The precursor 5-aminolevulinic acid of the photosensitizer protoporphyrin IX in the heme bio-synthetic pathway has been used in photodynamic therapy for treatment of variuos skin lessions for over than two past decades. The skin penetration studies require suitable method for determination of ALA in skin samples and other biological tissues. However only several methods were used for those purposes, such as HPLC which requires common derivatization reactions in order to determine typical amino acids. The penetrated amount of this prodrug can also be determined by autoradiographic and scintillation analysis already reported by Woolfson et al. (X), which responds to beta-particles or other harmful radiation not found in pharmaceutical preparations. Only a few direct determination methods have been reported. In reverse phase chromatography, no sufficient retention is achievable for nonderivatized ALA, because of its high polarity. For that reason, ion-exchange chromatography [5,6] or ionpairing chromatography [7] may be used. Capillary electrophoresis (CE) could be efficient method for the determination of charged substances. Its use for the simultaneous determination of ALA and porphobilinogen in biological samples by micellar electrokinetic chromatography (MEKC) has already been reported by Luo et al. [8]. Also a capillary electrophoresis method for simultaneous quantification of 5-aminolevulinic acid (ALA) and its degradation products was reported by Bunke et al. (X). In comparison to liquid chromatographic methods, capillary electrophoresis offers several advantages, e.g. high resolution and efficiency, short analysis time as a requirement for routine analysis, small sample volumes, small volumes of running buffers, inexpensive column, direct injection without sample pretreatment. CE of drug-related impurities has been reviewed by Altria [9]. However only few methods were reported for skin sample analysis using CE. Al-Otaibi et al. has reported quantification of tetracaine in skin using CE [X]. In addition, penetration studies require suitable method for extraction of substances from skin. The extraction from non-aqueous gels has already been reported by McCarron et. al. [X]. The extraction from skin with hydrochloric acid solution has also been reported by Shing-Chuan Shen and extraction with methanol was reported by M.B.R.Pierre et al. [X]. High attention should be paid to ALA stability as it degrades readily at alkaline pH. The chemical stability of ALA has been assessed in previous studies by visual color changes during the observation period (Chang et al., 1996) and by absorption spectroscopy (Novo et al., 1996) following the change in absorbance caused by the formation of a mixture of condensation products of ALA. The stability of 5-aminolevulinic acid in aqueous solution has already been reported by Elfsson et al. [X]. The stability of 5-aminolevulinic acid in solution has also been reported by Gadmar et al. [X].
The objective of this work was to develop an analytical CE method, which permits simultaneous quantification of ALA in skin samples. ICH guidelines describe the validation of analytical methods. We report on the validation of a CE method to demonstrate its suitability for ALA assay testing in skin.
2. Experimental
2.1. Chemicals
Reagents used for experiment: di-sodium tetraborate dekahydrate (purity - 99,0 %, Lachema, Czech Republic); sodium dihydrophospfate dihydrate (purity - 99,0 %, Lachema, Czech Republic); sodium dodecylsulphate (purity - 99,0 %, Carl Roth GmbH, Karlsruhe); sodium hydroxide (purity - 98,5 % Lachema, Czech Republic) methanol (purity - 99,5 %, Barta a Cihlar, Czech Republic); acetonitrile (purity - 99,9 %, Sigma-Aldrich, Vokietija); bidistilled water was produced in our laboratory; nylon filters (pore size - 0,45 μm).
2.2. Instrumentation
Analyses were performed using capillary electrophoresis apparatus Biofocus 3000 (BIORAD Laboratories); ultrasound bath Cole-Parmer (Cole-Parmer Instrument Company, Vernon Hills, Illinois, 60061, USA); centrifuge WIFUG (CHEMICO, Stockholm, Sweden); weighting-machine Sartorius Basic ( ); bidistilation aparatus ( ); 200 μl, 1000 μl, 5000 μl Eppendorf pipetes; 500 μl and 1500 μl test-tubes suited for microcentrifuges (BIO-RAD Laboratories)); analytical data was obtained by Biofocus 3000 software, electroferograms were analyzed by Biofocus Integrator software; statistical calculations were performed by MS Excell software.
2.3. Methods
Samples were injected hydrodynamically (10 psi·s). The analysis voltage was 24 kv and the temperature was 30 °C. Analyte was detected by UV absorbance at 200 nm. The capillary was rinsed for 2 min with 1M NaOH solution and for 2 min with bidistilled water.
The running buffer was 62 mM Borate (Na2B4O7) + 8% MeOH (v/v). pH was not adjusted. Buffer solution was filtered through 0.45 μm pore-size nylon filter and degassed by vacuum.
Acetone was used to measure electroosmotic flow (EOF). ALA was quantified relative to external standard. Separation parameters were calculated in accordance to European Pharmacopoeia. Detection (DL) and quantitation (QL) limits were calculated in accordance to ICH guidelines (X).
3. Results and discussion
3.1. Method optimization
The guideline for optimizing CE method for ALA determination in human skin extracts was previously mentioned CE method described by Bunke et al [x] .CE conditions for ALA assay were optimized using human skin extracts. In order to ensure ALA stability, isotonic phosphate buffer solution was used for extraction. ALA stability in acid and aqueous solutions has been already described by Elfsson et al. [x].
Table 1.
