Optimization of the extraction conditions and quantification by RP-LC analysis of three alkaloids in Zanthoxylum nitidum roots

Abstract

Context: A classic traditional Chinese medicine, Zanthoxylum nitidum (Roxb.) DC. widely used in China, exhibits anticancer, anti-inflammatory and antianalgesic activities. Alkaloids are one of the main bioactive components. It is urgent to develop a simple and reliable method to determine the main alkaloids in Z. nitidum roots.

Objective: To determine the three alkaloids in Z. nitidum roots, a reversed-phase liquid chromatographic (RP-LC) method combined with an optimum extraction condition was established.

Materials and methods: A method involving four-factor-three-level orthogonal array design including the extracting solvent and the RP-LC condition was assayed. Twenty batches were collected from different areas of the Guangxi Province at different harvesting times. The determined alkaloids were nitidine chloride (NC, 1), ethoxychelerythrine (2) and liriodenine (3). The stable mobile phase was a C₁₈ packing, and the mobile phase was acetonitril-aqueous phosphoric acid-triethylamine-buffer solution.

Results: The optimum extraction and detection conditions have been determined in the process of quantification of Z. nitidum root alkaloids. The three alkaloids were detected simultaneously in the 20 batches of samples. The results clearly showed that alkaloid concentrations differed significantly among Z. nitidum collected from various collection areas.

Discussion and conclusion: We have established an optimum extraction and detection conditions in the process of quantification the three alkaloids in Z. nitidum roots. From this research, the most influenced factor on Z. nitidum roots was the collecting location, and the next factor was the harvesting time. The collecting location and the harvesting time should be considered as the high-quality medicinal herbs factors.

• Column liquid chromatography
• orthogonal array design

Introduction

Zanthoxylum nitidum (Roxb.) DC. (Rutaceae), also named “DI QIN NIU”, “MAN JIAO” and “SHUANG MIAN ZHEN”, is a classic traditional Chinese medicinal (TCM) plant distributed chiefly in Guangxi, and also widely distributed in Guangdong, Fujian, Sichuan and Zhejiang provinces of China. The root of the plant is widely used in China to treat rheumatism, toothache, neuralgia and swelling of the throat. Its main bioactive components are alkaloids, coumarins and lignanoids (Yao & Hu, 2004). Recently, one of the authors has reported the standard operating procedures (SOP) for this plant (Lai et al., 2011). Pharmacological studies show that Z. nitidum roots have marked pharmacological antibacterial (Huang et al., 2013; Shi et al., 2005), antigastric ulcer, anticancer (Liu & Liu, 2010), anti-inflammation (Feng et al., 2011; Xu et al., 2010; Zhou et al., 2012) and antioxidation activities (Pang et al., 2007; Xie, 2000), protects liver and improves liver function (Pang et al., 2006), and so on. Apart from extensively use in the pharmaceutical industry, the roots are widely applied in light industry. Therefore, extractive conditions and quantitative determination methods in the roots were urgent.

Several qualitative and/or quantitative analytic methods have been reported for alkaloids (Liang & Zhang, 2005), neoherculin (Liu et al., 2005) and lignanoids (Zhang et al., 2002), such as HPLC fingerprinting analysis (Yan et al., 2006). However, many current techniques are time-consuming or unable to provide comprehensive chemical assessment of Z. nitidum roots. In this article, we were interested in developing a reversed-phase liquid chromatographic (RP-LC) method for rapid and accurate quantification of specific alkaloids in Z. nitidum roots from different harvest of medicinal plants. Satisfactory results were obtained. The simple procedure afforded efficient separation and quantification of compounds NC (1), ethoxychelerythrine (2) and liriodenine (3) in Z. nitidum roots extract. The contents of compounds 1–3 in Z. nitidum roots have been determined using a simple, stable chromatography and brief run times for the first time.

Materials and methods

Plant material

The 20 batches of Z. nitidum roots were collected from different producing areas and at different harvesting times of Guangxi Province of China (Table 1). Plant samples were stored at room temperature (25 °C) until needed for analysis. The species were identified by Prof. Mao-xiang Lai, one of the authors in this paper, and the voucher specimens were deposited at the School of Pharmaceutical Sciences, Guangxi Medical University, China.

Table 1. The specific information of the 20 samples of Z. nitidum roots.

