Clinica Chimica Acta 429 (2014) 90–95

Contents lists available at ScienceDirect

Clinica Chimica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /c l inch im

HPLC determination of D-3-hydroxybutyric acid by derivatization with abenzofurazan reagent and fluorescent detection: Application in theanalysis of human plasma

Giorgio Cevasco a,b,⁎, Anna Maria Piątek a,1, Sergio Thea a

a Department of Chemistry and Industrial Chemistry, University of Genoa, Via Dodecaneso, 31-16146 Genoa, Italyb CNR, Institute of Chemical Methodologies, Rome, Italy

⁎ Corresponding author.E-mail addresses: giorgio.cevasco@unige.it (G. Cevasco

(A.M. Piątek), sergio.thea@unige.it (S. Thea).1 Present address: Chemessentia-SRL, Via G. Bovio, 6-28

0009-8981/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.cca.2013.11.030

a b s t r a c t

a r t i c l e i n f o

Article history:Received 2 October 2013Received in revised form 14 November 2013Accepted 26 November 2013Available online 4 December 2013

Keywords:Biological samplesClinical/biomedical analysisFluorescence/luminescenceHPLC

A simple and sensitive new method for the determination of D-3-hydroxybutyric acid (D-3-HBA) in humanplasma after derivatization is described. The proposed method is based on the reaction of (2S)-2-amino-3-methyl-1-[4-(7-nitro-benzo-2,1,3-oxadiazol-4-yl)-piperazin-1-yl]-butan-1-one (NBD-PZ-Val) with D-3-HBA in the presence of O-(7-azobenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate(HATU) and N-ethyldiisopropylamine (DIEA) to produce a fluorescent derivative. The formed derivativewas monitored fluorimetrically at λex = 489 nm and λem = 532 nm. The HPLC analysis was carried out byuse of a C18 analytical column (Synergy Hydro 150 mm × 3 mm, i.d., 4 μm)with a binary gradient elution pro-gram of 0.1% aqueous trifluoroacetic acid versus methanol. The method showed satisfactory linearity(r2 = 0.9997) in the range from 20 to 500 μmol/L. The limit of detection (LOD) of the method was 7.7 μmol/L,while the limit of quantitation (LOQ) was 25.8 μmol/L. The analytical method was successfully applied tohuman plasma samples from normal healthy subjects.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Serum concentration of D-3-hydroxybutyric acid (D-3-HBA) is an in-dicator of the status of hepatic fatty acid oxidation that, when enhancedresults in the increased production of acetoacetic acid which is con-verted into D-3-HBA by a reaction catalyzed by D-3-hydroxybutyratedehydrogenase. D-3-Hydroxybutyric acid represents a major form ofketone bodies, especially in such metabolic situations as starvation[1,2] and diabetes mellitus [2], by which hepatic fatty acid oxidationis enhanced.

For this purpose, several enzymatic assays for the determination ofthis compound have been reported [3–9]. The principle of these assaysis based on the enzymatic method of Williamson and coworkers [3],where the bidirectional reaction of D-3-hydroxybutyrate dehydrogenaseis used for reversibly converting D-3-hydroxybutyrate to acetoacetateand the increase or decrease in absorbance at 340 nm is monitored[3–5]. The enzyme-spectrophotometric method specifically determinesD-3-HBA and is simple and accurate; however this method is not sensi-tive enough and, therefore, it requires a large volume of the serum sam-ple to obtain a reliable value.

), anna.piatek@chemessentia.it

100 Novara, Italy.

ghts reserved.

Subsequently Yamato et al. developed a spectrophotometricmethod for the differential determination of ketone bodies employingp-nitrobenzene diazonium fluoroborate (diazo reagent) as a color-developing reagent [10]. Unfortunately, the diazo reagent lacks the re-quired specificity for certain clinical applications. Diazo reagent canreact with compounds such as oxaloacetate, antidiabetes mellitusdrugs and other drugs in current medical usage.

More recently, a method for analyzing D-3-HBA by gaschromatography–mass spectrometry (GC–MS) has been reported[11] that is applicable to blood and urine samples containing this com-pound at a concentration higher than 20 μmol/L. However, this methodhas only a moderate sensitivity and, furthermore, it requires specificand expensive instrumentation.

