Spatially resolved inteRASomics: Can MS-imaging decipher RAS mutational status in thyroid cancer looking at RAS interacting proteins?

Introduction Thyroid cancer is one of the first tumours where activating RAS mutations were discovered [1]. In particular, RAS are the most frequently mutated oncogenes in follicular patterned lesions, and they have been associated with all three mutant isoforms of the RAS gene (NRAS, HRAS and KRAS). Among the follicular patterned lesions, the non-invasive follicular tumours with papillary- like features (NIFTPs) is a challenging lesion that does not have a clear cut benign or malignant morphology both in the pre-operative/cytological setting, and even after the analysis of the surgical specimens, thus maintaining its indeterminate status [2]. While the NIFTPs are often characterized by RAS-type mutations, the NIFTP category is becoming a basket that include also a large number of questionable cases which are however RAS wild-type. Moreover, the RAS mutational status information alone is not sufficient to give an accurate and correct indication regarding patient follow-up. The aim of the present study was to further characterize RAS- mutated and RAS wild-type NIFTPs lesions through their proteomic signature using a spatial proteomics approach. Matrix-Assisted Laser Desorption/Ionization (MALDI)–Mass Spectrometry Imaging (MSI) represents a cutting- edge technology to detect, directly in situ, small cell subpopulations based on their different molecular profiles, allowing the integration of molecular and histological information on tissue samples to point out early molecular alterations, even within regions that are indistinguishable at the microscopic level. Methods Archived FFPE samples from ten NIFTP (n=6 RAS-mutated and n=4 RAS-wild type) were retrieved from the archives of the Department of Pathology, IRCCS Fondazione San Gerardo dei Tintori (University of Milano-Bicocca (UNIMIB)), Italy and were genetically characterized for their RAS mutational status in collaboration with the Euopean Institute of Oncology using a Next Generation Sequencing (NGS) approach. For MALDI-MSI analysis, tissues were trypsin digested and α-cyano-4- hydroxycinnamic acid (CHCA, (10 mg/μL in 70% ACN, 30% H2O and 1% TFA) matrix was deposited onto the slides. MALDI-MSI proteomics analysis using a rapifleX MALDI TissuetyperTM MALDI-TOF/TOF MS (Bruker Daltonik GmbH, Bremen, Germany) equipped with a Smartbeam 3D laser operating at 2kHz frequency was performed. Mass spectra were acquired in reflectron-positive mode within the m/z 750 to 3000 mass range and images were acquired with a 50 x 50 μm spatial. After the analysis, MALDI matrix was removed and digested peptides were identified by nLC-ESI- MS/MS. Finally, tissue samples were stained with haematoxylin and eosin (H&E) thus allowing the integration of proteomic and morphological data. Regions of interest (ROIs) of NIFTP nodules were annotated by the pathologist. Results A total of 425 signals were detected in the MALDI-MSI dataset and all features were included in a pixel-by-pixel segmentation using an unsupervised hierarchical clustering approach highlighting: i) the presence of the nodular lesions that arose in the context of normal thyroid parenchyma, with the spectra deriving from the nodule and the parenchyma being clustered under separate nodes, ii) under the nodule node, a further separation enlightened the presence of both NIFTP nodular lesions and hyperplastic lesions, with the spectra deriving from the NIFTP nodules and the hyperplastic nodules being clustered under separate nodes. However, this untargeted approach could not highlight differences in the NIFTP nodules following their RAS mutational status. As a consequence, we introduced a novel inteRASomics approach that investigated spatially resolved RAS interacting proteins to decipher RAS mutational status. Hence, our data were combined with the protein- protein RAS interaction network obtained using the IntAct database. RAS interactome highlighted 526 and 559 interactors with NRAS and HRAS, respectively, and in our data, 20 out of 727 proteins identified by LC–ESI-MS/MS were common interactors with NRAS and HRAS. Among these, only four proteins (PPIA, ATP1A1, CANX, BCAP31) were identified with an error lower than 100 ppm in our MALDI–MSI analysis and their signals were used for an unsupervised pixel-by-pixel automatic segmentation. This targeted MALDI–MSI proteomic approach showed that the NIFTPs were stratified in two groups, corresponding to RAS-mutant and wild- type, thus highlighting the potential role of spatially resolved proteomics tool to interrogate the protein interactomes of RAS in order to gain more insight into RAS oncogene in thyroid cancer. Conclusions These results underlined the unique capability of spatial proteomics to detect the proteomic signatures of RAS-mutated and RAS-wild-type NIFTP lesions, highlighting that RAS-mutated nodules have different proteomic signatures from RAS wild-type, and that the latter share similarities with hyperplastic lesions. This targeted inteRASomics MALDI–MSI approach may represent an added value in the comprehension of NIFTPs, and might have a potential role to support traditional pathology. References 1. Suárez HG, Du Villard JA, Caillou B, Schlumberger M, Tubiana M, Parmentier C, et al; Oncogene, 2 (1988), pp 403–406. 2. Chu YH, Sadow PM; Seminars in Diagnostic Pathology. 37 (2020), pp 213-218. Acknowledgement This research was funded by Regione Lombardia: Programma degli interventi per la ripresa economica: sviluppo di nuovi accordi di collaborazione con le università per la ricerca, l’innovazione e il trasferimento tecnologico: NephropaThy and Ricerca Finalizzata GR-2019-12368592.