Accepted Manuscript Title: Solid Phase Microextraction fills the gap in tissue sampling protocols Author: Barbara Bojko Krzysztof Gorynski German Gomez-Rios Jan Matthias Knaak Tiago Machuca Vinzent Nikolaus Spetzler Erasmus Cudjoe Michael Hsin Marcelo Cypel Markus Selzner Mingyao Liu Shaf Keshjavee Janusz Pawliszyn PII: DOI: Reference: S0003-2670(13)01124-0 http://dx.doi.org/doi:10.1016/j.aca.2013.08.031 ACA 232789 To appear in: Analytica Chimica Acta Received date: Revised date: Accepted date: 18-6-2013 12-8-2013 17-8-2013 Please cite this article as: B. Bojko, K. Gorynski, G. Gomez-Rios, J.M. Knaak, T. Machuca, V.N. Spetzler, E. Cudjoe, M. Hsin, M. Cypel, M. Selzner, M. Liu, S. Keshjavee, J. Pawliszyn, Solid Phase Microextraction fills the gap in tissue sampling protocols, Analytica Chimica Acta (2013), http://dx.doi.org/10.1016/j.aca.2013.08.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. - In vivo SPME sampling was used during lung and liver transplantation. - Selection of the probe, transportation, storage conditions and analyte coverage were discussed. Ac ce pt e d M an us cr ip t - Organ preservation procedures were investigated using principal component analysis. Page 1 of 21 us cr ip t Barbara Bojko 1 , Krzysztof Gorynski 1 , German Gomez-Rios 1 , Jan Matthias Knaak 2 , Tiago Machuca 3 , Vinzent Nikolaus Spetzler 2 , Erasmus Cudjoe 1 , Michael Hsin 3 , Marcelo Cypel 3 , Markus Selzner 2 , Mingyao Liu 3 , Shaf Keshjavee 3 and Janusz Pawliszyn 1 1Department of Chemistry, University of Waterloo, Waterloo ON, Canada of Surgery, Multi Organ Transplant Program, Toronto General Hospital, Toronto, ON, Canada 3Latner Thoracic Surgery Research Laboratories, Toronto General Research Institute, University Health Network and Department of Surgery, University of Toronto, Toronto, ON, Canada an 2Department Keywords: in vivo SPME, metabolomics, tissue sampling, surgery d M Abstract Metabolomics and biomarkers discovery are an integral part of bioanalysis. However, untargeted tissue analysis remains as the bottleneck of such studies due to the invasiveness of sample collection, as well as the laborious and timeconsuming sample preparation protocols. In the current study, technology integrating in vivo sampling, sample preparation and global extraction of metabolites – solid phase microextraction was presented and evaluated during liver and lung transplantation in pig model. Sampling approaches, including selection of the probe, transportation, storage conditions and analyte coverage were discussed. The applicability of the method for metabolomics studies was demonstrated during lung transplantation experiments. pt e 3 4 5 6 7 8 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 36 37 38 39 40 41 ce 2 Solid Phase Microextraction fills the gap in tissue sampling protocols Introduction Tissue analysis is one of the most challenging, labor-extensive and time-consuming analytical tasks. It is also less frequently utilized when compared to blood, plasma, urine, saliva or stool analysis because of the difficulty in accessing the sample. However, the abovementioned easily accessible matrices carry only global information about the condition of the organism. In order to obtain organ-specific or spatially resolved data (e.g. tumor vs. healthy tissue, different organ regions), an analysis of tissue sample is necessary. The area of metabolomics studies has gathered increasing interest over the past years, becoming an integrated part of the ‘-omics’ family. In particular, a rising interest in tissue analysis for global screening of small molecules and biomarkers discovery has been observed. While the use of nuclear magnetic resonance (NMR) voids the need for a sample preparation step, mass spectrometry coupled to Ac 1 Page 2 of 21 ce pt e d M an us cr ip t different separation techniques (liquid or gas chromatography, capillary electrophoresis) requires effective sample clean-up and extraction of analytes. Several articles providing tissue preparation protocols and comparisons between the effectiveness of these extraction techniques can be found in the literature [1–4]. Generally, all these methods are based on removal of a piece of tissue (few milligrams to few grams) or the entire organ (depending on application; human or animal, in