Following a different approach, Tsai et al. [12] and Hsu et al. [13]have developed a pre-column fluorescence derivatization methodfor the enantioselective determination of both enantiomers of 3-hydroxybutyrate in the tissues of normal and diabetic rats, employinga column-switching HPLC system.

Given these limitations and on the basis of our previous experi-ence in this field [14], we have developed and validated a sensitiveand convenient high-performance liquid chromatography (HPLC)method for the quantitative determination of the D enantiomer of3-hydroxybutyric acid.

Although its significance from a clinical standpoint is lower, L-3-hydroxybutyric acid is also present in biological fluids; thus, for the

91G. Cevasco et al. / Clinica Chimica Acta 429 (2014) 90–95

determination of the D enantiomer in serum, separation of the twoindividual enantiomers is required.

Two strategies havemainly evolved for the separation of a pair of en-antiomers by HPLC. The first one is the “direct method” using a chiralstationary phase (CSP) [15]. The second one is the “indirect technique”based upon diastereomer formationwith a suitable chiral derivatizationreagent [16–24]. This indirect method involving a derivatization step issuitable for the trace analysis of enantiomers in biological samples, suchas blood and urine, because a highly sensitive detection can be per-formed with the option of coupling analytes with suitable chiral re-agents which have a high molar UV–Vis absorptivity (ε), or highfluorescence (FL) quantum yield (Φ). The labeling of chiral compoundswith a reagent absorbing in the UV or Vis region is the most popularmeans of derivatization. Therefore, many UV labels have been appliedto the tagging of various functional groups. However, in some real sam-ples the absorptivity of the so-obtained derivatives is not high enoughto ensure an adequate sensitivity. Very recently, a sensitive methodfor the detection of trace amounts of enantiomers of lactic and 3-hydroxybutyric acids in human saliva has been reported [24]. However,the employed derivatization reaction is quite slow, 90 min, and themethod requires the use of UPLC–MS/MS technique, a quite expensiveinstrumentation often out of reach for a number of laboratories.

Thus, a further selective and sensitive detection system is requiredfor the analysis of small quantities of chiral compounds in complex ma-trices such as biological samples. The FL labeling is one of themost effec-tive techniques for quantitative determinations in biological specimens,in terms of sensitivity and/or selectivity.

We have previously synthesized a new chiral, fluorescent deriva-tization reagent, having the benzofurazan structure, (2S)-2-amino-3-methyl-1-[4-(7-nitro-benzo-2,1,3-oxadiazol-4-yl)-piperazin-1-yl]-butan-1-one (NBD-PZ-Val), which reacts easily and quantitativelywith D- and L-lactic acid to form the corresponding stable, highly fluo-rescent diastereomeric amides [14]. In this paper, we describe a highlysensitive RP-HPLC method for the determination of D-3-HBA inhuman plasma with fluorescence detection after pre-column derivati-zation with NBD-PZ-Val (Fig. 1) in the presence of a coupling agent(HATU/DIEA).

2. Experimental section

2.1. Chemicals

4-Nitro-7-piperazinobenzofurazan (NBD-PZ) andN-ethyldiisopropyl-amine (DIEA) were obtained from Fluka; O-(7-azobenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU) waspurchased from Merck (Germany); D,L-3-hydroxybutyric and D-3-hydroxybutyric acid sodium salts, (S)-2-(Boc-amino)-3-methylbutyricacid (Boc-L-valine; enantiomeric purity ≥99.0%), O-(benzotriazol-1-

Fig. 1. Derivatization of D-3-hydroxybutyric acid by me

yl)-N,N,N′,N′-tetramethyl-uronium hexafluorophosphate (HBTU),trifluoroacetic acid (TFA), N,N-dimethylformamide (DMF) and hy-drochloric acid were from Sigma–Aldrich. NBD-PZ-Val was synthe-sized in our laboratory [14]. All other reagents were of the highestpurity available.

2.2. Plasma samples

Plasma samples were obtained through a standardized procedure.Human whole blood was collected into tubes containing EDTA asan anticoagulant. Plasma was obtained immediately by the centrifu-gation of the blood at 2000 g for 15 min, and samples were stored at−80°C until analysis.

