Ambient Ionization Mass Spectrometry for Cancer Diagnosis and Surgical Margin Evaluation

Transcription

1 Papers in Press. Published November 10, 2015 as doi: /clinchem The latest version is at Clinical Chemistry 62: (2016) Reviews Ambient Ionization Mass Spectrometry for Cancer Diagnosis and Surgical Margin Evaluation Demian R. Ifa 1 and Livia S. Eberlin 2* BACKGROUND: There is a clinical need for new technologies that would enable rapid disease diagnosis based on diagnostic molecular signatures. Ambient ionization mass spectrometry has revolutionized the means by which molecular information can be obtained from tissue samples in real time and with minimal sample pretreatment. New developments in ambient ionization techniques applied to clinical research suggest that ambient ionization mass spectrometry will soon become a routine medical tool for tissue diagnosis. CONTENT: This review summarizes the main developments in ambient ionization techniques applied to tissue analysis, with focus on desorption electrospray ionization mass spectrometry, probe electrospray ionization, touch spray, and rapid evaporative ionization mass spectrometry. We describe their applications to human cancer research and surgical margin evaluation, highlighting integrated approaches tested for ex vivo and in vivo human cancer tissue analysis. We also discuss the challenges for clinical implementation of these tools and offer perspectives on the future of the field. SUMMARY: A variety of studies have showcased the value of ambient ionization mass spectrometry for rapid and accurate cancer diagnosis. Small molecules have been identified as potential diagnostic biomarkers, including metabolites, fatty acids, and glycerophospholipids. Statistical analysis allows tissue discrimination with high accuracy rates ( 95%) being common. This young field has challenges to overcome before it is ready to be broadly accepted as a medical tool for cancer diagnosis. Growing research in new, integrated ambient ionization mass spectrometry technologies and the ongoing improvements in the existing tools make this field very promising for future translation into the clinic American Association for Clinical Chemistry The groundbreaking development of ambient ionization mass spectrometry (MS) 3 techniques has revolutionized the means by which molecular information can be obtained in real time from tissue samples. For decades, diagnosis of tissue samples in the clinical environment has been performed by skilled pathologists using light microscopy techniques. Hematoxylin and eosin (H&E), the routinely used dye for staining tissue sections, is powerful for highlighting the cell nuclei, cytoplasm and extracellular matrix, and therefore offers the pathologist a powerful view of the tissue structure and composition. Pathologists evaluate tissue cell morphology, structure, and organization within their neighboring environments, to derive a diagnosis of a variety of diseases, most notably cancer. With the advent of genomics and proteomics technologies, a variety of immunohistochemistry protocols were developed to stain specific protein markers and bring more specificity and sensitivity to tissue diagnosis. As a result, the role of immunohistochemistry in routine diagnostic pathology substantially increased, especially for the identification of cell lineages within a specific organ and for typing neoplasms (1). Genetic sequencing approaches have also been implemented in the clinic within the last decade with the goal of identifying specific mutations and chromosomal translocations in clinical samples. Characterization of inherited and tumor-specific genetic alterations in cancer not only assists in tumor diagnosis but also allows safer and more effective therapy choices (2). In the last 2 decades, MS imaging (MSI) has been gaining increased attention from the biological and medical communities as a powerful approach for tissue imaging and diagnosis (3 5). Secondary ion MS (SIMS) has been in use for surface analysis for decades (6). Yet, the inability of SIMS to detect intact large biomolecules and the complex instrumentation re- 1 Department of Chemistry, York University, Toronto, ON, Canada; 2 Department of Chemistry, The University of Texas at Austin, Austin, TX * Address correspondence to this author at: The University of Texas at Austin, 1907 Alguno Rd., Austin, TX Fax ; liviase@utexas.edu. Received August 17, 2015; accepted September 28, Previously published online at DOI: /clinchem American Association for Clinical Chemistry 3 Nonstandard abbreviations: MS, mass spectrometry; H&E, hematoxylin and eosin; MSI, MS imaging; SIMS, secondary ion MS; DESI, desorption electrospray ionization; DART, direct analysis in real time; LAESI, laser desorption ESI; EASI, easy ambient sonic-spray ionization; PESI, probe ESI; SPA-nanoESI, solid-probe assisted nanoesi; TS, touch spray; REIMS, rapid evaporative ionization MS; LASSO, least absolute shrinkage and selection operator; IDH1, isocitrate dehydrogenase 1; PCA, principal component analysis; HER2, human epidermal growth factor receptor 2; CUSA, cavitron ultrasonic surgical aspiration; V-EASI, Venturi-EASI. 1 Copyright (C) 2015 by The American Association for Clinical Chemistry

2 Reviews quired originally hindered its potential for routine tissue analysis. With the development of MALDI imaging, a bright future for tissue analysis by MS was recognized (7). Despite its requirements for matrix application to tissue samples and analysis under vacuum conditions, several studies have demonstrated MALDI s outstanding potential for routine cancer diagnosis based on the detection of proteomic signatures (5). Most recently, MALDI has become more accessible for routine and rapid clinical use through the development of matrix precoated glass slides, which greatly increases throughput in sample preparation (8). SIMS regained its potential for biological tissue analysis with the development of ion cluster sources, which allowed softer ionization of biomolecules from tissues (9, 10). Notably, SIMS has become the technique of choice for high spatial (lateral) resolution ( 1 m) tissue imaging by MS (11). It was not until the development of ambient ionization MS that a tangible path existed for routine use of MS technologies for tissue analysis in the clinical setting. What makes ambient ionization MS compatible for clinical use? The 2 major features are operational ease of use and real-time assessment of tissue molecular information. Unlike MALDI and SIMS, ambient ionization MS allows sample analysis in an open environment, at atmospheric pressure. Moreover, the majority of ambient ionization sources are operationally simple and relatively easy to assemble and to implement. Commercially available ambient ionization sources also tend to run at lower costs when compared to MALDI, SIMS, and other traditional high-performance MS technologies. Most notably, ambient ionization MS allows real-time assessment of tissue