Samples no.	Producing area^a	Collecting time	Specimen no.
1	Daxin County	August 2005	20051001
2	Qinzhou City	July 2006	20061001
3	Ningming County	August 2006	20061002
4	Tiandong County	September 2006	20061003
5	Wuming County	June 2007	20071001
6	Tiandeng County	July 2007	20071002
7	Lingshan County	September 2007	20071003
8	Nanning City	October 2007	20071004
9	Baise City	October2007	20071005
10	Bama County	October 2007	20071006
11	Napo County	October2007	20071007
12	Fushui County	June 2008	20081001
13	Longan County	July 2008	20081002
14	Chongzhuo County	July 2008	20081003
15	Lanxu County	August 2008	20081004
16	Yongning County	September 2008	20081005
17	Shangsi County	October 2008	20081006
18	Lingshan County	November 2008	20081007
19	Mashan County	December 2008	20081008
20	Yongning Datang County	November 2009	20091001

^aAll samples were collected from the main place of production of Z. nitidum, Guangxi Province, China.

Chemicals, reagents, and standard solutions

Methanol (MeOH) and acetonitrile (MeCN) were of LC grade (Fisher Company, Dallas, TX), phosphoric acid and triethylamine were of analytical grade (Beijing Chemical Corporation, Beijing, China). Deionized water was used for all analysis and was prepared using a Millipore (Billerica, MA) water purification system. All solvents and solutions were filtered through a Millipore filter (0.45 µm) before used. Reference standards of 1 and 2 were all purchased from the National Institutes for Food and Drug Control (Beijing, China). Compound 3 was kindly provided by the Key Laboratory for the Chemistry and Molecular Engineering of Medicinal Resources, Ministry of Education of China, Guangxi Normal University, whose structure was fully characterized by nuclear magnetic resonance (NMR) and MS spectroscopy. Their structures are shown in Figure 1. LC analysis indicated approximately 98.0% purity for all three compounds. Stock solutions of 1–3, 241, 391, 153 µg/mL, respectively, were prepared by accurate weighing and dissolution in MeOH in 10 mL volumetric flasks, and then stored at 4 °C. Standard solutions and calibration solutions were serially diluted with MeOH to prepare the appropriate concentrations prior to use.

Figure 1. The structures of the alkaloids from the Z. nitidum roots. 1: nitidine chloride (NC); 2: ethoxychelerythrine; 3: liriodenine.

Preparation of sample solutions

Twenty samples of dried Z. nitidum roots were collected as described in Table 1. The roots were pulverized into powder, passed through a 0.9 mm sieve, and then stored in a desiccator until required for assay. An accurately weighed amount (1.0 g) of each sample was extracted with 25 mL of MeOH solution twice under reflux for 2.0 h total (25-fold excess, v/w, 1.0 h each). The extract solution was filtered till hot. The filtrates were combined and concentrated in vacuo to afford a dried residue, which was carefully quantitatively transferred into a 25 mL volumetric flask with MeOH and a portion of the resultant solution was injected into an LC system for analysis of alkaloids. All resulting solutions were filtered through a 0.45 µm Millipore filter before injecting 10 µL for LC analysis.

Apparatus and LC conditions

The LC system was carried out using a Shimadzu LC-20 A HPLC system (Kyoto, Japan), which consisted of a binary gradient pump (model LC-20 A), a SPD-20MA diode array detector, an SiL-20 A auto sampler, a DGU-20A3 degasser and CTO-20 A column oven. The apparatus was interfaced to a Dell PC compatible computer using LC solution software. The RP-LC was carried out on an analytical Diamonsil™ ODS C₁₈ column (250 × 4.6 mm i.d.; 5 μm particles, China) with a Dikma Easy Guard C₁₈ insert (20 × 4.6 mm i.d.). The isocratic mobile phase consisted of a gradient of MeCN (component A) and 0.2% (v/v) aqueous phosphoric acid-triethylamine-buffer solution (PTBS, pH: 2.07, component B) at a flow rate of 1.0 mL/min. The mobile phase consisted of A:B (25:75, v/v). An aliquot (10 μL) of sample was injected into the LC and an analysis was carried out at 35 °C.