2.3. Sample preparation

Fortymicroliters of sample (either human plasma or 20–500 μmol/Lsolutions of sodium D-3-hydroxybutyrate or D,L-3-hydroxybutyratein water) was mixed with 40 μL of water and with 120 μL of thedeproteinizing solvent CH3CN, and the mixture centrifuged for6 min at 14,500 ×g. Following this, 40 μL of the resulting solutionwere added to 40 μL of 6 mmol/L NBD-PZ-Val in CH3CN in the pres-ence of 40 μL of HATU and DIEA (both 8 mmol/L in CH3CN), andthe resulting solution was incubated at 30 °C for 5 min, then acidi-fied with 80 μL of 1 mol/L HCl.

2.4. Apparatus and chromatographic conditions

HPLC analyses were carried out on a Agilent 1100 system (AgilentTechnologies, Inc.) equipped with an online vacuum degasser, ahigh-pressure gradient quaternary pump, a manual sample injector(loop 5 μL), a column oven and a fluorescence detector. Data analy-ses were performed using Agilent ChemStation software (AgilentTechnologies).

The HPLC separations of the derivatized racemic 3-HBA and D-3-HBA were performed on a reversed phase C18 column SynergyHydro (Phenomenex, Torrance, CA, USA) 150 mm × 3 mm (i.d.)with 4 μm particle size, protected by a Phenomenex C18 securityguard column (4.0 mm × 3.0 mm). 0.1% aqueous trifluoroaceticacid and MeOH were the mobile phases A and B, respectively. Theseparation was carried out at a flow rate of 600 μL/min, equilibratingthe column at 30% B; the gradient increased linearly up to 50% of B in20 min, then it reached 100% of B in 5 min and was further maintainedat 100% B for 2 min for washing the column. Re-equilibration was car-ried out over 8 min; total analysis timewas 35 min. The column heaterwas set at 30 °C. Typically, 5 μL of samplewas injected onto the column.Derivatized D-3-HBA was detected by fluorescence with excitation at489 nm and emission at 532 nm.

ans of NBD-PZ-Val to afford D-3-HBA-NBD-PZ-Val.

Fig. 2. Extracted ion chromatograms of derivatized samples of racemic and D-3-HBA (top)and full scan spectra (bottom).

Fig. 4. Time courses of the derivatization reactions of racemic 3-hydroxybutyric acid withNBD-PZ-Val: ♦ D-3-HBA; ■ L-3-HBA.

92 G. Cevasco et al. / Clinica Chimica Acta 429 (2014) 90–95

LC–MS analyses were carried out by means of an Agilent 1200 SLliquid chromatograph. The mass spectrometer was an Agilent 6430Triple Quadrupole equipped with ESI interfaces. ESI conditions forpositive ionization were: drying gas flow (nitrogen) of 12 L/min,capillary potential of 4000 V, nebulizer pressure of 50 psi and dryinggas temperature of 350 °C.

2.5. Calibration and linearity

Five standard solutions of D-3-HBA covering the concentrationrange 20–500 μmol/L were the calibration samples employed.These were analyzed in triplicate every day for four days. Linear re-gression analysis of the D-3-HBA peak area plotted against the nom-inal D-3-HBA concentration afforded the calibration parameters forD-3-HBA.

Fig. 3. Excitation and emission spectra of (A) N

2.6. Recovery and precision

Recovery of D-3-HBA from plasma was assessed by analyzinghuman plasma obtained from normal healthy subjects after additionof samples of D-3-HBA at various concentrations. Concentrations inspiked biological samples were determined in triplicate.

The intraassay (within-day) reproducibility of the methods wasestablished by replicate analyses (n = 3) of sample. The interassay(between-day) was established by replicate analyses of the samesample over four days.

3. Results

3.1. Synthesis of chiral derivatization reagent

Chiral derivatization reagent was synthesized as previously de-scribed [14]. In brief, the chiral reagent is obtained easily by reactingthe commercial reagents NBD-PZ and Boc-L-valine in the presence ofHBTU/DIEA at room temperature for 60 min.