Parameters
X
X/ALA
ALA
ALA/Y
Y
Effective mobility
2.101*10-4
-
2.238*10-4
-
2.312*10-4
Separation factor
-
1.085595
-
1.048077
-
Resolution
-
7.45
-
3.36
-
Symmetry factor
1.44
-
1.88
-
2.48
3.2. Method validation
3.2.1. Specificity
There were some issues at optimizing the specificity. The separation conditions used at the beginning of the research (50 mM Borate, pH not adjusted, voltage 20 kV, uncoated silica capillary, internal diameter 50 µm, 57,8 cm effective length, 62,3 cm total length) showed interference between ALA and one of the matrix (extract of derma) spikes. pH adjustment did not result in sufficient resolution. Addition of acetonitrile resulted in better separation, but sufficient resolution was not achieved. Better results were obtained when 5% (v/v) was added to the running buffer.
Figure 1. ...
After addition of 7 % (v/v) MeOH to running buffer it was spotted that some traces of unknown compound still interfere with ALA. It was decided to increase MeOH concentration. As addition of high quantity of MeOH to running buffer is still limited agent at improving resolution because of rapid buffer exhaustion it was decided to increase ionic strength also. Buffer concentration was raised to 62 mM and MeOH addition was raised to 8%. The sufficient value of resolution has been achieved and spikes has been separated completely ( fig.). Voltage also has been increased to 24 kV with the aim to shortening analysis time.
Figure 2. ...
Table 3. Effect of methanol concentration on resolution of ALA and matrix spikes
MeOH added, % (v/v)
ALA/Y
X/ALA
7
8
9
6.45
2.17
7.45
3.36
5.20
2.67
Thanks to good migration time reproducibility achieved, this method allowed unambigous discrimination of ALA matrix peaks. Best results were achieved for relative migration times related to the ALA peak.
Table 4.
Intra day (n=15)
Extra day (n=30)
Parameter
RSD, %
C.I., %
RSD, %
C.I., %
Migration time
1.15
±0.58
2.58
±0.92
Epidermis did not show any interference between matrix and ALA peak.
3.2.2. Accuracy
The accuracy of analytical procedure was evaluated by determining correacted peak area (Acorr=AR/MT) of 5 diferent concentration samples by calculating recovery, RSD and C.I. The obtained data shows sufficient accuracy for assay testing of ALA.
Table 4. Accuracy of analytical procedure
Sample
Recovery (%)
1
99.98
2
99.81
3
98.33
4
100.01
5
101.62
S.N., % (n=3)
1.1688503
P.I., % (n=3)
1.0245236
3.2.3. Precision
The precision of analytical procedure was evaluated by determining correacted peak area (Acorr=AR/MT) of 3 diferent concentration samples by calculating, RSD and C.I. The obtained data shows sufficient accuracy for assay testing of ALA.
Table 5. Precision of analytical procedure
RSD (%)
C.I. (%)
Repeatability (n=3)
1.68
±1.91
Intermediate precision (n=3)
1.93
±2.19
3.2.4. Linearity and range
ALA was quantified relative to external standard. 6 different concentrations were used to calculate correlation coefficient within range of 0.3 mg/ml and 0.025 mg/ml ALA. It was achieved good correlation coefficient showing suffient linearity for ALA assay.
3.2.5. Limits of detection and quantitation
Limits of detection and quantitation were calculated in accordance to equations indicated in ICH guidelines [x]:
DL - Detection limit;
QL - Quantitation limit;
σ - Standard deviation of background response;
S - Slope of calibration curve (139103.55126)
Standard deviation of background response has been measured by injection of blank sample (isotonic phosphate buffer : MeOH 1:1 (v/v)). The response obtained at ALA migration time was the factor for standard deviation calculation (n=25)
The values of these limits shows, that developed method is suitable for low ALA concentrations to be determined. Calculations shows: DL=1.04 µg/ml; QL=3.18 µg/ml.
We have tried different injection volumes. It seems that electrostacking effect occurs because of addition of MeOH.
This figure shows that electrostacking effect is effective and peaks do not get too large for accurate determination.
3.2.6. Sample preparation and extraction efficiency
9 skin samples from 3 different donors were investigated in order to obtain information about extraction efficiency. Epidermis was separated from derma using heat separation method [x]. ALA stability issue was resolved appealing to Blois et al. [x]. Those researchers state, that 0.1 % ALA solution, pH=5, half life at 63 oC is 438 h. Then samples were spiked with 3 different concentrations including lowest and highest points of calibration curve. Derma and epidermis samples were then extracted with 0.5 ml isotonic (0.067) mM phosphate buffer for 45 min using ultrasonication (40 kHz) at 25 oC. Proteins were precipitated by adding of MeOH (50% v/v). Precipitated samples were centrifuged for 2 min at 6000 rpm and filtered through 0.45 μm nylon filters. Extraction efficiency was evaluated by calculating mean recovery, RSD and C.I.
Extraction efficiency, %
Calculated for n=9
Derma
Epidermis
Mean recovery, %
96.44
96.12
RSD, %
3.06
1.74
C.I., %
±2.00
±1.14
3.2.7. Robustness
The electrophoretic resolution of ALA peak and closely adjoining peak were used to evaluate the method under modified conditions. Sufficient resolution under all separation conditions has been shown demonstrating sufficient robustness.
Robustness
Parameter
Rs
Temperature
28
2.18
30
3.36
32
3.19
Voltage
22
3.33
24
3.36
26
4.85
Buffer concentration (mM)
60
1.98
62
3.36
64
3.63
3.2.8. Discussion
The optimized parameters conform requirements with accordance to ICH guidelines for method to be validated. Electrostacking technique with ACN did not show statistically different results comparing it to obtained ones with MeOH. MeOH was chosen for cost effectiveness. The developed method can be used for wide range of ALA concentrations assay in extracts by simply regulating injection pressure and time.