Preparation of standard solutions and calibration

Quantitative analysis was carried out using an external standard method. Standards of the alkaloids 1–3 were accurately weighed and dissolved in 10 mL volumetric flasks with MeOH as stock solution with final concentrations of 241, 391, 153 µg/mL, respectively. The stock solutions were used to prepare standard solutions, which were stored at 4 °C and remained stable for at least two months (verified by re-assaying the standard solutions). Calibration curves were established based on six points for each alkaloid with concentrations of 7.53, 15.06, 30.12, 60.25, 120.50 and 241.00 μg/mL for 1; 12.21, 24.43, 48.87, 97.75, 195.50 and 391.00 μg/mL for 2; and 4.78, 9.56, 19.12, 38.25, 76.50 and 153.00 μg/mL for 3, respectively.

Results

Optimization of the extraction conditions

The experimental extraction conditions, for example extraction method, solvent and time, can easily influence extraction efficiency. In fact, as a result, it is necessary to estimate and optimize the factors affecting extraction to achieve maximum recovery. Owing to the excellent extraction of the three alkaloids under reflux, a method involving a four-factor-three-level orthogonal array design [L₉ (3⁴) OAD] including the components of extracting solvent, solvent volume, extraction duration and times of reflux was developed for the optimization of the extraction of the alkaloids from Z. nitidum roots [used as sample 20: dried Z. nitidum roots], followed by LC analysis. In this study, the main factors considered in the alkaloids optimization process are displayed in Table 2. Nine pulverized crude dried drug powder samples (a same batch sample 20, 1.0 g) were very carefully weighed and extracted according to the L₉ (3⁴) orthogonal array design. The variables used in this optimization were: (1) extraction solvents (factor A); (2) extraction number (factor B); (3) extraction method (factor C); and (4) volume of extraction solvents (factor D) {20 mL [20-fold excess, volume of added extraction solvent/crude drug weight (v/w)], 25 mL (25-fold excess, v/w) and 30 mL (30-fold excess, v/w), respectively}. All the levels and factors are shown in Table 2.

Table 2. Main factors and their level settings used for optimization of the extraction of the alkaloids.

Level	Extraction solvents (A)	Extraction number (B)	Extraction methods (C^a)	Volume of solvents (D, mL)
1	MeOH	2	Ultrasound	20
2	70% MeOH	3	Reflux	25
3	20%MeOH	4	SOAKING	30

^aAt room temperature, and the extraction times were 1 h.

Table 3 shows the experimental design of the conditions for optimization of the extraction of the alkaloids and the results obtained from each experimental trial. An L₉ (3⁴) matrix with nine treatments (described in detail above) was used in this table to assign the variables and the corresponding level settings considered, and the variables were varied with the level settings values in different experimental trials. The total peak area of the three alkaloids was calculated. It could be used as a response function because it can take into consideration the effect of changes in the variables on the extraction efficiency of this technique. After the nine experimental trials (n = 3) were implemented, the corresponding total peak area of the three alkaloids for the variables set from each experimental trial was calculated and then tabulated (Table 3). In addition, the responses of each factor at different levels (1, 2, 3) were given in Table 3 too. In Table 4, the R showed that the effects of the four factors on the extraction efficiencies in the order was A > C > D > B, the best combination of levels was A₁ B₁ C₂ D₂. The conclusion was further confirmed by variance analysis in Table 4.

Table 3. L₉ (3⁴) Matrix as the experimental design for optimization of the extraction of the three alkaloids.

	Column no.
Experimental trial	A^a	B^a	C^a	D^a	Total peak area of the alkaloids
1	1^b	1	1	1	9 223 528
2	1	2	2	2	10 112 133
3	1	3	3	3	7 827 213
4	2^b	1	2	3	10 504 364
5	2	2	3	1	5 663 974
6	2	3	1	2	8 977 292
7	3^b	1	3	2	1 244 095
8	3	2	1	3	1 348 094
9	3	3	2	1	1 597 171
	9 054 291.667	6 990 662.567	6 516 305.100	5 494 891.567	–
	8 381 876.867	5 708 067.200	7 404 556.233	6 777 840.100	–
	1 396 453.667	6 133 892.433	4 911 760.867	6 559 890.533	–
R^c	7 657 838.000	1 282 595.367	2 492 795.366	1 282 948.533	–

^aFour main factors.

^bThree levels for each main factor.

^cAverage response for each level.

Table 4. Analysis of variance.

