQFM

Grade	1	2	3	4	5	6	7	8
Sm ≥	0.95	0.90	0.85	0.80	0.70	0.60	0.50	<0.50
Pm (%) ∈	95–105	90–110	80–120	75–125	70–130	60–140	50–150	0–∞
α ≤	0.05	0.10	0.15	0.20	0.30	0.40	0.50	>0.50
Quality	Best	Better	Good	Fine	Moderate	Common	Defective	Inferior

---

2.5 Data analysis

MassLynx 4.1 software (Waters, USA) was used for UHPLC/Q-TOF/MS data processing. An in-house developed software “Digitized Evaluation System for Super-Information Characteristics of TCM Fingerprints 4.0” (software certificate no. 0407573, China) was used to evaluate the fingerprints. SIMCA-P⁺ software (Version 13.0, Umetrics, Umea, Sweden) was used for chemometric analysis.

3. Results and discussion

3.1 UHPLC/Q-TOF-MS analysis of the ISHI sample

Under the UHPLC/Q-TOF-MS conditions described in Section 2.3.1, totally 38 compounds (Fig. 1A), including 16 organic acids, 10 flavonoids, 8 sesquiterpene lactones, 2 nucleosides and 2 phenols, were identified in ISHI. Structures of them are summarized in Fig. 2. Information on the t_R (min), observed mass (m/z), calculated mass (m/z), mass error (in ppm), molecular formula, fragment ions, identification and maximum UV absorption wavelength (UV λ_max) are summarized in Table 2. Components with reference standards (compounds 3, 4, 16, 17, 18, 28, 29, 31 and 38) were unambiguously identified by comparing the retention time, accurate mass (within 5 ppm error) and MS/MS data. Components without reference standards were tentatively characterized based on accurate mass measurements (within 5 ppm error), tandem MS behavior and related literature.

Fig. 1 Total ion chromatogram of ISHI in negative ion mode (A), the MS/MS spectra and proposed fragmentation pathways of chicoric acid (B), luteolin-7-o-β-D-glucuronide (C), ixerin Z (D) and guanosine (E) in the negative ion mode.

Fig. 2 Chemical structures of 38 identified compounds.

Table 2 Identification of the chemical constituents of the ISHI samples using UHPLC/Q-TOF/MS in negative ion mode

No.	tR (min)	Observed mass (m/z)	Calculated mass (m/z)	Error (ppm)	Formula	Fragment ions	Identification	UV λmax (nm)
a [M + HCOO]^−.b Compounds be identified in ISHI for the first time.
1	0.92	191.0559	191.0556	1.57	C7H12O6	191, 173, 127	Quinic acid	250, 280sh
2	1.26	167.0345	167.0344	0.60	C8H8O4	167, 123	Vanillic acid^b	235, 284
3	1.37	243.0617	243.0617	0.00	C9H12N2O6	243, 200, 111	Uridine	200, 260
4	1.94	282.0836	282.0838	−0.71	C10H13N5O5	282, 150, 133	Guanosine	254, 280sh
5	4.54	315.0509^a	315.0505	1.27	C15H10O5	315, 269, 241, 151	Apigenin	268, 336
6	5.19	153.0187	153.0188	−0.65	C7H6O4	153, 109	3,4-Dihydroxybenzoic acid	221, 259, 295
7	5.83	181.0867	181.0865	1.10	C10H14O3	181, 163, 135	3-(4′-Hydroxy-3′-methoxyphenyl)-1-propanol^b	225, 275
8	7.08	353.0869	353.0873	−1.13	C16H18O9	353, 191, 179, 135	3-Caffeoylquinic acid	246, 290sh, 324
9	7.36	151.0399	151.0395	2.65	C8H8O3	151, 123	Vanillin^b	215, 260, 290
10	7.67	179.0339	179.0344	−2.79	C9H8O4	179, 161, 135	(E)-2,5-Dihydroxy cinnamic acid^b	245, 300sh, 328
11	8.08	536.2138	536.2132	1.12	C26H35NO11	536, 421, 259	11,13-Dihydro-13-prolyl-ixerin Z	—
12	8.33	277.1085	277.1076	3.25	C15H18O5	277, 259, 249, 233	Sonchifolactone E^b	270
13	8.57	329.0669^a	329.0661	2.43	C16H12O5	329, 283, 268, 151	Acacetin	255, 350
14	9.40	353.0873	353.0873	0.00	C16H18O9	353, 191, 179, 173, 135	4-Caffeoylquinic acid	246, 294sh, 323
15	9.46	341.0671	341.0661	2.93	C18H14O7	341, 179, 151, 135	Caffeic anhydride	230, 288sh, 328
16	9.81	353.0873	353.0873	0.00	C16H18O9	353, 191, 179, 135	5-Caffeoylquinic acid	242, 292sh, 323
17	10.02	179.0345	179.0344	0.56	C9H8O4	179, 135	Caffeic acid	244, 294sh, 325
18	10.21	193.0505	193.0501	2.07	C10H10O4	193, 179, 149, 133	Ferulic acid	251, 290sh, 327
19	11.45	305.1029^a	305.1025	1.31	C15H16O4	305, 215, 197	3-Hydroxydehydroleucodin^b	273
20	11.86	439.1610	439.1604	1.37	C21H28O10	439, 421, 259	Sonchifolactone D	276
21	12.26	515.1185	515.1190	−0.97	C25H24O12	515, 353, 335, 191, 179	1,3-Dicaffeoylquinic acid	244, 300sh, 328
22	13.05	609.1469	609.1456	2.13	C27H30O16	609, 447, 285, 151	Luteolin-7-O-β-D-gentiobioside	254 346
23	13.66	421.1504	421.1499	1.19	C21H26O9	421, 259, 241, 215, 197	Ixerin Z	272
24	14.21	425.1822	425.1812	2.35	C21H30O9	425, 263, 245, 201	11,13-Dihydroixerinoside^b	275
25	14.37	423.1643	423.1655	−2.84	C21H28O9	423, 261, 217	11,13α-Dihydroixerin Z	274
26	14.51	593.1505	593.1506	−0.17	C27H30O15	593, 285, 151	Luteolin-7-O-rutinoside	252, 339
27	14.65	423.1645	423.1655	−2.36	C21H28O9	423, 243, 199	11β-Hydroxyleucodin-11-O-β-glucopyranoside isomer	274
28	14.94	447.0939	447.0927	2.68	C21H20O11	447, 285, 151	Luteolin-7-O-β-D-glucoside	254, 347
29	15.54	461.0726	461.0720	1.30	C21H18O12	461, 285, 151	Luteolin-7-O-β-D-glucuronide	254, 349
30	16.64	515.1182	515.1190	−1.55	C25H24O12	515, 353, 335, 191, 173	3,4-Dicaffeoylquinic acid	240, 290sh, 325
31	16.84	473.0733	473.0720	2.75	C22H18O12	473, 311, 293, 179, 149	Chicoric acid	252, 300sh, 327
32	17.02	431.0967	431.0978	−2.55	C21H20O10	431, 269, 151	Apigenin-7-O-β-D-glucoside	253, 325
33	17.25	515.1198	515.1190	1.55	C25H24O12	515, 353, 191, 179	3,5-Dicaffeoylquinic acid	240, 300sh, 330
34	17.50	445.0781	445.0771	2.25	C21H18O11	445, 269, 151	Apigenin-7-O-β-D-glucuronide	266, 326
35	17.99	515.1197	515.1190	1.36	C25H24O12	515, 353, 335, 191, 173	4,5-Dicaffeoylquinic acid	240, 290sh, 330
36	18.51	591.1723	591.1714	1.52	C28H32O14	591, 283, 268, 151	Acacetin-7-O-β-D-rutinoside	252, 326
37	18.71	325.0562	325.0560	0.62	C14H14O9	325, 193, 163, 149	Feruloyltartaric acid	252, 326
38	20.50	285.0393	285.0399	−2.11	C15H10O6	285, 241, 151	Luteolin	253, 346

