Untargeted Lipidomics

Materials and Method

Reagents and Internal Standards: High-performance liquid chromatography (HPLC) grade acetonitrile and dichloromethane were purchased from Sigma-Aldrich (St. Louis, MO), isopropanol (Optima – LC/MS grade) was purchased from Fisher (New Jersey, NJ), methanol (LC-MS grade) was from J.T. Baker. Water was obtained from a Millipore high purity water dispenser (Billerica, MA). The following mass spectrometry-grade lipid standards were obtained from Sigma-Aldrich: 1-heptadecanoyl-2-hydroxy-sn-glycero-3-phosphocholine LPC (17:0/0:0), 1,2-diheptadecanoyl-sn-glycero-3-phosphocholine PC (17:0/17:0), 1,2-diheptadecanoyl-sn-glycero-3-phosphoethanolamine PE (17:0/17:0), 1,2-diheptadecanoyl-sn-glycero-3-phospho-L-serine (sodium salt) PS (17:0/17:0), N-heptadecanoyl-D-erythro- sphingosylphosphorylcholine 17:0 SM (d18:1/17:0), cholest-5-en-3ß-yl heptadecanoate 17:0 cholesteryl ester, 1-palmitoyl-2-oleoyl-sn-glycerol 16:0-18:1 DG, 1-heptadecanoyl-rac-glycerol 17:0 MG, 1,2,3-triheptadecanoyl-glycerol Triheptadecanoate 17:0TAG, N-heptadecanoyl-D-erythro-sphingosine C17 Ceramide (d18:1/17:0), 1,2-diheptadecanoyl-sn-glycero-3-phosphate (sodium salt) 17:0 PA, 1,2-diheptadecanoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (sodium salt) 17:0 PG, 1-heptadecanoyl-2-(5Z,8Z,11Z,14Z-eicosatetraenoyl)-sn-glycero-3-phospho-(1’-myo-inositol) (ammonium salt) 17:0-20:4 PI, 1,3(d5)-dinonadecanoyl-2-hydroxy-glycerol DG d5-(19:0/0:0/19:0) and Glyceryl tri(palmitate-d31) TG d31.

Sample preparation: Lipids were extracted form biological using a modified Bligh-Dyer method [1] using a 2:2:2 ratio volume of methanol: water:dichloromethane at room temperature after spiking internal standards(described above) The organic layer was collected and completely dried under nitrogen. Before mass spectrometry analysis, the dried lipid extract was reconstituted in 100 μL of Buffer B (10:85:5 ACN/IPA/H2O) containing 10mM ammonium acetate and subjected to LC/MS. Internal Standards and Quality Controls: QC samples were prepared by pooling equal volumes of each sample and injected at the beginning and the end of each analysis and after every 10 sample injections to provide a measurement of the system’s stability and performance as well as reproducibility of the sample preparation method.

Two kinds of controls were used to monitor the sample preparation and mass spectrometry. To monitor instrument performance, 10 μL of a dried matrix-free mixture of the internal standards reconstituted in 100 μL of buffer B (85% IPA:10%ACN:5% H2O in 10mM NH4OAc) was analyzed. As additional controls to monitor the profiling process, an equimolar mixture of 13 authentic internal standards and a characterized pool of human plasma and test pool (a small aliquot from all plasma used in this study) (extracted in tandem with plasma samples) were analyzed along with the plasma samples. Each of these controls, were included several times into the randomization scheme such that samples preparation and analytical variability could be monitored constantly.

Data Dependent Liquid chromatography-mass spectrometry (LC-MS/MS) for measurements of lipids:

