Qualification of a multiplexed tissue imaging assay and detection of novel patterns of HER2 heterogeneity in breast cancer

Abstract

Emerging data suggests that HER2 intratumoral heterogeneity (ITH) is associated with therapy resistance, highlighting the need for new strategies to assess HER2 ITH. A promising approach is leveraging multiplexed tissue analysis techniques such as cyclic immunofluorescence (CyCIF), which enable visualization and quantification of 10-60 antigens at single-cell resolution from individual tissue sections. In this study, we qualified a breast cancer-specific antibody panel, including HER2, ER, and PR, for multiplexed tissue imaging. We then compared the performance of these antibodies against established clinical standards using pixel-, cell- and tissue-level analyses, utilizing 866 tissue cores (representing 294 patients). To ensure reliability, the CyCIF antibodies were qualified against HER2 immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) data from the same samples. Our findings demonstrate the successful qualification of a breast cancer antibody panel for CyCIF, showing high concordance with established clinical antibodies. Subsequently, we employed the qualified antibodies, along with antibodies for CD45, CD68, PD-L1, p53, Ki67, pRB, and AR, to characterize 567 HER2+ invasive breast cancer samples from 189 patients. Through single-cell analysis, we identified four distinct cell clusters within HER2+ breast cancer exhibiting heterogeneous HER2 expression. Furthermore, these clusters displayed variations in ER, PR, p53, AR, and PD-L1 expression. To quantify the extent of heterogeneity, we calculated heterogeneity scores based on the diversity among these clusters. Our analysis revealed expression patterns that are relevant to breast cancer biology, with correlations to HER2 ITH and potential relevance to clinical outcomes.

Fig. 1. Overview of fluorescent CyCIF antibody qualification against antibodies used in the clinical laboratory.

To qualify breast cancer-related antibodies HER2, ER, PR, AR, and p53, the BC03 tissue microarray (TMA), which represents 16 breast tumors in duplicate, was used. Multiple CyCIF antibodies were compared to a single antibody commonly used in clinical practice as a reference. a Schematic representation of the different levels of fluorescent antibody validation using the CyCIF method, starting from tissue staining (lowest level of validation) towards patient-level (highest level) inter-assay comparison (i.e., direct comparison of each patient tissue to itself between assays). b–f Representative CyCIF images of antibodies tested by CyCIF on the BC03 TMA. Asterisks indicate clinical antibodies (*) and qualified CyCIF antibodies (**) for each target. g Representative CyCIF image of HER2 (TF-MA5-14509; sp3) and ER (CS98710) staining, showing the majority of tumor cells are ER+, and some showing strong, membrane staining for HER2. Left image is a full TMA core (36× mag.); the right image corresponds to the left image (74× mag.).

Fig. 2. Core-to-core comparison of clinical and CyCIF antibodies against ER, PR, and HER2.

To qualify breast cancer-related antibodies, the BC03 TMA, representing 16 breast tumors in duplicate was used. a–c CyCIF was performed using the qualified CyCIF antibody against a single antibody commonly used in clinical practice as a reference for ER (a), PR (b), and HER2 (c). The left graph depicts a single-cell dot-plot between the clinical clone on the x axis and the validated CyCIF antibody on the y axis. Each dot represents single-cell fluorescent intensity values from the two antibodies. Dashed lines indicate the gating cutoffs. The middle graph shows the corresponding mean log intensity of the core-to-core analysis of the clinical and CyCIF antibodies. The single-cell data were collected for individual TMA cores, with a binary gate applied to obtain the positive signal of each core (range from 0–1). The X- & Y axis represent the positive score calculated from either clinical or CyCIF antibodies, respectively. The right graph shows positivity scores (number of positive cells over total cells) for the clinical and CyCIF antibodies by TMA case. d, e Cross-assay comparison of the clinical and CyCIF antibodies analyzed by CyCIF compared to the clinical antibody analyzed by IHC using Aperio software for ER (d) and HER2 (e). Left, dot-plot representation of two different scores obtained from CyCIF and from IHC-Aperio. CyCIF of clinical (green dots) and CyCIF antibodies (blue dots) were used on the same section, while IHC was performed on a different section from the same TMA block. Each dot represents a single core from BC03 TMA. CyCIF scores are plotted on y axis as positive ratio of immunofluorescence, IHC scores on x axis are plotted as the percent of positive cells. Right graph, quantitative assessment of ER and HER2 IHC versus CyCIF staining. IHC scores by Aperio were used to stratify (0–24, 25–49, 50–74, 75–100) different TMA cores/cases, and the mean intensities of CyCIF antibody staining from each TMA core are shown using boxplot analysis. CyCIF antibodies: ER (CST 74244 S) and HER2 (ab225510).