---

3.1.1 Identification of organic acids in the ISHI sample. Organic acids in the ISHI sample could be classified as simple organic acids and acylated organic acids according to their chemical compositions, where simple organic acids referring to unacylated single carboxylic acids. Compounds 17 and 18 were unambiguously attributed to caffeic acid and ferulic acid by comparison with the reference compounds. In the negative ion mode, caffeic acid contained an [M − H]⁻ at m/z 179.0345 (0.56 ppm, C₉H₇O₄), and fragment at m/z 135.0449 (2.22 ppm, C₈H₇O₂) corresponding to [M − H − CO₂]⁻; ferulic acid exhibited an [M − H]⁻ at m/z 193.0505 (2.07 ppm, C₁₀H₉O₄), and fragments at m/z 178.0271 (2.81 ppm, C₉H₆O₄), m/z 149.0608 (3.36 ppm, C₉H₉O₂) and m/z 134.0373 (3.73 ppm, C₈H₆O₂), corresponding to [M − H − CH₃]⁻, [M − H − CO₂]⁻, [M − H − CH₃ − CO₂]⁻, respectively.

Based on the analysis of the standards above and related literature,^15,25 simple organic acids were found easily lose neutral units of CH₃ (15 Da), H₂O (18 Da), CO (28 Da) and CO₂ (44 Da) under the negative ion mode. According to these fragmentation patterns, compounds 1, 2, 6 and 10 were identified as quinic acid, vanillic acid, 3,4-dihydroxybenzoic acid and (E)-2,5-dihydroxy cinnamic acid, respectively. After a thorough literature search, vanillic acid (compound 2) and (E)-2,5-dihydroxy cinnamic acid (compound 10) were identified in ISHI for the first time. Briefly, vanillic acid contained an [M − H]⁻ at m/z 167.0345 (0.60 ppm, C₈H₇O₄) in the negative ion mode, and fragment at m/z 123.0449 (2.44 ppm, C₇H₇O₂) corresponding to [M − H − CO₂]⁻; (E)-2,5-dihydroxy cinnamic acid exhibited an [M − H]⁻ at m/z 179.0339 (−2.79 ppm, C₉H₇O₄) in the negative ion mode, and fragments at m/z 161.0236 (−1.86 ppm, C₉H₅O₃) and m/z 135.0445 (−0.74 ppm, C₈H₇O₂) corresponding to [M − H − H₂O]⁻ and [M − H − CO₂]⁻, respectively.

