A Comparison of Tissue Spray and Lipid Extract Direct Injection Electrospray Ionization Mass Spectrometry for the Differentiation of Eutopic and Ectopic Endometrial Tissues
Recent research revealed that tissue spray mass spectrometry enables rapid molecular profiling of biological tissues, which is of great importance for the search of disease biomarkers as well as for online surgery control. However, the payback for the high speed of analysis in tissue spray analysis is the generally lower chemical sensitivity compared with the traditional approach based on the offline chemical extraction and electrospray ionization mass spectrometry detection. In this study, high resolution mass spectrometry analysis of endometrium tissues of different localizations obtained using direct tissue spray mass spectrometry in positive ion mode is compared with the results of electrospray ionization analysis of lipid extracts. Identified features in both cases belong to three lipid classes: phosphatidylcholines, phosphoethanolamines, and sphingomyelins. Lipids coverage is validated by hydrophilic interaction liquid chromatography with mass spectrometry of lipid extracts. Multivariate analysis of data from both methods reveals satisfactory differentiation of eutopic and ectopic endometrium tissues. Overall, our results indicate that the chemical information provided by tissue spray ionization is sufficient to allow differentiation of endometrial tissues by localization with similar reliability but higher speed than in the traditional approach relying on offline extraction.
KeywordsAmbient mass spectrometry Tissue spray Extractive electrospray ionization Lipidomics Endometriosis
Ambient mass spectrometry (MS) enables the analysis of complex samples under atmospheric pressure without or with minimal sample pretreatment. This allows fast screening of molecular composition, which is essential for high-throughput biomarker search, classification of tissue pathologies based on molecular composition, including cancer staging, as well as for the progress in personalized medicine [1, 2, 3, 4, 5]. A large selection of ambient methods has been developed since the advent of desorption electrospray ionization (DESI) by Cooks' group in 2004 [1, 6, 7, 8]. A sub-group of ambient ionization methods is known as “substrate spray” methods . In substrate-spray methods, chemical extraction of a sample occurs online during the analysis inside an ion source. Lipids are readily extracted under such conditions and typically yield the most abundant peaks in mass spectra [9, 10, 11, 12, 13], which can be used for fast lipid screening. Substrate-spray methods include paper spray , leaf spray [15, 16], a whole organism spray , and tissue spray with different modifications [11, 12, 17, 18, 19, 20]. The latter method has been shown to be efficient in the analysis of brain and endometrium tissues and various cancer tumors [10, 11, 18]. Although increasing the speed of analysis, the online extraction in substrate-spray methods may also lead to the loss of essential information. This problem has been addressed in the recent publication devoted to the comparison of DESI-MS and ultraperfomance liquid chromatography/electrospray ionization-mass spectrometry (UPLC/ESI-MS) approach . It has been shown that the glycerophospholipid profile detected by DESI-MS is congruent to UPLC/ESI-MS.
There are two main approaches for the comprehensive analysis of global lipidome. The first approach is shotgun lipidomics , in which no separation step is used, and the whole lipid extract is analyzed. The second approach is liquid chromatography with mass spectrometry (LC/MS) platform [22, 23]. Shotgun lipidomics accounts for over 90% of lipids directly from the lipid extracts of biological samples, but many of low abundant lipid molecular species remain undetected by this technique . The time-limiting stage in shotgun lipidomics is sample preparation, which takes about 10–20 min. LC/MS lipidomics with reversed-phase, normal-phase, or hydrophilic interaction liquid chromatography (HILIC) allows detection of more lipid species [22, 23, 24, 25, 26] because of the separation of isobaric compounds and elimination of suppression effect for different lipid classes, but at the cost of the longest analysis time, which can last for 0.5–1 h. It should be noted that a fast method relying on the separation by supercritical fluid chromatography has been developed that provides high-throughput and comprehensive analysis of 24 lipid classes within 6 min . This brings the LC/MS platform closer to the shotgun platform as far as the duration of the analysis is concerned while keeping all the benefits of chromatographic separation. However, this method is not as widespread as other chromatographic methods. Therefore, a trade-off clearly exists between the speed and chemical depth of lipid profiling.