molecular information directly on tissue samples, without the need for sample preparation. A chemical mass spectrum displaying hundreds of compounds can be obtained from a selected region of a tissue sample placed under the ionization source in less than 1 s, depending on the choice of mass analyzer. These features were unimaginable 2 decades ago when the field of clinical MS was dominated by powerful yet time-consuming LC-MS approaches. Here, we provide a brief review of the historical progression of ambient ionization MS for tissue analysis and human cancer diagnosis, from the first attempts of cancer tissue imaging to the most recent uses of ambient ionization MS tools for in vivo cancer diagnosis. We also highlight interesting biochemical insights in cancer biology obtained with ambient ionization MSI of cancerous tissues, and present logical suggestions for approaches that would enable successful translation of this technology to the clinic. Ambient Ionization MS for Human Cancer Tissue Imaging and Diagnosis Ambient ionization MS techniques originated with the introduction of desorption electrospray ionization-ms (DESI-MS) in 2004 and direct analysis in real time (DART) in Since then, more than 30 ambient ionization techniques have been developed, including the commonly used laser desorption electrospray ionization (LAESI) (12) and easy ambient sonic-spray ionization (EASI) (13). These techniques often present small technical distinctions from each other and are classified according to their mechanisms of desorption and ionization. Furthermore, only a few provide imaging capability. For a detailed explanation and classification of the ambient ionization MS techniques, including major strengths, limitations, and figures of merits, very good articles and reviews can be found in the literature (14 18). DESI-MS has been the most extensively used ambient ionization MS technique for chemical imaging and diagnosis of tissue samples. In DESI-MS a spray of charged droplets is directed toward a tissue sample, allowing chemicals to be dissolved at the sample surface, ionized by mechanisms similar to ESI, and transferred into a mass spectrometer, where the mass-to-charge ratios (m/z) of molecular ions and their abundances are measured (19). The chemistry of the desorption process is similar to a solvent extraction experiment, and depending on the solvent of choice it is very effective for the desorption and ionization of small molecules (20). By rastering the tissue sample underneath the DESI-MS spray spot and collecting a mass spectrum at every point, it is possible to determine the distribution of numerous molecular species with a moderate spatial resolution. DESI has a typical lateral resolution of 200 m (21, 22); however, lower resolutions can be reached (23, 24). Probe ESI (PESI), solid-probe assisted nanoesi (SPAnanoESI) and touch spray (TS), introduced in 2007 (25), 2012 (26), and 2014 (27), respectively, are closely related techniques that use a probe (usually a sharp metal needle) to sample the tissue. An electrical potential is then applied to the probe to desorb, ionize, and transfer the tissue components to the mass spectrometer. These techniques have a lateral resolution in the range of 100 mto 1.0 mm (25 27) and have been successfully used to analyze tissue, including tissue pieces and frozen sections (Fig. 1). Rapid evaporative ionization MS (REIMS), introduced in 2009, was the first ambient ionization MS technique developed to integrate a routinely used surgical modality with MS in a single probe (28). The method employs commonly used thermal ablation surgical methods (electrosurgery and infrared laser surgery), which produce large amount of tissue-derived gaseous ions. A custom built electrosurgical handpiece, including the 2 Clinical Chemistry 62:1 (2016)

3 Ambient Mass Spectrometry for Cancer Diagnosis Reviews Fig. 1. (A), DESI-MS schematic diagram and application to brain tumor. Top spectrum shows the low tumor cell concentration region infiltrating into grey matter. Bottom spectrum shows DESI-MS of dense tumor region with high tumor cell concentration. DESI-MS ion images show the distribution of specific glycerophospholipids, and optical image of H&E-stained tissue section. [Figs. reproduced with permission from Balog et al. (29) and Dill et al. (42)]. (B), PESI-MS schematic diagram and application to colon cancer analysis. Top and bottom spectra show the presence of specific glycerophospholipids in normal and cancerous colon tissue, respectively. [Figs. reproduced with permission from Morrish et al. (49)]. (C), TS-MS schematic diagram and application to prostate cancer analysis. The mass spectrum shows the presence of specific glycerophospholipids in the region outlined which corresponds to prostate malignancy as determined by histopathological evaluation [Figs. reproduced with permission from Hiraoka et al. (25) and Agar et al. (61)]. Clinical Chemistry 62:1 (2016) 3

4 Reviews electrosurgical unit producing ions and a cutting blade, is connected to a 1 2-m long polytetrafluoroethylene tube to transmit the ions to a mass spectrometer for mass analysis through the use of a Venturi gas jet pump. REIMS has a lateral resolution in the range of mm (29). The first demonstration of ambient ionization MS for cancer diagnosis and profiling of marginal regions between cancer and normal tissue was reported a decade ago using a sample of metastatic human liver adenocarcinoma tissue (30). Different abundances of sphingomyelin lipids were observed in the cancerous regions of the tissue when compared to adjacent normal tissue by positive ion mode DESI-MS. These results showcased the ability of direct tissue analysis by ambient ionization MS of providing molecular information for cancer diagnosis. The profiling experiments were followed by a full demonstration of the imaging capabilities of DESI-MSI on a mouse brain tissue section (21, 31). In 2009, a small set of breast tissue samples was used in the first application of DESI-MS for human cancer tissue imaging (32). Differences in the relative intensities and spatial distributions of various lipids were observed between samples of benign ductal carcinoma and invasive ductal carcinoma. These preliminary experiments using ambient ionization MS for cancer tissue diagnosis were followed by larger studies using banked sets of cancer and adjacent normal tissue samples. In the earlier studies, histological analysis of an H&E stained tissue section adjacent to the one analyzed by ambient ionization MSI was performed by an expert pathologist to verify and compare tissue diagnosis with MSI data. In 2011, ambient ionization MSI of tissue sections without damage to the morphological structure of tissue cells was demonstrated though the development of nondestructive, histologically compatible solvent systems (33). This advance allowed the same tissue section to be chemically imaged and sequentially subjected to other tissue analysis protocols, including