Acylated organic acids are important organic acids in the ISHI samples. Compounds 16 and 31 were unambiguously attributed to 5-caffeoylquinic acid and chicoric acid by comparison with the reference compounds. Chicoric acid, a typical acylated organic acid in the ISHI samples, was used to characterize the fragmentation pathway (Fig. 1B). An [M − H]⁻ at m/z 473.0733 (2.75 ppm, C₂₂H₁₇O₁₂) was found in the negative ion mode, which could successively lose one caffeoyl group (162 Da) or two caffeoyl groups (2 × 162 Da) to form fragments at m/z 311.0415 (3.86 ppm, C₁₃H₁₁O₉) and m/z 149.0091 (3.36 ppm, C₄H₅O₆), respectively. Fragment at m/z 311.0415 further lose one H₂O (18 Da) or one tartaric acid (132 Da) to form fragments at m/z 293.0301 (1.37 ppm, C₁₃H₉O₈) and m/z 179.0349 (2.79 ppm, C₉H₇O₄), respectively. 5-Caffeoylquinic acid exhibited an [M − H]⁻ at m/z 353.0873 (0.00 ppm, C₁₆H₁₇O₉) in the negative ion mode, and fragments at m/z 191.0552 (−2.09 ppm, C₇H₁₁O₆), m/z 179.0343 (−0.56 ppm, C₉H₇O₄) and m/z 135.0440 (−4.44 ppm, C₈H₇O₂), corresponding to [M − H − caffeoyl]⁻, [M − H − quinic]⁻, [M − H − quinic − CO₂]⁻, respectively.

Based on the analysis of the standards above and related literature,^15,23,25,30 acylated organic acids in ISHI exhibited specific diagnostic fragments, such as tartaric acid ([C₄H₅O₆]⁻ at m/z 149), caffeic acid ([C₉H₇O₄]⁻ at m/z 179) and quinic acid ([C₇H₁₁O₆]⁻ at m/z 191), under the negative ion mode. According to these fragmentation patterns, compounds 8, 14, 15, 21, 30, 33, 35 and 37 were identified as 3-caffeoylquinic acid, 4-caffeoylquinic acid, caffeic anhydride, 1,3-dicaffeoylquinic acid, 3,4-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, 4,5-dicaffeoylquinic acid and feruloyltartaric acid, respectively.

3.1.2 Identification of flavonoids in the ISHI sample. Compounds 28, 29 and 38 were unambiguously attributed to luteolin-7-O-β-D-glucoside, luteolin-7-O-β-D-glucuronide and luteolin by comparison with the reference compounds. Luteolin-7-O-β-D-glucuronide, a typical flavonoid in the ISHI samples, was used to characterize the fragmentation pathway (Fig. 1C). An [M − H]⁻ at m/z 461.0726 (1.30 ppm, C₂₁H₁₇O₁₂) was found in the negative ion mode, which could successively lose one gluconic acid group (GlcA, 176 Da) to form fragment at m/z 285.0615 (1.75 ppm, C₁₂H₁₃O₈). Then C₁–C₃ of the A ring at m/z 285.0615 easily underwent the retro Diels–Alder (RDA) reaction and produced fragment ion at m/z 151.0035 [^1,3A⁻, 2.65 ppm, C₇H₃O₄]⁻.

Based on the analysis of the standards and related literature,^15,23,30 flavonoids in the ISHI samples could be divided into glycoside and aglycone. Glycoside compounds in the ISHI samples easily lose glycosyl (Glc, 162 Da), gluconic acid (GlcA, 176 Da) and rutinose (Rut, 308 Da) and converted to aglycone, while aglycone in the ISHI samples containing luteolin ([C₁₅H₉O₆]⁻ at m/z 285), apigenin ([C₁₅H₉O₅]⁻ at m/z 269) and acacetin ([C₁₆H₁₁O₅]⁻ at m/z 283), and usually underwent the RDA reaction and produced the same fragment ion at m/z 151 [^1,3A⁻, C₇H₃O₄]⁻ in the negative ion mode. According to these fragmentation patterns, compounds 5, 13, 22, 26, 32, 34 and 36 were identified as apigenin, acacetin, luteolin-7-O-β-D-gentiobioside, luteolin-7-O-rutinoside, apigenin-7-O-β-D-glucoside, apigenin-7-O-β-D-glucuronide and acacetin-7-O-β-D-rutinoside, respectively.

3.1.3 Identification of sesquiterpene lactones in the ISHI sample. Compound 23 was tentatively presumed to be ixerin Z.^31,32 The proposed fragmentation pathway of compound 23 in the negative ion mode is illustrated in Fig. 1D. There was an [M − H]⁻ at m/z 421.1504 (1.19 ppm, C₂₁H₂₅O₉) in the negative ion mode, and fragments at m/z 259.0975 (1.93 ppm, C₁₅H₁₅O₄), m/z 241.0871 (2.49 ppm, C₁₅H₁₃O₃), m/z 215.1077 (2.33 ppm, C₁₄H₁₅O₂) and m/z 197.0970 (2.03 ppm, C₁₄H₁₃O), corresponding to [M − H − Glc]⁻, [M − H − Glc − H₂O]⁻, [M − H − Glc − CO₂]⁻ and [M − H − Glc − CO₂ − H₂O]⁻, respectively.

Based on the analysis of related literature,^24,31–33 sesquiterpene lactones in the ISHI samples were found easily lose glycosyl (Glc, 162 Da) and neutral units, such as H₂O (18 Da) and CO₂ (44 Da) under the negative ion mode. According to these fragmentation patterns, compounds 11, 12, 19, 20, 24, 25 and 27 were identified as 11,13-dihydro-13-prolyl-ixerin Z, sonchifolactone E, 3-hydroxydehydroleucodin, sonchifolactone D, 11,13-dihydroixerinoside, 11,13α-dihydroixerin Z and 11β-hydroxyleucodin-11-O-β-glucopyranoside isomer, respectively.