Factors	Sum of squares of deviations	Degrees of freedom	F value	p Value^a	Significance
A	107 890 761 780 601.375	2	41.138	0.01 < p < 0.05	Yes
B (as error)	2 560 433 065 310.313	2	–	–	No
C	9 577 581 012 666.125	2	3.741	–	No
D	2 827 681 751 793.500	2	1.104	–	No

^aF_0.1(2,2) = 9.0; F_0.05(2,2) = 19.0; F_0.1(2,2) = 99.0.

Thus, the optimum conditions for extraction of the three alkaloids from the powdered Z. nitidum roots were as follows: dried and pulverized sample 20 (1.0 g) was extracted in 25 mL MeOH solution by refluxing twice for 2.0 h (25-fold excess, v/w, 1.0 h each).

Validation of the method

Calibration and linearity

The linearity of the plot of concentration (x, μg/mL) for each alkaloid against the peak area (y) was investigated. The results were expressed as the value of the correlation coefficient (r) (Table 5).

Table 5. Regression equations and the respective correlation coefficients for quantitative determination of the three alkaloids by LC.

Reference substance	Correlation coefficient (r)	Regression equation^a	Linear range (μg/mL)
1	0.9995	y = 49386x−186 544	7.53–241.00
2	0.9999	y = 38902x−117 918	12.21–391.00
3	0.9995	y = 49563x−137 027	4.78–153.00

^ay denotes peak area, x denotes concentration of alkaloid compounds (μg/mL).

Determination of the limits of detection and quantification

Based on a detector signal-to-noise ratio of 10:1, the average limit of quantification (LOQ) calculated from chromatograms of the blank feed standard compounds was 3.00 µg/mL for 1, 5.00 µg/mL for 2 and 2.00 µg/mL for 3. Similarly, based on a detector signal-to-noise ratio of 3:1, the average limit of detection (LOD) was 1.50 µg/mL for 1, 2.00 µg/mL for 2 and 1.00 µg/mL for 3, respectively.

Determination of precision, accuracy and stability

The precision of the method was validated by both intra- and inter-day precision. Five experiments were carried out by injecting the same standard solutions into the LC five times during one day, and one assay was performed each day for three consecutive days to determine intra-day and inter-days precision, respectively. The results were shown in Table 6, indicating that the accuracy and precision of the method was sufficient for determination of the three alkaloids in Z. nitidum roots (sample 20 in Table 1).

Table 6. Intra- and inter-day precision of the LC method for determination of the alkaloids.

	1		2		3
Precision	Peak area^a	RSD (%)	Peak area^a	RSD (%)	Peak area^a	RSD (%)
Intra-day^b	4 994 862 ± 38 135	0.76	5 756 849 ± 49362	0.82	883 397 ± 5322	0.60
Inter-day^c	5 006 658 ± 77 222	1.54	5 763 829 ± 35 549	0.62	882 514 ± 5988	0.68

^aMean ± SD.

^bSample analyzed five times on 1 day, n = 5.

^cSample analyzed each day on three consecutive days, n = 3.

As for the stability and repeatability assay, the same sample (sample 20) was analyzed at different intervals (0, 12, 24, 36 and 48 h) at room temperature (25 °C), and the sample solution was found to be stable (RSD values of the peak area were lower than 2%). These results are listed in Table 7. The repeatability of analysis of the three alkaloids (Table 7) was determined by analysis of five samples (sample 20) extracted and processed, as described above, in parallel.

Table 7. Stability and repeatability of the LC method for determination of the three alkaloids.

	1		2		3
Property	Peak area^a	RSD (%)	Peak area^a	RSD (%)	Peak area^a	RSD (%)
Stability	5 133 341 ± 68 626	1.34	5 755 657 ± 71 829	1.25	876 249 ± 10 461	1.19
Repeatability	5 117 142 ± 63 772	1.25	5 773 785 ± 47 219	0.82	87 7218 ± 11 929	1.36

^aMean ± SD, n = 3.

Spiked recovery

To determine the recoveries of the three alkaloids (Table 8), the sample powder (sample 20) was spiked with standards at three different concentration levels and then extracted by the procedure represented above. One crude drug sample, for which the contents of 1–3 had already been determined, was spiked by adding known amounts of standards of 1–3 at low (one half of the known amounts), medium (same as the known amounts) and high (one and half of the known amounts) levels. Then, the spiked samples were extracted, processed and quantified in accordance with the methods represented above. All the process was repeated three times for each level and recovery was calculated with the following equation:

Table 8. Recoveries of the LC method for determination of the alkaloids.