histochemistry for pathologic evaluation, and thus enabled unambiguous correlation between chemical and morphological information. Ultimately, it also allowed integration of DESI-MS into a routine clinical tissue analysis workflow, and was adopted by several research groups (33). Several types of human cancers have been investigated using ambient ionization MSI (Fig. 2). DESI- MSI has been successfully applied to investigate liver, breast, brain, kidney, prostate, bladder, stomach, colon, and rectum cancer tissues (Table 1). In the majority of the cases, a collection of lipid species encompassing different classes, such as fatty acids and glycerophospholipids, were observed as the discriminating molecules when comparing cancer and normal Fig. 2. Ambient ionization MS for human cancer tissue diagnosis. Different ambient ionization MS techniques have been used to successfully analyze and diagnose various types of human cancer tissues based on lipid and metabolic information. For more information, see Table 1. tissues. However, other specific metabolites such as cholesterol sulfate and 2-hydroxygluterate have been identified as potential individual biomarkers in human prostate and brain cancers, respectively. In a recent study, gadoteriol, a contrast agent used in MRI to target tumor sites, was mapped by DESI-MSI to reveal tumor margins and vasculature in human breast cancer tumors grown in mice. This work demonstrates that DESI-IMS can be applied for tissue discrimination based on the distribution of exogenous compounds (34). The probe-based techniques, PESI, SPA-nanoESI, and TS, have been applied to human kidney, colon, and prostate cancers. These techniques also identified lipids such as glycerophospholipids and triglycerides as discriminating molecules. REIMS has been applied to analyze cancer tissue from several organs, including stomach, colon, breast, liver, lung, and others, on the basis of lipid information, although not always in the imaging mode (Table 1). Ambient ionization MS provides a wealth of chemical information from a single pixel within a tissue sample, and thus calls for the use of statistical tools for appropriately evaluating the multidimensional data. A variety of supervised and unsupervised statistical methods and more advanced machine learning algorithms have been used to evaluate ambient ionization MS data from cancer tissue. Table 1 summarizes in chronological sequence the ambient ionization MS techniques that have been used to investigate human cancers, the main compounds de- 4 Clinical Chemistry 62:1 (2016)

5 Ambient Mass Spectrometry for Cancer Diagnosis Reviews Table 1. Chronological sequence of the ambient ionization MS techniques used to investigate human cancers, the type of tissue sample analyzed, the main compounds detected, and the statistical tools for data evaluation, when applied. Year Technique Type of tumor/organ Tissue sample type a Discriminating molecules b Statistical analysis c Reference 2005 DESI Adenocarcinoma/human liver Frozen tissue PC, SM No Wiseman et al. (30) 2009 DESI-MSI Adenocarcinoma/human liver; human breast Frozen tissue GP No Dill et al. (32) 2010 DESI-MSI Astrocytoma/human brain Frozen tissue PC, GalCer, SM No Eberlin et al. (45) 2010 DESI-MSI Papillary and renal cell carcinoma/human kidney 2010 DESI-MSI Adenocarcinoma and prostatic intraepithelial neoplasia/ human prostate 2011 DESI-MSI Transitional cell carcinoma/ human bladder 2011 DESI-MSI Oligodendroglioma, astrocytoma, and oligoastrocitoma/human brain Frozen tissue PI, PS, PG, FA PLS-DA, O-PLS Dill et al. (43) Frozen tissue CS PCA Eberlin et al. (71) Frozen tissue PI, PS, FA PCA, PLS-DA, Dill et al. (42) O-PLS Frozen tissue PI, PS, ST, FA SVM Eberlin et al. (46) 2011 LDI-MS Metastasis in liver/human colon Fresh ex-vivo Not assigned PCA Schaefer et al. (60) tissue 2011 DESI-MSI Seminoma/human testis Frozen tissue PI, PS, SL, CA No Masterson et al. (41) 2011 CUSA-MS Glioblastoma/human brain Fresh ex-vivo tissue Peptides, PCA, LDA Schaefer et al. (62) carbohydrates 2012 PESI-MS Cell carcinoma/human kidney Frozen tissue PC, TAG No Mandal et al. (72) 2012 DESI-MSI Colorectal adenocarcinoma Frozen tissue PI, PG, PE, PEp PCA, LDA Gerbig et al. (51) and liver metastasis/human liver, colon, and rectum 2013 REIMS Several types/several organs (stomach, colon, cecum, liver, lung, breast, brain, others) Fresh tissue ex-vivo and in-vivo GP, ST, SM PCA, LDA Balog et al. (58) 2013 DESI-MSI Glioblastoma, necrotic tumor/ human brain 2013 DESI-MSI Gliomas and meningiomas/ human brain Frozen tissue Not assigned SVM, PCA, plsa Calligaris et al. (73) Banked frozen PI, PS, PEp SVM Eberlin et al. (22) tissue 2013 PESI-MS Adenocarcinoma/human colon Frozen tissue PS, PI PCA Mandal et al. (53) 2013 SPA-ESI Renal cell carcinoma/human Frozen tissue PC, TAG PCA Mandal et al. (26) kidney 2014 DESI-MSI Gliomas/human brain Frozen tissue and fresh smeared tissue 2-HG No Santagata et al. (47) 2014 DESI-MSI Lymph node metastases/ human stomach 2014 DESI-MSI Human lymphoma/thymus Banked Frozen tissue 2014 DESI-MSI Stomach Banked frozen tissue 2014 DESI-MSI Invasive ductal carcinoma/ human breast 2014 DESI, TS Adenocarcinoma/human prostate 2015 DESI, TS Adenocarcinoma/human prostate 2015 DESI-MSI Infiltrating ductal carcinoma and infiltrating lobular carcinoma/human breast Frozen tissue PE, PI, PS PCA Abbassi-Ghadi et al. (74) GP, CL, acyl-pg SAM Eberlin et al. (48) GP, Gln, Glu, FA LASSO Eberlin et al. (44) Frozen tissue GP, FA PCA Calligaris et al. (55) Frozen tissue Not assigned No Kerian et al. (27) Frozen tissue PI PCA, LDA Kerian et al. (63) Frozen tissue PE, PC, PI, FA RMMC Guenther et al. (56) Continued on page XX Clinical Chemistry 62:1 (2016) 5

6 Reviews Table 1. Chronological sequence of the ambient ionization MS techniques used to investigate human cancers, the type of tissue sample analyzed, the main compounds detected, and the statistical tools for data evaluation, when applied. (Continued from page XX) Year Technique Type of tumor/organ 2015 REIMS imaging Colon rectal adenocarcinoma metastases in human liver 2015 DESI-MSI Fibroblastic and meningothelial meningiomas/human brain 2015 DESI-MSI Xenografts of human breast cancer tumor grown in mice Tissue sample type a Discriminating molecules b Statistical analysis c Reference Frozen tissue GP PCA Golf et al. (59) Frozen tissue GP PCA Calligaris et al. (54) Frozen tissue GDT (contrast agent) No Tata et al. (34) a Frozen tissues were stored at 80 C before cryosection and fresh tissues were analyzed without freezing. b PC, glycerophosphocholines; SM, sphingomyelins; GP, glycerophospholipids; GalCer, galactosylceramides; PI, glycerophosphoinositols; PS, glycerophosphoserines; PG, glycerophosphoglycerols; FA, fatty acids; CS, cholesterol sulfate; ST, sulfatides; SL, seminolipids; CA, citric acid; TAG, triacylglicerols; PE, glycerophosphoethanolamines; PEp, glycerophosphoethanolamines plasmalogen; 2HG, 