Agilent Technologies, Inc. · NYSE:A
Agilent Technologies — Technical Presentation (December 2017)
43 slides from Agilent Technologies's technical presentation, Dec 14, 2017
Slide 1: Addressing Sample Stability Concerns in Large-Scale LC-MS Metabolomics Studies
Addressing Sample Stability Concerns in Large-Scale LC-MS Metabolomics Studies Dr. Morgan J. Cichon Research Scientist - Metabolomics Discovery Themes _ The Ohio State University fi THE OHIO STATE UNIVERSITY
Slide 2: Themes Initiative
Themes Initiative .osu.edu) © Institution-wide strategic planning effort to enhance excellence in a number of key research areas, as accomplished through faculty hiring, student training opportunities, and industry partnerships. ¢ Brain Injury * Humanities and the Arts ¢ Data Analytics * Infectious
Slide 3: What Will the Food and Nutritional Metabolomics Initiative Discover?
What Will the Food and Nutritional Metabolomics Initiative Discover? ¢ A better understanding of individual biochemical variability in response to diet- and food-based interventions through: - Identifying metabolic signatures of individuals and diet - Discovering new molecular biomarkers that define
Slide 4: Outline
Outline ¢ Exposomics and human variability ¢ Sources of analytical variability in large clinical studies ¢ Post-extraction stability of urine and plasma ¢ Instrument stability across large analytical blocks of urine samples Discovery Themes For Research Use Only. Not for use in diagnostic procedures
Slide 5: Exposomics
Exposomics Centers for Disease Control- “The exposome can be defined as the measure of all the exposures of an individual in a lifetime and how those exposures relate to health.” Health outcomes Metabolome Biomarkers of exposure, biomarkers of response, patterns of biomarkers rocedures.
Slide 6: Human variability can be great
Human variability can be great * Humans are unique in their metabolomes due to genetics, microbiota, age, lifestyle, etc. * Differences in how people metabolize drugs, dietary components and other compounds from the environment * Can make it challenging to assess exposure * Leads to heterogeneity in
Slide 7: Human variability can be great: Nicotine Metabolism
Human variability can be great Nicotine Metabolism J * 70-80% of nicotine is metabolized to cotinine * Considerable interindividual differences in the rate of nicotine metabolism have been & ” observed * Differences in metabolism might Half \ Hal \ help explain/predict cy IEMA cy N ie) toatl | Aldeh
Slide 8: Human variability can be great: Isoflavone Metabolism
Human variability can be great Isoflavone Metabolism WO. = 2, Cluster 4| Cluster 3 Cluster 2 Cluster 1 daidzein £806 | 005 | 805 | Dos | 002 | B02 | 00s | Do C08 | 807 | 808 | Coz | B01 | Cor | ° wo eo oo oo oo oo oo oo ob ob ob ian oo1 O80 ovo oro ozo GHG+!eq yj VO Gg Ionb3 ¢ Isoflavones in the die
Slide 9: Challenges with large clinical metabolomics studies
Challenges with large clinical metabolomics studies * Consistency and standardization are key in metabolomics studies — The larger clinical studies become, the harder it can be to limit variability ¢ Sources of variability include- — Sample collection — Sample handling/storage — Sample prep and data
Slide 10: Variability in clinical metabolomics studies: Sample Collection
Variability in clinical metabolomics studies Sample Collection ¢ Simple and non-invasive collection procedures are preferred ¢ Challenges exist with multi-site studies ¢ Need to consider how samples are collected, when samples are collected, materials used in the collection process ¢ Pilot studies c
Slide 11: Sample Collection Example: Plasma Metabolomics
Sample Collection Example 1 Plasma Metabolomics * Random selection of samples were : : : : i ‘ pulled from a large clinical study on smoking (PI: Peter Shields) * Pilot metabolomics experiment conducted to evaluate the extraction and LC-MS method for plasma * Replicates and pooled QCs were included
Slide 12: Sample Collection Example
Sample Collection Example 1 ESITIC Son Smoker a ESITK San Smoker x197 | ESI TIC San .| Non-smoker x107 |-ESI TIC San Non-smoker Discovery Themes For Research Use Only. Not for use in diagnostic procedures.
