Determination Of Levetiracetam In Dried Blood Spot

Published Date: 02 Nov 2017

performance liquid chromatography tandem mass spectrometry 2

Background:A simple LC-MS/MS method was developed and validated for the 7

quantification of levetiracetam (LEV, Keppra®), a broad-spectrum antiepileptic drug 8

(AED) in dried blood spots (DBS). LEV was simply extracted with methanol spiked with 9

adenosine (ADE) as internal standard before LC-MS/MS analysis. The correlation 10

between the DBS and plasma concentrations of LEV was also determined. 11

Result:Linearity was from 0.067-60 μg/mL for LEV in DBS samples. The intra- and inter- 12

day accuracy and precision of the assay met validation acceptance criteria. DBS 13

concentrations were well correlated to the plasma concentrations (R

2

=0.9399), 14

asfraction of LEV bound to blood cells remains veryconstant (0.466 ± 0.041) over a wide 15

concentration range. Conclusion:The study illustrated that DBS could be used as 16

alternative matrixfor monitoringLEV in preclinical and clinical studies. 17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

3

Key Terms: 37

Levetiracetam: 38

Keppra®, (S)-α-ethyl-2-oxo-1-pyrrolidine acetamide,a second-generation, broad- 39

spectrum anticonvulsant medication used to treat epilepsy 40

Dried blood spots: 41

Technique for collecting small whole blood samples,typically in 5–50 µl, on filter paper 42

LC-MS/MS: 43

Liquid chromatography -tandem mass spectrometry that combines two powerful 44

techniques giving the chemical analyst the ability to analyze virtually any molecular 45

species; including, thermally labile, non-volatile,and high molecular weight species 46

Therapeutic drug monitoring: 47

A branch of clinical chemistry and clinical pharmacology that specializes in the 48

measurement of medication concentrations in blood 49

Preclinical Study: 50

A study to test a drug, a procedure, or another medical treatment in animals for support 51

of the safety and suitabilityof the new treatment 52

fBC: 53

The fraction of analyte bound to blood cells, assumed constant, that can be used to 54

predicate the plasma concentration of analyte 55

56

57

58

4

59

Levetiracetam(LEV, Keppra

®

), (S)-α-ethyl-2-oxo-1-pyrrolidine acetamide, is a novel, 60

broad-spectrum antiepileptic drug (AED)that is structurally unrelated to existing AEDs 61

[1]. Thus far, its mechanism of action has not beenunderstood completely.It has been 62

postulated to bind to the synaptic vesicle protein 2A (SV2A) in the brain [2,3], thereby 63

decreasing neurotransmission in epileptic circuits.LEV is rapidly and almost completely 64

absorbed following oral administration (> 95%). In addition, It displays a linear 65

pharmacokinetic profile over a wide range of therapeutic doses , insignificantplasma 66

proteins binding (< 10%) as well asreadily and adequately crosses the blood brain 67

barrier[4]. The drug’s minimal protein binding and non-involvement of hepatic 68

metabolismincur no clinically significant drug interactions that confer upon it an 69

exemplary safety profile. All these make it potentially useful for infants, young children 70

and adult with epilepsy [5]. Nevertheless, despite its desirable features, the relationship 71

between plasma concentrations of LEV and its clinical effect on different seizure types 72

and intractability has not been clearly determined [6-9]. Hence, it has been 73

recommended that therapeutic drug monitoring(TDM) of LEV be performed, 74

particularly in renal-impaired patients[5,6,10-13],elderly and children, where half-life of 75

the drug is extended [14] and shortened [15], respectively. 76

TDM as well as evaluation of systemic drug exposurein preclinical and clinical PK studies 77

have traditionally relied onanalysing plasma samples. However,there are recent 78

advocates on replacing this conventional sample matrix with dried blood spot(DBS) – 79

whole blood blottedand dried on paper, asDBS is simpler to collect, manage, transport 80

and store. Moreover, it reduces the blood volume required and the risk of HIV/AIDS and 81

other infectionduring long term monitoring [16,17].Therefore, immense interest has 82

been shown in the possible use of DBS as an alternative matrix to plasma with TDM of 83

several drugs already using DBS as matrices [18-20].Nevertheless, the use of DBS 84

monitoring of LEV has not been established, despiteseveral analytical methods reported 85

for the measurement of LEV in plasma and blood)[21]using gas chromatography (GC) 86

with nitrogen-phosphorus detection [21], GC-mass spectrometry (GC-MS) [22], 87

microemulsionelectrokinetic chromatography [23] andnumerous high performance 88

liquid chromatography (HPLC) techniques [24-27]. There are few reportedmethods [11, 89

