Quantification of peptides in human synovial fluid using liquid chromatography–tandem mass spectrometry

Quantification of peptides in human synovial fluid using liquid chromatography–tandem mass spectrometry.

Araceli Garcia-Ac

_{, Sung Vo Duy}

_{, · Sébastien Sauvé}

_{· Florina Moldovan}

_{· V Gaëlle Roullin}

_{· and Xavier Banquy}

1

_{Canada Research Chair in Bio-inspired Materials and Interfaces, Faculty of Pharmacy, Université de Montreal, Montreal, QC}

H3C 3J7, Canada.

2

_{Department of Chemistry, Université de Montréal, Montreal QC, H3C 3J7.}

3

_{Centre Hospitalier Universitaire St. Justine, Université de Montréal, Montréal, Québec,}

4

_{Pharmaceutical Nanotechnology Laboratory, Faculty of Pharmacy, Université de Montréal, Montreal QC, H3C 3J7, Canada.}

* Corresponding author:

xavier.banquy@umontreal.ca

.

Abstract A method to explore the stability of two anti-inflammatory peptides in human synovial fluid (HSF) has been

developed and validated using liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). The two peptides

are BQ123 Cyclo(-D-Trp-D-Asp-L-Pro-D-Val-L-Leu, Mw = 610.7) and R-954 (AcOrn[Oic

_{,( αMe)Phe}

_{, DβNal}

_,

Ile

_]desArg

_{-bradykinin, Mw = 1194.4). Human synovial fluid samples were analyzed after a protein precipitation step with}

acetonitrile and dilution with mobile phase. DMSO was used as anti-adsorptive agent. We used an octyl silane column with

formic acid (0.1%, v/v) in water as the aqueous mobile phase and acetonitrile isopropanol-formic acid (20:80, 0.1 v/v) as the

organic mobile phase and 0.7 mL/min flow rate. The peptides CY-771 and pepstatin A were used as internal standards.

Selective detection was performed by tandem mass spectrometry with a heated electrospray source (HESI), operated in

positive ionization mode and in selected reaction monitoring acquisition (SRM). The method limit of quantification (injection

volume = 10 µL) was 0.17 ng and 1.2 ng, corresponding to 28 and 102 nmol L

_{for BQ123 and R-954 respectively in human}

synovial fluid. Calibration curves obtained using matrix-matched calibration standards and internal standard were linear from

20 to 1000 nmol L

_{. Precision values (%R.S.D.) were ≤14% in the entire linear range. Accuracy measured at a low and a high}

concentration level ranged from 93.1 % to 102 %. The recoveries (at 800 nmol L

_{) were 96.4 % for BQ123 and 102.0 % for}

R-954. The method was successfully applied to follow the degradation kinetics of both peptides in human synovial fluid from

arthritic patients during 72 hours.

Keywords Quantitative bioanalysis · Tandem mass spectrometry ·Peptide quantification ·B1 receptor· Endothelin receptor

antagonist · Synovial fluid · Knee osteoarthritis · Distribution coefficient

1

2

3

4

5

6

7

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

Introduction

Due to increasing insights in the diverse roles of peptides in physiological processes, peptide-based pharmaceuticals

are expanding as a relatively novel class of strategies for treatment. Indeed, peptides represent an attractive point for the design

of novel therapeutics. Their specificity has been observed to translate into excellent safety, tolerability, and efficacy profiles in

humans. The biological functions of peptides are multiple, including carrying signals between glands and cells, structural and

mechanical roles, controlling body functions, acting as chemical transporters, process regulators and many others [1]. They are

also used as drugs for the treatment of diseases, typical examples being insulin, growth hormones, and therapeutic antibodies

[2]. The unprecedented number of marketing approvals in 2012 for peptide therapeutics could be a harbinger for the innovative

peptide-based drugs [3]. New synthetic strategies for limiting metabolism and alternative routes of administration have

emerged in recent years and resulted in a large number of peptide-based drugs that are now being marketed [4]. In parallel, the

increasing importance of biopharmaceuticals necessitates performance improvements in bioanalytical techniques. In particular,

their quantification in complex biological samples such as human synovial fluid (HSF) requires acute sensitivity and

selectivity, as all biological matrices contain a countless number of proteins, all made of the same 20 amino acids (AA)

precursors. In the present study, we focused on two peptides of interest for the treatment of osteoarthritis, namely; BQ123

(Cyclo(-D-Trp-D-Asp-L-Pro-DVal- L-Leu) and R-954 (AcOrn[Oic

_{,( αMe)Phe}

_{, DβNal}

_{, Ile}

_]desArg

_{-bradykinin).}

Osteoarthritis is the most common form of arthritis, affecting millions of people in the United States and around the world. It is

a complex disease whose etiology bridges biomechanics and biochemistry [5]. No curative treatment exists currently and only

antiinflammatory drugs are routinely prescribed for patient relief. Recently, the therapeutic effects of peptides on arthritis have

been associated with reduction of the autoimmune and inflammatory responses [6, 7]. In that context, both BQ123 and R-954

have shown promising activities.

BQ123 is an endothelin receptor A (ETA) antagonist [8] which blocks the ETA receptor,

resulting in a decrease of lipid peroxidation products [9], increase of the reduced glutathione (GSH) level, and enhanced super

oxidase dismutase (SOD) activity [10]. Moreover, the blockage of the ETA receptor alleviates the lipopolysaccharide

(LPS)-induced oxidative stress [11]. LPS is the main causative agent inducing sepsis, stimulating macrophages to excrete large

amounts of inflammatory biomarkers.