Slide 13: Sample Collection Example
Sample Collection Example 1 x197 |-ES1 TIC Son 25-| a Smoker ZZ ESITK San Smoker “a m/z 1226.5341 sO 100 150 20 20 20 350 40 450 50 S50 6 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1900 1350 1600 1450 1500 1550 1600 1650 1700 Couns vs, Mass-to- Charge (m/z)
Slide 14: Sample Collection Example
Sample Collection Example 1 Metabolite database search for m/z 1226.5341 4IMices==|¢ Spectra Search um Discovery A ui For Research Use
Slide 15: Sample Collection Example
Sample Collection Example 1 CH3 In-source fragments and sy adducts are consistent HN ° with diatrizoate. \ ' S100 180 200 25 300 380 400 450 SOD 550 600 650 70 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1900 1350 1400 1450 1500 1550 1600 1650 1700 Couns vs. Mass-to- Charge (m/z) Discovery The
Slide 16: Sample Collection Example
Sample Collection Example 1 Diatrizoate tructure Chemis Forma Average eect Woh 1 OH MoneicepeMeleuar Weight IUPAC Name i" 3¢ 2 rraenat PAC Name CAS Reg number simues car ome incni Key Discovery BS
Slide 17: Sample Collection Example
Sample Collection Example 1 & BD Vacutainer® CPT™ Cell Preparation Tube with Sodium Citrate For the Separation of Mononuclear Cells from Whole Blood Sterile Interior Contains: ¢ 0.45 mL of 0.1 Molar Sodium Citrate Solution (Top Fluid Layer) j ‘ ¢ 1.8 gm of Polyester Gel (Middle Layer) e a ution fPol
Slide 18: Sample Collection Example
Sample Collection Example 1 PCA colored by smoking status Discovery Themes For Research Use Only. Not for use in diagnostic procedures.
Slide 19: Sample Collection Example
Sample Collection Example 1 PCA colored by smoking status PCA colored by tube type PCA Mapping (38.9%) Tube type is driving separation Discovery Themes For Research Use Only. Not for use in diagnostic procedures.
Slide 20: Sample Collection Example: Bronchoalveolar Lavage Fluid Metabolomics
Sample Collection Example 2 Bronchoalveolar Lavage Fluid Metabolomics *« Samples collected from a clinical trial at OSU (PI: Peter Shields) to study the effects of electronic cigarettes on the lungs ¢« Untargeted metabolomicsis being used to identify biomarkers related to e-cigarette usage * Smoking
Slide 21: Sample Collection Example
Sample Collection Example 2 ESI+ BPC of bronchoaveolarlavage fluid extract 020406 08 1 1214 1618 2 22 24 26 28 3 32 14 36 18 4 42 44fh6 48 5 52 S4 5658 6 62 64 6668 7 72747678 & 82 84 86 88 9 G2 94 96 BB 10 102 104 106 108 Cowes vs. Acqusitan Time (min) Mass spectrum of saturating peak * 235.1992 .
Slide 22: Sample Collection Example
Sample Collection Example 2 Metabolite database search for m/z 235.1806 ! .) 08 TMC messes | Soomro eee rsuccomennn zl Discovery Themes
Slide 23: Sample Collection Example
Sample Collection Example 2 Metabolite database search for m/z 235.1806 _| Lidocaine is used as an anesthetic during the bronchoscopy procedure € TIC SS | eer Discovery Themes For Research Use rocedures.