13, 21, 22] for the analysis of LEV using LC-MS/MS,two of which [13, 21] employed solid 90

phase extraction (SPE), an elaborate and tiresome procedure. 91

Therefore,the objective of this study was to develop and validate a rapid, reliable, 92

selective, sensitive, and accurate LC-MS/MS method to quantify LEV in preclinical study, 93

5

to assess DBS suitability by fBCand relationship between the LEV levels in DBS and 94

plasmaas an alternative matrix for TDM of LEV. 95

Experimental 96

Chemicals and reagents 97

Levetiracetam (LEV) was purchased from Cell Molecular Pharmaceutical R&D Co., Ltd. 98

(Xi’an, China) and the internal standard (IS), adenosine from Sigma-Aldrich Co. 99

(Singapore). Water was purified through the use of Millipore

TM

Direct-Q 3 UV Water 100

Purification System (Singapore), while other chemicals and reagents were of analytical 101

grade and solvents were of HPLC grade. 102

Instrumentation and Chromatographic Conditions 103

TheLC-ESI-MS/MS system used was composed of a modelShimadzu UFLC system 104

(Shimadzu Scientific Instruments, Columbia, MD) coupled to a Q Trap

TM

3200 hybrid 105

triple quadrupole linear ion trap mass spectrometer(Applied Biosystems/MDS Sciex, 106

Concord, Ontario, Canada). Data processing was performed with Analyst

TM

1.4.2 107

software package (Applied Biosystems, MA., USA). 108

109

Chromatographic separation was performed on aZorbaxEclipse Plus C18 column (2.1 x 110

100 mm, I.D., 3.5 µm, Agilent Technologies, Palo Alto, CA, USA) with a Security Guard 111

Cartridge (3.0 X 4 mm, Agilent Technologies, Palo Alto, CA, USA). The column 112

temperature was ambient, the flow rate was 0.25 mL/min and the injection volume was 113

2 µL. The mobile phase A was 0.1% formic acid (FA) in Millipore

TM

water and mobile 114

phase B was 0.1% FA in methanol. The mobile phase gradient was as follows: initial 10% 115

B linear increased to 98% B in 0.3 min and held for2.6 min, finally dropped to 10% B 116

within 0.2min and equilibrated for 1.9 min. The total run time was 5min. 117

118

The mass spectrometer was operated using ESI sourcein the positive ion detection 119

mode for LEV determination. Acquisition was performed in multiple reaction monitoring 120

(MRM) mode using m/z171→126 and m/z268→136 for LEV and IS, respectively.The 121

optimized instrument parameters for monitoring theanalytes by mass spectrometry are 122

as follow: source temperature (TEM), 450°C; turbo spray voltage (IS), 5400 V; curtain gas 123

6

(CUR), 10 psi; Nebulizing gas (GS1), 40 psi; turbo ion spray gas (GS3), 40 psi; collision gas 124

(CAD), medium; and dwell time 200 ms. 125

Stock standards, calibration standard and quality control samples 126

Primary stock solution of LEV was prepared by dissolving an accurately weighted 127

amount in methanol: water (1:1) to yield 10 mg/mL. The stock solution was then further 128

diluted in methanol: water (1:1) to give working standard solutions with LEV 129

concentrations of 1, 3, 30, 100, 300, 720 and 900µg/mL. DBScalibration standards of LEV 130

were prepared by spiking 42µL fresh blank SD rat blood with 3 µL of a standard LEV 131

working solution, producing LEV blood concentrationof 0.067, 0.2, 2, 6.67, 20, 48 and 60 132

µg/mL. Aliquots (15 μl) of the calibration standards samples were spotted onto punched 133

out discs of Guthrie DBS paper (U.K. Neonatal Screening Laboratories Network- 134

Whatman 903

®

) and allowed to dry overnight at room temperature.The DBS calibration 135

standards and QC samples were used immediately uponcompletion of drying. The 136

quality control (QC) samples were similarly prepared in blank blood at the 137

concentrations of 0.2, 6.67 and 48µg/mL. The stock solution of the IS wasprepared in 138

methanol: water (1:1) in 1mg/mL and were further diluted in methanol to yield a 139

working stand solution of 3 µg/mL. 140

141

For assay of LEV in plasma samples, the working standard solutions with concentration 142

of 0.05, 0.15, 1.5, 5, 15, 36 and 45 µg/mL of LEV in methanol were prepared. Calibration 143

standards of LEV were prepared by spiking 20µL fresh blank SD rat plasma with 20µL of a 144

standard LEV working solution, producing calibration samples as equivalent to the 145

plasma concentrations of 0.05, 0.15, 1.5, 5.0, 15, 36 and 45 µg/mL.The quality control 146

