In situ enrichment and determination of 6 kinds of caine-type anesthetics in cosmetics and rat serum by thin layer chromatography-Raman spectroscopy

2.1 Materials

Tetracaine hydrochloride (CAS 136–47-0), procaine hydrochloride (CAS 51–05-8), ropivacaine hydrochloride (CAS 132112–35-7), lidocaine hydrochloride (CAS 6108–05-0), mepivacaine hydrochloride (CAS 1722–62-9) and dibucaine (CAS 85–79-0) were purchased from Shanghai Jingchun Biochemical Technology Co., Ltd. TLC silica gel 60F₂₅₄ (SKU 1055540001) was purchased from Millipore (Canada) Ltd. Cyclohexane and methanol were supplied by Fuchen (Tianjin) Chemical Reagent Co., Ltd. Triethylamine was obtained from Tianjin Kemio Chemical Reagent Co., Ltd. Cosmetics, namely INTRAL Redness Relief Soothing Serum (Darphin, 5 mL, REF.D320-43, UK, abbreviated as essence) was purchased from Darphin Official Flagship Store (Taobao (China) Software Co., Ltd). Mild Exfoliating Moisturizing Gel (Doctor Li, 100 g) was purchased from local cosmetics store. Cemoy Revival cream (10 mL) were obtained from Internet (Nicomama App, Hangzhou Zhicong Network Technology Co., Ltd). The above reagents were stored in refrigerator at 4℃. All these chemicals were analytical-grade reagents and used directly without further purification. The purified water purchased from Wahaha Group Co., Ltd was used throughout the experiment.

2.2 Preparation of anesthetics standard solutions

The standard solution of each anesthetic and their mixed standard solution were prepared with methanol. All solutions were stored at 4℃ for further use.

2.3 Preparation of rat serum samples

15 SD male rats were selected and raised in the common experimental environment. All rats were allowed to drink water through free access, and the rats were placed in humid cages for 2 weeks. On the last day, all rats should be fasting for 12 h before orbital blood collection. Next, blood was dripped into the 10 mL EP tube and stewed for 30 min at room temperature, then centrifuged at 4 °C, 3000 rpm for 5 min. The supernatant was stored at −80 °C.

Before use, the rat serum samples were thawed in an ice-water bath. 1 mL rat serum was added with 3 mL methanol, then vortexed for 1 min, ultrasonically dispersed for 5 min and finally centrifuged at 12000 rpm for 10 min. The supernatant was directly used for spotting as negative sample. The simulated positive serum samples were prepared as follows, negative rat serum was spiked with the mixed standard solution and treated as above, making the final concentrations were 2 mg mL⁻¹ for Pro, Tet, Dib and 8 mg mL⁻¹ for Mep, Lid, Rop, respectively.

2.4 Preparation of cosmetics samples

1.0 g of essence (gel) was accurately weighed and placed in 15 mL centrifuge tube, vortexed for 30 s with 2 mL methanol, diluted to 10 mL with methanol, and ultrasonically extracted for 20 min, then centrifuged at 5000 rpm for 5 min. 1.0 g of cream was accurately weighed and placed in 15 mL centrifuge tube, vortexed for 1 min with 8 mL 1% trichloroacetic acid and 2 mL methanol, centrifuged at 5000 rpm for 5 min. The supernatant and subnatant were directly used as negative samples for spotting in our study. The positive samples were prepared by adding mixed standard solution to cosmetics in the centrifuge tube and then pre-treated following the above steps. The final concentrations added in the centrifuge tube were 0.5 mg mL⁻¹ for Pro, Tet and Dib, and 2 mg mL⁻¹ for Mep, Lid and Rop.

2.5 TLC separation and in situ enrichment of anesthetics

The test solutions were pipetted with 2 uL quantitative capillary and spotted on the silica gel thin layer plate. Then the spotted thin layer plate was pre-saturated for 15 min with 5 mL cyclohexane-triethylamine (v:v = 7:3) solution as the developing solvent in the chromatographic tank. After the developing solvent was fully expanded, the spots were marked under 254 nm UV light irradiation and the R_f values were then calculated. Finally, the spots were enriched in situ with methanol and measured directly by Raman spectrometer with 532 nm excitation wavelength on TLC plates. The sampling volumes were 2 μL for standarad and rat serum samples, 4 μL for cosmetics samples.