3.2. Product identification

Derivatized samples of both racemic and D-3-HBA were analyzedby LC–MS to confirm that formation of the expected diastereomershad actually occurred. Full scan mass spectra were acquired usingelectrospray ionization (ESI) in positive ion mode (Fig. 2). The ex-tracted ion chromatogram (EIC, Fig. 2, top) displayed the signals ofthe separated diastereoisomers at 18.03 min (D-HBA-NBD-PZ-Val)and 19.48 min (L-HBA-NBD-PZ-Val), respectively. The mass spectrum

BD-PZ-Val and (B) D-3-HBA-NBD-PZ-Val.

Fig. 5. Chromatogram of derivatized enantiomers of 3-hydroxybutyric acid in water (blue: baseline).

93G. Cevasco et al. / Clinica Chimica Acta 429 (2014) 90–95

of each diastereomer (Fig. 2, bottom) showed the corresponding pro-tonated molecular ion at 435 m/z, the sodium and the potassium ad-duct at 457 m/z and 473 m/z respectively.

3.3. Fluorescence properties

Excitation and emission spectra of the chiral reagent and its deriva-tive with D-3-HBA in acetonitrile are reported in Fig. 3.

3.4. Time course study

Time course study on the derivatization reaction of racemic 3-HBAsodium salt with NBD-PZ-Val was carried out at 30 °C in the 1–60 min

Fig. 6. Chromatograms of derivatized D-3-hydroxybutyric acid: (a) in water, 50°

range; sample preparation has been already described in Section 2.3.The highest increase of the peak signal was observed in the vicinity of5 min. The peak area did not increase significantly even prolongingthe reaction time up to 60 min. Results of this study are displayed as agraph in Fig. 4. Thus, for the rest of this study, the conveniently shorttime of 5 min was adopted.

3.5. Chromatographic separation

Chromatogram of standard solution of the derivatized racemic 3-HBA in water is illustrated in Fig. 5 showing an excellent separation,whereas those of the derivatized enantiomer of D-3-HBA in water,human plasma, and human plasma spiked with standard solution

μmol/L, (b) in human plasma and (c) the same as b, spiked with 50° μmol/L.

Table 1Recovery of D-3-hydroxybutyric acid from human plasma.

Concentration of D-3-hydroxybutyric acid (μmol/L)

Spiked Measured ± SD % rec

0 44.24 ± 10.44 –

25 62.96 ± 10.19 90.950 81.98 ± 9.96 87.0100 120.17 ± 9.63 83.3200 211.70 ± 9.44 86.7

Table 3Determination of D-3-hydroxybutyric acid in human plasma of healthy subjects.

Subject Population Age Concentration of D-3-HBA (μmol/L)a ± SD

1 F 35 48.66 ± 9.692 F 28 25.83 ± 10.053 F 32 32.72 ± 9.944 F 26 30.08 ± 9.985 M 61 22.77 ± 10.10

a Calculated from calibration curves in water.

94 G. Cevasco et al. / Clinica Chimica Acta 429 (2014) 90–95

of D-3-HBA are shown in Fig. 6. The eluted peak related to the D-3-HBA was well separated with a retention time of 15.1 min.

3.6. Linearity, limit of quantification (LOQ) and limit of detection (LOD)

Calibration curve was obtained for D-3-HBA in the concentrationrange of 20–500 μmol/L. The regression equation was y = 2.16x +6.70 (r2 = 0.9997).

The LOD and LOQ values were reckoned by dividing three or tentimes the standard deviation of the blank to the slope of calibrationcurve, respectively. Standard deviation of the blank was estimatedby the regression residual standard deviation. The so-calculatedLOD value was checked by analyzing a sample spiked at the LODconcentration to ensure visual detection of the analytes. The LODand LOQ values were 7.7 μmol/L and 25.8 μmol/L respectively.

3.7. Recovery and precision

Analytical recoveries were assessed by measuring the concentra-tions of the analyte in human plasma spiked with four different con-centrations of standard solution of D-3-HBA, and were in the range83.3–90.9% (Table 1).