To our knowledge, this is the first report on the occurrence of sonchifolactone E (compound 12), 3-hydroxydehydroleucodin (compound 19) and 11,13-dihydroixerinoside (compound 24) in ISHI. Briefly, sonchifolactone E contained an [M − H]⁻ at m/z 277.1085 (3.25 ppm, C₁₅H₁₇O₅) in the negative ion mode, and fragments at m/z 259.0975 (1.93 ppm, C₁₅H₁₅O₄), m/z 249.1131 (2.49 ppm, C₁₄H₁₇O₄) and m/z 233.1183 (2.15 ppm, C₁₄H₁₇O₃) corresponding to [M–H–H₂O]⁻, [M − H − CO]⁻ and [M − H − CO₂]⁻, respectively; 3-hydroxydehydroleucodin contained an [M + HCOO]⁻ at m/z 305.1029 (1.31 ppm, C₁₆H₁₇O₆) in the negative ion mode, and fragments at m/z 215.1077 (2.33 ppm, C₁₄H₁₅O₂) and m/z 197.0970 (2.03 ppm, C₁₄H₁₃O) corresponding to [M − H − CO₂]⁻ and [M − H − CO₂ − H₂O]⁻, respectively; 11,13-dihydroixerinoside contained an [M − H]⁻ at m/z 425.1822 (2.35 ppm, C₂₁H₂₉O₉) in the negative ion mode, and fragments at m/z 263.1501 (2.28 ppm, C₁₂H₂₃O₆), m/z 245.1396 (2.86 ppm, C₁₂H₂₁O₅) and m/z 201.1499 (3.98 ppm, C₁₁H₂₁O₃) corresponding to [M − H − Glc]⁻, [M − H − Glc − H₂O]⁻ and [M − H–Glc − H₂O − CO₂]⁻, respectively.

3.1.4 Identification of nucleosides in the ISHI sample. Compounds 3 and 4 were unambiguously attributed to uridine and guanosine by comparison with the reference compounds. Guanosine, a typical nucleoside in ISHI samples, was used to characterize the fragmentation pathway (Fig. 1E). An [M − H]⁻ at m/z 282.0836 (−0.71 ppm, C₁₀H₁₂N₅O₅) was found in the negative ion mode, which could successively lose one ribose (rib, 132 Da) to form fragment at m/z 150.0412 (−2.77 ppm, C₅H₄N₅O), then further lose one neutral unit NH₃ (17 Da) to form fragment at m/z 133.0148 (−1.50 ppm, C₅HN₄O). Uridine exhibited an [M − H]⁻ at m/z 243.0617 (0.00 ppm, C₉H₁₁N₂O₆) in the negative ion mode, and fragments at m/z 200.0564 (2.50 ppm, C₈H₁₀NO₅) and m/z 111.0199 (3.60 ppm, C₄H₃N₂O₂), corresponding to [M − H − NHCO]⁻ and [M − H − rib]⁻, respectively.

3.1.5 Identification of phenols in the ISHI sample. Compound 7 contained an [M − H]⁻ at m/z 181.0867 (1.10 ppm, C₁₀H₁₃O₃) in the negative ion mode, and fragments at m/z 163.0763 (2.45 ppm, C₁₀H₁₁O₂) and m/z 135.0815 (3.70 ppm, C₉H₁₁O) corresponding to [M − H − H₂O]⁻, [M − H − H₂O − CO]⁻, respectively. Accordingly, compound 7 was tentatively presumed to be 3-(4′-hydroxy-3′-methoxyphenyl)-1-propanol.³⁴ Compound 9 contained an [M − H]⁻ at m/z 151.0399 (2.65 ppm, C₈H₇O₃) in the negative ion mode, and fragment at m/z 123.0448 (1.63 ppm, C₇H₇O₂) corresponding to [M − H − CO]⁻. Accordingly, compound 9 was tentatively presumed to be vanillin.³⁴ To the best of our knowledge, 3-(4′-hydroxy-3′-methoxyphenyl)-1-propanol (compound 7) and vanillin (compound 9) were reported in ISHI for the first time.

According to above studies, it was concluded that chemical constituents in the ISHI samples mainly consist of organic acids, flavonoids, sesquiterpene lactones, nucleosides and phenols. This is in agreement with previous studies.^15,22,23,32 The information on the structures of chemical constituents in ISHI will help in developing research strategies for bioactivity and quality control studies.

3.2 UV spectroscopic/fusion fingerprint analyses

3.2.1 Methodology validation of fingerprint analysis. Sample S1 as described in Section 2.2 was used to perform the following methodology validation experiments. Sample stability was evaluated by analyzing a single sample solution stored at room temperature for 0, 2, 4, 8, 16, or 24 h. Method repeatability was determined by analyzing six individual sample solutions under the same experimental procedure. Instrument intra-day and inter-day precision were validated by nine replicate injections of the same sample solution three times a day over three consecutive days.

In the UV fingerprint analysis, unseparated chromatograms (namely single chromatogram peak within 1 minute. Figures were not shown here) at 326 nm and UV spectra of samples in the region 200–400 nm were recorded at the same time. The analytical principle using FIA is shown in Fig. 3A, and the typical 3D chromatogram plot of one ISHI sample (S1) is shown in Fig. 3B. The stability, repeatability and precision were estimated by the retention time and peak area of S1 with the unseparated chromatogram at 326 nm. The obtained results showed that, for stability, the relative standard deviations (RSD) were less than 1.1% and 2.1%, respectively; for repeatability, the obtained values did not exceed 0.9% and 1.8%, respectively; for precision, the obtained values were not exceed 0.5% and 1.4%, respectively.

Fig. 3 FIA analytical principle plot for UV spectra (A), 3D chromatogram plot (min × nm × absorbance) for one ISHI sample (B), typical UV spectra for 28 ISHI samples (C) and PCA score plot for 28 ISHI samples (D).