2-hydroxygluterate; CL, cardiolipin; Gln, glutamine; Glu, glutamate; GDT, gadoteridol. c PLS-DA, partial least square discriminant analysis; O-PLS, orthogonal-partial least square; PCA, principal component analysis; SVM, support vector machine; LDA, linear discriminant analysis; plsa, probabilistic latent semantic analysis; SAM, significance analysis of microarrays; LASSO, least absolute shrinkage and selection operator; RMMC, recursive maximum margin criterion. tected, the type of tissue sample analyzed, and the statistical tools for data evaluation, when applied. Biochemical Insights into Cancer Metabolism by Ambient Ionization MS A trademark of ambient ionization MS for tissue analysis has been the favorable detection of small molecules, including lipids and metabolites (35). Thus, ambient ionization MS provides as an exciting opportunity to explore these underused molecules as diagnostic biomarkers. Cancer is largely caused by genetic abnormalities in protooncogenes and tumor suppressor genes that concomitantly regulate processes of cellular growth, metabolism, energy production, and lipid metabolism (36 38). Small metabolites involved in key regulatory steps of glucose transport, glutaminolysis, aerobic glycolysis (Warburg effect), and the Krebs cycle, and larger metabolites such as fatty acids and complex lipids are of great interest as their abnormal expression has been reported in various types of adenocarcinomas (39, 40). Aside from the broadly used HPLC-MS approaches, ambient ionization MSI is now one of the main tools for studying metabolic and lipid composition and distribution in neoplastic and other biological tissues. Ambient ionization MS is most powerful for identifying diagnostic molecules from biological tissues when used in combination with high-resolution mass analyzers and tandem MS for molecular characterization, and with biostatistical approaches for proper validation of diagnostic signatures. Because quantification has not been possible using ambient ionization MS, studies have effectively relied on variations in the relative abundances of lipids of a specific lipid class, or a variety of lipids species from many classes as means of identifying diagnostic signatures. Although the relative abundances of specific molecular markers obtained by ambient ionization MS are not an absolute picture of the composition of the tissue, the high reproducibility in molecular patterns reported by several groups has allowed successful tissue diagnosis. For example, DESI-MSI of testicular cancers revealed that lipids from a specific class exclusive to germ line cells, the seminolipids [seminolipid (16:0/16:0) and seminolipid (30:0)], were present in normal tubule testis tissue while undetectable in seminoma tissue (41). In bladder and kidney cancer, variations in the relative intensities of lipids, including fatty acids, glycerophosphoserines, glycerophosphoinositols, and glycerophosphoglycerols, were observed in cancer tissue when compared to adjacent normal tissue (42, 43). Using DESI-MSI and the least absolute shrinkage and selection operator (LASSO) biostatistical tool, several lipid and metabolic markers were identified in gastric cancer, normal epithelial, and stroma gastric tissue. High relative abundances of cardiolipins were observed in normal epithelial tissue, while sphingomyelin lipids were seen in higher relative abundance in normal stroma tissue. Peaks assigned as glutamine, glutamate, and palmitoylglycine, as well as polyunsaturated fatty acids, glycerophosphoserines, and glycerophosphoinositols were observed in high relative abundances in gastric cancer tissue (44). When DESI-MSI was applied to investigate brain cancers (45), variations in the abundances of lipids were observed as a function not only of disease state, but also cancer subtype, grade and tumor cell concentration. For example, sulfatides and galactoceramides, observed in low-grade astrocytomas (histological grades II and III), were undetectable in high-grade 6 Clinical Chemistry 62:1 (2016)

7 Ambient Mass Spectrometry for Cancer Diagnosis Reviews 4 Human genes: KRAS, Kirsten rat sarcoma viral oncogene homolog; NF2, neurofibromin 2 (merlin). astrocytomas (histological grade IV, or glioblastomas) by DESI-MSI (46). DESI-MSI has the potential to be used to pinpoint possible genetic mutation status and cancer molecular subtypes on the basis of metabolic signatures and lipid profiles. In gliomas, the oncometabolite 2-hydroxygluterate was detected by DESI-MSI from glioma tissue in tumors presenting a genetic mutation of the isocitrate dehydrogenase 1 (IDH1) enzyme. Overproduction and accumulation of 2-hydroxygluterate has been associated with a genetic mutation of the IDH1 enzyme, an indicator of better prognosis for patients. Detection of deprotonated 2-hydroxygluterate in the negative ion mode using high mass resolution DESI- MSI allowed accurate determination of IDH1 status in gliomas, even when immunohistochemistry was not useful for determining the genetic mutation (47). In a study using animal models and human lymphoma tissue, a correlation between increased intensities of various glycerophosphoglycerols and the overexpression of the MYC oncogene was observed using DESI-MSI and the statistical tool significance analysis of microarrays (48). The MYC oncogene has been known to regulate key genes involved in in glycolysis and lipid metabolism and has been strongly associated with the clinical aggressiveness of human cancers (49, 50). In human colorectal adenocarcinomas and liver metastasis, high mass resolution DESI-MSI allowed the identification of many lipids with increased abundance in colorectal adenocarcinoma tissue, including glycerophosphoetanolamines, glycerophosphoglycerols, and glycerophosphoinositols. Using principal component analysis (PCA) and linear discriminant analysis, the authors identified tissuespecific markers within different histological tissue types. Most notably, the lipidomic data obtained allowed identification of the Kirsten rat sarcoma viral oncogene homolog (KRAS) 4 genetic mutation in colorectal carcinomas with a 90% agreement with data from molecular biology approaches (51). KRAS has been reported to activate a number of enzymes playing important roles in lipid biosynthesis (52). Similar findings in terms of lipid distribution were reported in colon cancer in a smaller study using PESI-MS (53). In meningiomas, distinction between subtypes and identification of NF2 [neurofibromin 2 (merlin)] genetic aberrations was possible through DESI-MSI lipid data and statistical analysis (54). In breast cancer, a study using tissue samples obtained from patients with ductal carcinomas revealed an increase in the relative intensities of a variety of fatty acids in breast cancer tissue compared to adjacent normal tissue, including oleic, palmitic, arachidonic, and nervonic acids, as well as other glycerophospholipids (55). Interestingly, PCA clustering was observed between the estrogen receptor positive and progesterone