(QC) samples were similarly prepared in blank plasma at the concentrations of 0.15, 5 147

and 36 µg/mL. The stock solution of the IS were prepared in methanol: water (1:1) in 148

1mg/mL and were further diluted in methanol to yield a working stand solution of 6 149

µg/mL. 150

151

All the solutions were stored at -20°C and brought to room temperature before use. 152

Pharmacokinetic (PK) study 153

The in vivostudy was carried out according to the "Guidelineson the Care and Use of 154

Animals for Scientific Purposes" (National AdvisoryCommittee for Laboratory Animal 155

Research, Singapore, 2004). The animal handling procedures were reviewed and 156

7

approved by the Institutional Animal Care and Use Committee of the National University 157

of Singapore (NUS). 158

159

Six male Sprague-Dawley rats, weighing 290 to 320g (9 weeks old) were purchased from 160

Comparative Medicine, Centre for Life Sciences of NUS. The rats were kept at a specific 161

pathogen-free animal facility (24°C, 60% relative humidity) and maintained on a 12-hour 162

light/dark cycle with free access to a rodent autoclavable diet and water. At 24 hours 163

before the study, the animals were anaesthetized, and a polyethylene tube was inserted 164

into the right jugular vein for blood sampling. Following 1 day of post-surgical recovery, 165

catheter patency was ensured prior to the intravenous administration of LEV constituted 166

in 0.9% saline (25.0 mg/ml) and dosed at 40 mg/kg. Serial blood samples (approximately 167

300 μL) from each rat were collected at 0, 5, 15, 30, 45min and 1, 2, 4, 6, 8, 24h. All 168

samples were placed into heparinized tubes. To prepare a DBS sample, 15 µL of blood 169

was spotted onto the disc (6 mm) punched out from the Guthrie paperand allowed to 170

dry overnight at room temperature prior to storage at -80°C. At the same time the rest 171

of the blood taken was centrifuged and the plasma was stored and afterwards used to 172

determine the plasma concentration of the same drugwith the developed and validated 173

LC-MS/MS assay described in previous studies [5,13,22]. The assay was further validated 174

by comparing the LEV levels in DBS and plasma samples collected from Sprague Dawley 175

rats. Correlation between the DBS and plasma levelsand fraction of LEV bound to blood 176

cells (f

BC) were computed. 177

Sample preparation 178

Prior to assay, calibration standards, QC samples and frozen rat samples were thawed at 179

ambient temperature. For DBS samples, extraction was carried out by adding 200 μL of 180

methanol containing the IS (3 μg/mL) to each Eppendorf tube containing a DBS 181

specimen. The samples were vortexed for 60 s, then sonicated for 5 min, followed by 182

centrifugation at 10 000 rpm (4°C) for 5 min. 150 μl of the clear supernatant was 183

transferred and 2 μL injected into the LC-MS/MS system. 184

185

For assay of LEV in the plasma sample, protein precipitation was carried out by mixing 186

20-µL aliquot of plasma sample and 20 µL methanol: water (1:1) with 160 µL of 187

methanol containing the IS (6 µg/mL). The samples were vortexed for 30s, followed by 188

centrifugation at 10,000 rpm for 10min. 120µL of the clear supernatant was transferred 189

8

into the 96-wells plate at 4°C. Then, a 2-µL aliquot was injected into the HPLC-MS/MS 190

system for analysis. 191

192

Method Validation 193

Assay validation was performed according to the FDAguideline[23]. Selectivity was 194

assessed by comparing six individuals DBS blank samples with those obtained from 195

spiking blank blood with LEV at the lower limit of quantification (LLOQ). Quantitative 196

analysis of LEV in DBS and plasma samples was performed using IS method. Calibrations 197

were built from peak area ratios of analyte vs. IS and linearity was assessed by weighted 198