On the other hand, R-954 is a potent, selective and stable peptide antagonist of the

inducible G protein-coupled receptors (GPCR) kinin B1. The kallikrein–kinin system is composed of two major proteolytic

systems that are responsible for the release of pro-inflammatory substances, particularly bradykinin. R-954 shows potential

applications for the treatment of several diseases, including cancer, neurological disturbances of diabetes [12] and

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

osteoarthritis[13]. Previous work showed that both endotelin-1 (ET-1) and bradykinin play a crucial role in vivo in OA

pathophysiology, through their catabolic functions in cartilage degradation and inflammation involving metalloproteases

MMPs (MMP-1-and MMP-13), nitric oxide (NO), and prostaglandins (PGE

)[13, 14]. From a structural standpoint, both

peptides display significant chromophore groups for UV detection with conventional HPLC methods. Intrinsic peptide

fluorescence, originated with the aromatic amino acids tryptophan (TRP) and phenylalanine (PHE) [15] from BQ123 and

R-954 respectively, precludes analyses with spectrofluorometric methods. Nevertheless, those techniques are not appropriate

when only small sample volumes (≤ 200 µL) are available, such as the ones retrieved during in vivo studies in small rodents.

Therefore, the lack of a significant amount of synovial fluid led us to use LC-MS/MS. Determination of dietary supplements

and prescription drugs have been studied in HSF in a single class of compounds[16] [17, 18] or lipid study profiles[19].

However, to the best of our knowledge, no bioanalytical method for the determination of the BQ123 and R-954 peptides in

synovial fluid has ever been developed. One of the major difficulties faced when developing this analytical method was the

HSF composition. The presence of hyaluronic acid (HA) in HSF and the low concentration of proteins (10-30 mg/mL) [20],

which, unlike plasma (66-89 mg/mL) [20-22], can cause adsorption and loss of the studied analytes. The most abundant

macromolecules in synovial fluid are the sodium salt of hyaluronic acid (3 mg.mL

_{) and blood plasma proteins (albumin, ca.}

11 mg.mL

_{and globulins, ca. 7 mg.mL}

_{) [23] [24]. In this article, the nonspecific binding of the analytes to the wall of the}

peptide containers was assessed. Studies on peptides have indicated considerable adsorption to solid surfaces even to

siliconized glass and polypropylene materials [25]. Such losses may be due to the hydrophobicity of the peptide, the

physicochemical properties of the plastic, and laboratory manipulations [26-28]. This article describes the development of a

sensitive, accurate, and reliable label-free procedure, based on liquid chromatography coupled to tandem mass spectrometry

(LC-HESI-MS/MS) is d for the selective and simultaneous quantification of two peptides in human synovial fluid. The overall

goal of our research was to monitor the stability of the R-954 and BQ123 peptides in a new intra-articular nanoscale drug

delivery system for the treatment of knee osteoarthritis (OA). In this context, a sensitive analytical method capable of detecting

concentrations in the low nanogram–per-mililiter range is essential for the monitoring of the target peptides in human synovial

fluid.

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

Experimental methods

Chemicals

BQ123, Cyclo(-D-Trp-D-Asp-L-Pro-D-Val-L-Leu), 95.51% pure was purchased from China peptides Co. R-954

(AcOrn[Oic

_{,( αMe)Phe}

_{, DβNal}

_{, Ile}

_]desArg

_{-bradykinin) (purity >99% by HPLC) was kindly supplied by IPS Pharma,}

Sherbrooke Canada. Two internal standards were used, named pepstatin A (PEP A), Isovaleryl-Val-Val-Sta-Ala-Sta purity ≥

90% obtained from Sigma–Aldrich Canada Ltd (

Oakville, ON,

Canada ) and CY-771 (CY),

Val-Trp-Glu-Leu-Leu-Ala-Pro-Pro-Phe-Asn-Glu-Leu-Leu-Pro (purity 98.60%) supplied by China Peptides Co (Shanghai, China). Formic acid (F.A., 98%

pure) and dimethyl sulfoxide (DMSO) were purchased from Sigma–Aldrich Canada Ltd. LC–MS grade acetonitrile (ACN),

H

O and 0.1% F.A. in H

O were purchased from J.T. Baker (Phillipsburg, NJ, USA). Human synovial fluid was obtained from

the Musculo-Skeletal tissues Bank, Centre Hospitalier Universitaire CHU Sainte Justine (project No 2252). The synovial fluid

was collected during aseptic arthrocentesis, arthroscopy or total knee replacement surgery of arthritic patients. For this study

we used five synovial fluids, collected and preserved in distinct tubes, anonymized and pooled before we used it. Samples were

centrifuged at 2,000×g for 10 min to remove any dispersed solid matter, and then frozen at −20 ◦C in polypropylene tubes as

five separate 800 µL aliquots.

Standard solution preparation

Stock solutions with a nominal concentration of 1 mg mL

_{of R-954 in methanol-water-DMSO (30/69/1 v/v %), BQ123 in}

DMSO and PEP A in methanol-water-F.A. (30/69/1 v/v %) were prepared. A stock solution of the internal standard CY was

prepared at a nominal concentration of 0.7 mg mL

_{in methanol. To obtain the calibration curve and quality control (QC)}

samples in neat solution, six intermediate solution standards of BQ123 and R-954 ranging from 0.02-10 µmol L

_were

prepared in acetonitrile containing 1.7% DMSO. The internal standard (IS) working solution contained 5 µmol L

_{of CY and}

PEP A in acetonitrile. These solutions were used to create a seven-point matrix-free calibration curve ranging from 0.02 to 1

µmol L

_{, by mixing 100 µL of intermediate solutions of BQ123 and R-954 (or 100 µL of acetonitrile for zero point) with 100}

µL of internal standard solution (CY and PEP A) at 5 µmol L

_{with 800 µL of formic acid 0.1% in water.}

The second calibration curve matrix-based was prepared with BQ123 and R-954 at concentrations ranging from 0.1 to 5 µmol

L

_{in HSF containing 1.7% DMSO and spiked with a volume of stock standard solution lower than 3% of the total HSF}

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

volume. Working standards in HSF at 0.5 and 4 µmol L

_{were used to prepare QC samples.}

_{The calibration curve and QC}

samples were prepared by mixing spiked HSF with acetonitrile to precipitate proteins. To that extent, 40 µL of synovial fluid

containing a known quantity of peptide were precipitated with a total of 120 µL of acetonitrile including 20 µL of internal

standard. The sample was vortexed and centrifuged at 1,400 rpm for 10 min in a microcentrifuge Spectrafuge 16M from

LabNet international, New Jersey. Then, 80 µL of the supernatant were transferred into a disposable autosampler glass vial and

20 µL of 0.1% F.A. in water were added prior LC-HESI-MS/MS analysis. Final concentrations injected into LC-MS/MS were

0.02 to 1 µmol L

_{and 0.1 and 0.8 µmol L}

_{for the calibration curve and QC samples, respectively. HESI source optimization}

solutions of 5 µmol L

_{of each compound were prepared in F. A 0.1%- acetonitrile (50/50 v/v %). All stock and working}

solutions in acetonitrile were stored at −20 ◦C in glass vials for no longer than 6 weeks, whereas the working solutions in HSF

and standard calibration curves were freshly prepared each day.