receptor positive tumor center samples vs estrogen receptor negative and progesterone receptor negative samples, whereas no apparent separation was observed for human epidermal growth factor receptor 2 (Her2)-positive and Her2- negative tumor specimens. A different study using DESI- MSI to analyze breast cancer tissue biopsies reported similar findings (56). Among the highly significant peaks identified in high relative intensities in tumor tissue compared to morphologically normal glandular tissue were several fatty acids, glycerophosphoethanolamines, glycerophosphocholines, and glycerophosphoinositols. Using recursive maximum margin criterion analysis, diagnosis of breast cancer was achieved with an accuracy of 98.2%. Furthermore, a probability for correct discrimination of human receptor positive and human receptor negative tumors of 96% was observed, allowing diagnosis of the different molecule subtypes. Interestingly, an increase in the relative intensities of lactate and calcidiol in tumor-associated stroma tissue was observed when compared to stroma from histologically normal tissue, which could be due to the differentiation of cancer-associated fibroblasts and be associated with tumor development and progression into surrounding normal tissue (57). Clinical Studies Using Ambient Ionization MS Tools The field of cancer diagnosis by ambient ionization MS evolved with studies performed using banked tissue samples, which showcased the potential of ambient ionization MS for clinical use. Within the last 5 years, a few clinical studies were conducted through regularized patient recruitment and prospective collection of tissue samples from surgical practice. The main research goal that has led to this effort has been the use of ambient ionization MS for surgical margin evaluation, in particular, for evaluating frozen tissue sections obtained from surgery during or after surgical intervention, as well as fresh tissue pieces ex vivo and in vivo. Ambient ionization MS is particularly valuable for intrasurgical applications, because the invasive nature of the operation is compatible with the requirements of ambient ionization MS analysis and the diagnostic molecular information is highly desirable for rapid tissue diagnosis. REIMS was reported as a new ambient ionization technique in 2009 and offers the exciting possibility of in vivo, in situ MS tissue analysis (Fig. 3A) (28). Despite employing a different ionization mechanism, the mass spectra produced by REIMS are similar to those by obtained by DESI-MS, displaying the abundances of a variety of lipid markers. The online method, initially employed in ex vivo fresh porcine samples, was used in vivo during canine oncological surgery and enabled discrimi- Clinical Chemistry 62:1 (2016) 7

8 Reviews Fig. 3. Ambient ionization MS approaches used for in vivo and ex vivo human cancer tissue diagnosis. (A), A schematic representation of the online REIMS setup being using during surgery. REIMS has been tested for in vivo, intrasurgical diagnosis of a variety of human cancers [Figs. reproduced with permission from Calligaris et al. (54)]. (B) DESI-MS has been used in an offline approach to analyze ex vivo human glioma tissues intrasurgically. A timeline showing the speed of the DESI-MS method used during surgical workflow in comparison with histological approaches is shown. Visualization of 2-hydroxygluterate over three-dimensional MRI reconstruction of the tumor volume provides information on surgical margins [Figs. reproduced with permission from Dll et al. (43)]. (C), LDI-MS has been employed in surgery by coupling surgical lasers to a mass spectrometry online, and was shown to allow ex vivo cancer diagnosis (Figs. reproduced with permission from reference 56). (D), An online approach combining CUSA and V-EASI has been tested for ex vivo for human brain cancer diagnosis. A schematic representation of the CUSA system coupled to a Venturi pump and mass spectrometer inlet is shown [Figs. reproduced with permission from Balog et al.(58)) Balog. Although CUSA/V-EASI-MS and LDI-MS have so far only been tested in ex vivo human tissue, both are suitable for in vivo use. nation between a variety of tissue types and diagnosis in combination with PCA (29). In 2013, the REIMS setup was renamed as the intelligent knife (iknife) to describe tools that couple electrosurgical devices to MS and multivariate statistical techniques for rapid data analysis and patient diagnosis (58). The iknife was methodically employed to analyze a variety of ex vivo human cancer and normal tissues from which the authors constructed a spectral reference library. The complete system was then tested intrasurgically to analyze tissue in vivo, in cases for which electrosurgery was already being used for dissection and coagulation by the surgeon, and also ex vivo tumor tissue within the surgical suite. The entire procedure including sampling, sample transfer, chemical analysis, data processing, and audiovisual feedback takes approximately s, depending on the configuration of the instrumentation and the surgical procedure. From the 81 surgical procedures investigated with the iknife, tissue identification from the MS approach matched postoperative histopathology in 96.2% of the cases. Notably, in 15 cases in which the preoperative diagnosis and postoperative histopathological tissue identification were conflicting, the iknife technique was able to correctly identify the corresponding tissue in 11 cases, as corroborated by the postoperative definitive histological diagnosis. To date, REIMS is the only online approach that has been tested in vivo, 8 Clinical Chemistry 62:1 (2016)

9 Ambient Mass Spectrometry for Cancer Diagnosis Reviews during human cancer surgery. Although most of these analyses were performed in the profiling, point-to-point mode, the REIMS probe has recently been used in the imaging mode to produce spatially resolved images directly from fresh tissue samples (59). Another surgical modality coupled to MS is laser surgery (Fig. 3C) (60). The system was developed with commercially available surgical laser devices (both infrared and ultraviolet lasers), which produce ions for MS analysis by a mechanism similar to laser desorption ionization MS. The surgical lasers were connected to a mass spectrometer though polytetrafluoroethylene tubing and a Venturi gas jet pump, similarly to the iknife setup. The method was used to characterize and diagnose in vivo tissue during canine oncological surgery and ex vivo fresh human colon carcinoma samples. Two different approaches have been developed to couple cavitron ultrasonic surgical aspiration (CUSA) with previously developed ambient ionization MS techniques. CUSA is commonly used in brain surgery and employs mechanical disintegration of tissues and vacuum suction to remove resulting tissue debris and to avoid excessive tissue damage. CUSA was first combined offline with DESI-MSI to diagnose brain cancer (Fig. 3B) (61). Stereotactically