Compounds	Concentration	Peak area^a	RSD (%)	Recovery (%)
1	Low	2 801 147 ± 126 99	0.45	101.59
	Medium	5 791 216 ± 13 312	0.99	101.63
	High	8 828 274 ± 108 314	1.23	102.17
2	Low	3 062 977 ± 40 327	1.32	98.42
	Medium	6 111 971 ± 59 926	0.98	96.38
	High	9 173 489 ± 104 608	1.14	95.83
3	Low	486 777 ± 52 241	1.07	100.21
	Medium	1 102 818 ± 7931	0.72	99.58
	High	1 657 730 ± 11 575	0.70	96.10

^aMean ± SD, n = 3.

Quantitative determination of three alkaloids in Z. nitidum roots

The contents of 1–3 in 20 batches of the Z. nitidum roots were measured with the method developed above. The representative LC chromatograms of three standard alkaloids and the sample of Z. nitidum roots (sample 20) are shown in Figures 2 and 3, respectively. The amounts of 1–3 were calculated from the regression equations obtained from calibration curves, and the results were expressed as contents of alkaloids in the medicinal material (mg/g crude drug), which are listed in Table 9.

Figure 2. LC profile obtained from a mixture of the three alkaloids: 1–3 = same as the structures in Figure 1.

Figure 3. Typical LC profile of an extract from the Z. nitidum roots (sample: 20): 1–3 = same as the structures in Figure 1.

Table 9. Alkaloids content of the Z. nitidum roots.

		Amount (mg/g crude drug)
No. Source		1	2	3	Total
1	Daxin County	1.120 ± 0.771	3.554 ± 1.409	0.819 ± 0.219	5.493
2	Qinzhou City	1.028 ± 0.327	2.838 ± 1.474	0.356 ± 0.154	4.222
3	Ningming County	1.059 ± 0.462	5.944 ± 1.161	0.374 ± 0.102	7.376
4	Tiandong County	1.294 ± 0.848	8.347 ± 3.288	0.331 ± 0.158	9.972
5	Wuming County	2.532 ± 1.337	4.685 ± 2.741	0.251 ± 0.085	7.468
6	Tiandeng County	2.689 ± 1.190	4.287 ± 1.804	0.348 ± 0.097	7.324
7	Jiuzhou Lingshan County	2.022 ± 0.345	5.746 ± 1.623	0.547 ± 0.145	8.315
8	Nanning City	1.644 ± 0.760	3.851 ± 2.340	0.325 ± 0.086	5.820
9	Baise City	1.058 ± 0.321	3.806 ± 1.322	0.309 ± 0.243	5.174
10	Bama County	2.271 ± 1.707	5.019 ± 3.459	0.266 ± 0.081	7.629
11	Napo County	3.377 ± 2.096	8.929 ± 5.780	0.544 ± 0.359	12.850
12	Fushui County	2.024 ± 1.029	6.143 ± 3.886	0.278 ± 0.140	8.445
13	Longan County	2.074 ± 0.940	4.326 ± 1.719	0.363 ± 0.186	6.763
14	Chongzhuo County	2.692 ± 1.442	4.584 ± 2.140	0.673 ± 0.299	7.949
15	Lanxu County	1.285 ± 0.575	4.448 ± 1.080	0.282 ± 0.089	6.015
16	Yongning County	1.812 ± 0.964	4.451 ± 0.564	0.349 ± 0.178	6.611
17	Shangsi County	2.931 ± 0.882	5.588 ± 1.017	0.374 ± 0.104	8.893
18	Shaping Lingshan County	2.868 ± 1.756	5.161 ± 2.449	0.401 ± 0.156	8.430
19	Mashan County	2.540 ± 1.433	3.269 ± 1.238	0.285 ± 0.053	6.094
20	Yongning Datang County	2.549 ± 1.001	6.162 ± 1.488	0.341 ± 0.200	9.051

1: Nitidine chloride (NC); 2: ethoxychelerythrine; 3: liriodenine.