Intraoperative potential of online direct tissue analysis has been demonstrated in rapid evaporative ionization MS (REIMS) and implemented in a commercial product (iKnife) for online routine tissue analysis during surgery . Endometriosis is an abundant gynecological pathology, which is characterized by the extra-uterine presence of endometrial glands and stroma. The only reliable way to diagnose the pathology is surgical laparoscopy for the moment and intraoperative tissue identification methods are of high demand. Information about the molecular composition of eutopic (inside uterine) and ectopic (extra-uterine) endometrium can shed more light on the mechanisms of the disease and be used for tissue type determination [29, 30, 31]. Recently we demonstrated application of direct tissue analysis for the rapid differentiation between endometriotic tissues of different foci . A comparison of lipid profile obtained by tissue spray and electrospray ionization mass spectrometry of lipid extracts has not been provided so far for the best of our knowledge.
The present work compares the performance of tissue spray MS and lipid extract ESI-MS for the differentiation between endometriotic tissues of different foci. Endometrial tissues (50 samples) of eutopic endometrium and ovarian endometriotic lesions from 25 patients are studied.
Endometrial tissues (50 samples) of eutopic endometrium and ovarian endometriotic lesions from 25 patients were obtained from the Department of Surgery, V. Kulakov Research Center for Obstetrics, Gynecology and Perinatology (Moscow, Russia). All patients have read and signed informed consent approved by the Ethical Committee of the V. Kulakov Research Center for Obstetrics, Gynecology, and Perinatology (Moscow, Russia). Part of a sample is studied histologically to confirm tissue type and then separated in two parts for tissue spray MS and lipid extraction, frozen in liquid nitrogen immediately after surgery and stored under –75 oC until the analysis.
Methanol, acetonitrile, 2-propanol, chloroform, and formic acid of HPLC grade were purchased from Sigma-Aldrich (St. Louis, MO, USA). Deionized water was purchased from Panreac (Barcelona, Spain).
Tissue Lipid Extracts for Direct Injection ESI-MS and HILIC-LC/MS Analysis
Lipid extracts are prepared according to a modified Folch method . Briefly, 40 mg of tissue is homogenized in liquid nitrogen, 4 mL of chloroform–methanol (2:1, v/v) mixture is added to the sample, and the mixture is incubated for 10 min. The homogenate is filtered using coarse filter paper; 800 μL of 1 mol/L NaCl is added, and the mixture is centrifuged at 3000 rpm for 5 min at the ambient temperature. The organic bottom layer containing lipids is evaporated with a stream of nitrogen and redissolved in acetonitrile-2-propanol (1:1, v/v) mixture for analysis.
Tissue Spray MS Conditions
MS analysis of tissue samples is performed on Maxis Impact qTOF (Bruker Daltonics, Bremen, Germany) with the in-lab designed ion source for tissue spray MS [10, 11]. Mixture of H2O/methanol 1/9 is used for online tissue extraction and following spraying . The solvent is supplied to the tissue with flow rate of 1 mL/min by Dionex binary pump (Thermo Scientific, Germering, Germany). The Taylor cone on the sample tissue and electrospray is formed under 4.2 kV potential applied between tissue and inlet capillary. Distance between a sample and MS inlet is about 5–10 mm. The distance, high voltage potential, and solvent flow rate are optimized to supply stable spray. The signal of a tissue is saved during 3 min after total ion current (TIC) equilibration. Analysis schedule is as follows: 3 min in positive ion mode following by data dependent analysis (DDA). Mass spectra are registered at a 2 Hz frequency resulting in 360 spectra for 3 min. The mass range is m/z 400–1000.
Tandem MS (MS/MS) is done using DDA with the following characteristics. Three most abundant peaks are chosen after full mass scan and subjected to MS/MS analysis. Collision induced dissociation is 35 eV. Mass exclusion time is 1 min.
The lipid extracts are analyzed in triplicate on the Dionex UltiMate 3000 liquid chromatograph (Thermo Scientific, Germering, Germany) coupled to the Maxis Impact qTOF analyzer with ESI (Bruker Daltonics, Bremen, Germany) with a modified method described earlier [24, 34]. Three μL of the sample is injected onto a Spherisorb Si column (150 × 2.1 mm, 5 μm; Waters, Milford, MA, USA). Lipids separation is performed at a flow rate of 50 μL/min using acetonitrile as solvent A and 5 mM aqueous ammonium acetate as solvent B by a linear gradient from 6% to 23% (v/v) of solvent B over 25 min. The column temperature is 40 oC.