registered tissues removed with the CUSA device were collected from a surgical case. The tissues were frozen, sectioned and analyzed by DESI-MS in the laboratory. Variations in the relative abundance of various glycerophosphocholines were tentatively associated with changes in tumor cell concentration in the positions of the brain tumor volume sampled. In a different approach, online coupling of CUSA with ambient ionization MS was performed by introducing the effluent of the CUSA device directly into a Venturi-EASI (V- EASI) source for ionization (Fig. 3D) (12, 62). A database comprising 284 histologically assigned spectra of human brain tissues and brain tumors was used to develop an identification algorithm. The method was tested ex vivo using porcine and human brain cancers in the laboratory, and besides commonly observed fatty acid and complex lipids, allowed the detection of other tissue constituents such as carbohydrates and peptides. The identification performance of the online CUSA/V-EASI system for glioblastoma and healthy brain samples was 100%, which was partially due to the limited number of samples. Offline coupling of CUSA with DESI-MSI was reported in a clinical study using 32 tissue samples collected from 5 research participants who underwent brain tumor resection, and the samples were analyzed in the laboratory (22). DESI-MSI results of stereotactically registered samples were correlated to preoperative MRI through neuronavigation and visualized over segmented 3-dimensional MRI tumor volume reconstruction. The molecular information from DESI- MSI provided accurate diagnosis in comparison to final histopathologic diagnosis, and valuable information on tumor cell concentration to define tumor boundary. A database built from previously collected mass spectra from banked tissue samples was used for classifying surgical tissue. Interestingly, in a case of oligodendroglioma grade III tumor, visualization of the tumor cell concentration results obtained by DESI-MSI over 3-dimensional MRI showed high tumor cell concentration in 3 positions in the tumor margin, raising the concern that residual tumor remained at the resection margin. This case clearly emphasized the need for real-time tissue characterization modalities and the potential that ambient ionization MS approaches have to minimize recurrence by tumor resection beyond the radiographic margins. The offline DESI-MSI approach was later used intraoperatively to analyze ex vivo brain cancer tissue during 2 brain cancer surgeries (47). In the study, a DESI-MS system was installed in the operating suite, and detection of 2-hydroxygluterate allowed real-time assessment of IDH1 genetic mutation status and surgical margin evaluation. Other offline DESI-MSI approaches have been used to analyze ex vivo human tissue obtained from surgical practice but analyzed in the laboratory for gastric, meningioma, breast, and prostate cancers. In gastric cancer, a statistical database built using the LASSO statistical approach and data from 68 human cancer samples ( mass spectra) were used to classify proximal and distal margin sections prospectively collected from 9 surgical cases (44). Notably, in a surgical case for which frozen section analysis by histopathology was very difficult, resulting in disagreement between pathologic readings, DESI-MSI identified both proximal and distal margins as negative for cancer, which corroborated the final diagnosis. This is a powerful example of how offline DESI-MSI could assist as a complementary tool for margin assessment from frozen sections. In meningioma, ex vivo DESI-MSI of 9 stereotactically registered surgical samples and 3 autopsy samples was performed, and the results for all samples indicated a strong correlation with histopathologic diagnosis (54). Two studies in breast cancer also employed offline DESI-MSI of prospectively collected surgical tissues to evaluate margin status. In one study, 61 breast tissue samples encompassing tumor and adjacent tissue were obtained from 14 patients with breast cancer who had undergone mastectomy and were collected from the tumor center, tumor edge, and normal breast tissue, as well as from contralateral breast tissue when bilateral mastectomies were performed (55). DESI- MSI allowed discrimination between normal and cancerous tissue, and hence delineation of tumor margins. In the second study, intraoperative biopsies were meticulously collected during surgery for the study from Clinical Chemistry 62:1 (2016) 9

10 Reviews Fig. 4. Offline and online approaches employing ambient ionization MS for cancer diagnosis surgical margin evaluation in clinical practice. Offline approaches rely on ex vivo analysis of tissue samples intrasurgically, or nearby the surgical suite. Rapid tissue processing may be employed prior to analysis by ambient MS. Mass analysis and statistical classification tools are employed providing valuable diagnostic information to the medical professionals (<5 min). Online approaches could employ commercially available as well as novel surgical handpieces for in vivo sampling and/or ionization of tissue samples. Tubing systems are employed to transfer samples and/or analytes from the surgical bed to the mass spectrometer. Mass analysis and statistical classification tools are employed to provide real time (<10 s) tissue diagnosis to the surgeon. the lesions (28 patients, 28 samples) and tumor bed (22 patients, 98 samples) and were analyzed by DESI- MSI (56). Using DESI-MSI, differentiation between tumor tissue and morphologically normal breast tissue was possible with an overall accuracy of 98.2%. DESI-MSI was also used to examine 100 samples of prostate cancer obtained from 12 human patients who underwent radical prostatectomy surgery (63). A database built using PCA linear discriminant analysis presented 95% accuracy in cross-validation. Besides DESI-MSI, the authors also used TS-MS to analyze 70 prostate biopsy samples from 6 patients, resulting in 110 mass spectra. This novel method provided prediction success rates measured against histopathology of 93%. Although both DESI-MS and TS were used offline on tissue sections, the study showcases the fact that both methodologies could be useful in differentiating tumor and normal prostate tissue at surgical margins intraoperatively. Perspectives and Conclusions Ambient ionization MS approaches provide the speed and ease of use together with outstanding molecular information on tissue composition that allows disease diagnosis and subtyping. As described here, many pilot studies in the laboratory and clinical studies have been performed to investigate its potential benefits, especially for intrasurgical margin evaluation. This young field has still many challenges to overcome and improvements to achieve before it is widely accepted as a medical tool for routine clinical use in cancer diagnosis. As surgical margin evaluation continues to be one of the main research focuses of clinical ambient ionization MS, a continuous effort will be made in developing surgical devices that can be used in situ, in vivo, during surgical intervention. We foresee 2 main approaches, online and offline, being used in the development of these tools (Fig. 4). The first online method would employ already commercially available surgical handpieces/ probes for tissue sampling/ionization coupled or not to tubing systems to transport samples/analytes from patient to the mass spectrometer for ionization and analysis. This approach is likely to be more easily acceptable to surgeons and medical staff who are already acquainted with their current surgical tools. From a logistic standpoint, it is also more straightforward for researchers to perform modifications on currently regularized medical 10 Clinical Chemistry 62:1 (2016)