Chromatographic separation was performed on a Shimadzu CTO-20A Nexera X2 UHPLC systems equipped with a degasser, binary pump, thermostatted autosampler, and column oven (all components manufactured by Shimadzu (Canby, OR, USA). The column heater temperature was maintained at 55oC and an injection volume of 5 μL was used for all analyses. For lipid separation, the lipid extract was injected onto a 1.8 μm particle diameter, 50 × 2.1 mm id Waters Acquity HSS T3 column (Waters, Milford, MA). Elution was performed using acetonitrile / water (40:60, v/v) with 10 mM ammonium acetate as solvent A and acetonitrile / water / isopropanol (10: 5: 85 v/v) with 10 mM ammonium acetate as solvent B. For chromatographic elution we used a linear gradient beginning with 60% Solvent A and 40% Solvent B. The gradient was ramped in a linear fashion to 98% Solvent B over the first 10 minutes and was held at 98%B for 7 minutes. Thereafter the composition was returned to 40% Solvent B and 60% Solvent A and held for 3 minutes. The flow rate used for these experiments was 0.4 mL/min and the injection volume was 5μL. The column was equilibrated for 3 min before the next injection and run at a flow rate of 0.400uL/min for a total run time of 20 min.

Mass spectrometry data acquisition for each sample was performed in both positive and negative ionization modes using a TripleTOF 5600 equipped with a DuoSpray ion source (AB Sciex, Concord, Canada). Column effluent was directed to the ESI source and voltage was set to 5500V for positive ionization and 4500V for negative ionization mode. The declustering potential (DP) was 60 V and source temperature was 450oC for both modes. The curtain gas flow, nebulizer, and heater gas were set to 30, 40, and 45, respectively (arbitrary units). The instrument was set to perform one TOF MS survey scan (150 ms) and 15 MS/MS scans with a total duty cycle time of 2.4 s. The mass range of both modes was 50-1200 m/z. Acquisition of MS/MS spectra was controlled by the data dependent acquisition (DDA) function of the Analyst TF software (AB Sciex, Concord, Canada) with application of following parameters: dynamic background subtraction, charge monitoring to exclude multiply charged ions and isotopes, and dynamic exclusion of former target ions for 9 s. Collision energy spread (CES) of 20V was set whereby the software calculated the CE value to be applied as a function of m/z.

A DuoSpray source coupled with automated calibration system (AB Sciex, Concord, Canada) was utilized to maintain mass accuracy during data acquisition. Calibrations were performed at the initiation of each new batch or polarity change.

Data Processing: The raw data was converted to mgf data format using proteoWizard software [3]. The NIST MS PepSearch Program was used to search the converted files against LipidBlast [4,5] libraries in batch mode. We optimized the search parameters using the NIST11 library and LipidBlast libraries and comparing them against our lipid standards. The m/z width was determined by the mass accuracy of internal standards and was set 0.001 for positive mode and 0.005 for negative mode. The minimum match factor in used in the PepSearch Program was set to 200. The MS/MS identification results from all the files were combined using an in-house software tool to create a library for quantification. All raw data files were searched against this library of identified lipids with mass and retention time using Multiquant 1.1.0.26 [6]. (ABsciex, Concord, Canada). Quantification was done using MS1 data. The QC samples were also used to remove technical outliers and lipid species that were detected below the lipid class-based lower limit of quantification. QC samples evenly distributed along analytical runs of the study were analyzed.

Filtering data based on Missing values

First step in data analysis is filtering features based on missing values.There are two types of QC samples run along with the experimental samples.Sample ID’s CS00000001 are called “Test.Pool” samples and CS00000004 are called “Pooled.Plasma”.”Pooled.Plasma” is a red cross plasma pool run as an internal QC sample in all studies. “Test.Pool” is made by taking small aliquot from each experimental sample and pooling it together to create a QC sample which is run after every few runs to get assess drifts and other variation caused by the run.

Missing patterns are plotted and carefully examined for groupwise missingness. The data is imputed using KNN algorithm with knn.impute function from bnstruct package

