A

wwiCcwoqwf©

K

1

atohr[plo

0d

Analytica Chimica Acta 577 (2006) 46–51

Validated high-performance thin-layer chromatography method fordetermination of trigonelline in herbal extract

and pharmaceutical dosage form

Shruti Chopra ∗, Farhan J. Ahmad, Roop K. Khar, Sanjay K. Motwani,Saiqa Mahdi, Zeenat Iqbal, Sushma Talegaonkar

Department of Pharmaceutics, Faculty of Pharmacy, Jamia Hamdard, Hamdard Nagar, New Delhi 110062, India

Received 5 March 2006; received in revised form 31 May 2006; accepted 10 June 2006Available online 21 June 2006

bstract

A new, simple, sensitive, selective, precise and robust high-performance thin-layer chromatographic (HPTLC) method for analysis of trigonellineas developed and validated for the determination of trigonelline in herbal extracts and in pharmaceutical dosage forms. Analysis of trigonellineas performed on TLC aluminium plates pre-coated with silica gel 60F-254 as the stationary phase. Linear ascending development was carried out

n twin trough glass chamber saturated with mobile phase consisting of n-propanol-methanol-water (4:1:4, v/v/v) at room temperature (25 ± 2 ◦C).amag TLC scanner III was used for spectrodensitometric scanning and analysis in absorbance mode at 269 nm. The system was found to giveompact spots for trigonelline (Rf value of 0.46 ± 0.02). The linear regression analysis data for the calibration plots showed good linear relationshipith r2 = 0.9991 ± 0.0002 in the concentration range 100–1200 ng spot−1 with respect to peak area. According to the International Conferencen Harmonization (ICH) guidelines the method was validated for precision, recovery, robustness and ruggedness. The limits of detection and

uantification were determined. The trigonelline content of herbal extracts quantified and estimated from the formulation was found to be wellithin limits (±5% of the labeled content of the formulations). Statistical analysis of the data showed that the method is reproducible and selective

or the estimation of trigonelline.2006 Elsevier B.V. All rights reserved.

eywords: Trigonelline; High-performance thin layer chromatography; Method validation; Quantitative analysis; Herbal extract; Gels

am

mHrh

o

. Introduction

Trigonella foenum-graecum from the family Fabaceae, anromatic herbaceous plant is widely cultivated in Medi-erranean countries and Asia [1]. The literature records a varietyf therapeutic actions of Trigonella foenum-graecum includingypocholesterolaemia [2–4], hypoglycaemia [5–7], antibacte-ial [8], antiviral [9], anti-inflammatory activities [1], antioxidant10], appetite stimulant [11], etc. Phytochemically, different

arts of the plant contain several constituents such as alka-oids, proteins, flavonoids, saponins, etc. Several reports focusn trigonelline, a major active constituent having hypoglycemic

∗ Corresponding author. Tel.: +91 11 26059688; fax: +91 11 26059663.E-mail address: shrutichopra21@yahoo.com (S. Chopra).

imnlmab

003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.aca.2006.06.034

ctivity, hypocholesterolemic, antiseptic, antimigrane, antitu-or, mutagenic and osmoregular properties [7,12,13].Several analytical methods have been published for the deter-

ination of trigonelline (Fig. 1) based on UV [14], HPLC [15],PTLC, TLC [16] and OPLC [17,18], but there are very few

eports on analytical methods for estimation of trigonelline inerbal extracts and its dosage form.

HPTLC is a powerful analytical technique due to its meritsf reliability, simplicity, reproducibility and speed. This methods economical as it utilizes smaller amounts of solvents with

inimum sample clean up. Also, in a short duration, a largeumber of samples are simultaneously analyzed. HPTLC has no

imitation on the choice of the mobile phase and unlike HPLC,

obile phases having pH 8 and above can be employed. Directpplication of suspensions, dirty or turbid samples are possi-le. Furthermore, it permits a simultaneous assay of several

S. Chopra et al. / Analytica Chim

c[

ctw

2

2

STfDoI

2

wm(swssodst(mpDIu

2

i1tt

tr

2

2

p(p(oo

2

taewa

2

ofs3itd

2

ardt

2

at

2

tsttd(

Fig. 1. Chemical structure of trigonelline.

omponents in a multicomponent formulation or herbal extracts19–21].