Intra-day and inter-day precisions were assessed by analyzing stan-dard solution of the derivatized enantiomer of D-3-HBA in human plas-ma (Table 2): intra-day values were below 1% and inter-day valueswere lower than 2.13%.

3.8. Analysis of D-3-hydroxybutyric acid derivative in human plasma

The concentrations of D-3-hydroxybutyric acid in human plasmafrom 5 healthy subjects were determined to assess the applicability ofthe proposed method to this biological fluid. Results are displayed inTable 3.

4. Discussion

The present study describes a useful indirect RP-HPLC method withfluorescence detection for the quantitative determination of D-3-hydroxybutyric acid in human plasma employing a new chiral derivati-zation reagent. The synthesis of this reagent is very simple and, owing tomild reaction conditions adopted, no racemization occurs. As a furtherevidence of this fact, the reaction of enantiomerically pure D-3-HBAwith the chiral reagent produced a single chromatographic peak, thusindicating that the chiral derivatization reagent was optically pure and

Table 2Intra- and inter-day reproducibility of D-3-hydroxybutyric acid in human plasma.

Intra-day (n = 3) Inter-day (n = 4)

Concentration ± SD RSD (%) RSD (%)

D-3-HBA (μmol/L) 99.83 ± 9.79 0.78 ± 0.07 2.13 ± 0.20

that racemization did not occur during the synthesis of this reagent orduring the derivatization reaction.

The proposed HPLC method, possessing quite satisfactory linearityand precision, has been successfully applied for the determination ofD-3-HBA in very small volumes (40 μL) of human plasma samples.

The diastereomer of L-3-HBA, which is well separated from the Ddiastereomer in buffer (Fig. 5), is hidden in plasma by a large peak(Fig. 6) most likely due to lactic acid.

As shown by data reported in Table 3 our values are in goodagreement with those from previous work (mean value = 27.5 ±19.5 μmol/L [10]).

It is worth noting that the present method is significantly moresensitive than any of the other few methods proposed for the deter-mination of D-3-HBA in plasma, in particular than that based on GC–MS, where LOD value was reported as high as 20 μmol/L [11].

As for the method where NBD-PZ and a rather complex column-switching HPLC system with chiral columns were used [12,13], wenote that the present procedure is definitely more convenient, as itdoes not require costly CSP columns and long incubation times, i.e.5 min versus 3 h. Unfortunately, no detection limits have been re-ported in these papers thus making any comparison on this pointimpossible.

5. Conclusion

This method provides a simple, fast, sensitive and efficient assayfor detection of D-3-HBA in plasma that avoids the need for high-priced CSP columns (or very expensive instrumentation such asGC–MS or LC–MS) and therefore could be introduced into the routineclinical laboratory practice.

Acknowledgments

The research grant awarded to one of us (A.M.P.) by FondazioneCARIGE (Genova, Italy) is gratefully acknowledged. Thanks are alsodue to the Compagnia di San Paolo (Torino, Italy) for the acquisitionof a gradient NMR probe.

References

[1] Ozand PT, Reed WD, Girard J, et al. Hypoketonaemic effect of L-alanine. Specificdecrease in blood concentrations of 3-hydroxybutyrate in the rat. Biochem J1977;164:557–64.

[2] Bates MW, Krebs HA, Williamson DH. Turnover rates of ketone bodies in normal,starved and alloxan-diabetic rats. Biochem J 1968;110:655–61.

[3] Williamson DH, Mellanby J, Krebs HA. Enzymic determination of D(−)-β-hydroxybutyric acid and acetoacetic acid in blood. Biochem J 1962;82:90–6.

[4] Brashear A, Cook GA. A spectrophotometric, enzymatic assay for D-3-hydroxybutyratethat is not dependent on hydrazine. Anal Biochem 1983;131:478–82.

[5] Harano Y, Ohtsuki M, Ida M, et al. Direct automated assay method for serum or urinelevels of ketone bodies. Clin Chim Acta 1985;151:177–83.

[6] Uno S, Takehiro O, Tabata R, et al. Enzymatic method for determining ketone bodyratio in arterial blood. Clin Chem 1995;41:1745–50.

