Increased asymmetric dimethylarginine (ADMA) in human plasma has been associated with reduced generation of nitric oxide, leading to atherosclerotic diseases. ADMA may therefore be an important biomarker for cardiovascular disease. In the present study, three sample preparation techniques were compared regarding the quantification of L-arginine and ADMA in human plasma: (A) protein precipitation (PP) based on aqueous trichloroacetic acid (TCA), (B) PP using a mixture of ammonia and acetonitrile, and (C) solid-phase extraction (SPE). The samples were analysed by using high-performance liquid chromatography with fluorescence detection (HPLC-FLD). The analytical performance of (A) was comparable with that of (C), demonstrating recoveries of >90%, coefficient of variations (CVs, %) of <8, and a resolution (
Nitric oxide (NO) is important in numerous biological processes including the relaxation of smooth muscles and inhibition of platelet aggregation [
A sample preparation step is often required before chromatographic analysis in order to separate the analytes of interest from proteins and other possible interferences in plasma [
After sample preparation, the separation and subsequent detection of L-arginine, ADMA, and SDMA are often performed using HPLC-FLD after pre- or postcolumn derivatization of L-arginine and ADMA with a fluorescent tag. In most studies, the fluorophore
Since current analytical methods described in the literature for quantification of L-arginine and ADMA involve relatively complicated and time-consuming sample preparation procedures, the aim of the present study was to investigate whether simple and fast PP before quantification of the biomarkers L-arginine, ADMA, and SDMA in human plasma may result in acceptable analytical performance evaluated by the FDA guideline for bioanalytical validation [
All separations were performed on a Dionex Ultimate 3000 (Thermoscientific, Waltham, MA, USA) coupled to a Dionex fluorescence detector containing a micro flow cell (2
L-arginine (>99%), NG,NG′-dimethyl-L-arginine di(
Calibration stock solutions of 2.0 mM of L-arginine and 6.05 mM of ADMA and SDMA were prepared in 60% methanol and stored at −20°C until use. Working calibration standards of L-arginine, ADMA, and SDMA were prepared in 10 mM HCl in concentrations ranging from 0.5
Three different sample preparation techniques (A), (B), and (C) were evaluated by using human plasma obtained from one person. The procedures are briefly described in the following: (A) PP was performed on 50
Recoveries in % were determined for each sample preparation technique by fortifying the concentration of L-arginine and ADMA in a human plasma sample with calibration standard and measuring the concentration in the spiked sample in relation to the concentration measured in the unspiked sample plus the added calibration standard concentration. The retention times for L-arginine, ADMA, and SDMA were verified in each sample preparation procedure by 10 times fortifications with calibration standards.
In order to detect L-arginine, ADMA, and SDMA, precolumn derivatization was performed automatically by mixing 3
The chromatographic conditions used in the current method were inspired by a previously published method for the quantification of L-arginine, ADMA, and SDMA [
Robustness was initially tested for procedure (A) applying different time intervals (2, 8, 10, and 12 min) for performing PP. Thereafter, the selected sample preparation technique, procedure (A), combined with the developed chromatographic method was validated with respect to linearity, recovery, precision, and sensitivity (taken as LLOQ) on a human plasma sample [
The stability for extracted human plasma samples and calibration standards was tested for L-arginine and ADMA at 4°C. The long-term stability was further assessed for human plasma at −80°C and for calibration stock solutions at −20°C. All stability tests were performed in triplicate. Acceptable stability means that at least 90% of the initial amount was found at a given time point.
Blood samples were obtained from six apparently healthy individuals after informed consent but without collection of any data on the individuals. The use of human blood for quality control purposes is not subject to ethical approval in Denmark. The blood was collected in VacutainerTM collection tubes containing K3-EDTA from Becton Dickinson (Franklin Lakes, NJ, USA) and centrifuged at 2000 g for 5 min at 4°C. The resulting plasma was stored in aliquots at −80°C and used for verification of the analytical method.
Preliminary studies of different PPs were performed. Zinellu et al. showed that the recoveries of L-arginine and ADMA were low using 100% acetonitrile [
In the development of the SPE cleanup, (C), adding PBS buffer (pH 7.4) to plasma in a ratio of 3 : 1 as described in previous studies was found to clog the Oasis MCX column, and therefore, human plasma was diluted with 4% phosphoric acid as described by the manufacturer’s protocol. According to a recent publication by de Jong and Teerlink [
Working calibration standards of L-arginine and ADMA dissolved in 10 mM HCl were used to construct a calibration curve for the comparison of the three sample preparation techniques (A), (B), and (C). Recoveries in % for L-arginine and ADMA were determined and are shown in Table
Determined average concentrations of L-arginine and ADMA in human plasma ±SD (
Human plasma, mean ± SD ( |
Procedure A | Procedure B | Procedure C | |||
---|---|---|---|---|---|---|
Added ( |
Measured, mean ± SD ( |
Recovered, mean ± SD (%) | Recovered, mean ± SD (%) | Recovered, mean ± SD (%) | ||
L-arginine | 65.3 ± 2.09 | 20 | 94.3 ± 10.1 | 113 ± 8.21 | — | — |
40 | 115 ± 3.20 | 109 ± 2.78 | 68.4 ± 5.50 | 118 ± 3.62 | ||
60 | 138 ± 8.67 | 110 ± 6.92 | — | — | ||
ADMA | 0.40 ± 0.013 | 0.1 | 0.47 ± 0.030 | 91.5 ± 3.03 | — | — |
0.2 | 0.59 ± 0.070 | 98.0 ± 7.07 | 73.5 ± 4.22 | 94.7 ± 8.46 | ||
0.3 | 0.70 ± 0.032 | 98.9 ± 4.55 | — | — |
Figures
Representative chromatograms obtained from the analysis of human plasma using (a) procedure (A), PP with aqueous TCA; (b) procedure (B), PP with a mixture of acetonitrile and ammonia (90 : 10); and (c) procedure (C), SPE cleanup with Oasis MCX. Insets: zoom of the ADMA and SDMA peaks at 18.1 and 19.0 min, respectively.