vivo or post mortem study etc.), followed by tissue disruption, centrifugation and solvent extraction. The complexity of the subsequent steps depends on the procedure used, such as various modifications of bead beating, grinding, homogenization, as well as various approaches to solvent addition (e.g. stepwise, two-step, all-in-one). The use of different solvents and different configurations of solvents results in different analyte coverage. While this precludes comparison of data originating from samples subjected to different analytical protocols, on the other hand, it also complements the information, allowing flexibility in the selection of the appropriate tools for a given application. Considering the ultimate goal of metabolomics, analysis of the entire metabolome, the chosen method should yield the best possible coverage. However, other factors also need to be taken into consideration; in the recent protocol developed for global metabolite profiling of animal and human tissues via UPLC-MS [4], a variety of issues were brought up, such as losing labile metabolites during the period between sample collection and preservation, shortage of material collected for multiple diagnostic purposes, and the heterogeneity of the sample, among others. An alternative to the invasive sample collection-based methods is microdialysis (MD); however, MD is usually used for targeted analysis only, due to limitations in the extraction of hydrophobic species as well as severe matrix effects, prohibiting non-targeted determinations. Recent metabolomics analyses conducted with the use of MD confirmed the biased nature of this methodology [5, 6]. Another setback with MD analysis is timing: it is usually required to give patients time to recover after a probe implantation. In addition, the MD probe must be constantly connected to tubing and supplied with power, making it inconvenient for some applications. Solid phase microextraction (SPME) is an analytical technique already validated in many different areas of bioanalysis, including ex vivo as well as in vivo sampling for targeted and untargeted studies [7–10]. Particularly for in vivo sampling, the technology offers some unique features, such as a lack of sample collection and extraction of unstable species [11–13]. It should be emphasized that by using one of the many calibration approaches available, fully quantitative data is ensured [14], while the wide variety of extraction phase chemistries permits selection of the most suitable phase for a specific application. The applicability of in vivo SPME for blood sampling was reported for pharmacokinetics [15–18] and metabolomics study [19], while direct tissue extraction so far has mainly focused on the determination of pharmaceutical concentrations in fish muscle, and neurotransmitter and drug levels in rat brain [20]. The availability of biocompatible SPME fibers opens the door for low-invasive, sample-free tissue analysis, also in the view of global metabolomics. In this study, we would like to present for the first time the use of SPME for untargeted direct extraction from lung and liver performed in a pig model. Ac 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 Page 3 of 21 us cr ip t Experimental Materials and methods Acetonitrile, methanol and water (all LC/MS Optima grade) were purchased from Fisher Scientific (Ottawa, Canada). Prototypes of biocompatible SPME mix-mode probes (C18 with benzenesulfonic acid, 45 µm thickness, 4, 7 and 15 mm length of coating) were provided by Supelco (Bellefonte, PA, USA). The standards used for instrumental QC (test mixture) preparation were purchased from Sigma-Aldrich (Dphenylalanine, L-tryptophan, cholic acid, deoxycholic acid) and CDN Isotopes (Lphenyl-d5-alanine). Normothermic Ex Vivo Liver Perfusion (NEVLP) and Normothermic Ex Vivo Lung Perfusion (EVLP) an Information about Normothermic Ex Vivo Liver Perfusion (NEVLP) and Ex Vivo Lung Perfusion (EVLP) procedures can be found in [21] and [22], respectively. pt e d M In vivo extraction from liver and lung Prior to use, all fibers underwent preconditioning for a minimum of 60 min exposure to methanol:water 1:1 v/v. Additionally, for the lung sampling, sterilization of the fibers was conducted by 30 min steam autoclaving at 121 oC. The preconditioning