Search

The Supplementary information file contains: (i) Supplementary Methods, concerning PCR primers, antibodies and explanation of the chemometric approach. (ii) Supplementary Figures 1-4, which provide extended data for the multivariable models of Figure 2 and 3, MS/MS analyses, and further gene expression data supporting Figure 5. (iii) Supplementary Tables 1-4, which provide demographics of the cohorts and univariate lipid analyses from Cohort 1 and 2.

Table S1: Glycerophospholipid relative abundances compared between lymph node metastases (LNM) and primary tumor by means of ANOVA. Table S2: Glycerophospholipid relative abundances compared between malignant tissue (lymph node metastases and primary tumor) and non-malignant lymph node tissue types by means of ANOVA. Table S3: Contingency table comparing immunohistochemistry reference test and mass spectrometry image tissue classification prediction images. Table S4: Contingency Table comparing immunohistochemistry reference test and mass spectrometry image tissue classification prediction images (TCPI) - with re-classification of positive capsular identification on TCPI as normal lymph nodes. Figure S1: DESI-MS of normal lymph node with carbon deposits from smoking. Figure S2: DESI-MS of lymph node with macro-metastases. Figure S3: Internal cross validation - lipidomic profile of lymph node metastases versus lymph node metastases response to chemotherapy. Figure S4: Mass spectrometry image tissue classification prediction images compared to matched AE1/AE3 immunohistochemistry images of 90 lymph node (LN) data set. Figure S5: DESI-MS of mouse brain cryo-section with a spatial resolution of 20μm per pixel. Figure S6: DESI-MS of mouse brain cryo-section with a spatial resolution of 75μm and 30 scans per second. Supplementary Figure S7: Image alignment of histology image and Mass Spectrometry Image (MSI) for tissue specific pixel labeling. Supplementary Figure S8: Overview for the creation of a composite database of tissue specific mass spectra. Tissue specific mass spectra extracted from representative pixels of Mass Spectrometry Images from individual samples are used to populate combined data units of that tissue type with further sub-classification of its sample origin.

Histopathological assessment of lymph node metastases (LNM) depends on subjective analysis of cellular morphology with inter-/intraobserver variability. In this study, LNM from esophageal adenocarcinoma was objectively detected using desorption electrospray ionization-mass spectrometry imaging (DESI-MSI). Ninety lymph nodes (LN) and their primary tumor biopsies from 11 esophago-gastrectomy specimens were examined and analyzed by DESI-MSI. Images from mass spectrometry and corresponding histology were coregistered and analyzed using multivariate statistical tools. The MSIs revealed consistent lipidomic profiles of individual tissue types found within LNs. Spatial mapping of the profiles showed identical distribution patterns as per the tissue types in matched IHC images. Lipidomic profile comparisons of LNM versus the primary tumor revealed a close association in contrast to benign LN tissue types. This similarity was used for the objective prediction of LNM in mass spectrometry images utilizing the average lipidomic profile of esophageal adenocarcinoma. The multivariate statistical algorithm developed for LNM identification demonstrated a sensitivity, specificity, positive predictive value, and negative predictive value of 89.5%, 100%, 100%, and 97.2%, respectively, when compared with gold-standard IHC. DESI-MSI has the potential to be a diagnostic tool for perioperative identification of LNM and compares favorably with techniques currently used by histopathology experts. Cancer Res; 76(19); 5647–56. ©2016 AACR.

Full methodological details of data analysis with respect to: tissue specific mass spectra extraction; data pre-processing (mass range selection, peak alignment, normalization, de-noising, data averaging); multivariate statistical models; glycerophospholipid annotation and individual glycerophospholipid comparison between tissue types.