Stability of BQ123 and R-954 in human synovial fluid and buffer solutions

Stability studies of BQ123 and R-954 were carried out in human synovial fluid at 37°C. To investigate the stability, the

reaction mixture was treated with 1.7% v/v of DMSO and spiked with aliquots of 2 and 4 µL of stock solutions (1 mg mL

_{) of}

BQ123 and R-954 to reach a final concentration of 8.5 µmol L

_{, respectively. The mixture was incubated at 37°C in a water}

bath (Isotemp 210, Fisher Scientific). At specific time intervals (0, 5, 24, 29, 48, 53 and 72 h)

,

a 20.0-µL aliquot was removed

and processed for BQ123 and R-954 quantification as follows. 80 µL of IS working solution in acetonitrile were added to the

removed aliquot to precipitate proteins. The sample was then vortexed and centrifuged at 1,400 rpm for 10 min in the above

microcentrifuge. Thereafter, 30 µL of the supernatant were mixed with 70 µL of H

O- 0.1% F.A. and transferred into

disposable 1.5-mL autosampler vials (fitted with an insert) to achieve a 0.5 µmol L

_{final concentration of the corresponding}

peptide and IS for HPLC–ES-MS/MS analysis.

A similar study was carried out in three buffer solutions, 0.15 M sodium citrate buffer, pH 5 and 0.15 M phosphate buffers, pH

6 and 7.4 [29] at 37 °C. Reaction solutions were prepared with the described buffer solutions by adding 1.7% v/v of DMSO

and diluting 1 mg mL

_{stock solutions of BQ123 and R-954 with appropriate volumes of 0.15 M sodium citrate and 0.15 M}

phosphate solutions to a final concentration of 8.5 µmol L

_{. The resulting spiked buffer solutions with BQ123 and R-954 were}

incubated separately in a 37°C water bath as previously described. At 0, 5, 24, 29, 48 and 53 h following incubation start, a

10.0-µL aliquot was removed, diluted 20 times with acetonitrile-H

O-FA (30:69.9:0.1 v/v), containing 0.5 µmol L

of internal

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

standard and achieving a 0.4 µmol L

_{final concentration for the peptides. Three calibration curves in 0.15 M sodium citrate}

buffer, pH 5, 0.15 M phosphate buffers, pH 6 and 7.4 were also performed at the same time.

Instrumentation, LC–HESI-MS/MS analysis

Analysis was performed using a liquid chromatography-tandem mass spectrometry (LC–MS/MS) from Thermo Fisher

Scientific (Waltham, MA). It consisted of a LC Accela system (HTC Thermopal autosampler and Accela 1250 quaternary

pump) coupled with a heated electrospray ionization (HESI) source and a TSQ Quantiva™ Triple Quadrupole Mass

Spectrometer. The analytical column was a Luna C8 100 Å (5 µm particle size, 30 mm × 2.0 mm i.d.) protected by a

Phenomenex security guard cartridge C8 (4 mm × 2.0 mm i.d.); both columns were from Phenomenex (Torrance, CA). Mobile

phases were F.A. (0.1%, v/v) in water (solvent A) and acetonitrile-isopropanol (20:80 v/v) containing 0.1% F.A. (solvent B).

Separation was achieved at a 0.7 mL min

_{flow rate under gradient elution conditions with 5% of eluent B during 1.0 min and}

rising from 5 to 98% in 1.5 min, holding for 1 min and then returning to 5% for re-equilibration during 1.4 min. Total run time

was 5 min for each injection. Injected sample volume was 10 µL in a full-loop mode. The analytical column was maintained at

28⁰C to improve reproducibility and retention time values. A typical chromatogram of BQ123, R-954 and their internal

standards showing both MRM transitions can be found in Fig. 1. Measure of the exact mass was performed using a Q-Exactive

Orbitrap mass spectrometer controlled by the Xcallibur 2.3 software (Thermo Fischer Scientific, Waltham , MA, USA).

Orbitrap parameters were set as follows: AGC (maximum capacity in C-trap) was set at 5x10

_{, maximum injection time at 50}

ms and resolution at 70, 000 FWHM at 200 m/z. The mass scan range was set at 150-1000 m/z (full scan MS mode).

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

Figure 1. (A) Chromatogram of a blank solvent injected immediately after the 1 µmol L

_{BQ123 and R-954 standard and their}

respective internal standards at a concentration of 0.5 µmol L

_{showing the absence of carry-over after the injection of the}

highest calibration curve level. (B) Chromatogram showing both MRM transitions of a solution standard in mobile phase

containing 0.25 µmol L

_{BQ123 and R-954 and their respective internal standards at a concentration of 0.5 µmol L}

_eluted

with the proposed gradient and mobile phase. (C) Chromatogram showing both MRM transitions of a standard solution in HSF

spiked with 0.25 µmol L

_{BQ123 and R-954, after the proposed extraction method by protein precipitation.}

Mass spectrometry parameters

Heated electrospray ionization (HESI) was performed in positive mode using a spray voltage of 2.7 kV. The sheath and

auxiliary gas (N

) were set to 35 and 5 arbitrary units, respectively. Ion transfer capillary temperature was set to 350 °C,

vaporizer temperature 400 °C, and ion sweep gas pressure to 2 arbitrary units. The tandem mass spectrometer was operated in

selected reaction monitoring (SRM) mode.