mixture (methanol:water 1:1 v/v) was also sterilized by filtering, using 0.45 µm pore size sterile syringe filters. The extraction time was 20 min for the lung, 30 min for the liver. Immediately after sampling, fibers were rinsed with deionized water, dried with Kimwipe and placed on dry ice in the empty vials. After being transported to the laboratory, all fibers were placed at -80 oC until analysis. ce Evaluation of transportation and storage condition In order to evaluate storage and transportation conditions, nine 7 mm fibers were divided into three groups, and each group was subjected to a different protocol. Six of the fibers were used for in vivo extraction from the liver; immediately after the 30 min exposure to the tissue, fibers were removed from the organ, quickly rinsed with deionized water and gently dried with Kimwipe. Three fibers were placed in the empty 0.3 mL polypropylene vials and three others in identical vials containing 0.3 mL desorption solvent (acetonitrile:water, 1:1 v/v). All vials were placed on dry ice and transported to the laboratory, where they were stored at -80 oC for two weeks. For comparative analysis, a piece of liver was collected at the beginning of the in vivo extraction and immediately placed in dry ice together with the fibers used for the in vivo study. Similarly to the fibers, the liver tissue was kept frozen at -80 oC for two weeks, and then thawed on ice and used for 30 min SPME extraction using the last three fibers. Desorption was performed simultaneously for all nine fibers; the desorption time was 90 min with 1000 rpm vortex agitation (model DVX-2500, VWR International, Mississauga, ON, Canada). Extracts were further injected to the LC-MS system for analysis. Ac 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 Page 4 of 21 Liquid chromatography-mass spectrometry analysis (LC/MS) The samples were analyzed by reverse phase liquid chromatography (Accela, Thermo Scientific) coupled to bench top orbitrap mass spectrometer (Exactive, Thermo Scientific). The details and conditions of the LC/MS method can be found elsewhere [19, 23] M an us cr ip t Data processing and statistical analysis The raw data files .raw obtained from Thermo Xcalibur were converted to .mzXML using MSConvert software. Subsequently, MZmine software version 2.10 was used for data processing. For filtering, the Savitzky-Golay algorithm was used, and mass detection was conducted with exact mass method using 5e 3 noise level. For the chromatogram builder m/z tolerance of 0.001 or 5 ppm was used. The chromatogram deconvolution was performed using the Savitzky-Golay method and for the alignment, the ‘join aligner’ option was selected. After data processing, the first minute (0-1 min) corresponding to column void volume and the last five minutes of column re-equilibration (35-40 min) were excluded from the analysis. Given chromatograms were subjected to manual peak picking and all signals with unacceptable shape, as well as those present in blank were removed from the analysis. The final dataset containing signal intensities for all molecular features defined by accurate mass and retention time were exported to .cvs file and used for statistical analysis using SIMCA-P software version 12.0.1. The Pareto scaling was applied in principal component analysis and accurate masses were used as variables ce pt e d Results and discussion Tissue sampling with SPME fibers The sampling approach involved direct insertion of biocompatible mix mode fibers to liver and lung. The fibers were assembled in a 24 ga hypodermic needle, which served as guide cannula when placing the fiber inside the sampling area. Figure 1 presents the sampling of liver and lung performed in pig model during organ transplantation. Ac 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 Page 5 of 21 t us cr ip Figure 1. SPME sampling of pig liver (A) and lung (B) with biocompatible mix-mode fibers Ac 166 167 168 169 170 171 172 173 174 175 176 177 178 179 ce pt e d M an 164 165 Effects of fiber length on sampling efficacy The length of the fiber coating can be adjusted to the requirements of the particular application. To achieve the best sensitivity of the method, the longest possible coating should be used; however, the length of the coating should