The aim of this work was to develop an accurate, spe-ific, repeatable and robust method for the determination ofrigonelline. The proposed method was validated in complianceith ICH guidelines [22,23].

. Experimental

.1. Materials

Standard trigonelline hydrochloride was purchased fromigma–Aldrich Chemicals Pvt., Ltd. New Delhi, India.rigonella foenum-graecum was procured as a fresh materialrom the local suppliers and was assessed biologically by theepartment of Botany. All chemicals and reagents used weref analytical grade and were purchased from Merck Chemicals,ndia.

.2. Instrumentation and chromatographic conditions

The samples were spotted in the form of bands of width 6 mmith a Camag microlitre syringe on precoated silica gel alu-inium plate 60F-254 (20 cm × 10 cm) with 200 �m thickness

E. Merck, Germany) using a Camag Linomat V (Switzerland)ample applicator. A constant application rate of 150 nL s−1

as employed and space between two bands were 10 mm. Thelit dimension was kept at 5 mm × 0.45 mm and 20 mm s−1

canning speed was employed. The mobile phase consistedf n-propanol-methanol-water (4:1:4, v/v/v). Linear ascendingevelopment was carried out in a twin trough glass chamberaturated with mobile phase. The optimized chamber satura-ion time for the mobile phase was 30 min at room temperature25 ± 2 ◦C) at relative humidity of 55 ± 5%. The length of chro-atogram run was 80 mm. Subsequent to the development, TLC

lates were dried in a current of air with help of an air-dryer.ensitometric scanning was performed on Camag TLC scanner

II in the absorbance mode at 269 nm. The source of radiationtilized was a deuterium lamp.

.3. Calibration curve of trigonelline

A stock solution of trigonelline (100 �g mL−1) was prepared

n methanol. Different volumes of stock solution 1, 2, 4, 6, 8,0, 12 �L, were spotted on the TLC plate to obtain concen-rations of 100, 200, 400, 600, 800, 1000, 1200 ng spot−1 ofrigonelline, respectively. The data of peak areas plotted against

2

h

ica Acta 577 (2006) 46–51 47

he corresponding concentrations were treated by least-squareegression analysis.

.4. Method validation

.4.1. PrecisionRepeatability of the sample application and measurement of

eak area were carried out using six replicates of the same spot600 ng spot−1 of trigonelline) and was expressed in terms ofercent relative standard deviation (%R.S.D.) and standard errorS.E.). The intra- and inter-day variation for the determinationf trigonelline was carried at three different concentration levelsf 400, 600 and 800 ng spot−1.

.4.2. Robustness of the methodBy introducing small changes in the mobile phase composi-

ion, mobile phase volume, duration of mobile phase saturationnd activation of prewashed TLC plates with methanol, theffects on the results were examined. Robustness of the methodas done in triplicate at a concentration level of 600 ng spot−1

nd the %R.S.D and S.E. of peak areas was calculated.

.4.3. Limit of detection and limit of quantificationIn order to estimate the limit of detection (LOD) and limit

f quantitation (LOQ), blank methanol was spotted six timesollowing the same method as explained in Section 2.2 and theignal-to-noise ratio was determined. LOD was considered as:1 and LOQ as 10:1. LOD and LOQ were experimentally ver-fied by diluting the known concentrations of trigonelline untilhe average responses were approximately 3 or 10 times the stan-ard deviation of the responses for six replicate determinations.

.4.4. Recovery studiesThe pre-analyzed samples were spiked with extra 50, 100

nd 150% of the standard trigonelline and the mixtures wereeanalyzed by the proposed method. The experiment was con-ucted six times. This was done to check for the recovery of therigonelline at different levels in the formulations.

.4.5. RuggednessA solution of concentration 1000 ng spot−1 was prepared and

nalyzed on day 0 and after 6, 12, 24, 48 and 72 h. Data werereated for %R.S.D. to assess ruggedness of the method.

.4.6. SpecificityThe specificity of the method was ascertained by analyzing

he standard drug and extract. The spot for trigonelline in theample was confirmed by comparing the Rf values and spec-ra of the spot with that of the standard. The peak purity ofhe trigonelline was assessed by comparing the spectra at threeifferent levels, viz. peak start (S), peak apex (M) and peak endE) positions of the spot.

