Spatially-resolved deep biomarker profiling of single cells in FFPE tissue samples through CODEX®

Maria Elena Gallina1, Jaskirat Singh1, Atri Choksi1, Nadya Nikulina1, Gajalakshmi Dakshinamoorthy1, Joseph Kim1, Sejal Mistry1, Julia Kennedy-Darling1

1Akoya Biosciences, Menlo Park, CA

Spatially resolved, deep biomarker profiling of tissuespecimens is crucial for investigating the architectureand cell diversity in complex matrices, such as the tumormicroenvironment (TME). The correlation of these twoparameters with the progression of a variety ofpathologies, ranging from cancer to autoimmunediseases, is still vastly unknown due to the lack oftechnologies that enable the detection of both highcontent and spatial, resolution. Formalin-fixed paraffin-embedded (FFPE) tissues constitute the ideal sample forthese investigations as they retain tissue morphology atroom temperature for long periods of time, which isconvenient and cost-effective and are available in vastspecimen archives.Akoya Biosciences, Inc. is commercializing CODEXTM (CO-Detection by indEXing), a multi-parametric imagingplatform that performs the concurrent detection andquantification of dozens of protein targets with single-cell resolution within a tissue specimen. The CODEXplatform uses a proprietary Barcodes library to labelantibodies and iterative cycles of adding and removingconjugated dye-labeled Reporters to reveal the stainingpattern for a subset of the target markers per cycle. Theentire data acquisition process is fully automated usingthe CODEX instrument, which integrates with existingmicroscopes.To support high-multiplexing capabilities, more than 45different antibody clones have been screened andvalidated for the detection of target epitopes in humanFFPE secondary lymphoid tissues for both healthy andcancerous specimens using the CODEX platform. Proofof-concept data have also been obtained for differentcancer types, including melanoma and head and neckcancer in both normal and TMA formats.An analysis software has been developed to extractmeaning from CODEX datasets and is included as part ofthe CODEX platform. The analysis pipeline includesimage drift compensation, deconvolution andsegmentation to measure integrated fluorescenceintensity for each cell across tens of parameters. In thismanner, infiltrating lymphocytes and other immune cellscan be identified within the TME. With the highmultiplexing capabilities, a variety of T cells can beidentified to detect the ratios of bothimmunosuppressive and immune activating subtypes.Analysis of FFPE tissues using the CODEX platformdemonstrates unparalleled opportunities forsimultaneously detecting tens of markers with single-cellresolution and spatial information within a tissuespecimen.

CODEX Analysis: Revealing the Spatial Distribution ofrelevant cellular phenotypes

CODEX: a Highly Versatile Imaging Platform

FFPE Melanoma

CODEX assays can be performed on a variety of tissue sections,ranging from FFPE to fresh-frozen samples. CODEX assays arealso non-destructive: tissues can be cleared and stored afterone or more CODEX runs. Here representative images fromCODEX multicycles on FFPE human tonsil, melanoma and breastcancer tissues are reported. CODEX antibody staining is alsoshown in tissue biopsy cores from tissue microarrays (TMAs) asproof of concept.

FFPE Tonsil

Ki67DAPI

TMA: stage II oral cancer

TMA: laryngeal cancer

Ki67CD7

FFPE Melanoma FFPE Breast CancerFFPE Tonsil

Overview of CODEX workflow

Deep phenotyping with spatial information of FFPE Human Melanoma Tissue

Segmented cells were clustered on the basis of the expression levels of 12 biomarkers toidentify 8 different cellular phenotypes. The spatial location of different cell types arerevealed at single-cell resolution

Spatial correlations between multiple phenotypes can now be investigated and compared region-to-region

Microscopic interactions between single cells are revealed for a dozen of different phenotypes with a single CODEX run

CODEX technology allows unprecedented possibilities forinvestigating the microenvironment of FFPE tissue samples bycombining deep biomarker profiling with microscopic spatialinformation.

CD34CD4CD3CD8ECadki67

VimentinCD68CD31

PancytokeratinCD11cCD20