The first and third

quadrupoles

of the triple quadrupole MS/MS instrument were

operated at 0.7 FWHM

. In order to select precursor ions for each peptide, chromatograms were initially recorded in full scan

mode. In all cases, the pseudo-molecular [M+H]

_{or [M+2H]}

_{ion was selected. Once precursors were selected, full product}

ion scans were recorded, and collision energies were optimized to get the maximum intensity of the obtained fragment ions.

The pressure of the collision gas in the collision cell was set to 1.5 mTorr. The most intense SRM transition (SRM#1) was

selected for quantification and the second most intense (SRM#2) was used for confirmation. SRM transitions, collision energy

(CE) and RF lens were compound-specific and are provided in Table 1

.

Table 1. MS/MS transitions parameters for the different peptide analytes.

Estimation of compound adsorption and lipophilicity constants

To investigate the loss of the target peptides during the solution preparation, the extent of adsorption and lipophilicity constants

(Log D) were calculated. First, adsorption was assessed in HSF, plasma, phosphate buffer (PBS), and mixes of organic

solvent-water (10% v/v) such as acetonitrile, acetone, and DMSO. Adsorption was measured at a low and a high concentration

level (0.05 and 1 µmol L

_{) of peptide by a simple quantitative experiment consisting of seven consecutive steps, transferring}

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

the test solution from one polypropylene vial (0.5 mL) to another followed by LC–MS quantification. Three 20-µL aliquots of

each solution were pipetted from each test tube, diluted with mobile phase and followed by LC-MS/MS quantitation. HSF and

plasma samples were processed with acetonitrile precipitation and their corresponding dilution with mobile phase, as described

in the stability study section.

Theoretical calculation for the prediction of physical chemical properties for organic compounds can be used to estimate

lipophilicity constants. Nevertheless peptides widely vary, with large ranges of physical size and formal charge. In the case of

peptides, their physicochemical properties are determined by the analogous characteristics of the amino acids which constitute

them. The huge combination possibilities of these building blocks and their structural modifications cause hard-to-predict

behaviors with regard to physical properties, including solubility, pKa and adsorption

.

In a recent study comparing the

prediction capacity of different modelling tools, no statistically significant relationship between the length of peptide and their

respective lipophilicity descriptor was evidenced. The affinity constant values for cyclic peptides were under-predicted and

those of unblocked peptides were over-predicted [30]. Since R-954 and BQ123 may coexist in different forms (ionized or not)

according to the pH, the values of their molecular physical property log D were experimentally assessed. Direct measurement

by the shake flask method [31] in octanol/water was performed to measure solubility. Briefly, octanol and an appropriate

phosphate buffer in water were the standard solvents for the equilibrium experiments. Three stock bottle of high purity

analytical grade n-octanol with a sufficient quantity of the each tested buffers (pH 1.2, 4.5 and 6.8) were saturated. Both

solvents were shaken for 24 hours on a mechanical shaker. Solvents were then allowed to rest for two days to facilitate phase

separation. The active ingredients at a concentration of 25 mg mL

_{were perfectly solubilized in DMSO, 20 μL of this solution}

were dissolved in 480 μL of n-octanol in 1.5 mL eppendorf tubes. They were stirred for 5 min on a Vortemp at 500 RPM,

followed by centrifugation for 5 min at 2000 RPM. A volume of 500 μL of buffer (pH 1.2, 4.5 and 6.8) was added and placed

on a mechanical shaker (Vortemp) for 72h at 25°C ± 1°C. The separation of the two phases was achieved by centrifugation

without temperature control. The resulting liquids were left for equilibration at 23°C ± 1°C for at least one hour before

analysis. The aqueous phase was sampled by minimizing the risk of including traces of n-octanol. To sample the aqueous

phase and avoid cross contamination with the organic phase a syringe initially filled with air was used. The concentrations of

the test substances in both phases were determined by HPLC-UV, using substance-specific methods. The total quantity of the

substance present in both phases was calculated and compared with the quantity originally introduced.

Sample preparation for the matrix effect study

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

Despite the use of internal standards to compensate the unexpected interferences from HSF with the analyte response, the

impact of matrix effects was specifically assessed. Suppression or enhancement of the signal was investigated using the same

mass spectrometric method, for each peptide. Six standards in neat solution (30 v/v acetonitrile in formic acid 0.1%) containing

an increasing concentration of BQ123 and R-954 (0.02, 0.2, 0.4, 0.6, 0.8 and 1 µmol L

_{) were injected in parallel with six}

matrix standards spiked with the same increasing amounts of BQ123 and R-954 in HSF and following the proposed method

using acetonitrile to effect the deproteinization. A final concentration of 0.5 µmol L

_{of the internal standards was fixed in all}

standards (in neat solution and matrix). Both calibration curves were injected in triplicate, two at the beginning and one at the

end of the sample list. The matrix effect was then determined by comparing the slopes of the matrix-matched calibration

curves to those in the neat solution[32].

Method validation

The validation was performed according to the EMA 2012 guideline on bioanalytical method validation [33]. During method

validation and real sample analysis, a blank of HSF was spiked with the two peptides of interest using them as references to

prepare calibration standards and quality control samples. In addition, internal standards for each of the peptides were added

during sample processing of the chromatographic runs. Selectivity, carry-over, method limit of detection (LOD), method limit

of quantification (LOQ), calibration range (calibration curve performance), accuracy, precision, and matrix effects were

evaluated. Selectivity of the method was proved by preparing and evaluating six different blank synovial fluid samples to

investigate potential interferences of endogenous compounds. The chromatogram of blank synovial fluid sample was compared

with those of HSF samples spiked with BQ123, R954 and their internal standards. The error tolerated on the relative retention

time of each of the tested peptides was ±0.5 min. Carry-over was addressed during method development; it was assessed by

injecting blank samples after a calibration standard at the upper limit of quantification. LOD and LOQ were determined as

respectively 3 and 10 times the standard error of the y-intercept divided by the slope of the calibration curve.