2.6 Instruments and measurement

Raman spectra were measured with the Thermo Scientific DXR3xi Raman Imaging Microscope equipped with an excitation laser wavelength of 532 nm (laser power: 10.0 mW) at room temperature. All Raman spectra were recorded by a 50 × microscope objective. The laser power, exposure time and scanning times of target anesthetics were different and listed in Table S1. The pinhole was 25 μm. The obtained Raman spectra were baseline corrected by NGS LabSpec 5, and plotted with Origin 8.5. Other instruments included BYLAB three by ultraviolet analyzer (Beijing Bingyang Technology Co., Ltd, 254 and 365 nm), CNC ultrasonic cleaner (KQ-500DE, Kunshan Ultrasonic Instrument Co., Ltd), High speed centrifuge (TG16, Shanghai Huxiangyi Centrifuge Instrument Co., Ltd), vortex mixer (Qun'an Instrument Co., Ltd), 2 μL quantitative spotting capillary (Tianjin Enlida Technology Co., Ltd) and chromatographic tank (Shanghai Huake Laboratory Equipment Co., Ltd). The HPLC system was Aglient 1260 (Palo, Alto, CA, USA) equipped with a sample injection pumb, an automatic sampler, an ultraviolet (UV) detector and a thermostat. Detailed HPLC conditions refer to the Supporting Information.

3.1 Selection of developing solvent

Different developing systems were explored for separation of six kinds of caine-type anesthetics, including chloroform–methanol, chloroform–methanol-cyclohexane, cyclohexane-methanol, cyclohexane-methanol-aqueous ammonia, chloroform-acetone-methanol–water, ethyl acetate–methanol-aqueous ammonia, ethyl acetate–methanol-diethylamine, cyclohexane-diethylamine and cyclohexane-triethylamine systems. The six analytes can not be mutually separated except using cyclohexane-triethylamine system. Optimization on the proportion of developing solvent (5:5, 6:4, 13:7, 7:3, 3:1 and 4:1) was conducted (Fig. S1). The separation effect and spot shapes of cyclohexane-triethylamine system (7:3) were better, thus finally it was used as the developing agent in this experiment. In Fig. 1, the six anesthetics were separated effectively with R_f values of 0.09 (Pro), 017 (Tet), 0.41 (Dib), 0.50 (Mep), 0.62 (Lid) and 0.73 (Rop), respectively. The R_f values of the analytes were recorded and the mixed standard solution was spotted on every TLC plate, so that the TLC spots could be traced when they were invisible under UV irradiation.

Close

Fig. 1

TLC plates under 254 nm UV light irradiation for separation of six anesthetics with R_f values of 0.09 (Pro), 017 (Tet), 0.41 (Dib), 0.50 (Mep), 0.62 (Lid) and 0.73 (Rop), respectively. Developing system: cyclohexane-triethylamine system (7:3). The concentrations were 2 mg mL⁻¹ for Pro, Tet, Dib, and 5 mg mL⁻¹ for Mep, Lid, Rop, respectively.

Export to PPT

3.2 In situ enrichment and Raman spectra of spots

After TLC separation, the Raman spectra of spots were directly obtained. However, the sensitivity was still largely limited to the visibility of TLC spots. In order to improve the detection sensitivity, an in situ enrichment procedure was applied (Fig. 2). The spots could be manually concentrated to much smaller points through the promotion of methanol with spotting capillary from different directions. Even the originally invisible spots became visible and detectable. As expected, the Raman responses were improved dramatically, especially for Tet, Mep and Rop (Fig. 3), which benefited from in situ enrichment. Thus, the Raman spectra of six anesthetics were all collected on concentrated sample spots, which were basically consistent with intrinsic Raman signals of their solids. Furthermore, the Raman spectra were distinguishable which was attributed to the specificity of Raman spectroscopy. The strong peaks at 1609 cm⁻¹ for Pro and 1611 cm⁻¹ for Tet were assigned to ring C = C stretching vibration. The bands at 1701 cm⁻¹ (Pro) and 1692 cm⁻¹ (Tet) corresponded to the C = O stretching vibration. The peaks observed at 1275 cm⁻¹ (Pro) and 1279 cm⁻¹ (Tet) were from C-O stretching[Sherlin et al., 2018]. For Lid, the very strong one observed at 1265 cm⁻¹ corresponded mainly to CH bending and C–C stretching. At 1597 cm⁻¹, the peak was assigned to HNC scissoring and NC stretching modes. At 2930 cm⁻¹, there was a very intense line in the Raman spectrum, which was due to the methylene CH₂ sretching in the two ethyl groups[Badawi et al., 2015]. For Mep and Rop, there was more abundant fingerprint information compared with the others.