Maxis Impact qTOF is used in the direct injection ESI-MS and HILIC-LC/MS methods (Bruker Daltonics, Bremen, Germany). Mass spectra are obtained in positive ion mode in the mass range m/z 400–1000 with the following setting: capillary voltage 4.1 kV in positive ion mode (3.0 kV in negative ion mode), pressure of the nebulizing gas 0.7 bar, drying gas flow rate 6 L/min, and temperature of the drying gas 200 oC.
Obtained tissue spray mass spectra from each sample is averaged over 3 min and saved in m/z – Intensity tables using DataAnalysis software (Bruker Daltonics, Bremen, Germany). ESI-MS is averaged over the time of a sample injection. Thus obtained data is processed with scaling on TIC and peak alignment by MALDIquant R package . Multivariate data analyses is performed using orthogonal projections onto latent structures discriminant analysis (OPLS-DA) method  implemented in ropls package . The quality of statistical models are estimated by R2 and Q2 parameters, where R2 describes fraction of data that the model can explain using the latent variables, and Q2 describes part of data predicted by the model according to the cross-validation. Venn diagrams are built with VennDiagram package .
Automated peak annotation for tissue spray and ESI-MS experiments is provided with in-lab created R code, which compares measured accurate m/z values with theoretical computer-generated values. The code searches a record within 10 ppm from the experimental m/z. Protonated, sodiated, and potassiated ion adducts are considered. More precise identification is done based on the MS/MS data for the peak under consideration, if it undergone MS/MS analysis.
Lipid assignment for HILIC-LC/ESI-MS experiment is made combining retention time information with the MS data. The HILIC column is used to separate different classes of lipids based on the polar head groups. The HILIC-LC/ESI-MS data is summed across all scans for each lipid class under consideration. Peaks in thus obtained averaged spectra are annotated in a similar way as described previously but, additionally, information about lipid class is taken into account.
Results and Discussion
Comparative Lipid Composition Analysis
Tissue Differentiation by Tissue Spray and Tissue Extract ESI-MS
R2 for the tissue spray data is 0.32 and 0.47 for the ESI-MS of lipids extract. Predictive capability of the models is estimated by Q2, which is 0.73 for tissue spray and 0.82 for the extracts. This parameter is obtained with leave-one-out cross-validation (LOOCV). Therefore, tissue spray data are quite well described by the model and have weaker but still satisfactory predictive ability [44, 45]. This can be due to the generally noisier and less stable signal in tissue spray as compared to ESI-MS.
Lipid Species with high VIP Scores in OPLS-DA Analysis Grouped by Presence in Different Experimental Methods
Tissue spray MS and extract ESI-MS
Tissue spray MS
Differential lipid species mainly belong to the class of PC lipids, but several PE and few SM are present as well. PC is an essential source of polyunsaturated fatty acids, which can be transformed into eicosanoids, and the latter are involved in many metabolic pathways. Distortion in PC metabolism may cause various pathologies [46, 47]. Sphingolipids participate in signal transduction and cell fate determination . PE are the most widespread lipids on the cytoplasmic membrane. PE are involved in different cellular activities, e.g., cell cycle, membrane fusion, autophagy, and apoptosis . Several publications have been devoted to the investigation of endometriotic lipidome, but most of these studies are based on plasma, serum, or peritoneal fluid as an object [50, 51, 52]. PC, PE, and SM lipid classes were found to be altered in endometriosis patients compared with healthy individuals in those studies. In the investigation of Bi-Cheng Yang and coworkers , PC 38:4 and SM 34:1 are also among featured lipids in Vouk’s paper . In Dutta’s work lipids from endometriosis mice serum and liver were profiled . It was found that PEs were downregulated, whereas SMs, PCs, lysoPCs, lysoPEs, and plasmeny-PEs were upregulated in endometriosis mice .