The calibration standards were prepared in HSF as the matrix of the study samples by spiking the blank matrix with known

concentrations of R-954 and BQ123 (range tested: 30-1194 ng mL

_{and 15-610 ng mL}

_{). Six calibration concentration levels}

were used, in addition to the blank sample (processed matrix sample without analyte and without IS) and a zero sample

(processed matrix with IS). Each calibration standard was prepared and analyzed in triplicate. The blank and zero samples

were not taken into consideration to calculate the calibration curve parameters. For each analysis, the calibration curve was

prepared using freshly-spiked standards. The accuracy was assessed on samples spiked with known amounts of the analyte, a

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

low level concentration 0.1 µmol L

_{and a high level concentration equal to 80% the maximal concentration of the calibration}

curve, they were also used as QC samples. These QC samples were spiked independently from the calibration standards, using

separately prepared stock solutions. They were analysed against the calibration curve, and the obtained concentrations were

compared with the expected value. The precision, expressed as the coefficient of variation (CV), described the reproducibility

of individual measures of the analytes. As well, intra-day and inter-day precisions were evaluated for a low QC sample

concentration (0.1 µmol L

_{). For the validation of the within-run precision, three QC samples at a low level concentration were}

prepared and injected the same day in a single run. For the validation of the between-run precision, the same process was

followed by performing evaluations during three different days. These samples were measured in triplicates on 3 different days

to assess linearity, accuracy, precision (intra-day and inter-day), LOD, LOQ. Each replicate was injected once in each series of

injections together with a six-point based calibration curve ranging from 0.02 to 1 µmol L

_{. The BQ123 and R-954}

concentration was calculated based on the peak area

ratio (analyte/IS)

of each SRM transition and using the mentioned

calibration curve. Resulting concentrations were computed using the embedded Xcalibur software 3.0 (Thermo Corp., USA).

3. Results and discussion

Method Development

Sample pre-treatment

Endogenous proteins were successfully eliminated through protein precipitation resulting in good recoveries,

meaning that the therapeutic peptides were not totally eliminated in the precipitated protein pellet, with recoveries

more than 96% of the initial mass. HSF was mixed with different parts of a water-miscible solvent, acetonitrile or

methanol, followed by sample centrifugation. The best results were obtained using three parts of acetonitrile to one

part of sample, producing a 4 and 12 fold higher response than using methanol for BQ123 and R-954, respectively.

Protein precipitation with acetonitrile, the best choice according to the composition and protein content of synovial

fluid, was additionally easy to perform, and required inexpensive chemicals and reagents.

MS/MS optimization

The full scan MS and ion product spectra of BQ123 and R-954 are shown in Fig. 2. Addition of 0.1% (v/v) formic

acid to the mobile phase improved the response of the four peptides compared to other mobile phase additives. For

this study, the LOQ for R-954 and BQ123 had to be as low as possible and MS settings were especially optimized

for these two analytes. Therefore, multiple product ions were selected for SRM analysis and the Q1 resolution was

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

maintained at 0.7 in order to keep the high absolute intensity at the LOQ concentration level. Heating of the

ESI-probe successfully helped to evaporate the solvent without degradation of the compounds and improved the signal of

the peptides at the applied flow rate of 700 µL/min. BQ123 generated the single charged ion as the most abundant

peak, [M+H]

_{at m/z = 611.3 in full scan MS (fig. 1A, left). The}

_{C peaks (insert) with a 1 unit of difference on the}

m/z scale showed that the charge state of the peptide ion that produce this peak cluster was 1. Following the selection

of the precursor, the collision energy was optimized and the full product ion scan was recorded (fig.1A, right). The

fragment ions obtained with maximum intensity were at m/z = 169.2 and m/z= 197.2. The ion product mass

spectrum of the protonated molecule showed fragments arising from a five cyclic ring-opened peptide. A stepwise

fragmentation of BQ123 by loss of aspartic acid (D), triphtophane (W) and leucine (L), produced the asylum

H-Proline (P)-Valine (V)

_{ion, m/z = 197.1. The spectrum was dominated by the m/z = 169.2 ion formed from the loss}

of the elements of CO of the fragment m/z = 197.1.

R-954, with monoisotopic mass 1193.6 u, gave the double-charged ion [M+2H]

_{as the most abundant peak at m/z =}

597.9, figure 1B-left. The insert shows the typical isotopic pattern for small peptides and their double-charged ions

with a distance between the main peak and the isotope of 0.5 amu. Biological molecules as R-954 of mass 1000 units

and larger show a considerable contribution of the

_{C isotope [34], due the naturally occurring}

_C/

_{C ratio (98.89%}

_{C, 1.11%}

_{C). Fragment ions with maximum intensity obtained after collision energy optimization were at m/z =}

170.0, and m/z= 134.3. Both fragments are immonium ions derived from the detachment, breakdown and the loss of

CO of the individual amino acids naphthyl-Alanine (naphthyl-A) and methyl-phenylalanine (methyl-F), respectively.

Due to the good charge stabilizing properties of the conjugated pi-electron systems, the most stable products ions are

those coming from the two amino acids with aromatic rings, giving rise to intense peaks [35]. This was confirmed by

measuring the accurate mass of the fragments using the Q-Exactive Orbitrap mass spectrometer and calculating the

mass deviations reporting the single mass measurement error [36] of 2.1 ppm for m/z= 170.1 and 2.0 ppm for m/z=

134.3.

Figure 2. Full scan MS (left) and ion product (right) mass spectra of the protonated ion under positive mode for

BQ123 [M+H]

_{(A) and the doubly charged R-954 [M+2H]}

_{(B). Inserts showing the isotope distribution of}

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

BQ123 and R-954 taking into account the naturally occurring

_C/

_{C ratio, (monoisotopic mass 610.3 u , C}

31

H

N

O

and 1193.6 u, C

H

N

O

, respectively). MS settings were as listed in table 1.