11 Ambient Mass Spectrometry for Cancer Diagnosis Reviews devices to include ambient ionization MS capabilities. As different surgical modalities are employed depending on the cancer being treated and on practice-pattern variations, various surgical tools will be developed and employed in a variety of institutions. The iknife approach, for example, has a great potential to be successfully used in vivo during a variety of surgical interventions; nevertheless, the electrocauterization process employed results in thermal and mechanical damage of the tissue, which is undesirable and unused in the treatment of certain cancers types. In a second online approach, new surgical handpieces/tools used for the purpose of providing samples and/or analytes for MS analysis would be developed. These tools could be designed as a variation of already currently available ambient ionization MS modalities, such as DESI-MS, PESI, PS-MS, and others. However, many technical variations and regulatory approvals would be required for their in vivo use. For example, DESI-MS employs high gas flow rates and organic solvents, which are not readily compatible with living organs. Operational changes would be necessary to make DESI-MS usable for direct, in vivo tissue analysis. A probe based on a modified form of DESI-MS has been designed for possible in vivo use, although it has not been tested for human cancer tissue diagnosis (64). Sampling techniques such as PESI, TS, and SPA-nanoESI could be more easily implemented because they provide the material needed for ambient ionization MS analysis in a less aggressive form. New applications of solid-phase extraction (65) and cold laser (66) to tissue sampling have been emerging and offer new possibilities for the development of ambient ionization MS tools for in vivo cancer diagnosis. Note that because in vivo approaches rely on the transport of samples and/or analytes from the surgical bed to the mass spectrometer, in cases in which tubing systems are used, possible issues with carryover and contamination must be systematically evaluated. Although integrated surgical probes are an exciting possibility for in vivo use of ambient ionization MS techniques, offline approaches offer similar clinical benefits while requiring a slightly longer processing time. We anticipate that DESI-MS analysis of tissue prepared from ex vivo tissue samples during surgery will become a common clinical tool, complementary to standard histopathology of H&E stained sections. Although it does not deliver real-time diagnosis, the molecular information obtained from tissues provides valuable diagnostic information, especially when histological assessment is difficult and there is disagreement between the pathologic readings. Other ambient ionization MS techniques are likely to be implemented in the offline approach. Direct analysis of tissue smears and touch preps quickly prepared from ex vivo surgical tissue by ambient ionization MS are likely to become routinely applied in the operating room. As the field continues to mature, researchers and medical institutions should strive to share ambient ionization MS results to unify and validate the approaches being developed. Dedicated pathologic support and software tools are necessary for accurately correlating molecular and histological information and defining diagnostic patterns (67). Concomitantly, an effort should be made to assure the most appropriate statistical planning approaches and tools are used for interpreting and classifying mass spectra. For example, although many studies have relied on PCA for grouping spectra by similarity and demonstrating discrimination, more sophisticated statistical approaches for classification and feature selection may yield better results with similar computational requirements (44, 68, 69). Another critical factor that has been underexplored is the sensitivity of ambient ionization MS approaches in detecting cancer cells at a low tumor cell concentration, a common scenario in the evaluation of samples at the margins of the tumor. Future studies should seek to demonstrate the diagnostic capabilities on tissues of high cellular heterogeneity and low tumor cell concentrations. As efforts are pursued to use the molecular information obtained by ambient ionization MS to better understand cancer biochemistry, statistical approaches to properly identify and validate possible biomarkers using large cohorts of samples should be employed. Furthermore, the inherent matrix effects issues of direct tissue analysis by MS should always be considered when evaluating molecular findings, and follow-up studies with standard HPLC-MS approaches are important for validation (70). Because matrix effects hinder ionization of certain classes of molecules, targeted analyses of specific diagnostic compounds are not readily possible using ambient ionization MS techniques and may prevent specific clinical applications. New approaches that provide a more comprehensive analysis of the molecular components of tissue samples are needed to increase diagnosis capabilities of ambient ionization MS techniques. Similarly, methodologies that would allow absolute quantification of specific molecules from tissue would expand the clinical uses of the techniques (18). Nevertheless, we anticipate that studies with the goal of correlating mass spectra patterns with molecular and histological subtypes of diseases will be increasingly pursued and will provide insights into disease markers and aberrant biological pathways to improve therapeutic approaches. Integration of ambient ionization MS tools with other imaging modalities already used in the clinic, such as MRI and PET (positron emission tomography), is important to provide comprehensive molecular evaluation of tissue to medical professionals and ultimately assist doctors in making critical medical decisions. Future stud- Clinical Chemistry 62:1 (2016) 11