(1/x

2

) least-squares analysis. During validation, the calibration curves were defined on 199

three different days based upon assays of the spiked duplicate samples, and QCs were 200

determined in five replicates on the same day. Accuracy and precision were also 201

assessed by determining QC samples at three concentration levels on three validation 202

days. The accuracy was expressed by relative errors(RE) and precision by relative 203

standard deviation (RSD). 204

The lower limit of quantification (LLOQ), defined as the lowest concentration at which 205

both precision and accuracy were less than or equalto 20%, were evaluated by 206

analyzing samples which were prepared in five replicates. 207

The extraction recoveries of LEV were determined bycomparing the peak areas of QC 208

samples to those of LEV added post-extraction in blank DBS sample. To evaluate the 209

matrix effect, a post-extraction addition method was also utilized. 210

The analyte stability in DBS samples under processing was evaluated at the 211

concentrations of QC samples. The bench-top stability was determined DBS samples at 212

ambient temperature for 24 h. The freeze-thaw stability was studied after three 213

successive freeze-thaw cycles at -80°C. Similar experiments were performed for long- 214

term stability evaluation after 30 days storage at -80°C. Stability was also assessed after 215

processed DBS samples in the instrument autosamplerat approximately4°C. The results 216

were compared with those QC samples freshly prepared. 217

218

9

Results and discussion 219

Method validation 220

Selectivity 221

Potential interference from endogenous compounds was investigated by analyzing rat 222

plasma of six different subjects. FIGURE1 shows the typical chromatograms of a blank, a 223

spiked DBS sample with LEV at lower limit of quantification (LLOQ) and the internal 224

standard. No significant interference or ion suppression from endogenous substances 225

was observed at the retention time of the analyte and IS. The retention times were 226

approximately 2.64 min and 2.82min for IS and LEV, respectively. 227

Linearity of calibration curves 228

The linear regressions of the peak area ratios versus concentration were fitted over the 229

concentration range of 0.067-60 μg/mL for LEV in DBS samples. The typical equation of 230

the calibration curves was as follows: y=2.3e

3

x+0.00434, r=0.9969 Where y represents 231

the peak ratio of analyte to IS and x represents the concentration of the analytes in 232

plasma. The correlation coefficient (r) exceeded 0.99, showing a good linearity among 233

the concentration range. 234

Precision and accuracy, LLOQ, Extraction recovery and matrix effect 235

The intra- and inter-day precision and accuracy forLEV in QC and LLOQ samples are 236

given inTABLE1. The lower limit of quantification (LLOQ) was 0.067µg/mL for LEV. The 237

intra and inter-day RE and RSD were less than 7%at LLOQ level. For all the QC samples, 238

the intra and inter-day precision was below 9% and the RE was from 2.8% to 7.5%. 239

Therefore, the method presented acceptable accuracyand precision. 240

The mean extraction recoveries were also shown in TABLE2. The RSD was less than 6% for 241

all recoveries throughout the entire standard concentration ranges, showing good 242

consistency. The matrix effects calculated were in the range of93% to 101%. Therefore, 243

ion suppression or enhancement from rat blood was negligible under the current 244

conditions. 245

Stability 246

DBS offers a simpler and better method for storage as it allows the samples to bemore 247

stable at room temperature.LEV in DBS samples proved to be stable at room 248

10

temperature, even for 24h (REs≤10%). Stability was also assessed using the processed 249

DBS samples in the instrument autosampler at approximately 4°C for 8h.The analyte 250

wasfound to bestable with percentage of LEV remaining at≥90%. Similar experimental 251

results were obtained for the three successive freeze-thaw cycles at -80°C andthe long- 252

term stability evaluation after 30 days storage at -80°C. 253

Application to PK study in the Sprague-Dawley rats 254

PK study 255

This validated method was successfully applied to the PK studies after intravenous 256

administration of 40 mg/kg LEV in Sprague-Dawley rats. Their concentration-time 257

profiles are shown in FIGURE2. Generally they showed linear PK profile, which conforms 258

with those found in other studies[11,21] 259

Correlation between DBS and Plasma 260

To determine a quantitative relationship between the LEV levels in DBS and plasma 261

samples, the corresponding DBS and plasma levels determined by the validated LC- 262

MS/MS method were compared (FIGURE3).The good correlation were shown(R

2

263

=0.9399) with the slope values higher than 1, whichindicated higher concentration 264

values determined from plasma samples. The DBS concentrations were 33% (SD 7.1%) 265

lower than the corresponding plasma concentrations.Since LEV binds minimally to 266

plasma proteins (< 10%)[4],the total plasma concentration of the drug will be 267

approximately equal to its unbound plasma concentration.As DBS concentration is a 268

measure of the whole blood concentration, the observation of the consistently lower 269