Slide 24: Sample Collection Example
Sample Collection Example 2 Lidocaine NNNN, Ow lon Counts 0.20.40.60.8 1 1.21.41.61.82 2.22426283 3.23.43.63.84 4.24.44.64.85 5.25.45.65.86 6.26.46.66.87 7.27.47.67.88 8.2848.68.89 9.29.49.69.8 10 10.20.40.60.8 Acquisition Time (min) * o Acetylcarnitine Hypoxanthine states Lidocaine is important for
Slide 25: Variability in clinical metabolomics studies: Sample handling and storage
Variability in clinical metabolomics studies Sample handling and storage * Time to freeze * Storage temperature ¢ Freeze-thaw cycles Discovery Themes For Research Use Only. Not for use in diagnostic procedures.
Slide 26: Variability in clinical metabolomics studies: Storage temperature
Variability in clinical metabolomics studies Sample handling and storage # of increased metabolites 55 50 45 40 * Storage temperature 25 wroms 30 @ 2<FC<4 — Significant time- and temperature- asc 20 1<FC<4, dependent changes in plasma 15 me metabolites have been observed re see figure = J (ses figur
Slide 27: Variability in clinical metabolomics studies: Freeze-thaw cycles
Variability in clinical metabolomics studies Sample handling and storage ¢ Freeze-thaw cycles — Should be avoided if possible, but at the very least should be consistent across samples — Noticeable changes by in the plasma metabolome by NMR have been reported after 3 freeze-thaw cycles (Pinto et al.
Slide 28: Variability in clinical metabolomics studies: Sample prep and data acquisition
Variability in clinical metabolomics studies Sample prep and data acquisition * Multiple analysts ¢ Sample prep procedure ° Time to analysis following extraction (i.e. time in autosampler) * Temperature of autosampler ¢ Instrument stability — Should be smallest source of variability Discovery Themes
Slide 29: Stability of biological samples post-extraction: Urine Autosampler Study
Stability of biological samples post-extraction Urine Autosampler Study 7 -Aliquots were loaded into -Feature extraction and data Sample Pre! —— autosampler at 4°C reduction using Agilent -Pooled urine -10 conditioning samples Profinder and MPP -10x dilution w/ 0.1% FA injected first -Filtered to re
Slide 30: Stability of biological samples post-extraction: Urine PCA
Stability of biological samples post-extraction Urine Autosampler Study PCA of Pooled Urine Injections (%Sr'S) Z UaUodWOD Component 1 (11.46%) Discovery Themes Separation of pooled samples based on length of time in the autosampler Injections cluster more closely toward end of batch For Research Us
Slide 31: Stability of biological samples post-extraction: Urine Autosampler Study
Stability of biological samples post-extraction Urine Autosampler Study Mass vs. RT Plot of Compounds Significantly Different (P < 0.05) between Initial and Final Injections ee Po Be he Observe both increasing and decreasing metabolites Many altered metabolites have small fold changes e . a vs a Sig
Slide 32: Stability of biological samples post-extraction: Urine Heat Map
Stability of biological samples post-extraction Urine Autosampler Study Heat map of PCA component 1 loadings (organized by injection “group”) § g § § g Biggest changes appear to be in a —l phospholipids (1.7-3.3 fold = ———— decreases) == ——- — Other changes include phenol, ——— ——————* bilirubin deri
Slide 33: Stability of biological samples post-extraction: Plasma Autosampler Study
Stability of biological samples post-extraction Sample Prep -Pooled plasma -MeOH extraction, centrifuged 4°C for 10 min -Died down and reconstituted in H,0 -Aliquoted into 50 HPLC vials Plasma Autosampler Study Data Acquisition -Aliquots loaded into autosampler at 4°C -10 conditioning samples -Agile
Slide 34: Stability of biological samples post-extraction: Plasma PCA