The robustness of procedure (A) was initially tested by performing PP on ice with different time points (2, 8, 10, or 12 min). The robustness study did not show any difference between the obtained concentrations of L-arginine and ADMA; thus, this step was not found to be critical.
Validation results can be found in Tables
Validation data: calibration curve equation, correlation coefficient, LLOQ, ULOQ, intra- and interday precision, and average concentration ± SD quality control samples (human plasma from 6 volunteers) for L-arginine and ADMA using the selected sample preparation (A) with the developed chromatographic method.
Calibration curve equation | Correlation coefficient ( |
LLOQ ( |
ULOQ ( |
Intraday precision (CV, %) ( |
Interday precision (CV, %) ( |
Quality control ( | |
---|---|---|---|---|---|---|---|
L-arginine |
|
0.999 | 0.14 | 15 | 3.2 | 5.0 | 64.1 ± 10.3 |
ADMA |
|
0.994 | 0.012 | 0.50 | 3.4 | 3.3 | 0.27 ± 0.02 |
SDMA |
|
0.989 | — | — | 4.2 | 3.8 | 0.57 ± 0.09 |
Table
Stability data for human plasma prepared by procedure (A) and a calibration standard at ULOQ as well as long-term data at −80°C for human plasma and at −20°C for a calibration stock solution. The data are calculated as average of three determinations.
4°C | −80°C | −20°C | ||||
---|---|---|---|---|---|---|
Prepared human plasma | Calibration standard | Human plasma | Calibration stock solution | |||
24 h | 48 h | 24 h | 48 h | 6 months | ||
L-arginine | 96.1% | 95.4% | 91.1% | 87.5% | 95.3% | 91.5% |
ADMA | 92.8% | 93,5% | 91.8% | 89.6% | 95.0% | 103% |
In all three sample preparation procedures, baseline resolution could not be obtained between ADMA and SDMA, and both ADMA and SDMA were surrounded by larger unknown peaks. Optimisation of the HPLC method could possibly have improved the resolution, especially in the chromatogram obtained after performing procedure (C). However, in order to compare the sample preparation procedures, it was considered necessary to use exactly the same chromatographic method for all three procedures. Since baseline resolution was not attained, there is a chance of experiencing interference from adjacent peaks. However, the average concentration of ADMA found in the quality control samples did not indicate interference, since the observed concentrations were in the lower end of the ADMA range stated in literature [
The selectivity was surprisingly not markedly increased by performing the SPE cleanup (Figure
The supernatants obtained after performing procedures (A) and (C) were clear, as opposed to the relatively cloudy solution obtained after procedure (B), which further lead to recoveries below 75%. Zinellu et al. used a 10 : 90 mixture of ammonia : acetonitrile for PP. However, they performed additional steps of filtration and concentration of their samples, which allegedly improved the recovery.
The main difference between sample preparation procedures (A) and (B) in relation to (C) is that laborious steps in procedure (C) have to be performed, resulting in a diminished throughput of samples per day. This is not desirable in a clinical laboratory, where research studies often comprise many samples. In addition, the probability of introducing errors and the risk of analyte loss during performance of the several steps in the SPE cleanup makes this approach less preferable. Furthermore, the cost-effectiveness of (C) with acquisition of SPE cartridges and an increased workload in the laboratory are estimated to be around 10 Euros per sample, which is less attractive when compared to (A) with an estimated cost of 3 Euros per sample.
One argument often stated for performing the laborious SPE cleanups is the decreased contamination of the HPLC column and/or detector. However, when using procedure (A) and the developed analytical method including a washing step of the column, the pressure increase was not considered to be critical (around 5 bars per 100 human plasma samples) and could be eliminated by changing the precolumn. The same HPLC column has been used for several hundred human plasma samples without deterioration of the resolution.
Overall, the validation results were acceptable. The recoveries for L-arginine were generally found to be above 100%. The reason for this could be that the calibration standard used for spiking in the recovery experiment was dissolved in 10 mM HCl, which apparently affected the observed recovery of L-arginine. However, the level of L-arginine in human plasma has not been observed to deviate from literature values; thus, the issue is considered to be related to the setup of the recovery experiment. Regarding LLOQ, the found recovery for L-arginine was just above the acceptance criterion of 20%. However, since the LLOQ for L-arginine is well below the expected concentration of L-arginine in human plasma, the marginally higher variation for L-arginine at LLOQ is not anticipated to have any practical importance.
L-arginine, ADMA, and SDMA concentrations for the quality controls can be found in Table
The analytical performance of sample preparation procedure (A) consisting of simple PP with TCA is considered to be superior to procedure (B) and similarly as good as performing a SPE cleanup, procedure (C) in quantifying L-arginine, ADMA, and SDMA in human plasma using HPLC-FLD. The developed HPLC-FLD method with sample preparation procedure (A) proved to be both accurate with recoveries ranging from 91.5 ± 3.03% to 113 ± 8.21% and precise with CVs of no more than 5%. Furthermore, the sensitivity was acceptable with an LLOQ of 0.14
An earlier version of this work was presented as a poster at Copenhagen Symposium on Separation Sciences 2016.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
The authors thank Joan Frandsen for the excellent technical assistance. Anne Marie Voigt Schou-Pedersen and Jens Lykkesfeldt are partly supported by the Lifepharm Centre for