ABSTRACT: GeoMx Digital Spatial Profiling (Nanostring) is a commercial spatial transcriptomics method to selectively analyze regions of interest within intact tissue sections. However, it is unclear how tissue decalcification affects the results of this assay, or if decalicified samples are comparable to untreated samples. For this purpose, we assembled a set of tissues that were subjected to progressive durations of incubation with ethylene-diamine-tetra-acetic (EDTA), and analyzed comparable regions of the tissue using GeoMx DSP. In the associated manuscript, we show that EDTA incubation attenuates signals, and distorts counts from a subset of probes. We determine the undisclosed full-length target sequences of probes used in the human whole transcriptome panel, updating target transcripts and genes, and identify that gene-promiscuous probes are particularly affected by normalization after EDTA incubation. For the purpose of this study, we prepared series of non-calcified tissues that were incubated with EDTA for increasing time periods, and matched control samples from the same source tissue, which were left untreated. Specimens submitted for routine tissue analysis were collected for research purposes after approval of the ethics commission of the Medical University of Vienna (2167/2021, registered at https://ekmeduniwien.at/core/catalog/2022/) as graphically shown in Figure 1A). In the course of routine pathology workup, 24 clinical tissue specimens (Table S1) of different organs stemming from 20 patients were fixed in 7.5% buffered formalin for 2.0 to 3.5 (mean 3.3, median 3.0) days according to the laboratory’s routine procedure. At macroscopic dissection, an excess piece of tissue was removed and divided into 4 macroscopically similar samples. Tissues were then either processed immediately in a Tissue-Tek VIP® 6 AI (Sakura Finetek) or incubated at room temperature in EDTA solution (Tritiplex, Glatt-Koller, 403212270) for 1, 3, and 7 days before processing them analogous to the untreated tissue. Samples were embedded in paraffin blocks according to standard protocols using low-melting point Paraffin (Histo-Comp, ATS-200856). Blocks were sectioned, stained with hematoxylin/eosin and scanned (Pannoramic 250 Flash II, 3DHistech) to confirm the content of equivalent histological structures among samples sourced from the same specimen. Areas for tissue microarray (TMA) cores were selected using the digital slide scans. Tissue microarrays were then assembled using an automated TMA device (TMA Grandmaster, 3DHistech) using 1mm punches and a recipient block of 5mm depth. TMA were tempered by alternating storage at 37°C, 4°C, and room temperature, for 5 times (2h each), following by a single incubation at 65°C for 10 minutes. Slides were sectioned using a rotating microtome at 5µm thickness. One section was used for DSP analysis as outlined below, while an adjacent section was used for validation of the TMA cores using hematoxylin & eosin stain. DSP was performed according to protocols provided by Nanostring using the following standard parameters. In brief, sections were deparaffinized and subjected to heat-induced protein epitope retrieval (using retrieval buffer at pH 9, Thermo Fisher Scientific, 00-4956-58) for 15 minutes at 100°C using a steam cooker. Retrieval for mRNA was performed using proteinase K (Thermo Fisher Scientific, 25530049) at 1µg/ml for 15 minutes. Slides were then incubated with the human Whole Transcriptome Atlas (WTA) probe panel (Nanostring, NA-GMX-RNA-NGS-HuWTA-4, Lot HWTA21003) at 37°C in a humidified environment overnight. Sections were washed according to standard protocols and direct immunofluorescence was performed with antibodies against CD34 conjugated to Alexa-647 (Novus Biologicals, clone QBEnd-10, NBP2-34713AF647), Pan-cytokeratin conjugated to to Alexa532 (Novus Biologicals, clone AE1/AE3, NBP2-33200), CD45 conjugated to Alexa594 (clone 2B11 + PD7/26, Novus Biologicals, NBP2-34528), and a nuclear counterstain (Syto-13, Nanostring) DSP device run and library preparation were performed according to the manufacturers’ protocol. The library was sequenced using an Illumina NextSeq 550 sequencer with the specifications provided by Nanostring (paired-end reads at length 27, index length 8) at a library concentration of 1.6 pM using a NextSeq 550 High Output 75-cycle kit (Illumina, 20024906). Nuclei counts of regions of interest were obtained via the default automatic nuclear segmentation settings of the DSP device software.
