Spatially Resolved, Multiplexed Digital Characterization of Protein Distribution and Abundance in Formalin-Fixed, Paraffin-Embedded (FFPE) Diffuse Large B Cell Lymphoma Tissue Sections Based on the Nanostring® Digital Spatial Profiling (DSP) Technology

Imaging-based techniques are powerful tools for analyzing protein and mRNA expression and localization within the tumor microenvironment. However, they typically suffer from a number of challenges, including lack of dynamic range, difficult quantitation, and labor intensive workflow for very limited multiplexing. We have developed a novel platform based on the nCounter® barcoding technology that enables spatially resolved, digital characterization of proteins and mRNA in a highly multiplexed assay. The assay relies upon probes coupled to photocleavable oligonucleotide tags which are released from discrete regions of the tissue using focused through-objective UV (~ 365nm) exposure. An automated prototype capable of imaging, selective illumination using a digital mirror device, and sample collection was developed by modifying a standard microscope. Cleaved tags are quantitated in an nCounter assay, and counts are mapped back to tissue location, yielding a spatially-resolved digital profile of analyte abundance.

To apply this technology to the characterization of hematological malignancies, we have developed the nCounter Vantage 3D™ Heme RNA panel that characterizes biology known to be relevant for these tumors, including specific profiling of MAPK, MYC signaling, NF-κB, PI3K-AKT, and B cell and T cell receptor signaling. We have also developed the Vantage 3D Heme panel of antibodies that focuses on common signaling pathways, with an emphasis on phosphorylated targets. We have used these panels to characterize differences in protein expression between ABC and GCB subtypes of diffuse large B cell lymphoma.

3 ABC and 3 GCB lymphoma samples, previously identified by the NanoString Lymphoma Subtyping Test, were profiled on the DSP platform. 5 um tissue sections were stained with fluorescent antibodies for CD45 (to illuminate tumor and infiltrating immune cells), CD3 (to identify infiltrating T cells) and SYTO-83 (to illuminate nuclei), along with the oligo labeled antibodies from either the heme panel or the ST signaling panel. Regions of interested (ROI) were selected based on abundance of CD45 and CD3 staining from the ST-panel stained tissue section, and the same ROI was selected on the heme panel stained section using image alignment.

In this study, we used these two Vantage 3D Heme panels to compare protein expression between the different subtypes of DLBCL and identify proteins that vary between subtype. We also identified heterogeneity of protein expression within the ABC subtype for the key tumor suppressors p53 and individual sample variation of expression in p-AKT within the GCB subtype. Furthermore, we observe strong concordance of signal from antibodies that are included in both antibody panels, suggesting technical reproducibility of the platform.

Data are analyzed on software that is being developed in parallel with the technology, which enables visualization of the individual ROI within a sample, comparison of ROI across samples, and comparison of the individual probe counts within and between ROI.

Conclusions

Nanostring digital spatial profiling provides a quantitative and reproducible digital readout for multiplexed protein expression from the same microscopic regions of individual tissue sections. In this study we demonstrate how DSP technology can help elucidate protein expression profiles within individual lymphoid tumor specimens. First we show how regions of interest with the tumor (ROI) are identified and profiled using two multiplex protein panels simultaneously. Second, we demonstrate high platform concordance between proteins present within different NanoString panel products. Finally, we present methodology and data analysis workflow capable of identifying protein expression differences both within individual tumors, and between ABC and GCB lymphoma subtypes. Application of this technology to the study of hematological malignancies may help to identify novel drug targets and/or characterize response to therapy, thus helping to define novel biology and accelerate the clinical development path.