Enabling MSI-Guided Laser Capture Microdissection

Introduction/Rationale: Coupling MALDI mass spectrometry imaging (MALDI-MSI) with Laser Capture Microdissection (LCM) allows for precise dissection of tissue regions based on molecular features [1]. Automated methods for alignment of the coordinate systems of the MSI and LCM platforms reduces errors associated with manual definition of ROI’s and increases throughput (a major bottleneck for LCM). Here we present the development of a method to transfer regions of interest from MALDI MSI images to an LCM platform, using consecutive tissue sections mounted on ITO conductive slides for MALDI MSI and on PEN-coated slides for LCM. Methods: The test system consists of a gelatin-embedded mouse liver. 12 µm slices were cut using a cryostat and two consecutive slices were mounted on ITO and PEN slides. The ITO slide was spray-coated with DHB (30mg/mL, MeOH 70%, water 30%, 0.2% TFA) and a MALDI image was acquired with an EP-MALDI source coupled to a Q-Exactive mass spectrometer. The MSI data was imported into MATLAB. The tissue mounted on the PEN slide was stained with hematoxylin and a high resolution optical image acquired using an Aperio Scanscope. The LCM instrument used was an Apotome 2 Axio Observer Z1 microscope equipped with a Palm Robomover LCM system (both Zeiss). Results: An image of an ion with a regular distribution on the tissue is used to align the MS image to the optical image of the hematoxylin-stained tissue section mounted on the PEN slide. The optical image of the PEN slide tissue section is imported in MATLAB and cropped to match the size of the MALDI image. An intensity-based co-registration algorithm is then used to align the MS image to the cropped optical image. The MS image is then rescaled to match to the original optical image. To obtain regions-of-interest to transfer to the LCM platform, the MSI data was TIC normalized and a k-means cluster analysis performed. The image of the cluster of interest was aligned to the PEN slide using the same transformations used for the whole MSI data, binarized and segmented to obtain the coordinates of the vertices of the cluster region. Vertex coordinates were expressed after setting the axes origin to a user-defined reference point on the slide. The coordinates of the origin in the Aperio reference system were then matched to the coordinates of the reference point in the Zeiss coordinate system and the same transformation applied. Coordinates were then formatted as an Element file readable by the LCM and exported as text files. Border coordinates were imported in the Zeiss PALMRobo software and regions of interest automatically dissected. Conclusions/Novelty: The presented method enables rapid transfer of coordinates from a MALDI image to an LCM instrument, increasing throughput and reducing errors due to freehand cutting. The method is applicable to consecutive tissue sections, and ROI’s can be defined either by MSI or via histopathological specification.