In the fusion fingerprint analysis, the macro qualitative similarity (S_m, as shown in eqn (1)) of the fusion fingerprints was used to estimate the stability, repeatability and precision. The obtained results showed that, for stability, the RSD was less than 0.9%; for repeatability, the obtained value did not exceed 0.7%; and for precision, the RSD was less than 0.3%. These results demonstrated that the developed UV spectroscopic and fusion methods met the fingerprint analysis requirements for the ISHI samples.

3.2.2 Analysis of UV spectra. Under the methods described in Sections 2.2 and 2.3.3, UV spectra of the 28 samples in the region 200–400 nm were recorded, as shown in Fig. 3C. Three strong absorption bands at around 236 nm, 286 nm and 326 nm were observed in both Fig. 3B and C.

Based on the maximum UV absorption (Table 2) and the structural analysis (Fig. 2) obtained by UHPLC/Q-TOF-MS, the main UV absorption bands of the ISHI samples were preliminarily assigned as follows. Organic acids in ISHI usually have a strong absorption in the region 220–250 nm, accompanied with a shoulder peak (sh) in the region 280–300 nm, as described in Table 2. Therefore, it is supposed that the absorption band at around 236 nm (Fig. 3B and C) are mainly due to π–π* electronic transitions of aromatic rings in the organic acids (Fig. 2); the absorption band at around 286 nm (Fig. 3B and C) might be attributed to n–π* electronic transitions of conjugated systems involving aromatic rings and C=O or C=C and C=O in the organic acids (Fig. 2). In addition, organic acids and flavonoids in ISHI showed strong absorption in the region 320–330 nm and 320–350 nm (Table 2), respectively. Hence, the absorption band at around 326 nm (Fig. 3B and C) are presumably due to n–π* electronic transitions of conjugated systems involving aromatic rings and C=O in the organic acids and flavonoids (Fig. 2).

Principal component analysis (PCA), a well known chemometrics method,^3,35,36 was carried out to evaluate the discriminating ability of the three main UV absorption bands (236 nm, 286 nm and 326 nm in Fig. 3B and C). It constructed a two-dimensional matrices (28 × 3), where 28 and 3 representing the sample number and wavelength type, respectively. A two-component PCA model was performed with a total variance of 99.1% explained (PC1 = 88.4% and PC2 = 10.7%). In the PCA score plot (Fig. 3D), 28 samples except two outliers (S21 and S23) were clearly divided into two clusters marked as group 1 and 2, respectively. Group 1 (S1–S20 and S22, from manufacturer A) was obviously different from group 2 (S24–S28, from manufacturer B). Accordingly, it could conclude that the products from the same manufacturer, qualitatively, had relatively good consistency.

3.2.3 Sample quality evaluation based on the UV fingerprints. The UV RFP was constructed by taking the average of the 28 corresponding spectra. As shown in Fig. 3B, the UV spectra fingerprints of the 28 ISHI samples were very similar. Thus, SQFM was adopted to describe the subtle spectral differences. S_m, P_m and α values were computed by importing all the UV spectral data into the in-house software mentioned above, and the obtained results are described in Table 3.

Table 3 The evaluation results for 28 ISHI samples based on UV fingerprints and fusion fingerprints by SQFM

Sample	UV				Fusion
	Sm	Pm (%)	α	Grade	Sm	Pm (%)	α	Grade
S1	0.991	94.6	0.007	2	0.950	98.1	0.017	1
S2	0.995	96.1	0.003	1	0.953	94.2	0.025	2
S3	0.997	93.9	0.003	2	0.962	98.1	0.041	2
S4	0.999	93.4	0.001	2	0.965	96.1	0.026	1
S5	0.999	98.0	0.003	1	0.961	92.3	0.027	2
S6	1.000	95.5	0.001	1	0.967	96.4	0.049	2
S7	0.993	102.0	0.003	1	0.979	98.5	0.006	1
S8	0.953	106.8	0.010	2	0.973	110.9	0.013	3
S9	0.992	91.5	0.003	2	0.968	93.0	0.021	2
S10	0.996	107.9	0.007	2	0.956	100.2	0.027	1
S11	0.961	96.7	0.003	1	0.953	99.1	0.002	1
S12	0.988	95.7	0.006	1	0.975	96.5	0.007	1
S13	0.995	102.7	0.000	1	0.978	99.3	0.001	1
S14	0.998	111.8	0.003	3	0.955	98.2	0.002	1
S15	0.997	98.2	0.002	1	0.973	103.6	0.006	1
S16	0.978	110.0	0.003	3	0.962	115.3	0.025	3
S17	0.997	107.7	0.000	2	0.974	106.3	0.015	2
S18	0.997	103.1	0.004	1	0.974	106.8	0.028	2
S19	0.999	86.5	0.000	3	0.966	112.3	0.020	3
S20	0.997	98.7	0.000	1	0.968	107.1	0.002	2
S21	0.990	78.4	0.026	4	0.952	78.8	0.033	4
S22	0.994	112.9	0.004	3	0.964	112.3	0.005	3
S23	0.983	84.2	0.024	4	0.975	78.1	0.029	4
S24	0.992	101.0	0.003	1	0.961	101.3	0.020	1
S25	0.988	104.8	0.002	1	0.970	104.8	0.003	1
S26	0.992	103.2	0.003	1	0.975	102.9	0.028	1
S27	0.986	98.9	0.001	1	0.970	96.5	0.017	1
S28	0.988	102.9	0.000	1	1.000	100.0	0.000	1