Stability of biological samples post-extraction Plasma Autosampler Study PCA of Pooled Plasma Injections Y-Axis Component 1 (31.26%) (%29'S) Z JUauodWo) Z-Axis The time to injection explains more of the variation inthe plasma data than the urine data (31% vs 11%) Injections again cluster more closel
Slide 35: Stability of biological samples post-extraction: Plasma Autosampler Study
Stability of biological samples post-extraction Plasma Autosampler Study Mass vs. RT Plot of Compounds Significantly Different (P <0.05) between Initial and Final Injections “ : 2 (tes * Be * eal ae so0e|g — Sor . hs enti : isle S ® ©, 7 a oa gS Be. 7 € 3 : | gs9 _ ae |e bed PD [ ‘| : eee =o - 8 : 8
Slide 36: Continuous data acquisition and batch size limitations
Continuous data acquisition and batch size limitations ¢ In LC-MS metabolomics experiments, running continuously is preferred — Avoid small shifts in retention time and instrument sensitivity — Minimize batch effects * However, data quality can suffer if running continuously for too long — Loss of s
Slide 37: Continuous data acquisition and batch size limitations: Urine Metabolomics Study
Continuous data acquisition and batch size Urine Metabolomics Study * Samples from an OSU-Center of Excellence in Regulatory Tobacco Science project ¢ Urine collected from adolescents with self-reported cigarette and smokeless tobacco usage (PI: Peter Shields) * Urine prep = centrifugation + dilutio
Slide 38: Urine data should be normalized to handle significant differences in urine volume
Urine data should be normalized to handle significant differences in urine volume between subjects Urine Osmolality 1400 Almost 25 gues e os e é ° fold 4 #200 Te — oe oe ; v= difference in i e eo @e L e e ¢ e & A 3} ee ° ° ° osmolali = 1000 #-. o_o SLs aS > 2%, re Se ty 5 ea oo * en? e 8% &.° ° “s w
Slide 39: Urine QCs cluster together in unsupervised multivariate analyses
Urine QCs cluster together in unsupervised multivariate analyses PCA — Positive Mode a abe Y-Axis prime Bs Y-Axis PCA — Negative Mode 9 ° 3 3 s re} S J > 2 = > N a a ot a ial @ Q = s S & & Vv @ fe} Z-Axis Z-Axis Component 1 (21.97%) Component 1 (23.83%) No significant differences in the individual b
Slide 40: Internal standards show little variability across study samples
Internal standards show little variability across study samples Negative mode internal standard Positive mode internal standard (4-nitrobenzoic acid) (8C-labeled phenylalanine) H) ti) Mab WN VAY Wt hay (wy M ‘4 1k ai i wee f " 1 hy ' ohh \ i, Ley YY 02 un | { oa}! II Normalized intensity Values e No
Slide 41: QCs can be used to assess consistency in the detection of various metabolites
QCs can be used to assess consistency in the detection of various metabolites oH "sO) sO nC or N 1. CYP2A6 ei N © CYP2A6 ‘e N yay Yo eine NZ CHs 2. Aldehyde 4 2 CH; 3 Hs Oxidase N Nicotine Cotinine trans-3'-Hydroxycotinine: 3-Hydroxycotinine Cotinine 3500000 1400000 3000000 Fy is © 2500000 - c 2 2 £
Slide 42: Summary
Summary ¢ Pilot metabolomics studies are useful when working with new sets of samples ¢ How biological samples are collected is a significant source of variation — Blood collection tube type must be standardized — It is difficult to adapt human samples collected from non- metabolomics focused studie
Slide 43: Acknowledgments
Acknowledgments OSUCCC The James Peter Shields — Professor, CCC Deputy Director Tae Onto State Universiry = Ping-Ching Hsu — Assistant Professor (now at CONPRENENSVECANGERCENTER LANA) ww THE OHIO STATE UNIVERSITY COLLEGE OF FOOD, AGRICULTURAL, Nutrient and Phytochemical Analytics Shared Resource — C
Agilent Technologies — Technical Presentation (December 2017) on 6ix