Close

Fig. 2

Sketch of in situ enrichment and Raman measurement of spots on TLC plates.

Export to PPT

Close

Fig. 3

Raman spectra of anesthetics solid (a), before (b) and after (c) in situ enrichment. The concentrations were 2 mg mL⁻¹ for Pro (A), Tet (B), Dib (C), and 5 mg mL⁻¹ for Mep (D), Lid (E),Rop (F), respectively.

Export to PPT

3.3 Concentration dependent Raman spectra of anesthetics

A series of anesthetic standard solutions in different concentrations were spotted on a TLC plate and the spots were in situ enriched under 254 nm UV light irradiation following by Raman measurements under 532 nm excitation wavelength. The TLC plates of anesthetics before and after in situ enrichment were shown in Fig. S2. Representative concentration dependent Raman spectra from enriched spots of anesthetics and the plot of Raman intensity versus spotting amounts were illustrated in Fig. 4. The Raman spectra were first baselined then analyzed by characteristic peak’s intensity of each anesthetic, 1609 (Pro), 1611 (Tet), 1381 (Dib), 1267 (Mep), 1265 (Lid) and 1267 cm⁻¹ (Rop), respectively. As can be seen from Fig. 4, the intensities of characteristic bands decreased gradually with the anesthetics concentrations coming down. The variation tendencies of Raman intensities with the sampling amounts of anesthetics were investigated. The error bars were made with at least five sets of data which represented the precision. Finally, after linear fitting by Origin 8.5, the linear equations, the correponding linear range and the minimum detection amounts for Pro, Tet, Dib, Mep, Lid and Rop were listed in Table 1.

Close

Fig. 4

Representative concentration-dependent Raman spectra of the anesthetics. The Raman intensities at 1609 (Pro), 1611 (Tet), 1381 (Dib), 1267(Mep), 1265(Lid) and 1267 (Rop) cm⁻¹ versus the target amounts sampled on TLC plates. Each error bar indicated the standard deviation of the Raman intensity. The error bars were made with at least five sets of data.

Export to PPT

In Fig. S2, the minimum amounts visible to the naked eyes were about 1.2, 2, 1.2, 2, 8 and 8 μg for Pro, Tet, Dib, Mep, Lid and Rop by TLC. Combining in situ enrichment TLC with Raman scattering technique, the minimum detection amounts were optimized to 0.4, 0.2, 1.2, 1, 1 and 1 μg, respectively. Additionally, the accuracy of identification can be further guaranteed by uniqueness of molecular vibration spectrum. Thus, our method not only utilized the TLC predominance of convenient sepration for mixtures, but also exhibited the superiority in fingerpringting of Raman spectroscopy.

3.4 Determination of cosmetics samples

Cosmetics samples were tested by means of standard addition method for exploring application fields of the proposed protocol. Image of TLC plate and corresponding Raman responses of spiked samples (essence, gel and cream) were presented in Fig. 5. For the negative sample of gel, an interference spot with the same R_f value as Pro appeared, but no Raman signal was observed. There were no background interferences observed from both TLC image and Raman spectra of negative samples (Fig. S3). Anesthetics between spiked samples and standard solution achieved consistency on the spots positions. Then we calculated the amounts of anesthetics via Raman intensity of characteristic peaks and their linear equations. As shown in Table 2, the calculated amounts of each anesthetic in this work coincided with true addition amounts basically. The recoveries were between 91.5% and 110.9%, the RSD values were all within 11%. Accordingly, our approach based on in situ concentration TLC and Raman scattering was proved to be feasible for anesthetics determination precisely and conveniently.