ensure good spatial resolution, and coating should be immersed in the desired sampling area only. For instance, the fiber length cannot extend beyond the studied region for organs with structures that are not homogenous, or anatomically divided into sections, lobs or surfaces such as lungs, liver or kidneys, or if sampling needs to be limited to altered tissue (e.g. malignant, infected, inflammatory). Also, it should be Page 6 of 21 t us cr ip an M d pt e ce 207 209 211 213 215 217 219 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 remembered that in order to achieve good reproducibility, the entire length of the coating must be covered by the tissue matrix. Taking into consideration the need for different coating lengths for different applications, mix mode fibers with three different coating lengths were tested during liver sampling. The extraction time used for all three coating types was 30 min. As expected, the 15 mm coating gave the highest sensitivity and coverage and the 4 mm coating the lowest (Fig. 2). Conversely, fibers with 4 and 7 mm coating lengths provided better reproducibility when compared with the 15 mm coating, which could be related to the structure of the liver. As a compromise between sensitivity and reproducibility, probes with 7 mm extraction phase were selected for further studies. Figure 2. The principal component analysis showing clustering of three types of fibers used for liver sampling: 4, 7 and 15 mm coating length. Ac 180 181 182 183 184 185 186 187 188 189 190 191 192 193 195 197 199 201 203 Sample preservation and transportation When using standard methods of tissue analysis, how samples are handled after tissue collection is a major concern. In any case, the collected tissue should be placed on ice, preventing loss of unstable analytes by slowing down tissue metabolism. If the sample cannot be immediately processed, it has to be snap frozen at a minimum of – 20 oC, and preferably at -80 oC. To further minimize the progress of metabolic reactions, and to accelerate freezing of the samples, it is recommended that obtained tissue is cut into small pieces, and stored this way. This procedure also helps avoid multiple freezing and thawing of the same sample, which can also influence results. For in vivo SPME, direct sampling of the metabolites from the Page 7 of 21 d M an us cr ip t living system voids the need for tissue collection. Because of the biocompatible nature of the coating, which restricts access for large molecules such as proteins (enzymes) to the coating, the extracted analyte is protected from enzymatic degradation. This feature allows analysis of short-lived species, and has been already discussed elsewhere for blood sampling [19]. This sample-free sampling gives us different options for transportation and storage of the analytes, either leaving analytes on the fiber until analysis, or transporting and storing as an extract. In both cases, the fiber needs to be quickly rinsed with deionized water immediately after sampling. In the first scenario, after the fiber is rinsed, it should be placed on ice securely hidden in the needle assembly. It can be transported and stored in such way at a low temperature (-30 or -80 oC). The second option involves desorption on site, with transportation and storage of the extract conducted in sealed vials. However, a principal component analysis plot (Fig. 3) shows that data points corresponding to three different extractions, transportation and storage approaches do not cluster, showing that there is an influence of the used protocol on the final results. The orientation of the clusters indicates that both principal components, PC1 and PC2, influence the separation of the samples, suggesting that different factors are responsible for the changes in the metabolic profile. The differences observed between the two in vivo sets of samples (followed by transport of the analytes on fiber and transport of the analytes in the extract after on-site desorption) can be explained by the partial evaporation of the desorption solvents and the preconcentration of the analytes in the extract with respect to in-lab desorption. 