---

For the UV fingerprints, the S_m and α values of the 28 ISHI samples were above 0.961 and below 0.026, respectively, indicating that all samples were similar to RFP. Based on qualitative parameters S_m and α, the quality grades of the 28 samples should be in grade 1. However, only 15 samples (S2, S5, S6, S7, S11, S12, S13, S15, S18, S20, S24, S25, S26, S27 and S28) met the grade, the remaining ones were in range grade 2–4 in combination with the quantitative similarity parameter P_m. For example, 7 samples (S1, S3, S4, S8, S9, S10 and S17) were judged as grade 2 with P_m in the range 91.5–107.9%; 4 samples (S14, S16, S19 and S22) were judged as grade 3 with P_m in the range 86.5–112.9%; 2 samples (S21 and S23) were judged as grade 4 with P_m 78.4% and 84.2%, respectively. These results indicated that although qualitative evaluation (S_m and α) was important, quantitative assessment (P_m) should not be ignored, because P_m as a quantitative parameter, describing the overall ingredient content in the samples, thus, has a great potential to be associated with medicinal efficacy in clinical situations. Generally, samples with grade ≤5 were recommended as qualified ones. Accordingly, in this study, the qualities of the 28 samples were all judged as qualified with the UV fingerprints.

3.2.4 Sample quality evaluation based on the fusion fingerprints. Based on the structural analysis (Fig. 2) and the maximum UV absorption (Table 2) obtained by UHPLC/Q-TOF-MS, it is noteworthy that the ISHI samples displayed strong UV absorption bands at around 236 nm, 286 nm and 326 nm (Fig. 3B and C), which corresponding organic acids and flavonoids in ISHI (as described in Section 3.2.2). Moreover, organic acids and flavonoids have been regarded as the major active substances in ISHI.^21,27 Accordingly, fusion fingerprints of the three wavelengths (236 nm, 286 nm and 326 nm) could realize the overall evaluation of the ISHI samples by importing the HPLC signal data into the in-house software mentioned above, and the obtained SQFM results are described in Table 3.

For the fusion fingerprints, the S_m and α values of the 28 ISHI samples were above 0.950 and below 0.049, respectively, demonstrating that all samples had similar chemical compositions. The quality grades of the 28 samples should be judged as grade 1 based on the qualitative parameters S_m and α. However, only 14 samples (S1, S4, S7, S10, S11, S12, S13, S14, S15, S24, S25, S26, S27 and S28) were judged as grade 1, the remaining ones were in range grade 2–4 in combination with the quantitative similarity parameter P_m. For example, 8 samples (S2, S3, S5, S6, S9, S17, S18 and S20) were judged as grade 2 with P_m in the range 92.3–107.1%; 4 samples (S8, S16, S19 and S22) were judged as grade 3 with P_m in the range 110.9–115.3%; 2 samples (S21 and S23) were judged as grade 4 with P_m 78.8% and 78.1%, respectively. Based on the criteria described in Table 1 and Section 3.2.3, the qualities of all the 28 samples were judged as qualified with the fusion fingerprints.

Compared with the UV fingerprint method, the fusion fingerprints exhibited almost the same results. For example, all of the 28 samples were judged as qualified in range grade 1–4; S24–S28 were judged as grade 1 in the method of UV fingerprints as well as fusion fingerprints; S21 and S23 were judged as grade 4 in both UV and fusion fingerprints method. However, acceptable slight differences were found between the two methods. For example, S1, S4 and S10 were judged as grade 1 in the fusion fingerprints but grade 2 in the UV fingerprints; S2, S5, S6, S18 and S20 were grade 2 in the fusion fingerprints but grade 1 in the UV fingerprints; S8 was judged as grade 3 in the fusion fingerprints but grade 2 in the UV fingerprints. All in all, the fusion of the three wavelengths (236 nm, 286 nm and 326 nm) HPLC fingerprints combined with UV fingerprints provide a reliable and feasible means to control the quality consistency of the ISHI samples, which was basically consistent with our previous study using microemulsion electrokinetic chromatography method.⁹

3.3 Relationship between HPLC fingerprints and antioxidant activities in vitro

According to the reported literature, flavonoids and organic acids in the ISHI samples possess excellent antioxidant abilities.^21,27 From the chemical components in Table 2, it appeared that the ISHI samples have excellent antioxidant abilities due to the presence of a quantity of flavonoids and organic acids. Furthermore, both flavonoids and organic acids in ISHI showed strong UV absorbance at 326 nm (as described in Table 2, Fig. 3B and C and Section 3.2.2). These information encourage us to assay the fingerprint-efficacy relationship between HPLC fingerprints at 326 nm and the antioxidant activities.

The HPLC fingerprints and antioxidant activities in vitro were measured under the conditions described in Section 2.3.2 and 2.3.4, respectively. Typical HPLC chromatograms at 326 nm with 40 co-possessing fingerprints for the 28 ISHI samples are shown in Fig. 4A. Correlation analysis between the HPLC fingerprints (as the descriptor matrix X) and the 1/IC₅₀ values (as the response matrix Y) was performed by the OPLS method.

Fig. 4 Typical HPLC chromatograms for 28 ISHI samples at 326 nm (A), OPLS model of standardized regression coefficient plot (B), regression coefficients plot of the calibration model (C) and validation model (D).