Close

Fig. 5

Image of TLC plate (A) and corresponding Raman responses of negative and spiked samples with essence (B), gel (C) and cream (D). The Raman intensity of lidocaine was multiplied by 5 times. “S” means “the mixed standard solution”. “+” and “-” represent positive and negative samples.

Export to PPT

As stated above, LLE and SPE are commonly used for extraction of anesthetics in cosmetics samples, which have adverse impact on experimenters and environment due to large amounts of organinc solvents, long analysis time and high costs. In our study, the treatments and separation processes for cosmetics were comparatively simpler and more affordable. Besides, the method would have great potential for on-site qualitative screening of anesthetics if portable Raman spectrometer was used.

3.5 Determination of rat serum simulated positive samples

In the same way, we applied the approach to rat serum samples. Negative sample, rat serum simulated positive sample and mixed standard solution were spotted on the same TLC plate. As can be seen in inset of Fig. 6, there was a one-to-one correspondence for spot positions between positive sample and mixed stardand solution. There presented no visible spots for the negative sample, and no peaks were detected for negative sample. It was demonstrated that rat serum matrix did not interfere with the identification of anesthetics we studied, which facilitated the detection of targets. The Raman spectra of six anethetics from positive sample and mixed stardard solution were measured and compared in terms of peak position and Raman intensity. For each anesthetic, the Raman intensities of characteristic peaks were substituted into respective standard curve. The calculated amounts and related results were listed in Table 3, which were basically in agreement with added amounts. The recoveries were between 94.6% and 108.1%, the RSD values were all within 12%.

Close

Fig. 6

Raman spectra of Pro (A), Tet (B), Dib (C), Mep (D), Lid (E) and Rop (F) for negative rat serum sample (a), standard solution (b) and positive sample (c). The inset images were corresponding TLC plates before and after in situ enrichment. “S” means “the mixed standard solution”. “+” and “-” represent positive sample and rat serum blank sample.

Export to PPT

3.6 Accuracy and reproducibility

To verify the accuracy and reproducibility of the method, the recovery tests on simulated positive samples (rat serum and cosmetics) in different concentrations for anesthetics within the linear range were conducted, obtaining corresonding data in Table 2 and 3. The recovery was calculated through the ratio of measured value to added value. The repetitive experiments were conducted by repeated measurements for more than 5 times, then calculating RSD based on the determined amounts. The results indicated that our approach was feasible for identification and quantification in both rat serum and cosmetics with good accuracy and reproducibility. Additionally, lt can also be seen from error bars of peak intensities in different amounts in Fig. 3 that the method was of acceptable repeatability. The repeatability was also confirmed by the Raman mapping image of TLC spots from Tet standard solution in Fig. S4. Thus, the established method exhibited good accuracy and reproducibility. Notably, spotting and enriching procedures were all conducted manually, so it would inevitably introduce certain deviations. In our future research, automatic TLC sampler can be used for getting narrower sample strips to improve the RSD values. Additionally, more Raman data, even all the Raman mapping data on one spot can be obtained for quantitative analysis to further improve the RSD values at the cost of a certain efficiency.

3.7 Specificity

As shown in Figs. 5 and 6, the spots of simulated positive sample on TLC plate had the same R_f values with that of the mixed standard solution. There were no visible spots appeared on TLC plate for negative sample, nor that discriminable Raman responses. It was indicated that the anesthetics in rat serum and cosmetics were separated and identified effectively without interferrence in our study. Combined with Raman spectroscopy, the specificity was greatly improved due to the fingerprint property of Raman spectra.

3.8 HPLC verification

To further confirm the results of this work, all the cosmetics samples were also analyzed by HPLC method (Fig. S5-S8), and similar results were obtained as shown in Table S2 including HPLC experimental details. The results also proved the accuracy and feasibility of the proposed approach. Although our method was less sensitive than HPLC, the simplicity, selectivity and low costs would make it a promising and practical platform for simple analysis of anesthetics. Meanwhile, we will focus on the enhancement of sensitivity ulteriorly.