.5. Detection of related impurities

The related unknown impurities were determined by spottingigher concentrations of the trigonelline. Trigonelline solution

4 Chimica Acta 577 (2006) 46–51

waotmsc

2

52m1Santoiw

2

1maaamf2ans

3

3

hr(b

F(

cact

3

scpbits

3

(baSs

TI

A

468

8 S. Chopra et al. / Analytica

as prepared at a concentration of 2000 �g mL−1 in methanol,nd this solution was termed as sample solution. One millilitref the sample solution was diluted to 40 mL with methanol andhis solution was termed as standard solution (50 �g mL−1). Two

icrolitres of both the standard (100 ng spot−1) and the sampleolution (4000 ng spot−1) were applied on HPTLC plate and thehromatograms were run as described in Section 2.2.

.6. Analysis of trigonelline in herbal extracts

To determine the content of trigonelline in herbal extract,00 mg was transferred into a 50 mL volumetric flask containing5 mL methanol, sonicated for 30 min and diluted to 50 mL withethanol. The resulting solution was centrifuged at 3000 rpm for

5 min and the supernatant was analyzed for the drug content.ix microliters of the filtered solution (10,000 �g mL−1) waspplied on the TLC plate followed by development and scan-ing as described in Section 2.2. The analysis was repeated inriplicate. The possibility of interference from other componentsf the extract in the analysis was studied. Trigonelline contentn herbal extracts from two different sources (Sources 1 and 2)as also studied.

.7. Analysis of trigonelline in prepared formulations

An accurately weighed quantity of gel equivalent to about00 ng of trigonelline, i.e., 8.5 g of gel was extracted with 25 mLethanol by sonication for 30 min. This extract was centrifuged

t 12,000 rpm for 15 min at 4 ◦C. The supernatant was filterednd the filtrate was dried to constant weight at room temper-ture. The residue was redissolved in 5 mL of methanol. Sixicroliters of the filtered solution was applied on the TLC plate

ollowed by development and scanning as described in Section.2. The analysis was repeated in triplicate. Placebo gels werenalyzed similarly to study the potential interference. Combi-ation of active components were also analyzed separately totudy the interference of each active component.

. Result and discussion

.1. Development of the optimum mobile phase

The TLC procedure was optimized with a view to quantify the

erbal extract. Initially n-propanol-methanol-water in varyingatios was tried. The mobile phase n-propanol-methanol-water4:1:2, v/v/v) gave good resolution with Rf = 0.46 for trigonellineut typical peak shape was missing. Finally, the mobile phase

liia

able 1ntra- and Inter-day precision of HPTLC method (n = 6)

mount (ng spot−1) Intra-day precision

Mean area S.D. %R.S.D.

00 1826.10 1.57 0.08600 2720.68 1.53 0.05600 3450.56 1.45 0.042

ig. 2. Chromatogram of standard trigonelline (1000 ng spot−1): peak 1Rf = 0.46 ± 0.02), mobile phase: n-propanol-methanol-water (4:1:4, v/v/v).

onsisting of (4:1:4, v/v/v) gave a sharp and well-defined peakt Rf = 0.46 (Fig. 2). Well-defined spots were obtained when thehamber was saturated with mobile phase for 30 min at roomemperature.

.2. Calibration curves

The developed HPTLC method for estimation of trigonellinehowed a good correlation coefficient (r2 = 0.9991 ± 0.0002) inoncentration range of 100–1200 ng spot−1 with respect to theeak area. Fig. 3 displays three-dimensional image of the cali-ration samples at 269 nm. The mean value (±S.D.) of slope andntercept were 4.1312 ± 0.0491 and 208.2135 ± 4.5092, respec-ively. No significant difference was observed in the slopes oftandard curves (ANOVA, P > 0.05).

.3. Method validation

The %R.S.D. for repeatability of sample application600 ng spot−1) and measurement of peak areas were found toe 0.09 and 0.15% respectively. The measurement of the peakrea at three different concentration levels showed low values of.E. and %R.S.D. (<1%) for inter- and intra-day variation, whichuggested an excellent precision of the method (Table 1). The

ow values of S.D., %R.S.D. and S.E. obtained after introduc-ng small deliberate changes in the developed HPTLC methodndicated the robustness of the method (Table 2). Detection limitnd quantification limit with signal-to-noise ratio of 3:1 and 10:1

Inter-day precision

S.E. Mean area S.D. %R.S.D. S.E.