277 280 281 282 283 284 285 286 Ac ce pt e 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 262 Figure 3. The influence of storage and transportation condition on PCA clustering; blue diamonds – ex vivo extraction from the lung tissue stored at -80 oC for two weeks; green triangles – in vivo extraction and on-fiber transportation and storage for two weeks; red dots – in vivo extraction, desorption on-site, transport and storage of extract. Page 8 of 21 an us cr ip t The differences between ex vivo and in vivo sampling are observed in the loss of fast turnover compounds in the case of the former approach, as reported previously for blood sampling [19]. Moreover, degradation products can be produced in the tissue after sample collection and before metabolism quenching, thus not being present in the in vivo extracts. Tentative identification of the compounds differentiating in vivo sampling, followed by on-fiber transport from ex vivo sampling of stored tissue, suggested the presence of several intermediate metabolites, such as itaconic acid, hypoxanthine, lactaldehyde or hydroxyacetone. The chromatogram and corresponding spectra with masses assigned as different adducts of lactaldehyde or hydroxyacetone and hypoxantine are presented in Figure 4. On the other hand, in the extracts obtained from ex vivo lung sampling, the dominant compounds, when compared with in vivo samples, were diacylglycerol. Their presence can be explained by the remaining enzymatic activity of phospholipase, and the degradation of the cell membrane. However, due to the fact that in global metabolomics, the analysis of unknown compounds does not permit prior evaluation of the stability of all metabolites of interest, the transportation/storageon-fiber protocol seems to be the most convenient method. ce Figure 4. Chromatograms and corresponding mass spectra indicating masses assigned as adducts of lactaldehyde or hydroxyacetone (A) and hypoxantine (B) Effects of positive vs. negative ionization mode of molecular analysis The analyses of molecular features detected in positive and negative ionization mode after 20 min extraction from liver and 30 min extraction from lung, using 7 mm mix-mode coating and separation on pentafluorophenyl column, are shown in Fig. 5A and C and 5B and D, respectively. Only signals that were present in at least two out of three replicates of the sample were taken into account,. For the liver sampling, the number of molecular features detected in negative mode was 311, and ranged between 100.07596 and 971.44040 m/z, while in positive, the number was 1580, and ranged between 80.96398 and 947.44186 m/z. Corresponding results obtained for the lung study showed the number of molecular features detected in negative and positive mode were 239 and 1041, respectively. The masses ranges for negative and positive modes were 86.948939 - 781.25821 and 110.5100 - Ac 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 pt e d M 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 Page 9 of 21 ce pt e d M an us cr ip t 997.77161, respectively. The range of detected metabolites support the observation that SPME provides a good and balanced coverage, as shown in the previous metabolomics study [19]. However, the currently used coating does not give as wide a yield of metabolites as solvent methods. This can be addressed by implementing alternative extraction phases based on SPE particles i.e. hydrophilic interaction liquid chromatography (hilic) , hydrophilic-lipophilic balance (HLB), or polystyrene divinylbenzene (PS-DVB), which can go with polarities as low as logP -3 and extract sugars or peptides. In terms of hydrophobic compounds, the limit is primarily dependent on the concentration of free molecules in the studied matrix. It should be noted that an SPME fiber interacts only with free molecules; its coating does not pick up the bound fraction of the analyte. Despite the fact that the percentage of bound hydrophobic metabolites is usually high, because of poor water solubility, the sorbents used for SPME extraction have a high affinity towards these compounds, therefore providing good extraction recovery. However, the determination of free concentrations has an important advantage over the total concentration extraction given by solvent-based extraction methods, since the free portion exhibits biological activity, and the information taken from SPME directly