The incidence of esophageal adenocarcinoma is rising, survival remains poor, and new tools to improve early diagnosis and precise treatment are needed. Cancer phospholipidomes quantified with mass spectrometry imaging (MSI) can support objective diagnosis in minutes using a routine frozen tissue section. However, whether MSI can objectively identify primary esophageal adenocarcinoma is currently unknown and represents a significant challenge, as this microenvironment is complex with phenotypically similar tissue-types. Here, we used desorption electrospray ionization-MSI (DESI-MSI) and bespoke chemometrics to assess the phospholipidomes of esophageal adenocarcinoma and relevant control tissues. Multivariate models derived from phospholipid profiles of 117 patients were highly discriminant for esophageal adenocarcinoma both in discovery (AUC = 0.97) and validation cohorts (AUC = 1). Among many other changes, esophageal adenocarcinoma samples were markedly enriched for polyunsaturated phosphatidylglycerols with longer acyl chains, with stepwise enrichment in premalignant tissues. Expression of fatty acid and glycerophospholipid synthesis genes was significantly upregulated, and characteristics of fatty acid acyls matched glycerophospholipid acyls. Mechanistically, silencing the carbon switch ACLY in esophageal adenocarcinoma cells shortened glycerophospholipid chains, linking de novo lipogenesis to the phospholipidome. Thus, DESI-MSI can objectively identify invasive esophageal adenocarcinoma from a number of premalignant tissues and unveils mechanisms of phospholipidomic reprogramming.Significance:These results call for accelerated diagnosis studies using DESI-MSI in the upper gastrointestinal endoscopy suite, as well as functional studies to determine how polyunsaturated phosphatidylglycerols contribute to esophageal carcinogenesis.

Supplementary Figure Legends

Background: Tumor-derived Extracellular Vesicles (EVs) are found in body fluids of cancer patients. These EVs include exosomes and microvesicles, which play an important role in tumor development and metastasis. Although EVs have great potential as liquid biopsies for early detection of breast cancer, the lipid composition of breast cancer-derived EVs is unknown. Here, we studied the lipid composition of EVs and its potential for the identification of body fluid-based biomarkers for breast cancer diagnosis. Methods: The lipid composition of EVs and their parental cells were evaluated for nine breast cancer and two non-cancerous mammary cell lines. EVs were isolated by size-exclusion chromatography (SEC). Furthermore, ten blood plasma samples were studied from women diagnosed with primary breast cancer before surgery, nine samples from patients with progressive metastatic breast cancer and ten samples from healthy women who neither had pre-existing medical conditions nor infections at the time the samples were collected or in the previous weeks. EVs were obtained from blood plasma samples by a combination of density gradient ultracentrifugation and SEC. Lipids were extracted from EVs and analyzed by reverse-phase liquid chromatography mass-spectrometry (LC-MS). Results: We found that EVs produced by breast cancer cell lines were enriched in sphingomyelins and ceramides when compared to their parental cells. Furthermore, phospholipids such as phosphatidylcholine and phosphatidylethanolamine were found to be abundant in these EVs. Multivariate analyses showed that EVs released by breast cancer cells can be distinguished from those released by non-cancerous cells based on their phospholipid and sphingolipid composition (AUCROC=0.99, p Conclusion: Our findings demonstrate the potential of the lipid content of cancer-derived EVs for the identification of lipid biomarkers for breast cancer diagnosis and its use for the development of non-invasive (or minimally invasive) diagnostic methods based on blood plasma analysis. Citation Format: Erika Dorado, M Luisa Doria, Anika Nagelkerke, James S McKenzie, Stefania Maneta-Stavrakaki, Charlotte Ion, Thomas Whittaker, Jeremy Nicholson, Molly M Stevens, R Charles Coombes, Zoltan Takats. Lipidomic analysis of extracellular vesicles and its potential for the identification of body fluid-based biomarkers for breast cancer diagnosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 459.