0.64 1898.58 1.54 0.081 0.630.62 2642.37 1.99 0.075 0.810.59 3420.80 2.91 0.056 0.78

S. Chopra et al. / Analytica Chimica Acta 577 (2006) 46–51 49

Fig. 3. Three-dimensional image of the

Table 2Robustness of the HPTLC method (n = 3, 600 ng spot−1)

Parameter S.D. of Peak Area %RSD S.E.

(1) Mobile Phase composition(n-propanol-methanol-water)

1.63 0.3857 0.1575

(2) Mobile phase volume (18, 20and 22 mL)

1.79 0.2857 0.1167

(3) Duration of saturation (20, 30and 40 min)

1.48 0.2644 0.1080

(4) Activation of prewashed TLC 1.26 0.1973 0.0806

wt

q

rvntppct

3

tspot from the standard solution. The sample solution showed

plates (2, 5 and 7 min)

ere found to be 2.27 and 7.58 ng, respectively, which indicates

he adequate sensitivity of the method.

The proposed method when used for extraction and subse-uent estimation of trigonelline from the formulation afforded

t1t

Fig. 4. Spectra comparison of stand

calibration spots for trigonelline.

ecovery of 98.91–101.34% as listed in Table 3. Low %R.S.D.alue of 0.0551 between the peak area values proved the rugged-ess of the method indicating that trigonelline is stable duringhe extraction procedure as well as during analysis. The peakurity of trigonelline was assessed by comparing the spectra ateak start, peak apex and peak end positions of the spot. Goodorrelation (r = 0.9992) was obtained between the standard andhe sample overlain spectra of trigonelline (Fig. 4).

.4. Detection of related impurities

The spots other than the principal spot for trigonelline fromhe sample solution were not more intense than the principal

wo additional spots at Rf = 0.05 and 0.91 having peak areas of69.60 and 1370.60 (Fig. 5), which were much less as comparedo the standard solution.

ard trigonelline and extract.

50 S. Chopra et al. / Analytica Chimica Acta 577 (2006) 46–51

Table 3Recovery Studies (n = 6)

Excess drug added to analyte (%) Theoretical content (ng) Amount found (ng) Recovery (%) %R.S.D. S.E.

0 400 395.64 98.91 0.254 1.8650 600 603.90 100.65 0.162 1.73

100 800 810.72 101.34 0.097 1.37150 1000 991.90 99.19 0.107 1.82

F1(

3

opctt(

Fpl

FP(

3

of

ig. 5. Chromatogram of standard trigonelline and its impurity peaks: peak(impurity) Rf: 0.05, peak 2 (standard trigonelline) Rf = 0.46 ± 0.02, peak 3

impurity) Rf: 0.91.

.5. Estimation of trigonelline in herbal extracts

A single spot at Rf = 0.46 was observed in the chromatogramf the trigonelline isolated from extract along with other com-onents. There was no interference in analysis from the other

omponents present in the extracts. These components appear inhe chromatogram at significantly different Rf valves. The totalrigonelline content was found to be 1.99% (w/w) and 2.10%w/w) in Sources 1 and 2, respectively (Fig. 6).

ig. 6. Chromatogram of Trigonella foenum-graecum extract (1200 ng spot−1);eak 1–7 belong to other components present in the extract. Peak 4 is of trigonel-ine (Rf: 0.46 ± 0.02); mobile phase: n-propanol-methanol-water (4:1:4, v/v/v).

itw

4

rateqhaees

R

ig. 7. Chromatogram of Trigonella foenum-graecum gel (107 ng spot−1):eaks 1–10 belong to other components present in the gel. Peak 5, trigonellineRf: 0.46 ± 0.02); mobile phase; n-propanol-methanol-water (4:1:4, v/v/v).

.6. Analysis of the prepared formulation

A single spot at Rf = 0.46 was observed in the chromatogramf the trigonelline extracted from gels. There was no interferencerom the excipients and the other active components presentn the herbal gel formulation (Fig. 7). The % recovery of therigonelline from the gel formulation was found to be 98% andas well within the limits (label claim ± 5%).