correlates to the actual response of the organ. It is worth mentioning that SPME can go beyond untargeted analysis of small molecules, and offer very selective extraction if a molecularly imprinted polymer or aptamer based coating is used. This should be considered as a useful tool for monitoring biomarkers, which is the ultimate goal of metabolomics study. By providing additional information not accessible from standard techniques, in vivo SPME may serve as a complementary or alternative method, particularly when sample collection is restricted, or when spatial or temporal resolution is required. The preconcentration of the analytes on the fiber and their desorption in a small volume ensures good sensitivity of the method (e.g. intensity for the presented liver study was within the range of 5e3 to 7.5e7 for positive mode and 5.3e3 to 3.7e6 in negative mode). The total time “from sampling to extract” would take 2-3 hours with the current experimental setup, versus over 15-20 hours of standard solvent based approaches, including 3-hour and overnight dryness of aqueous and organic extracts, respectively [3]. 353 Ac 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 Page 10 of 21 t us cr ip an M pt e d Figure 5. Ion map (m/z versus retention time) for lung (A, C) and liver (B, D) extracted with 7 mm mix-mode fiber during in vivo exposure followed by LC-MS analysis in positive (ESI +) (A and B) and negative (ESI -) (C and D) ionization mode. ce Detection of metabolites after different lung preservation condition The in vivo SPME-LC/MS study, followed by multivariate principal component analysis employed to the liver and lung transplantation experiment, allowed us to distinguish the clinically relevant separation of the investigated individuals, and indicate metabolites resolving the clustered data. Figure 6 presents the example of PCA plot for the lungs subjected to two different organ preservation procedures. The cluster on the left side of the plot represents a standard preservation method (Cold Standard Preservation after flush with perfadex), and the cluster on the right represents the samples collected from the lung after ex vivo perfusion (EVLP). It can be noticed that the first principal component distinguishes differences between the two preservation methods, while the second characterizes changes between different samples obtained during the certain experiment. Ac 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 Page 11 of 21 an Figure 6. PCA showing clustering of the samples collected under different clinical conditions of the lung (ESI -) ce pt e d M Conclusions The results of the present study indicate that in vivo SPME can be successfully used as an alternative approach to tissue sampling. The method offers enrichment, low invasiveness, balanced yield and the possibility of capturing labile compounds. Further, development of the extraction phases will not only extend the range of extracted analytes and make the technology more attractive for global screening, but also increase the chance of extraction biomarkers directly from the studied tissue by designing highly selective coatings (i.e. molecularly imprinted polymers, aptamers). The mentioned feature of low invasiveness and good spatial resolution provided by SPME fibers permits repeated sampling from the tissue and consequently makes the method suitable for monitoring of drug and biomarkers distribution and concentration over the period of time. This application of in vivo SPME can be also found as a useful tool in the research area, especially in animal studies, where minimum invasiveness during repeated sampling is a major concern. Although time resolution associated with relatively long exposure of the probe to the organ for optimum sensitivity for metabolomics remains the main drawback of the approach, it may be improved for targeted analysis, where extraction time can be optimized prior to the experiment. With its given simplicity and solventless nature, this technique can be an attractive solution for on-site analysis. For diagnostic purposes, ongoing studies are being carried out in our research group for improvement of previously reported direct couplings of SPME with mass spectrometer [24–29] for rapid bedside or intrasurgical analysis. Ac 386 388 390 392 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 us cr ip t 374 Reference Page 12 of 21 423 424 [1] P. Masson, A.C. Alves, T.M.D. Ebbels, J.K. Nicholson, E.J. Want, Anal. Chem. 82 (2010) 7779–7786 426 427 [3] E.J. Want, P. Masson, F. Michopoulos, I.D. Wilson, G. Theodoridis, R.S. Plumb, J. Shockcor, N. Loftus, E. Holmes, J.K. Nicholson, Nat. Protoc. 