The results clearly show there are substantial differences on the contents of the individual alkaloids in different harvesting area of Z. nitidum roots. For example, sample 11 had the greatest alkaloid concentration, with sample 4 having the second greatest amount. Among all the samples, the average content of compound 2 (ethoxychelerythrine) was highest and attained 5.06 ± 2.099 mg/g crude drug. These variations, too, might arise from differences in collection areas.

Discussion

The alkaloids are the bioactive compounds in Z. nitidum roots (Shi et al., 2005, 2010; Xu et al., 2010; Yao & Hu, 2004). As the major constituents of Z. nitidum, alkaloids have noteworthy biological activities such as anticancer, anti-inflammation, antihypertension, analgetic and antiemetic effects. Compounds 1, 2 and 3 have anticancer activity (Huang & Li, 1980; Suffness & Cordell, 1985; Xu et al., 2010; Ying et al., 2005), compound 3 has strongly antifungal activity (Nissanka et al., 2001), and so on. Therefore, the RP-LC method was established to determine the contents of alkaloids 1–3 in Z. nitidum roots.

The solvents, solvent volumes, method of extraction and extraction time seemed to be the most important conditions affecting the efficiency for the extraction of the three alkaloids. In this research, the optimum conditions for extraction of the three alkaloids from Z. nitidum roots were as follows: dried and pulverized Z. nitidum roots (1.0 g) were extracted with 25 mL of MeOH solution (25-fold excess, v/w) by refluxing twice (1.0 h each).

Some RP-LC methods have been developed for quantification of compounds 1 and compound 2 in Z. nitidum roots (Liang & Zhang, 2005), and of compounds 1 and 3, N-norchelerythrine, skimmianine and magnoflorine using a Waters Nova-Pak C₁₈ column (150 × 3.9 nm, 5 µm particles) (Shi et al., 2006). Other investigators (Qin et al., 2006) developed an RP-HPLC method to determine compound 1 in 10 different areas with a Kromasil C₁₈ column (250 × 4.6 nm, 5 µm particles). In this research, satisfactory results were obtained. The interference is always a problem in the identification and determination of target compounds in crude drugs, especially in the determination of trace compounds. As C₁₈ columns are widely used and offer excellent performance for alkaloids, a Diamonsil™ ODS C₁₈ column with a Dikma Easy Guard C₁₈ insert was chosen as the stationary phase. When the stationary phase was fixed, the composition of the mobile phase was a main factor for separating target compounds from interference. Initially, different proportions of the mobile phase system such as MeCN–H₂O, MeOH–H₂O, MeOH–H₂O–HCOOH, MeOH–MeCN–H₂O, MeOH–H₂O–glacial acetic acid and MeOH–H₂O–phosphoric acid were used as an isocratic mobile phase, but no significant resolution of compounds 1–3 could be achieved. Finally, the developed LC method employed an elution system of a gradient of MeCN-aqueous phosphoric acid (0.2%)-triethylamine-buffer solution (pH = 2.07) as the mobile phase. Reasonable retention time symmetrical peaks with good separation were obtained at a flow rate of 1.0 mL/min and a 285 nm as detection wavelength. The three alkaloids showed satisfactory separation from interfering compounds within 20 min. When peak purity was determined by acquiring UV spectra of all peaks of interest, no evidence of impurities was found. The recovery test of the method indicated that the treatment of the samples did not result in loss of the three alkaloids. The method also meets the requirements of convenience and time efficiency for evaluating the three alkaloids content with large quantities of Z. nitidum roots.

Conclusions

In summary, we have established an optimum extraction and detection conditions to quantify the three alkaloids in 20 samples of Z. nitidum roots. This method utilized simple, stable chromatography and short time run condition for the first time. From the contents of the alkaloids in Z. nitidum roots, it was shown clearly that there were essential differences on the contents of the individual alkaloids in different areas of Z. nitidum. Therefore, our studies suggest that the collection of high-quality medicinal herbs depends on the collecting location and harvesting time.

Declaration of interest

The authors explicitly declare no conflict of interest within this manuscript. This project was partly supported by the National Natural Science Foundation of China (30760302), the Chinese Herbal Medicine Support Fund Project by the Chinese National Development and Reform Commission (2007) 2706, the Key Scientific and Technological Projects of Guangxi, China (0992003A-21), and the Doctor Foundation of Guangxi Medical University.