Liquid chromatography

Obtaining a fast analysis with good chromatographic performance for a mixture of greatly different peptides, each

showing distinct characteristics and in presence of certain endogenous compounds from the matrix, was quite

challenging. The BQ123 peptide showed high retention and poor peak shape using a C18 analytical column, even at

higher percentages of organic solvent in either the initial mobile phase or the injection solvent. The hybrid

polymer-silica C8 column allowed for a good retention of the four compounds in comparison to the other stationary phases

used. The length of the chromatographic run was extended to five minutes, which is sufficient for the elution of the

most retained sample components including a 1-min re-equilibration of the column at the initial chromatographic

conditions. One minute of aqueous phase was run before starting the gradient, to eliminate interferences of

compounds with greater polarity eluting at the same time as the peptides of interest. Moreover, the compounds were

eluted in a mobile phase containing a large amount (80%) of organic solvent, thus ensuring a high ionization

efficiency. The LC–HESI-MS/MS analysis showed suitable reproducibility of the chromatographic separation, and

no shortening or variability of the retention times was observed even after injection of large sample sequences, up to

80–100 injections. In the optimized analytical conditions the mean retention time of BQ123 and R-954 were of 2.27

± 0.03 min and 1.95 ± 0.02, respectively (%R.S.D. = 0.6% and 0.9, n = 20).

Estimation of peptide adsorption to solid surfaces and lipophilicity constants

One of the pitfalls encountered during the early development of this methodology was the immediate adsorption of

the two target peptides to the materials commonly used in the laboratory manipulations. Numerous factors could

impact the compound observed concentrations, accuracy, and precision of a bioanalytical method, including, but not

limited to, matrix effects and losses by adsorption. The first considered factor was the nonspecific binding of the

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

analytes to the wall of the peptide containers. To overcome this aspect and avoid severe underestimation of the

compounds’ concentrations in the assays, an adsorption study containing different anti-adsorptive agents was carried

out. The applied approach was a simple quantitative procedure previously used and described above [37, 38].

Several types of anti-adsorption agents, such as BSA[39, 40], acetonitrile[37], tween 20[41], tween 80[42] and

DMSO [43] among others, have been used for blocking adsorption in HSF and urine assays. In the present study,

DMSO, acetone, and citric acid were tested as anti-adsorptive agents. As observed in figure 3 (6

_{pair of bars),}

DMSO was identified as the best anti-adsorptive as samples pre-treated with DMSO helped to avoid losses,

producing a relative recovery percentage of 96.4% for BQ123 and 102.0% for R-954.

Figure 3, displays ten pairs of bars showing the relative loss by adsorption in four untreated samples, the test with

four antiadsorption agents and two treated samples with DMSO. Each pair of bar correspond to one experiment in

one matrix showing the response of a first bar without transfer and a second one after seven transfer steps in

polypropylene tubes. All the test were performed at a low level concentration of 0.05 µmol/L, except for the second

pair of bars, it was spiked to 1 µmol/L named high concentration. The first four pairs of bars, correspond to HSF and

plasma prepared without the antiadsorptive agent DMSO. The next four, correspond to the tests with the four

antiadsorptive agents. The last two pairs of bars, are the results when the HSF and PBS pH 6.5 were treated with

DMSO.

The relative loss by adsorption (-35%) was pronounced in untreated HSF samples at low concentration (first pair of

bars), due to the limited binding capacity of the wetted solid surface area, , as previously reported [37] and also,

probably due to HSF high salts levels and relatively low protein concentration. HSF originates from plasma, with

small molecules and smaller proteins freely diffusing from the plasma into the HSF. Although larger molecular

weight proteins are excluded, total protein concentration is somewhat lower in plasma [44]. Another reason is the

high ionic strength of HSF (0.16M)[45]; the high concentration of electrolytes in HSF may produce a “salting out”

effect, driving the studied compounds further to the hydrophobic interface. This is confirmed by the results obtained

in the PBS pH 6.5 test (4

_{pair of bars), where the concentration of salts is high and there is no protein. Adsorption}

effects were not observed in plasma (3

_{pair of bars), which may be due to the strong interaction protein –peptides,}

such as hydrogen bonding and hydrophobic interactions [46].

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

Figure 3. Relative recovery of BQ123 (above) and R-954 (below) without transfer step and after seven transfer steps

in polypropylene tubes. Both peptides concentration 0.05 µmol L

_{, * high concentration 1 µmol L}

_{. After every 45}

min storage at 10°C, 200 µL were transferred from one reaction vial, the first and seventh transfer were measured by

LC-MS/MS. Initial concentration of DMSO: 1.7 % (v/v) in the treated samples.

To elucidate the propensity to partition between the polypropylene surface of the transfer vials and the aqueous

matrices, HSF or phosphate buffers, the distribution coefficient (D) was experimentally assessed. D is the

appropriate descriptor of lipophilicity of ionisable compounds that may exist in different forms (ionized or not) at

any given pH, such as R-954 and BQ123, determining their propensity to adsorb to hydrophobic surfaces such as

plastic [47].

Distribution was compared to percentage of loss. As showed in figure 4, a good correlation was observed between

the percentage loss and log D for both peptides at pH 1.2, 4.5 and 6.8. In our hands a log D of 1.6, at its neutral

form was experimentally obtained, showing very good agreement with Paladino et al[48], that have reported a log P

= 1.55 for BQ123.

Figure 4. Percentage loss and log D as a function of pH for BQ123 and R-954 in buffer phosphates pH=1.2, 4.5 and

6.8, not treated with DMSO. Each point represents the mean of triplicates. Full lines and dashed lines are the

corresponding results of BQ123 and R-954, respectively.

Matrix effects

As depicted in Fig. 5, the presence of a matrix effect was assessed by comparing the slopes of both neat solution and

HSF calibration curves (both treated with DMSO 1.7% v/v) by means of a Student’s t-test. By comparing the

obtained t values, 4.033 for BQ123 and 1.341 for R-954 with their tabulated t (2.003), the study showed that R-954

was not subjected to matrix effect, while BQ123 was susceptible to signal enhancement. To reduce matrix effects,

internal calibration and matrix-matched calibration curves using HSF standards were used for the rest of the

validation study.

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

Figure 5: Comparison of HSF (dotted line) and neat solution (continuous line) calibration curves of BQ123 and

R-954 obtained during the matrix effect assessment.

Method validation

Method validation results are summarized in Table 2

.

