Background: Tissue-based cyclic immunofluorescence (t-CyCIF) is a recently described technique for highly multiplexed immunofluorescence microscopy of formalin-fixed, paraffin-embedded (FFPE) specimens. Via an iterative process, successive four-channel images are collected from the same sample and then registered to each other to generate a high-dimensional representation that is used for visualization and analysis. This technique can be used to capture up to 60 different antigens on a single FFPE tumor section and permits quantification of cell lineage and state markers, intracellular signaling proteins, drug targets and immune cell antigens, thereby promoting biomarker discovery efforts that are fundamental to precision oncology. As with most technologies that utilize immunostaining, proper antibody validation is key to reliable performance. In this study we used t-CyCIF to evaluate multiple antibodies directed against proteins commonly used to characterize breast carcinomas and their associated microenvironment. Our goal was to validate these antibodies on the t-CyCIF platform prior to the more widespread use of this technology to propel novel discoveries on breast cancer initiation, progression and treatment.

Methods: To choose the optimal antibody candidate for each biomarker in t-CyCIF, we compared multiple fluorophore-conjugated antibodies for each of the following three proteins routinely evaluated in breast carcinomas: estrogen receptor (ER), progesterone receptor (PR) and HER2. For each of these, a single antibody commonly used in clinical practice was used as a reference. Analyses were performed at the level of pixels, cells and tissue cores. In addition, inter-assay analyses were performed comparing: (1) t-CyCIF vs. immunohistochemistry (IHC), the latter assessed both by digital pathology and by two independent pathologists; and (2) t-CyCIF vs. fluorescence in situ hybridization (FISH) for HER2. Following validation of these antibodies, we evaluated the expression of CD45, CD68, PD-L1, p53, Ki67 and androgen receptor along with ER, PR and HER2, to better understand both the tumor microenvironment and the cell identities/states in breast carcinomas.

Results: A total of 948 tissue cores were included in the study. In the first phase, 13 different antibodies were analyzed: three raised against ER and five each against PR and HER2. The pixel-to-pixel evaluation resulted in r scores using Pearson correlation equal to 0.86 for both ER markers tested; ranging from 0.88 to 0.93 for PR; and from 0.56 to 0.94 for HER2. The correlation scores in single-cell comparisons ranged from 0.76 to 0.88 for ER, 0.54 to 0.81 for PR, and from 0.56 to 0.76 for HER2. Comparisons were then performed at the tissue core level. In light of the data generated through these multiple levels of analyses, we identified fluorophore-conjugated candidates for use in t-CyCIF. Correlation scores on the tissue core level were high in the inter-assay analyses, i.e. t-CyCIF vs. IHC (e.g. r scores up to 0.91 for HER2 on t-CyCIF vs. IHC) and t-CyCIF vs. HER2 FISH (r scores up to 0.71). In the second phase, we characterized the tumor microenvironment and cell identities present in 260 breast carcinomas. With a qualified panel of antibodies, we performed single-cell analyses of 589,343 cells and identified unexpected patterns of PD-L1 expression in distinct populations of tumor cells.

Conclusion: This study is the first to evaluate the performance of breast cancer-specific antibodies in a highly multiplexed imaging platform such as t-CyCIF. This work demonstrates a step-by-step approach for qualifying reagents to be used in a multiplexed, spatially resolved tissue imaging modality. This validation study will facilitate the use of t-CyCIF for additional studies in breast cancer to evaluate both tumor elements and components of the microenvironment.

Citation Format: Ricardo G Pastorello, Jia-Ren Lin, Ziming Du, Shaolin Mei, Krishan Taneja, Deborah A Dillon, Stuart J Schnitt, Peter K Sorger, Elizabeth A Mittendorf, Sandro Santagata, Jennifer L Guerriero. Highly multiplexed tissue-based cyclic immunofluorescence in breast cancer for precision oncology [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-02-03.