After excluding the two outliers (S21 and S23) based on the t[1]–t[2] score plot, the remaining samples were randomly divided into two sets: one is the a training set to establish the OPLS model; another is the test set to validate the model (Table 4). As shown by the regression coefficients plot of the calibration model (Fig. 4B), 28 peaks (1, 2, 3, 4, 5, 7, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 22, 23, 24, 26, 28, 29, 33, 34, 35, 38, 39 and 40) out of 40 fingerprints in the HPLC chromatogram were positively correlated, while the remaining peaks were negatively correlated with 1/IC₅₀, indicating that the majority of the chemical components in ISHI possess strong antioxidant abilities. Furthermore, the peak 16, 17, 18, 22, 23, 28 and 35 in HPLC fingerprints were identified as 5-caffeoylquinic, caffeic acid, ferulic acid, luteolin-7-O-β-D-glucoside, luteolin-7-O-β-D-glucuronide, chicoric acid and luteolin, respectively, by comparing the retention time and on-line UV spectra with reference standards. The established calibration model (Fig. 4C) achieved an explained variance (R²) of 99.62%, a predictive ability (Q²) of 91.03%, a root mean square error of estimation and a cross-validation procedure value of 0.0012 and 0.0045, respectively, indicating that the obtained model was optimal.

Table 4 Overview of the experimental and predicted values for OPLS model

Obs ID (primary)	Ypred (1/IC50)	Yvar (1/IC50)	RE^c (%)
a Used for the calibration model.b Used for the validation model.c RE: relative error.
S1^a	0.202	0.202	0.42
S2^a	0.197	0.197	0.13
S3^a	0.212	0.213	0.32
S4^a	0.221	0.221	−0.22
S6^a	0.224	0.224	0.18
S7^a	0.229	0.230	0.44
S8^a	0.245	0.245	−0.15
S9^a	0.226	0.225	−0.48
S12^a	0.227	0.227	−0.20
S13^a	0.230	0.230	0.00
S14^a	0.219	0.219	−0.01
S15^a	0.239	0.237	−0.63
S16^a	0.248	0.251	1.14
S19^a	0.238	0.238	−0.07
S20^a	0.216	0.216	−0.30
S22^a	0.229	0.228	−0.35
S24^a	0.256	0.256	0.08
S25^a	0.244	0.243	−0.26
S5^b	0.221	0.216	−2.61
S10^b	0.229	0.221	−3.24
S11^b	0.230	0.230	0.06
S17^b	0.241	0.238	−1.41
S18^b	0.241	0.239	−0.85
S26^b	0.245	0.243	−0.76
S27^b	0.241	0.238	−1.41
S28^b	0.243	0.243	0.04

---

The antioxidant activity prediction was performed on the established validation model (Fig. 4D). As a satisfactory result, an explained variance (R²) of 95.15% and a root mean square error of prediction value of 0.0039 were obtained, indicating that the established OPLS model possessed a well predictive ability. From Table 4, no significant difference was observed between the measured and predicted 1/IC₅₀ values of all the samples.

4. Conclusions

In the present study, UHPLC/Q-TOF-MS was performed to investigate the chemical profiling of the ISHI sample. As a result, totally 38 compounds, including 16 organic acids, 10 flavonoids, 8 sesquiterpene lactones, 2 nucleosides and 2 phenols, were identified or tentatively characterized, providing important information on chemical substances for further bioactivity and quality control studies. It should be noted that 7 compounds, including 2 organic acids (vanillic acid and (E)-2,5-dihydroxy cinnamic acid), 3 sesquiterpene lactones (sonchifolactone E, 3-hydroxydehydroleucodin and 11,13-dihydroixerinoside) and 2 phenols (3-(4′-hydroxy-3′-methoxyphenyl)-1-propanol and vanillin) were identified in ISHI for the first time. Based on the structural analysis and the maximum UV absorption obtained by UHPLC/Q-TOF-MS, the three main UV absorption bands (236 nm, 286 nm and 326 nm) of the ISHI samples were preliminarily assigned from organic acids and flavonoids, which have been regarded as the major active substances in ISHI. Accordingly, UV spectroscopic fingerprints and multi-wavelength (236 nm, 286 nm and 326 nm) fusion fingerprints realized the overall evaluation of the ISHI samples. In fingerprint assessments, SQFM was established for scientific quality evaluation from qualitative and quantitative perspectives, by which the quality grades of the 28 samples from the two fingerprints methods exhibited almost the same results, and basically consistent with our previous study using microemulsion electrokinetic chromatography method, indicating that the fusion HPLC fingerprints combined with UV fingerprints provide a reliable and feasible means to control the quality consistency of the ISHI samples. Moreover, PCA has been applied to evaluate the discriminating ability of the three main UV absorption bands (236 nm, 286 nm and 326 nm), and results demonstrating that samples from the same manufacturer, qualitatively, had relatively good consistency. In addition, the fingerprint-efficacy relationship between HPLC fingerprints (at 326 nm: both flavonoids and organic acids in ISHI showed strong UV absorbance at 326 nm based on UHPLC/Q-TOF-MS) and antioxidant activities was established utilizing OPLS, providing important medicinal efficacy information for quality control. This study offered a scientific, rapid and effective analytical strategy for TCM/HM quality control, which would play an important role in ISHI factory production and clinical applications.

Acknowledgements

This work was supported by National Natural Science Foundations of China (81403094 and 81560695) and the general education department of Scientific Research Project of Liaoning Province (L2013393).