Linearity in neat solution and HSF was suitable for BQ123

with determination coefficients R

_{≥0.99 and acceptable for R-954 (R}

_{>0.98). The method intra-day variation for}

R-954 was lower than 18%, while inter-day precision was between 5-13 %. Extraction recovery achieved for both

peptides was suitable (>95%). The LOD ranged from 0.008 to 0.031 µmol L

_{, while the LOQ ranged from 0.028 to}

0.102 µmol L

_{. During preliminary experiments, external calibration without IS correction was used, and a higher}

degree of variability was observed (data not shown). Therefore, it was decided instead to consider a further IS

correction to reduce variability to acceptable values (<15%).

Table 2. Method validation parameters for the analysis of BQ123 and R-954 in HSF and neat solution.

1

_{Results obtained using a calibration curve with HSF standards and following all the process.}

_{Results obtained using a calibration curve in neat solution treated with 1.7% of DMSO (%v/v).}

_{Equation and coefficient of determination for the calibration curve using SRM#1}

b

_{Accuracy of samples (n=3) spiked for a final concentration of 0.1 µmol L}

_{for hsf and mobile phase the same day.}

_{Accuracy of samples (n=3) spiked for a final concentration of 0.8 µmol L}

_{the same day.}

d

_{Relative standard deviation of analytes in hsf and neat solution (n=3)3 spiked for a final concentration of 0.1 µmol}

L

_{the same day.}

e

_{Relative standard deviation of analytes in hsf and neat solution (n=3) spiked for a final concentration of 0.1 µmol}

L

_{on 3 different days (N=12).}

f

_{Sample spiked for a final concentration of 0.8 µmol L}

_(n=3).

g

_{Calculated using the standard line slopes in human synovial fluid and neat solution.}

Stability study of BQ123 and R-954 in HSF and buffer solution

To demonstrate the applicability of the developed method, the stability of the two peptides in HSF and in three

buffer solutions (0.15 M sodium citrate buffer, pH 5 and 0.15 M phosphate buffers, pH 6 and 7.5 at 37 °C) during

the proposed period of time was investigated. The results are summarized in Figure 6. BQ123 and R-954

concentrations in HSF both at 8.5 µmol L

_{were not impacted. Thus, both peptides were found to be stable in human}

393

394

395

396

397

398

399

400

401

402

403

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

synovial fluid. R-954 and BQ123 at 8.5 µmol L

_{were also found to be stable in buffers, with no decrease in}

concentration at 37° C for at least 53 h.

Figure 6. Effect of different physiological conditions on the in vitro stability of BQ123 and R-954 in HSF and

buffers pH 5, 6 and 7.4 incubated at 37 °C for 77 h in HSF and 48 h in buffers. Points represent the experimental

data and the solid lines were drawn using linear least-squares regression analysis. Each value represents the mean of

triplicate samples.

Conclusion

Protein precipitation provided a fast and simple technique for peptide analysis. The proposed LC-HESI-MS/MS

method using matrix-matched calibration curves yielded suitable results, without requiring analyte-HSF sample

cleanup.

The proposed LC–HESI-MS/MS method for determination of BQ123 and R-954 in synovial fluid fulfilled the

acceptance criteria generally established for bioanalytical assays in pharmaceutical analysis. In the explored

concentration range (up to 610 ng mL

_{BQ123 and 1194 ng mL}

_{R-954), the method proved to be selective,}

accurate, precise, and sensitive enough to allow analysis of BQ123 and R-954 in 20-µL human synovial fluid

samples. The method, which does not require pre-analytical derivatization, can be directly applied after a simple

protein precipitation clean-up step, thus reducing analytical variability and shortening sample processing. The

combined use of internal standardization and matrix-matched calibration allowed compensating ion signal

suppression and improving accuracy. Using the herein method, up to 60 samples per day can be analyzed with an

autosampler system. The LOQ value of the method is adequate to quantify BQ123 and R-954 in vitro. The validated

method has been applied to the determination of both peptides in synovial fluid and buffer solutions after a

degradation kinetics at 37°C for 53 consecutive hours.

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

Acknowledgments

The authors are grateful to M.Sc. Martin Jutras at the Platform of Biopharmacy, Dr. Alexandra Fürtös at the

Department of Chemistry, Université de Montréal, for their helpful advice on LC-MS analyses and M. S Gaelle

Maggliocco for her valuable help during this work. We thank the Canadian Foundation for Innovation for the

support for the mass spectrometry instrumentation. AGA is grateful for the financial support of the TransMedTech

Institute. XB acknowledges the financial support of CIHR and CRC.

450

451

452

453

454

455

456

457

Referencesovative peptide-based drugs

[1] T. Uhlig, T. Kyprianou, F.G. Martinelli, C.A. Oppici, D. Heiligers, D. Hills, X.R. Calvo, P. Verhaert, EuPA

Open Proteomics, 4 (2014) 58-69.

[2] A. Scherl, Methods, 81 (2015) 3-14.

[3] A.A. Kaspar, J.M. Reichert, Drug Discov. Today, 18 (2013) 807-817.

[4] P. Vlieghe, V. Lisowski, J. Martinez, M. Khrestchatisky, Drug Discov. Today, 15 (2010) 40-56.

[5] D.T. Felson, R.C. Lawrence, M.C. Hochberg, T. McAlindon, P.A. Dieppe, M.A. Minor, S.N. Blair, B.M.

Berman, J.F. Fries, M. Weinberger, K.R. Lorig, J.J. Jacobs, V. Goldberg, Ann. Intern. Med., 133 (2000)

726-737.

[6] M. Delgado, C. Abad, C. Martinez, J. Laceta, R.P. Gomariz, Nat. Med., 7 (2001) 563-568.

[7] E. Gonzalez-Rey, A. Chorny, M. Delgado, Nat. Rev. Immunol., 7 (2007) 52-63.

[8] A. Kowalczyk, A. Jelen, M. Zebrowska, E. Balcerczak, A. Goraca, Oxid. Med. Cell. Longev., (2016).

[9] S. Briyal, T. Philip, A. Gulati, J. Alzheimer's Dis., 23 (2011) 491-503.

[10] R. Ozdemir, H. Parlakpinar, A. Polat, C. Colak, N. Ermis, A. Acet, Toxicology, 219 (2006) 142-149.