Fig. 3. Inter-assay analysis of HER2 enriched TMAs (TMAs 226 and 227).

Following the selection of qualified ER, PR, and HER2 antibodies, two HER2-enriched TMAs, which included 567 tissue cores (representing 189 patients in triplicate), were used to further qualify CyCIF antibodies. a, b Percent of ER+ and HER2+ cells assessed through CyCIF (y axis) is compared to the score assigned by a clinical pathologist (x-axis) for each TMA. c Cross-assay comparison of the HER2 clinical and CyCIF antibodies analyzed by CyCIF compared to the clinical antibody analyzed by IHC using Aperio software. Left, dot-plot represents two different scores obtained from CyCIF and one obtained from IHC-Aperio. CyCIF of clinical (green dots) and CyCIF (blue dots) antibodies were used on the same section, while IHC was done on a different section from the same TMA block. Each dot represents a single core from BC03 TMA. CyCIF scores are plotted on y axis as positive ratio of immunofluorescence, IHC scores on x axis plotted as percent of positive cells. Right, quantitative assessment HER2 IHC versus CyCIF staining. IHC scores by Aperio were used to stratify (0–24, 25–49, 50–74, 75–100) different TMA cores/cases, and the mean intensities of CyCIF antibody staining from each TMA core are shown using boxplot analysis. d Clinically annotated HER2 FISH scores against IF/CyCIF staining using the SP3 antibody (Pearson r = 0.71) and HER2 FISH scores against IF/CyCIF staining using the CyCIF antibody, ab225510 (Pearson r = 0.65). Individual patients are shown in different colors, in triplicate. The triplicate cores tend to cluster together, indicating minimal variation.

Fig. 4. Overview of single-cell data from HER2+ enriched breast tumors.

Following the selection of qualified ER, PR, and HER2 CyCIF antibodies, the expression of selected antibodies was evaluated at a single-cell level in 567 HER2+ invasive breast cancer samples from 189 patients, and t-Distributed Stochastic Neighbor Embedding (t-SNE) is shown as a distribution of all single cells. a Selected markers were used to plot single cells. b t-SNE in a is shown colored by patient ID. c Visualization of markers within t-SNE plots.

Fig. 5. CyCIF single-cell analysis of HER2+ breast cancer reveals tumor populations with heterogenous HER2 expression.

Following the selection of qualified ER, PR, HER2, AR, and p53 CyCIF antibodies, the expression of selected antibodies was evaluated at a single-cell level in 567 HER2+ invasive breast cancer samples, representing 189 patients. a Single-cell clustering is shown, and b median expression for each antigen across each cluster is shown. Relative expression of HER2 is designated as high, moderate (mod), low, and negative among clusters 2, 5, 4, and 7, respectively. Tumor clusters are defined as: cluster 2 (HER2^highER^negPR^posAR^posPD-L1^highKi67^pos), cluster 5 (HER2^modER^negPR^posAR^posp53^highPD-L1^posKi67^pos), and cluster 4 (HER2^lowER^posPR^lowAR^pos)). One cluster was HER2-negative (cluster 7 (HER2^negER^highPR^highAR^high)). Cluster 3 and 6 represent immune/stromal populations as characterized by the expression of the leukocyte marker CD45. Cluster 1 has heterogeneous expression of most proteins and, therefore, did not form a distinct population of cells. Area in A and B refers to the nuclear area of segmented cells. c The 7 cell clusters are visualized using t-SNE. d Volcano plots of expression of each marker by cluster.

Fig. 6. HER2 Heterogeneity scores derived from clustering analysis reveal correlation to clinical outcome.

Following the selection of qualified ER, PR, HER2, AR, and p53 CyCIF antibodies, the expression of selected antibodies was evaluated at a single-cell level in 567 HER2+ invasive breast cancer samples, representing 189 patients. Tissues from HER2+ patients (n = 77) in which there were at least 500 cells pooled from the triplicate cores were used for ITH analysis. a HER2 expression was analyzed in single cells, and the coefficient of variation (C.V.) among patients was plotted (y axis) by recurrence status. b HER2 and Ki67 mean intensity expression measured by CyCIF. c Distribution of cells across all clusters (blue) and HER2 core number 113 (orange) and d representative tumor with low (HER2-5 and HER2-161) and high (HER2-164 and HER2-170) heterogeneity. e HER2 heterogeneity scores were generated by identifying cells from each tissue mapped to the entire t-SNE. A larger boundary corresponds with higher diversity. f Samples that have equal distribution of each cluster have high heterogeneity and are diamond-shaped in the boundary mapping. g GMM and t-SNE scores reveal an association with recurrence. h Patients treated with Trastuzumab were removed from the GMM and t-SNE score analysis.