12 Reviews ies regarding the direct impact of the use of ambient ionization MS in improving surgical outcomes and patient survival will help determine the most suitable and reliable tools, as well as the tangible benefits of ambient ionization MS technologies, beyond what is already achieved with histopathology. Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. References Authors Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: None declared. Research Funding: L.S. Eberlin, National Institutes of Health/National Cancer Institute (NIH/NCI) through the K99/R00 Pathway to Independence Award (grant 1K99CA ). Expert Testimony: None declared. Patents: D.R. Ifa, patent number WO A1; L.S. Eberlin, Patent numbers US A1, US A1, US A1, and WO A1. 1. Mino-Kenudson M, Chirieac LR, Law K, Hornick JL, Lindeman N, Mark EJ, et al. A novel, highly sensitive antibody allows for the routine detection of alk-rearranged lung adenocarcinomas by standard immunohistochemistry. Clin Cancer Res 2010;16: Jones S, Anagnostou V, Lytle K, Parpart-Li S, Nesselbush M, Riley DR, et al. Personalized genomic analyses for cancer mutation discovery and interpretation. Sci Transl Med 2015;7. 3. McDonnell LA, Heeren RMA. Imaging mass spectrometry. Mass Spectrom Rev 2007;26: Walch A, Rauser S, Deininger S-O, Hoefler H. Maldi imaging mass spectrometry for direct tissue analysis: A new frontier for molecular histology. Histochem Cell Biol 2008;130: Seeley EH, Caprioli RM. Molecular imaging of proteins in tissues by mass spectrometry. Proc Natl Acad Sci USA 2008;105: Braun RM, Blenkinsopp P, Mullock SJ, Corlett C, Willey KF, Vickerman JC, Winograd N. Performance characteristics of a chemical imaging time-of-flight mass spectrometer. Rapid Comm Mass Spectrom 1998;12: Sarsby J, Griffiths RL, Race AM, Bunch J, Randall EC, Creese AJ, Cooper HJ. Liquid extraction surface analysis mass spectrometry coupled with field asymmetric waveform ion mobility spectrometry for analysis of intact proteins from biological substrates. Anal Chem 2015;87: Yang J, Caprioli RM. Matrix precoated targets for direct lipid analysis and imaging of tissue. Anal Chem 2013; 85: Carado A, Kozole J, Passarelli M, Winograd N, Loboda A, Bunch J, et al. Biological tissue imaging with a hybrid cluster sims quadrupole time-of-flight mass spectrometer. Appl Surf Sci 2008;255: Kiss A, Smith DF, Jungmann JH, Heeren RMA. Cluster secondary ion mass spectrometry microscope mode mass spectrometry imaging. Rap Commun Mass Spectrom 2013;27: Klitzing HA, Weber PK, Kraft ML. Secondary ion mass spectrometry imaging of biological membranes at high spatial resolution. Meth Mol Biol 2013;950: Santos VG, Regiani T, Dias FFG, Romao W, Jara JLP, Klitzke CF, et al. Venturi easy ambient sonic-spray ionization. Anal Chem 2011;83: Haddad R, Sparrapan R, Eberlin MN. Desorption sonic spray ionization for (high) voltage-free ambient mass spectrometry. Rapid Commun Mass Spectrom 2006; 20: Alberici RM, Simas RC, Sanvido GB, Romao W, Lalli PM, Benassi M, et al. Ambient mass spectrometry: Bringing ms into the real world. Anal Bioanal Chem 2010;398: Wu C, Dill AL, Eberlin LS, Cooks RG, Ifa DR. Mass spectrometry imaging under ambient conditions. Mass Spectrom Rev 2013;32: Hsu C-C, Dorrestein PC. Visualizing life with ambient mass spectrometry. Curr Opin Biotecch 2015;31: Harris GA, Galhena AS, Fernandez FM. Ambient sampling/ionization mass spectrometry: Applications and current trends. Anal Chem 2011;83: Ellis SR, Bruinen AL, Heeren RMA. A critical evaluation of the current state-of-the-art in quantitative imaging mass spectrometry. Anal Bioanal Chem 2014;406: Takats Z, Wiseman JM, Gologan B, Cooks RG. Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science 2004;306: Badu-TawiahAK, EberlinLS, OuyangZ, CooksRG. Chem aspects of the extractive methods of ambient ionization mass spectrometry. Ann Rev Phys Chem 2013;64: Ifa DR, Wiseman JM, Song Q, Cooks RG. Development of capabilities for imaging mass spectrometry under ambient conditions with desorption electrospray ionization (desi). Int J Mass Spectrom 2007;259: Eberlin LS, Norton I, Orringer D, Dunn IF, Liu X, Ide JL, et al. Ambient mass spectrometry for the intraoperative molecular diagnosis of human brain tumors. Proc Natl Acad Sci USA 2013;110: Campbell DI, Ferreira CR, Eberlin LS, Cooks RG. Improved spatial resolution in the imaging of biological tissue using desorption electrospray ionization. Anal Bioanal Chem 2012;404: Kertesz V, Van Berkel GJ. Improved imaging resolution in desorption electrospray ionization mass spectrometry. Rapid Commun Mass Spectrom 2008;22: Hiraoka K, Nishidate K, Mori K, Asakawa D, Suzuki S. Development of probe electrospray using a solid needle. Rapid Commun Mass Spectrom 2007;21: Mandal MK, Yoshimura K, Saha S, Ninomiya S, Rahman MO, Yu Z, et al. Solid probe assisted nanoelectrospray ionization mass spectrometry for biological tissue diagnostics. Analyst 2012;137: Kerian KS, Jarmusch AK, Cooks RG. Touch spray mass spectrometry for in situ analysis of complex samples. Analyst 2014;139: Schaefer K-C, Denes J, Albrecht K, Szaniszlo T, Balog J, Skoumal R, et al. In vivo, in situ tissue analysis using rapid evaporative ionization mass spectrometry. Angew Chem Int Ed 2009;48: Balog J, Szaniszlo T, Schaefer K-C, Denes J, Lopata A, Godorhazy L, et al. Identification of biological tissues by rapid evaporative ionization mass spectrometry. Analy Chem 2010;82: Wiseman JM, Puolitaival SM, Takats Z, Cooks RG, Caprioli RM. Mass spectrometric profiling of intact biological tissue by using desorption electrospray ionization. Angew Chem Int Ed 2005;44: Wiseman JM, Ifa DR, Song QY, Cooks RG. Tissue imaging at atmospheric pressure using desorption electrospray ionization (desi) mass spectrometry. Angew Chem Int Ed 2006;45: Dill AL, Ifa DR, Manicke NE, Zheng O, Cooks RG. Mass spectrometric imaging of lipids using desorption electrospray ionization. J Chromatogr B Analyt Technol Biomed Life Sci 2009;877: Eberlin LS, Ferreira CR, Dill AL, Ifa DR, Cheng L, Cooks RG. Non-destructive, histologically compatible tissue imaging by desorption electrospray ionization mass spectrometry.chembiochem 2011;12: Tata A, Zheng J, Ginsberg HJ, Jaffray DA, Ifa DR, Zarrine- Afsar A. Contrast agent mass spectrometry imaging reveals tumour heterogeneity. Anal Chem 2015;87: Eberlin LS, Ferreira CR, Dill AL, Ifa DR, Cooks RG. Desorption electrospray ionization mass spectrometry for lipid characterization and biological tissue imaging. Biochim Biophys Acta Mol Cell Biol Lipids 2011;1811: Bishop JM. Molecular themes in oncogenesis. Cell 1991;64: Corda D, Zizza P, Varone A, Filippi BM, Mariggio S. The glycerophosphoinositols: Cellular metabolism and biological functions. Cell Mol Life Sci 2009;66: Soga T. Cancer metabolism: Key players in metabolic reprogramming. Cancer Sci 2013;104: Utsugi T, Schroit AJ, Connor J, Bucana CD, Fidler IJ. Elevated expression of phosphatidylserine in the outermembrane leaflet of human tumor-cells and recognition by activated human blood monocytes. Cancer Res 1991;51: Dobrzynska I, Szachowicz-Petelska B, Sulkowski S, Figaszewski Z. Changes in electric charge and phospholipids composition in human colorectal cancer cells. 12 Clinical Chemistry 62:1 (2016)