[7] Byrne HA, Dornan TL, Tieszen KL, New JP, Hollis S. Evaluation of an electrochemicalsensor for measuring blood ketones. Diabetes Care 2000;23:500–3.

[8] Laun RA, Rapsch B, Abel W, et al. The determination of ketone bodies: preanalytical,analytical and physiological considerations. Clin Exp Med 2001;1:201–9.

95G. Cevasco et al. / Clinica Chimica Acta 429 (2014) 90–95

[9] Harano Y, Kosugi K, Hyosu T, Uno S, Ichikawa Y, Shigeta Y. Sensitive and simplifiedmethod for the differential determination of serum levels of ketone bodies. ClinChim Acta 1983;134:327–36.

[10] Yamato S, Shinohara K, Nakagawa S, et al. High-performance liquid chromatographydetermination of ketone bodies in human plasma by precolumn derivatization withp-nitrobenzene diazonium fluoroborate. Anal Biochem 2009;384:145–50.

[11] Hassan HMA, Cooper GAA. Determination of β-hydroxybutyrate in blood and urineusing gas chromatography–mass spectrometry. J Anal Toxicol 2009;33:502–7.

[12] Tsai YC, Liao TH, Lee JA. Identification of L-3-hydroxybutyrate as an original ketonebody in rat serum by column-switching high-performance liquid chromatographyand fluorescence derivatization. Anal Biochem 2003;319:34–41.

[13] Hsu WY, Kuo CY, Fukushima T, et al. Enantioselective determination of 3-hydoxybutyrate in the tissues of normal and streptotocin-induced diabetic rats ofdifferent ages. J Chromatogr B 2011;879:3331–6.

[14] Cevasco G, Piątek AM, Scapolla C, Thea S. A simple, sensitive and efficient assay forthe determination of D- and L-lactic acid enantiomers in human plasma by high-performance liquid chromatography. J Chromatogr A 2011;1218:787–92.

[15] Toyo'oka T. Resolution of chiral drugs by liquid chromatography based upon diaste-reomer formation with chiral derivatization reagents. J Biochem Biophys Methods2002;54:25–56.

[16] Toyo'oka T. Recent progress in liquid chromatographic enantioseparation basedupon diastereomer formation with fluorescent chiral derivatization reagents.Biomed Chromatogr 1996;10:265–77.

[17] Sun XX, Sun LZ, Aboul-Enein HY. Chiral derivatization reagents for drugenantioseparation by high-performance liquid chromatography based upon

pre-column derivatization and formation of diastereomers: enantioselectivityand related structure. Biomed Chromatogr 2001;15:116–32.

[18] Srinvas NR, Igwemezie LN. Chiral separation by high-performance liquid chro-matography. I. Review on indirect separation of enantiomers as diastereomericderivatives using ultraviolet, fluorescence and electrochemical detection.Biomed Chromatogr 1992;6:163–7.

[19] Srinvas NR. Evaluation of experimental strategies for the development of chiralchromatographic methods based on diastereomer formation. Biomed Chromatogr2004;18:207–33.

[20] Toyo'oka T. Fluorescent tagging of physiologically important carboxylic acids, in-cluding fatty acids, for their detection in liquid chromatography. Anal Chim Acta2002;465:111–30.

[21] Ilisz I, Berkecz R, Péter A. Application of chiral derivatizing agents in the high-performance liquid chromatographic separation of amino acid enantiomers: a re-view. J Pharm Biomed Anal 2008;47:1–15.

[22] Bhushan R, Kumar V. Synthesis of chiral hydrazine reagents and their application forliquid chromatographic separation of carbonyl compounds via diastereomer forma-tion. J Chromatogr A 2008;1190:86–94.

[23] Bhushan R, Dixit S. Application of hydrazino dinitrophenyl-amino acids as chiralderivatizing reagents for liquid chromatographic enantioresolution of carbonyl com-pounds. Chromatographia 2011;74:189–96.

[24] Tsutsui H, Mochizuki T, Maeda T, et al. Simultaneous determination of DL-lactic acid and DL-3-hydroxybutyric acid enantiomers in saliva of diabetesmellitus patients by high-throughput LC–ESI-MS/MS. Anal Bioanal Chem2012;404:1925–34.