References

1. J. X. Wang, T. Hou, L. Wei, L. Y. Shi, J. He, N. Zhou, G. W. Sun, X. L. Zhang and X. M. Liang, RSC Adv., 2015, 5, 25768–25776.
2. S. T. Wang, W. Gao, Y. X. Fan, X. G. Liu, K. Liu, Y. Du, L. L. Wang, H. J. Li, P. Li and H. Yang, RSC Adv., 2016, 6, 27320–27328.
3. L. P. Yang, G. X. Sun, Y. Guo, Z. F. Hou and S. Chen, PLoS One, 2016, 11, e0148878.
4. G. X. Sun and J. X. Zhang, Chin. J. Chromatogr., 2009, 27, 318–322.
5. D. Yao, X. S. Meng, S. Wang, Y. R. Bao, Y. Pan and L. Han, China J. Chin. Mater. Med., 2013, 38, 1513–1517.
6. P. P. Ren, G. X. Sun and L. N. Sun, Chin. J. Pharm. Anal., 2009, 29, 411–415.
7. H. B. Zhu, C. Y. Wang, Y. Qi, F. R. Song, Z. Q. Liu and S. Y. Liu, Talanta, 2013, 103, 56–65.
8. D. Custers, M. Canfyn, P. Courselle, J. O. De Beer, S. Apers and E. Deconinck, Talanta, 2014, 123, 78–88.
9. L. P. Yang, X. M. Xie, J. Zhang and G. X. Sun, PLoS One, 2016, 11, e0157601.
10. X. Dai, H. Song, W. Liu, S. Yao and G. Wang, RSC Adv., 2016, 6, 10078–10085.
11. E. M. Tan, S. Amirjalayer, S. Smolarek, A. Vdovin, A. M. Rijs and W. J. Buma, J. Phys. Chem. B, 2013, 117, 4798–4805.
12. Z. L. Fang and S. H. Fang, Chin. J. Anal. Lab., 1991, 4.
13. G. X. Sun, X. W. Dou and L. F. Li, Cent. South Pharm., 2012, 10, 543–548.
14. Y. B. Li, Y. Wang, B. Yang, Y. M. Wang, Z. G. Hou, A. Z. Li, Y. Y. Xu, L. Ju, H. Y. Wu and Y. J. Zhang, RSC Adv., 2015, 5, 2209–2216.
15. L. Yuan, Z. Z. Zhang, Z. G. Hou, B. Yang, A. Z. Li, X. J. Guo, Y. M. Wang and Y. Li, Anal. Methods, 2015, 7, 5120–5127.
16. L. F. Lin, H. M. Lin, M. Zhang, X. X. Dong, X. B. Yin, C. H. Qu and J. Ni, RSC Adv., 2015, 5, 107623–107636.
17. H. J. Huang, H. Y. Chen, Y. S. Chang and C. Y. C. Chen, RSC Adv., 2015, 5, 6625–6635.
18. T. Prevc, N. Šegatin, N. P. Ulrih and C. Blaž, Talanta, 2013, 109, 13–19.
19. Y. T. Zhang, Q. Li, H. Xing, X. F. Lu, L. S. Zhao, K. K. Qu and K. S. Bi, Food Res. Int., 2013, 53, 847–856.
20. O. P. Sharma and T. K. Bhat, Food Chem., 2009, 113, 1202–1205.
21. Y. C. Zhang, F. F. Gan, S. B. Shelar, K. Y. Ng and E. H. Chew, Food Chem. Toxicol., 2013, 59, 272–280.
22. Y. C. Zhang, C. N. He and E. H. Chew, Chem. Biodiversity, 2013, 10, 1373–1391.
23. Y. Liu, J. Q. Lu, J. Y. Zhang, F. Wang, Q. Wang and Y. J. Qiao, China J. Chin. Mat. Med., 2013, 38, 2675–2681.
24. P. Y. Shi, Y. F. Zhang, H. B. Qu and X. H. Fan, Phytochem. Anal., 2011, 22, 66–73.
25. F. Wang, J. Y. Zhang, Q. Wang, Y. Liu, Z. J. Wang and J. Q. Lu, Cent. South Pharm., 2013, 11, 561–565.
26. F. Wang, J. Zhang, P. Yin, Z. Wang, L. Dong and J. Lu, Anal. Methods, 2014, 6, 3515.
27. C. G. Chen, H. L. Jia, S. X. Lv and C. Q. Xu, Chin. J. Clin. Pharmacol., 2012, 28, 196–199.
28. J. Trygg and S. Wold, J. Chemom., 2002, 16, 119–128.
29. P. Bhandari, N. Kumar, B. Singh and P. S. Ahuja, Indian J. Exp. Biol., 2010, 48, 323–328.
30. Y. Liu, J. Q. Lu, J. Y. Zhang, Q. Wang, F. Wang, Y. J. Qiao and Y. L. Zhang, Anal. Methods, 2012, 4, 4230.
31. L. Y. Dong, Y. Liu, J. Y. Zhang, W. Cai, R. R. Liu and J. Q. Lu, World Science and Technology/Modernization of Traditional Chinese Medicine and Materia Medica, 2014, 16, 2671–2675.
32. W. Cai, J. Y. Zhang, L. Y. Dong, P. H. Yin, C. G. Wang, J. Q. Lu and H. G. Zhang, J. Pharm. Biomed. Anal., 2015, 107, 290–297.
33. Y. C. Zhang, L. Zhou and K. Y. Ng, J. Asian Nat. Prod. Res., 2009, 11, 294–298.
34. N. Zhang, A. L. Lu, D. Wang, G. Chen, Q. Dang and Y.-H. Pei, J. Shenyang Pharm. Univ., 2007, 24, 549–551.
35. D. D. Wang, J. Liang, W. Z. Yang, J.-J. Hou, M. Yang, J. Da, Y. Wang, B.-H. Jiang, X. Liu, W.-Y. Wu and D.-A. Guo, J. Pharm. Biomed. Anal., 2014, 89, 130–141.
36. A. L. Pomerantsev and O. Y. Rodionova, J. Chemom., 2014, 28, 429–438.

---