Introduction A multi-modal analysis approach using desorption electrospray ionization (DESI-MSI) and RNA-Seq can envision a complete metabolic and genetic information from clinical specimens revealing tumour heterogeneity. Coupled with laser capture microdissection (LCM) provides a compelling opportunity for molecular sub-characterizing of the tumour tissues. The envisioned combination analysis raises special requirements, including short LCM time, to prevent RNA degradation during microdissection at room temperature. The isolated RNA must have sufficient quality and quantity to carry out RNA seq for transcriptomics. Objectives The aim of this study was developing a multi-modal analysis protocol obtaining metabolic clusters to get a more in-depth knowledge for tumour-heterogeneity from Patient-derived Xenografts (PDXs) and clinical specimens. Methods PDXs and primary tissue biopsies from patients with Breast cancer were cryosectioned at ten μm and mounted on PEN membrane glass slides, which are unique slides for LASER Capture Microdissection (LDM). The DESI imaging analysis area was obtained line-by-line using the DEFFI sprayer. The analyzed tissue sections were stained with H&E and annotated by a histopathologist to allow the alignment of optical and MSI images. Next, the areas of interest in the same slide were microdissected by Laser Capture Microdissection for LC-MS and RNA-seq (Leica LDM 7000). RNA was isolated with a commercial kit (Qiagen RNeasy Micro Kit). Finally, standardization of RNA quality control was done by the Agilent 2100 Bioanalyzer System and followed by RNA Sequencing. Results Preliminary results showed that the extracted samples from microdissected sections using Laser Capture Microdissected for LC-MS could be used to validate the metabolites and lipids, which already had been imaged by DESI-MSI. These DESI-MSI and LC-MS results, which obtained from specific areas on the tissue sections can be attributed to identifying metabolically different sub-clones in the adjacent tumour sections. The next identification method for sub-cloning is a transcriptomic approach. For the transcriptomic study, the results of Agilent showed that the RNA quality of samples was sufficiently competent to carry out downstream analysis, including RNA seq. RNA seq can identify specific gene expression of the pathways, which are related to the identified metabolic profiling by DESI-MSI and LC-MS. Conclusion: We found that developing a multi-modal analysis protocol coupled to Laser capture microdissection is a promising approach for the identification of metabolic heterogeneity in the cancerous specimens. Citation Format: Emine Kazanc, Evi Karali, Vincen Wu, Paolo Inglese, James McKenzie, Aurelien Tripp, Nikos Koundouros, Thanasis Tsalikis, Hiromi Kudo, George Poulogiannis, Zoltan Takats. A multimodal analysis in breast cancer: Revealing metabolic heterogeneity using DESI-MS imaging with Laser-microdissection coupled transcriptome approach [abstract]. In: Proceedings of the AACR Virtual Special Conference on Tumor Heterogeneity: From Single Cells to Clinical Impact; 2020 Sep 17-18. Philadelphia (PA): AACR; Cancer Res 2020;80(21 Suppl):Abstract nr PO-042.

Objectives Progression through the adenoma-adenocarcinoma sequence has been shown to demonstrate hallmark shifts in the tissue lipidome. The incorporation of certain phospholipid subclasses into cellular membranes has previously been found to correlate with susceptibility to external stressors through alteration of membrane biodynamics. On this basis, it is hypothesized that modulation of cellular lipid modulation may influence radioresistance in rectal cancer. A desorption electrospray ionization mass spectrometry imaging (DESI-MSI) platform was utilized to demonstrate lipidomic shifts reflective of tumor heterogeneity and modulation of cellular lipids as a radioresistance mechanism. Methods 26 specimens of fresh, human rectal cancer tissue were harvested at surgical resection from matched cohorts; 13 of each from those who had received RT, and those who proceeded straight to surgery (STS). DESI-MSI was performed using an Orbitrap Fourier transform mass spectrometer in negative ion mode (mass range 150-2000m/z). Lipid spectra acquired were spatially aligned to subsequent H&E staining of the tissue section in a chemometric toolbox in MatLab, annotating morphologically distinct regions of the tumour microenvironment. Statistical modeling was performed incorporating clinicopathological metadata. ANOVA identified spectral peaks with discriminating power between both RT and STS groups, and between clusters of patients across the range of Mandard tumor regression grades (TRG). Lipid classes and putative structures were determined using an existing MS/MS data reference library (MetLin). Results Significant clustering of lipid species was demonstrated in both RT and STS groups, for both adenocarcinoma and stroma. Leave-One-Out, all-pixels cross-validation (LOOCV) yielded high predictive values (>97.9%) for RT vs. STS adenocarcinoma pixels. Cancers subjected to RT demonstrated significantly increased abundance of phospholipids compared to the STS group, notably phosphatidylglycerols (PG), phosphatidylethanolamines (PE) and phosphatidylserines (PS) on univariate analysis of all spectra, with a defined p/q value of 0.05 and negative log2 fold change of 0.1. Three-class PCA and discriminant analysis demonstrated clustering of samples according to Mandard TRG 1-2, TRG 3, and TRG 4-5. Clear separation of spectra was observed on LOOCV between all three classes. Compounds in the range 240-450m/z were implicated in this shift, in addition to the phospholipids as described. Conclusion This work has demonstrated a putative role for phospholipids and fatty acids in radioresistance, in a human rectal cancer cohort. DESI-MSI provides a unique foundation from which to continue ‘bottom-up' pathway analysis - utilizing both immunohistochemistry and tissue genotyping – which will assist in elucidating the possible mechanisms of response to radiotherapy in rectal cancer. Citation Format: Liam Robert Poynter, Reza Mirnezami, Renata Soares, James McKenzie, Pranav Patel, Nadia Peppa, Alexander Mirnezami, James Kinross, Ara Darzi, Zoltan Takats. Modulation of cellular phospholipids correlates with tumor regression grade and radio resistance in rectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3203.