LEV DBS concentrations is probablyas a resultof thelower concentration of LEV in whole 270

blood. Indeed, erythrocytes often serve to dilute the drug in whole blood compared to 271

plasma due to the extra volume provided by red blood cell according to the literature 272

[24].This would explain the lowerLEV concentrationsin DBS samples and blood samples 273

when compared to the plasma concentration found in this study and other previous 274

study[21]. 275

Hematocrit is the most important determinant of whole blood viscosity. It impacts the 276

flux and diffusion properties of blood spotted ontoGuthrie paper,and subsequently, the 277

concentration of each sample [20,25,26]. This haematocrit effect was minimized in this 278

study, asexperimental inbred animals were used and they were expected to have fairly 279

11

uniform haematocrit value.Li et al.[24] modified a formula proposedby Eyles et 280

al.[27]and applied it to correlate the DBS analyte concentration,DBS

[analyte]

with 281

theplasma analyte concentration, plasma[analyte]

.The formula incorporates both 282

haematocrit and fraction of an analyte bound to blood cells(f

BC): 283

(DBS[analyte]/[1-haematocrit])×(1- fBC

) = plasma[analyte]

284

fBCwas originally assumed to be constant in this formula. As mentioned earlier, the 285

haematocrit should be fairly constant in our homogenous population of inbred 286

rats.Weherein chose to affix the rat haematocrit value at40%, as indicated by a previous 287

study[28]. With the fixed hematocrit value, the f

BCvalues were then computed. It was 288

found that the fBC

values remained very constant throughout the wide concentration 289

range, at0.466±0.041 (mean +SD), which supported the widely accepted assumption of 290

constantf

BC, thus further validating our study. 291

While we are accustomed to using plasma concentrations for TDM, it has been 292

recommended that the whole blood drug concentrationbe used since it gives a better 293

prediction of the therapeutic effectthan the plasmadrug concentration [18]. Clinically, 294

more variation in haematocrit values is anticipatedin patients, thusthe hematocrit levels 295

should be included when interpreting the correlation of LEV levels in DBS to plasma. 296

Constant f

BC

values of LEV wasfound in this study. Thus, when LEV DBS levels are 297

converted to their corresponding plasma levels, only thehematocrit valuesneed to be 298

included in the conversion. If the hematocrit values are anticipated to be fairly constant 299

in the study subjects, a population average value could be adopted in the calculation. 300

301

Conclusion 302

This study describes an alternative determination of LEV from DBS samples using LC-ESI- 303

MS/MS method. It has proved to be rapid, sensitive,selective, accurate and precise and 304

gave reliable results. The DBS concentrations were also found to be well correlated to 305

their plasma concentrations (R

2

=0.9399). In addition,f

BCremained constant over a wide 306

range of drug concentrations. Thus, it is noteworthy that when converting the DBS 307

concentrations to the plasma concentrations, the hematocrit levels need to be 308

considered in patients with anticipated fluctuationin the levels. Our findings illustrated 309

that DBS could be used as a matrix for quantification of LEV for PK studies and TDM. 310

311

Future perspective 312

12

The use of dried blood spots (DBS) as a sampling technique was introduced in the early 313

1960s. It has been routinely carried out on infantsand more recently, for the TDM of 314

several drugs. DBS offers a more patient-friendly and convenient alternative to 315

conventional venous blood sampling. DBS-LC-MS/MS, involving the use of highly 316

sensitive LC-MS/MS system to quantify DBS samples, is an emerging topic in the 317

pharmaceutical community and is expected to play anincreasing role in drug discovery 318

and TDM. Our study demonstrates the feasibility of such approach for quantifying LEV 319

by establishing the excellent correlation between DBS levels and conventional plasma 320

levels. This finding has opened up new avenues for the bioanalysis of LEV and will be 321

especially useful in the TDM of this anti-epilepticdrug. 322

323

Executive summary 324

The first paper to report a simple LC-MS/MS method for quantification of 325

levetiracetam in dried blood spot with a good LLOQ of 0.07 μg/mL. 326

The method developed was fully validated in terms of linearity, selectivity, accuracy, 327

precision and stability and proved to be robust andreliable. It was also successfully 328

applied to a preclinical study. 329

Good correlation between the LEV levels in DBS and plasma samples were found 330

(R

2

=0.9399). In addition, f

BC

(0.466 ±0.041; mean ± SD) remained very constant over a 331

wide drug concentration range. 332

The results support the use of DBS as a matrix for quantification of LEV for 333

pharmacokinetic studies and therapeutic drug monitoring. 334

335

336

13

337

338