Overall, our results indicate that the tissue spray enables endometrial tissue differentiation as well as extract ESI-MS analysis, but tissue spray is ca. four times faster (5 min per sample) than the classic approach based on offline extraction (20 min per sample). Tissue spray MS and extract ESI-MS yield similar spectra with the essential difference in preferential adducts type. In tissue spray MS the most intensive peaks correspond to sodiated species whereas in extract ESI-MS protonated species are readily formed. The presence of sodiated clusters adds complication to the process of spectral interpretation but is not crucial for the analysis by high-resolution mass spectrometry. Caution should be paid in biological interpretation of tissue differences obtained with tissue spray because correlation between the most essential lipids obtained by tissue spray and ESI-MS is not complete.
This work was supported by the Russian Science Foundation project (no. 16-14-00029).
N.S. is grateful to the Russian Ministry of Science and Education (no. MK-8484.2016.7) for the partial support of data analysis and sample collection. Z.W., K.C., and H.C. are grateful for financial support (in part of technical support of ion source development and implementation) from the International Science and Technology Cooperation Program of China (no. 2015DFA40290), the National Natural Science Foundation of China (no. 21520102007), and Science and Technology Planning Project at the Ministry of Science and Technology of Jiangxi Province, China (no. 20152ACB21013, 20161BBH80055).
- 10.Chagovets, V., Kononikhin, A., Starodubtseva, N., Kostyukevich, Y., Popov, I., Frankevich, V., Nikolaev, E.: Peculiarities of data interpretation upon direct tissue analysis by Fourier transform ion cyclotron resonance mass spectrometry. Eur. J Mass Spectrom. 22, 123–126 (2016)CrossRefGoogle Scholar
- 11.Kononikhin, A., Zhvansky, E., Shurkhay, V., Popov, I., Bormotov, D., Kostyukevich, Y., Karchugina, S., Indeykina, M., Bugrova, A., Starodubtseva, N., Potapov, A., Nikolaev, E.: A novel direct spray-from-tissue ionization method for mass spectrometric analysis of human brain tumors. Anal. Bioanal. Chem. 407, 7797–7805 (2015)CrossRefGoogle Scholar
- 28.Balog, J., Sasi-Szabó, L., Kinross, J., Lewis, M.R., Muirhead, L.J., Veselkov, K., Mirnezami, R., Dezső, B., Damjanovich, L., Darzi, A., Nicholson, J.K., Takáts, Z.: Intraoperative tissue identification using rapid evaporative ionization mass spectrometry. Sci. Transl. Med. 5, 194ra193–194ra193 (2013)CrossRefGoogle Scholar
- 31.Liu, H., Lang, J.H.: Is abnormal eutopic endometrium the cause of endometriosis? The role of eutopic endometrium in pathogenesis of endometriosis. Med. Sci. Monit: Int. Med. J. Experim. Clin. Res. 17, RA92–RA99 (2011)Google Scholar
- 32.Folch, J., Lees, M., Sloane Stanley, G.H.: A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226, 497–509 (1957)Google Scholar
- 39.Hsu, F.F., Turk, J.: Electrospray ionization with low-energy collisionally activated dissociation tandem mass spectrometry of glycerophospholipids: mechanisms of fragmentation and structural characterization. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 877, 2673–2695 (2009)CrossRefGoogle Scholar
- 41.Schiller, J., Suss, R., Petkovic, M., Hilbert, N., Muller, M., Zschornig, O., Arnhold, J., Arnold, K.: CsCl as an auxiliary reagent for the analysis of phosphatidylcholine mixtures by matrix-assisted laser desorption and ionization time-of-flight mass spectrometry (MALDI-TOF MS). Chem. Phys. Lipids. 113, 123–131 (2001)CrossRefGoogle Scholar
- 45.Worley, B., Powers, R.: Multivariate analysis in metabolomics. Curr. Metabolomics. 1, 92–107 (2013)Google Scholar
- 50.Dutta, M., Anitha, M., Smith, P.B., Chiaro, C.R., Maan, M., Chaudhury, K., Patterson, A.D.: Metabolomics reveals altered lipid metabolism in a mouse model of endometriosis. J. Proteome Res. 15, 2626-2633 (2016)Google Scholar