MYC, BCL2, and BCL6 are commonly used markers for immunohistochemistry of Diffuse large B-cell lymphomas (DLBCL). Coexpression of MYC and BCL2 in particular constitutes a subgroup of “double expressor lymphomas” (DEL) with a distinct poor clinical outcome. However, it is not known if MYC and BCL2/BCL6 coexpression occurs in the same cell or in different cells within the tumor, as traditional immunohistochemistry (IHC) is limited by the number of markers that can be simultaneously assessed within formalin-fixed, paraffin-embedded (FFPE) samples. We set out to discover the clinical significance of MYC, BCL2, and BCL6 colocalization at single-cell resolution using multiplexed quantitative immunofluorescence (qIF) based on sequential OPAL-TSA staining and spectral microscopy on the Vectra platform. The initial discovery cohort comprised 90 cases of DLBCL from NUH Singapore with adequate clinical follow-up after R-CHOP therapy. We stratified each DLBCL tumor into 8 “clonal fractions” based on the possible permutations of MYC (M), BCL2 (2), and BCL6 (6) colocalization: M+2+6+, M+2+6-, M+2-6+, M+2-6-, M-2+6+, M-2-6+, M-2+6-, and M-2-6-. Interestingly, even within cases that fit traditional IHC criteria for “positivity” of MYC, BCL2, and BCL6, only a subset of cells within each case expressed multiple markers concurrently. Using the fraction of each of these clones as a continuous variable, Cox regression analysis revealed that the percentage of M+2+6- cells in a case was most predictive of poor survival. Importantly, the same clonal fraction (M+2+6-) was a significant poor prognostic feature in 2 smaller validation cohorts from SGH Singapore (n=41) and MD Anderson Cancer Centre USA (n=36). The single-cell staining pattern of these markers revealed a stark contrast between healthy tonsil tissue and DLBCL tissue. In the tonsil, colocalization of each marker was nonrandom (mutually exclusive BCL2 positivity in B cells outside the germinal center and BCL6 positivity inside the germinal center), whereas in DLBCL samples the mutual exclusivity pattern noted in the tonsil was lost, leading to a random distribution of colocalization of MYC, BCL2, and BCL6. The random nature of this colocalization allowed us to mathematically predict the “extent” of these 8 subclones from any data set with quantitative data of each single marker (MYC, BCL2, and BCL6). We therefore attempted to evaluate this model in RNA expression datasets of DLBCL cases with clinically annotated data. Remarkably, in concordance with our IF data, the “predicted” M+2+ 6- subgroup consistently was associated with an unfavorable prognosis in 3 independent mRNA datasets (GSE10846 n=233, GSE117556 n=469, GSE32918 n=140). In summary, we have for the first time assessed the expression of MYC, BCL2, and BCL6 at the single-cell level in DLBCL. These results may explain the apparent protective function of BCL6 expression in prior cohort studies of DEL, and provide a quantitative tool for the identification of DLBCL cases with poor survival on R-CHOP.

Citation Format: Michal Hoppe, Shuangyi Fan, Patrick Jaynes, Phuong Mai Hoang, Liu Xin, Sanjay De Mel, Li Mei Poon, Esther Chan, Joanne Lee, Yen Lin Chee, Choon Kiat Ong, Tiffany Tang, Soon Thye Lim, Nicholas Francis Grigoropoulos, Sheng-Tsung Chang, Shih-Sung Chuang, Joseph Khoury, Hyungwon Choi, Wee Joo Chng, Siok-Bian Ng, Claudio Tripodo, Anand D. Jeyasekharan. Prognostic significance of MYC, BCL2, and BCL6 colocalization at single-cell resolution in DLBCL [abstract]. In: Proceedings of the AACR Virtual Meeting: Advances in Malignant Lymphoma; 2020 Aug 17-19. Philadelphia (PA): AACR; Blood Cancer Discov 2020;1(3_Suppl):Abstract nr PO-35.