8 (2013) 17–32. t 428 [2] F.M. Geier, E.J. Want, A.M. Leroi, J.G. Bundy, Anal. Chem. 83 (2011) 3730–3736. [4] H. Wu, A.D. Southam, A. Hines, M.R. Viant, Anal. Biochem. 372 (2008) 204–212. us cr ip 425 429 430 [5] C. Wibom, I. Surowiec, L. Mörén, P. Bergstro, M. Johansson, H. Antti, A.T. Bergenheim, J. Proteome. Res. 9 (2010) 2909–2919. 433 [7] H. Kataoka, K. Saito, J. Pharm. Biomed. Anal. 54 (2011) 926–950. 440 441 442 443 444 an M [10] B. Bojko, E. Cudjoe, M. Wasowicz, J. Pawliszyn, TrAC Trends Anal. Chem. 30 (2011) 1505–1512. d 438 439 [9] S. Risticevic, V.H. Niri, D. Vuckovic, Anal. Bioanal. Chem. 393 (2009) 781–795. pt e 437 [8] B. Bojko, E. Cudjoe, G. Gómez-Ríos, K. Gorynski, R. Jiang, N. Reyes-Garcés, S. Risticevic, É. S. Silva, O. Togunde, D. Vuckovic, J. Pawliszyn, Anal. Chim. Acta. 750 (2012) 132–151. [11] D. Vuckovic, S. Risticevic, J. Pawliszyn, Angew. Chem. Int. Ed. 50 (2011) 5618– 5628. [12] D. Vuckovic, Bioanalysis, 3 (2011) 1305–1308. ce 434 435 436 [6] O. Hrydziuszko, M.A Silva, M.T. Perera, D.A Richards, N. Murphy, D. Mirza, M.R. Viant, OMICS, 14 (2010) 143–150. [13] G. Ouyang, D. Vuckovic, J. Pawliszyn, Chem. Rev. 111 (2011) 2784–2814. [14] G. Ouyang, J. Pawliszyn, Anal. Chim. Acta. 627 (2008) 184–197. Ac 431 432 445 446 [15] H.L. Lord, R.P. Grant, M. Walles, B. Incledon, B. Fahie, J.B. Pawliszyn, Anal. Chem. 75 (2003) 5103–5115. 448 449 [17] F.M. Musteata, I. de Lannoy, B. Gien, J. Pawliszyn, J. Pharm. Biomed. Anal. 47 (2008) 907–912. 447 450 451 [16] F.M. Musteata, M.L. Musteata, J. Pawliszyn, Clin. Chem. 52 (2006) 708–715. [18] D. Vuckovic, I. De Lannoy, B. Gien, Y. Yang, F.M. Musteata, R. Shirey, L. Sidisky, J. Pawliszyn, J. Chromatogr. A, 1218 (2011) 3367–3375. Page 13 of 21 452 453 [19] D. Vuckovic, I. de Lannoy, B. Gien, R.E. Shirey, L.M. Sidisky, S. Dutta, J. Pawliszyn, Angew. Chem. Int. Ed. 50 (2011) 5344–5348. 455 456 457 [21] M.U. Boehnert, J.C. Yeung, F. Bazerbachi, J.M. Knaak, N. Selzner, I.D. McGilvray, O.D. Rotstein, O.A. Adeyi, S.M. Kandel, P. Rogalla, P.M. Yip, G.A. Levy, S. Keshavjee, D.R. Grant, M. Selzner, Am. J. Transplant. 13(2013) 1441-1449. t [20] E. Cudjoe, B. Bojko, P. Togunde, J. Pawliszyn, Bioanalysis, 4 (2012) 2605–2619. us cr ip 454 458 459 460 461 [22] M. Cypel, J.C. Yeung, M. Liu, M. Anraku, F. Chen, W. Karolak, M. Sato, J. Laratta, S. Azad, M. Madonik, C.-W. Chow, C. Chaparro, M. Hutcheon, L.G. Singer, A.S. Slutsky, K. Yasufuku, M. de Perrot, A.F. Pierre, T.K. Waddell, S. Keshavjee, New Engl. J. Med. 364 (2011) 1431–4140. 463 464 [24] Y. Wang, M. Walles, B. Thomson, S. Nacson, J. Pawliszyn, Rapid Commun. Mass Spectrom. 18 (2004) 157–162. 466 467 [26] S. Perera, A. Berthod, E. Dodbiba, D.W. Armstrong, Rapid Commun. Mass Spectrom. 26 (2012) 853–862. 469 470 [28] M. Walles, Y. Gu, C. Dartiguenave, F.M. Musteata, K. Waldron, D. Lubda, J. Pawliszyn, J. Chromatogr. A, 1067 (2005) 197–205. 471 472 473 474 475 476 477 478 479 an M d [27] J.K. Lokhnauth, N.H. Snow, J. Sep. Sci. 28 (2005) 612–618. pt e 468 [25] F.S. Mirnaghi, J. Pawliszyn, Anal. Chem. 84 (2012) 8301–8309. [29] H. Tong, N. Sze, B. Thomson, S. Nacson, J. Pawliszyn, The Analyst, 127 (2002) 1207–1210. ce 465 [23] V. Bessonneau, B. Bojko, J. Pawliszyn, Bioanalysis, 5 (2013) 783–792. Acknowledgment The authors want to acknowledge the Natural Sciences and Engineering Research Council (NSERC) of Canada and Supelco (Sigma-Aldrich) for financial support. The authors express their sincere gratitude to Dr. Marcin Wasowicz for his help in establishing the collaboration between Toronto General Hospital and University of Waterloo. We also thank Ms. Nathaly Reyes-Garcés for her help in SPME sampling. Ac 462 Page 14 of 21 Ac ce pt ed M an us c Page 15 of 21 Ac pt a M ed ce Page 16 of 21 Ac ce pt ed M an us c Page 17 of 21 Ac ce pt ed M Page 18 of 21 an us c Ac ce pt ed M an us cr ip t Page 19 of 21 Ac ce pt ed M an us cr ip t Page 20 of 21 Ac ce pt ed M an us cr ip t Page 21 of 21