Internal standards

Table 1.Internal standards Positive mode RSD%
RSD%
IS CE 17:0; [M+NH4]+ 25.40
IS Cer 35:1; [M+H]+@7.12 5.54
IS Cer 35:1;[M-H20]+ 10.07
IS D31-TAG 5.54
IS DG 38: 13.77
IS LPC 17:0;[M+H]+ 6.27
IS MG 17:0; [M+NH4]+ 8.17
IS PC 34:0; [M+H]+ 1.61
IS PE 34:0; [M+H]+ 2.74
IS PG 34:0;[M+NH4]+@6.1 13.74
IS PS 34:0;[M+H]+ 9.96
IS SM 35:1;[M+H]+ 4.18
IS TG 51:0; [M+NH4]+ 8.62
Table 2.Internal standards Negative mode RSD%
RSD%
IS Cer 35:1; [M+Hac-H]- 4.14
IS Cer 35:1; [M-H]- 10.61
IS LPC 17:0; [M+Hac-H]- 4.75
IS MG 17:0;[M+Hac-H]- 8.96
IS PA 34:0;[M-H]- NaN
IS PC 34:0;[M-Ac-H]- 4.78
IS PE 34:0;[M-H]- 9.91
IS PG 34:0;[M-H]- 4.80
IS PI 37:4;[M-H]- 7.16
IS PS 34:0;[M-H]- 13.55
IS SM 35:1; [M+Hac-H]- 9.69
Table 3.Descriptives for Internal standards Positive mode
nbr.val nbr.null nbr.na min max range sum median mean SE.mean CI.mean.0.95 var std.dev coef.var
IS CE 17:0; [M+NH4]+ 17 0 0 9239 27087 17848 275914 14891 16230 1137 2411 21986347 4689 0.29
IS Cer 35:1; [M+H]+@7.12 17 0 0 1084376 1612555 528179 22962492 1413073 1350735 46004 97524 35977816531 189678 0.14
IS Cer 35:1;[M-H20]+ 17 0 0 2505946 5023740 2517795 63397264 3866227 3729251 186601 395576 591936872392 769374 0.21
IS D31-TAG 17 0 0 517628 2445652 1928024 25939499 1991984 1525853 205559 435766 718327074410 847542 0.56
IS DG 38: 17 0 0 1398932 3248336 1849404 36892845 2096382 2170167 132986 281917 300648313967 548314 0.25
IS LPC 17:0;[M+H]+ 17 0 0 2163173 4218940 2055767 56311696 3546776 3312453 204024 432512 707639358165 841213 0.25
IS MG 17:0; [M+NH4]+ 17 0 0 83684 138689 55004 1856906 106875 109230 3293 6980 184318326 13576 0.12
IS PC 34:0; [M+H]+ 17 0 0 2432709 6894422 4461713 83401307 6360165 4905959 494193 1047642 4151849291187 2037609 0.42
IS PE 34:0; [M+H]+ 17 0 0 922889 1463560 540671 20231782 1071869 1190105 49437 104802 41548424557 203834 0.17
IS PG 34:0;[M+NH4]+@6.1 17 0 0 2717 180031 177314 2008246 116485 118132 9924 21037 1674186803 40917 0.35
IS PS 34:0;[M+H]+ 17 0 0 115474 178161 62687 2508572 145793 147563 4268 9047 309608833 17596 0.12
IS SM 35:1;[M+H]+ 17 0 0 1358533 2678270 1319737 34528158 2230845 2031068 94278 199861 151103154091 388720 0.19
IS TG 51:0; [M+NH4]+ 17 0 0 1686109 3530037 1843928 42325145 2401727 2489714 140483 297811 335504199384 579227 0.23
Table 4.Descriptives for Internal standards Negative mode
nbr.val nbr.null nbr.na min max range sum median mean SE.mean CI.mean.0.95 var std.dev coef.var
IS Cer 35:1; [M+Hac-H]- 17 0 0 11746378 14893779 3147402 224983713 13429927 13234336 227452 482176 879482540744 937807 0.07
IS Cer 35:1; [M-H]- 17 0 0 591854 893406 301551 11516757 666576 677456 17224 36514 5043604198 71018 0.10
IS LPC 17:0; [M+Hac-H]- 17 0 0 632096 1334764 702668 17224713 1085228 1013218 66268 140481 74654025572 273229 0.27
IS MG 17:0;[M+Hac-H]- 17 0 0 1172265 1939603 767338 26589881 1572628 1564111 64407 136536 70519425433 265555 0.17
IS PA 34:0;[M-H]- 16 0 1 468 175442 174974 542458 7474 33904 12763 27203 2606181654 51051 1.51
IS PC 34:0;[M-Ac-H]- 17 0 0 659693 1149691 489999 15647621 1005317 920448 49381 104683 41454403627 203604 0.22
IS PE 34:0;[M-H]- 17 0 0 2050674 3423743 1373069 47296161 2720103 2782127 111102 235526 209842727367 458086 0.16
IS PG 34:0;[M-H]- 17 0 0 818847 5370087 4551240 43208449 1366557 2541673 447976 949667 3411604490752 1847053 0.73
IS PI 37:4;[M-H]- 17 0 0 51522 148645 97123 1536886 76055 90405 8051 17068 1102035081 33197 0.37
IS PS 34:0;[M-H]- 17 0 0 492535 1329890 837355 13311100 720152 783006 63579 134781 68718564645 262142 0.33
IS SM 35:1; [M+Hac-H]- 17 0 0 298159 519319 221160 6775119 406974 398536 16737 35480 4762032079 69007 0.17