[11] A. Kowalczyk, P. Kleniewska, M. Kolodziejczyk, B. Skibska, A. Goraca, Arch. Immunol. Ther. Exp., 63

(2015) 41-52.

[12] F. Gobeil, P. Sirois, D. Regoli, Peptides, 52 (2014) 82-89.

[13] G.N. Kaufman, C. Zaouter, B. Valteau, P. Sirois, F. Moldovan, Arthrit. Res. Ther., 13 (2011).

[14] M.e. al, in.

[15] J.R. Lakowicz, Fluorophores, Springer US, Boston, MA, 2006.

[16] E. Pastorini, R. Rotini, M. Guardigli, S. Vecchiotti, S. Persiani, G. Trisolino, D. Antonioli, L.C. Rovati, A.

Roda, J. Pharm. Biomed. Anal., 50 (2009) 1009-1014.

[17] L.F. Huang, F.Q. Guo, Y.Z. Liang, B.Y. Li, B.M. Cheng, J. Pharm. Biomed. Anal., 36 (2004) 877-882.

[18] R. Turner, L.K. Stamp, A.J. Kettle, J. Chromatogr. B, 891 (2012) 85-89.

[19] M.K. Kosinska, S.C. Mastbergen, G. Liebisch, J. Wilhelm, R.B. Dettmeyer, B. Ishaque, M. Rickert, G.

Schmitz, F.P. Lafeber, J. Steinmeyer, Osteoarthritis Cartilage, 24 (2016) 1470-1478.

[20] D.K. Lloyd, J. Chromatogr. B, 866 (2008) 154-166.

[21] Y. Xu, L. Du, M.J. Rose, I. Fu, E.J. Woolf, D.G. Musson, J. Chromatogr. B Analyt. Technol. Biomed. Life.

Sci., 818 (2005) 241-248.

[22] D.D. Vanslyke, A. Hiller, R.A. Phillips, P.B. Hamilton, V.P. Dole, R.M. Archibald, H.A. Eder, J. Biol.

Chem., 183 (1950) 331-347.

[23] A.O. Bingoel, D. Lohmann, K. Pueschel, W.M. Kulicke, Biorheology, 47 (2010) 205-224.

[24] K.M.N. Oates, W.E. Krause, R.L. Jones, R.H. Colby, J. R. Soc. Interface, 3 (2006) 167-174.

[25] B.F. Lindberg, K.E. Andersson, Pharmacol. Toxicol., 68 (1991) 276-281.

[26] P. Hyenstrand, J.S. Metcalf, K.A. Beattie, G.A. Codd, Toxicon, 39 (2001) 589-594.

[27] P. Hyenstrand, J.S. Metcalf, K.A. Beattie, G.A. Codd, Water Res., 35 (2001) 3508-3511.

[28] M.J. Yahya AM, D’Arcy PF, Drug Metab. Drug Interact, 6 (1988) 1-45.

[29] V.S. Stoll, J.S. Blanchard, Methods in Enzymology, 182 (1990) 24-38.

[30] S.J. Thompson, C.K. Hattotuwagama, J.D. Holliday, D.R. Flower, Bioinformation, 1 (2006) 237-241.

[31] H.G. Vogel, F.J. Hock, J. Maas, D. Mayer, Drug discovery and evaluation: safety and pharmacokinetic

assays, Springer, 2006.

[32] G. Munoz, S.V. Duy, P. Labadie, F. Botta, H. Budzinski, F. Lestremau, J. Liu, S. Sauvé, Talanta, 152

(2016) 447-456.

[33] in, European Medicines Agency, 2011, pp. 23.

[34] K. Strupat, Molecular weight determination of peptides and proteins by ESI and MALDI, in: A.L.

Burlingame (Ed.) Mass Spectrometry: Modified Proteins and Glycoconjugates, 2005, pp. 1-36.

[35] J.H. Gross, Mass Spectrometry, 1 ed., Springer-Verlag Berlin Heidelberg, 2004.

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

[36] R. Zubarev, M. Mann, Mol. Cell. Proteomics, 6 (2007) 377-381.

[37] H. John, M. Walden, S. Schafer, S. Genz, W.G. Forssmann, Anal. Bioanal. Chem., 378 (2004) 883-897.

[38] J.Z. Wenkui Li, Francis L. S. Tse, Handbook of LC-MS Bioanalysis: Best Practices, Experimental

Protocols, and Regulations, John Wiley & Sons, 213.

[39] A. Gotoh, K. Uchida, Y. Hamano, S. Mashiba, I. Kawabata, Y. Itoh, Clin. Chim. Acta, 397 (2008) 13-17.

[40] M. Schwartz, W. Kline, B. Matuszewski, Anal. Chim. Acta, 352 (1997) 299-307.

[41] Y. Xu, L.H. Du, M.J. Rose, I. Fu, E.J. Woolf, D.G. Musson, J. Chromatogr. B, 818 (2005) 241-248.

[42] W.K. Li, S.Y. Luo, H.T. Smith, F.L.S. Tse, J. Chromatogr. B, 878 (2010) 583-589.

[43] R.N. Xu, B. Boyd, M.J. Rieser, T.A. Ei-Shourbagy, J. Sep. Sci., 30 (2007) 2943-2949.

[44] A.Y. Hui, W.J. McCarty, K. Masuda, G.S. Firestein, R.L. Sah, Wiley Interdisciplinary Reviews-Systems

Biology and Medicine, 4 (2012) 15-37.

[45] P.W.E. John A Kellum, Stewart's Textbook of Acid-Base, Lulu, 2009.

[46] M. Laznicek, A. Laznickova, J. Pharm. Biomed. Anal., 13 (1995) 823-828.

[47] G. Piraprez, M.F. Herent, S. Collin, Flavour Fragrance J., 13 (1998) 400-408.

[48] J. Paladino, N. Kucharczyk, A. Morris, M. Thibault-Naze, J.-P. Mahieu, B. Serkiz, J. Volland, C. Autissier,

J.-L. Fauchere, Drug Design Discov., 12 (1994) 121-128.

.

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524