. Conclusion

The developed HPTLC technique is a precise, specific, accu-ate and robust for the determination of trigonelline. Statisticalnalysis proves that the method is reproducible and selective forhe analysis of trigonelline. Since the proposed mobile phaseffectively resolves trigonelline, the method can be used forualitative as well as quantitative analysis of trigonelline inerbal extracts and to detect the related impurities to establishlkaloidal components’ content in products grown in differ-nt climatic conditions. Further the proposed method can bextended to study the degradation of trigonelline under differenttress conditions, as per the recommendations of ICH guidelines.

eferences

[1] A. Ahmadiani, M. Javan, S. Semnanian, E. Barat, M. Kamalinejad, J.Ethnopharm. 75 (2001) 283.

[2] M.A. Ajabnoor, A.K. Tilmisany, J. Ethnopharm. 22 (1988) 45.[3] A. Stark, Z. Madar, Br. J. Nutr. 69 (1993) 277.

Chim

[

[

[

[

[[

[[

[

[[

[

S. Chopra et al. / Analytica

[4] M. Al-Habori, A.M. Al-Aghbari, M. Al-Mamary, Phytother. Res. 12 (1998)572.

[5] G. Ribes, Y. Sauvaire, J.C. Baccou, G. Valette, D. Chenon, E.R. Trimble,M.M. Loubatiers-Mariani, Ann. Nutr. Metab. 28 (1984) 37.

[6] R.D. Sharma, T.C. Raghuram, N.S. Rao, Eur. J. Clin. Nutr. 44 (1990)301.

[7] J.A. Duke, CRC Handbook of Medicinal Spices, CRC Press, New York,2001, p. 297.

[8] M.A. Bhatti, M.T.J. Khan, B. Ahmed, M. Jamshaid, W. Ahmad,Fitoterapia 67 (1996) 372.

[9] S. Ghosal, R.S. Srivastava, D.C. Chatterjee, S.K. Dutta, Phytochemistry 13(1974) 2247.

10] A.M. El-Sokkary, M.A. Ghoneim, Ind. J. Dairy Sci. 4 (1951)123.

11] P. Petit, Y. Sauvaire, G. Ponsin, M. Manteghetti, A. Fave, G. Ribes, Phar-

macol. Biochem. Beh. 45 (1993) 369.

12] J. Mishkinsky, B. Joseph, F.G. Sulman, A.L. Goldschmied, Lancet 290(1967) 1311.

13] J. Barnes, L.A. Anderson, J.D. Phillipson, Herbal Medicines, second ed.,Pharmaceutical Press, 2002, p. 209.

[

[

ica Acta 577 (2006) 46–51 51

14] S.M.A. Wahab, M.A. Selim, Egypt J. Pharm. Sci. 26 (1986) 335.15] H.Q. Zhao, Y. Qu, X.Y. Wang, H.J. Zhang, F.M. Li, H. Masao, Zhongguo

Zhong Yao Za Zhi 27 (2002) 194.16] S. Scinto, R. Chillemi, M. Piattelli, J. Nat. Prod. 51 (1988) 322.17] E. Mincsovics, E. Sardi, I. Velich, G. Katay, E. Tyihak, J. Planar Chro-

matogr. 15 (2002) 280.18] B. Barna, A. Sarhan, N. Cong, Z. Kiraly, E. Tyihak, Comparative inves-

tigation of trigonelline level in tomato leaves by HPTLC and OPLC, in:Proceedings of the International Symposium on TLC with special Empha-sis on OPLC, Szeged, 1984.

19] E.A. Abourashed, J.S. Mossa, J. Pharm. Biomed. Anal. 36 (2004) 617.20] H. Agarawal, N. Kaul, A.R. Paradkar, K.R. Mahadik, J. Pharm. Biomed.

Anal. 36 (2004) 33.21] V.B. Patravale, S. D’Souza, Y. Narkar, J. Pharm. Biomed. Anal. 25 (2001)

685.

22] M.J. Ansari, S. Ahmad, K. Kohli, J. Ali, R.K. Khar, J. Pharm. Biomed.

Anal. 39 (2005) 132.23] Q2A, ICH, Q2A (R1), Validation of Analytical Procedures: Text

and Methodology, International Conference on Harmonization, Geneva,(November 2005),.(http://www.ich.org/LOB/media/MEDIA417.pdf).