Data normalization using internal standards

Each lipid was normalized using an Internal standard that minimized its RSD post-normalization.Each mode data was normalized separately.The file names for positive and negative normalized data are Positive_normalized.txt and Negative_normalized.txt.

Data quality with PCA and RSD disribution plot

Normalized data from both modes is combined, then repeats removed to give the final dataset(datasetcombined.txt).The data was then normalized to Protein measurement for each sample.

A QC PCA plot post-normalization

Bar plot for RSD distribution normalized data

Data analysis

For data analysis we use only features below 30% RSD.

A t.test was performed using variable Hormone between 2 groups.

After FDR correction we found deferentially significant(cutoff taken < .05)lipids.The results are in file named “t.test_results.txt”

**PCA plot for only FDR corrected significant lipids on factor Researcher_Subject_ID

Heatmaps for lipids significant for Hormone after FDR correction

After FDR correction, lipids were differentially significant for factor Hormone.The following heatmap is made with only differentially significant lipids.

## png 
##   2

Classwise data presentation and Analysis

Total sum of each class was calculated by adding up all the lipids in that class.The total sum for each class is reported in a csv file “EX01362_classsum.csv”.The data was log transformed and bar plot for total sum is created . An anova was fit on total sum for each class and results are reported in “t.test_for_total_class_lipid_results.txt” file. None of the class sum was differential in the t.test results after FDR correction.

Same way percentage of each class was calculated by adding up all the lipids in that class dividing by total lipids.The total percentage for each class is reported in a csv file EX01362_class_percentage.csv”. The data was log transformed .The bar plot for total percentage is created for each class. An anova was fit on total percentage for each class and results are reported in “t.test_for_total_percentage_class_lipid_results.txt” file. Few of the classes for Percentage were found significantly differential after FDR correction at 0.1. A heatmap for percentage class is created for only differential classes.

## png 
##   2

References

  1. Bligh, E. G.; Dyer, W. J. A Rapid Method of Total Lipid Extraction and Purification. Can. J. Biochem. Psysiol 1959, 37, 911–917.
  2. (Gika, H. G.; Macpherson, E.; Theodoridis, G. A.; Wilson, I. D. Evaluation of the repeatability of ultra-performance liquid chromatography−TOF-MS for global metabolic profiling of human urine samples. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2008, 871 (2), 299−305
  3. A cross-platform toolkit for mass spectrometry and proteomics. Chambers, M.C., MacLean, B., … Mallick, P. Nature Biotechnology 30, 918-920 (2012)
  4. Kind T, Meissen JK, Yang D, Nocito F, Vaniya A, Cheng YS, Vandergheynst JS, Fiehn O. Qualitative analysis of algal secretions with multiple mass spectrometric platforms. J Chromatogr A. 2012, 1244: 139-147
  5. Meissen JK, Yuen BT, Kind T, Riggs JW, Barupal DK, Knoepfler PS, Fiehn O. Induced pluripotent stem cells show metabolomic differences to embryonic stem cells in polyunsaturated phosphatidylcholines and primary metabolism. PLoS One. 2012, 7 (10): e46770.
  6. Ejsing CS, Duchoslav E, Sampaio J, et al. Automated identification and quantification of glycerophospholipid molecular species by multiple precursor ion scanning. Anal Chem. 2006;78:6202–6214 >