Patterns of Oncogene Coexpression at Single-Cell Resolution Influence Survival in Lymphoma

Cancers often overexpress multiple clinically relevant oncogenes, but it is not known if combinations of oncogenes in cellular subpopulations within a cancer influence clinical outcomes. Using quantitative multispectral imaging of the prognostically relevant oncogenes MYC, BCL2, and BCL6 in diffuse large B-cell lymphoma (DLBCL), we show that the percentage of cells with a unique combination MYC+BCL2+BCL6- (M+2+6-) consistently predicts survival across four independent cohorts (n = 449), an effect not observed with other combinations including M+2+6+. We show that the M+2+6- percentage can be mathematically derived from quantitative measurements of the individual oncogenes and correlates with survival in IHC (n = 316) and gene expression (n = 2,521) datasets. Comparative bulk/single-cell transcriptomic analyses of DLBCL samples and MYC/BCL2/BCL6-transformed primary B cells identify molecular features, including cyclin D2 and PI3K/AKT as candidate regulators of M+2+6- unfavorable biology. Similar analyses evaluating oncogenic combinations at single-cell resolution in other cancers may facilitate an understanding of cancer evolution and therapy resistance.

SIGNIFICANCE

Using single-cell-resolved multiplexed imaging, we show that selected subpopulations of cells expressing specific combinations of oncogenes influence clinical outcomes in lymphoma. We describe a probabilistic metric for the estimation of cellular oncogenic coexpression from IHC or bulk transcriptomes, with possible implications for prognostication and therapeutic target discovery in cancer. This article is highlighted in the In This Issue feature, p. 1027.

PLK1 inhibition selectively induces apoptosis in ARID1A deficient cells through uncoupling of oxygen consumption from ATP production

Inhibitors of the mitotic kinase PLK1 yield objective responses in a subset of refractory cancers. However, PLK1 overexpression in cancer does not correlate with drug sensitivity, and the clinical development of PLK1 inhibitors has been hampered by the lack of patient selection marker. Using a high-throughput chemical screen, we discovered that cells deficient for the tumor suppressor ARID1A are highly sensitive to PLK1 inhibition. Interestingly this sensitivity was unrelated to canonical functions of PLK1 in mediating G2/M cell cycle transition. Instead, a whole-genome CRISPR screen revealed PLK1 inhibitor sensitivity in ARID1A deficient cells to be dependent on the mitochondrial translation machinery. We find that ARID1A knock-out (KO) cells have an unusual mitochondrial phenotype with aberrant biogenesis, increased oxygen consumption/expression of oxidative phosphorylation genes, but without increased ATP production. Using expansion microscopy and biochemical fractionation, we see that a subset of PLK1 localizes to the mitochondria in interphase cells. Inhibition of PLK1 in ARID1A KO cells further uncouples oxygen consumption from ATP production, with subsequent membrane depolarization and apoptosis. Knockdown of specific subunits of the mitochondrial ribosome reverses PLK1-inhibitor induced apoptosis in ARID1A deficient cells, confirming specificity of the phenotype. Together, these findings highlight a novel interphase role for PLK1 in maintaining mitochondrial fitness under metabolic stress, and a strategy for therapeutic use of PLK1 inhibitors. To translate these findings, we describe a quantitative microscopy assay for assessment of ARID1A protein loss, which could offer a novel patient selection strategy for the clinical development of PLK1 inhibitors in cancer.

Abstract PO-35: Prognostic significance of MYC, BCL2, and BCL6 colocalization at single-cell resolution in DLBCL

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.

DNA damage signalling as an anti-cancer barrier in gastric intestinal metaplasia

OBJECTIVE

Intestinal metaplasia (IM) is a premalignant stage that poses a greater risk for subsequent gastric cancer (GC). However, factors regulating IM to GC progression remain unclear. Previously, activated DNA damage response (DDR) signalling factors were shown to engage tumour-suppressive networks in premalignant lesions. Here, we interrogate the relationship of DDR signalling to mutational accumulation in IM lesions.

DESIGN

IM biopsies were procured from the gastric cancer epidemiology programme, an endoscopic surveillance programme where biopsies have been subjected to (epi)genomic characterisation. IM samples were classified as genome-stable or genome-unstable based on their mutational burden/somatic copy-number alteration (CNA) profiles. Samples were probed for DDR signalling and cell proliferation, using the markers γH2AX and MCM2, respectively. The expression of the gastric stem cell marker, CD44v9, was also assessed. Tissue microarrays representing the GC progression spectrum were included.

RESULTS

MCM2-positivity increased during GC progression, while γH2AX-positivity showed modest increase from normal to gastritis and IM stages, with further increase in GC. γH2AX levels correlated with the extent of chronic inflammation. Interestingly, genome-stable IM lesions had higher γH2AX levels underscoring a protective anti-cancer role for DDR signalling. In contrast, genome-unstable IM lesions with higher mutational burden/CNAs had lower γH2AX levels, elevated CD44v9 expression and modest promoter hypermethylation of DNA repair genes WRN, MLH1 and RAD52.

CONCLUSIONS

Our data suggest that IM lesions with active DDR will likely experience a longer latency at the premalignant state until additional hits that override DDR signalling clonally expand and promote progression. These observations provide insights on the factors governing IM progression.

Therapeutic Strategy for the Prevention of Pseudorabies Virus Infection in C57BL/6 Mice by 3D8 scFv with Intrinsic Nuclease Activity

3D8 single chain variable fragment (scFv) is a recombinant monoclonal antibody with nuclease activity that was originally isolated from autoimmune-prone MRL mice. In a previous study, we analyzed the nuclease activity of 3D8 scFv and determined that a HeLa cell line expressing 3D8 scFv conferred resistance to herpes simplex virus type 1 (HSV-1) and pseudorabies virus (PRV). In this study, we demonstrate that 3D8 scFv could be delivered to target tissues and cells where it exerted a therapeutic effect against PRV. PRV was inoculated via intramuscular injection, and 3D8 scFv was injected intraperitoneally. The observed therapeutic effect of 3D8 scFv against PRV was also supported by results from quantitative reverse transcription polymerase chain reaction, southern hybridization, and immunohistochemical assays. Intraperitoneal injection of 5 and 10 μg 3D8 scFv resulted in no detectable toxicity. The survival rate in C57BL/6 mice was 9% after intramuscular injection of 10 LD50 PRV. In contrast, the 3D8 scFv-injected C57BL/6 mice showed survival rates of 57% (5 μg) and 47% (10 μg). The results indicate that 3D8 scFv could be utilized as an effective antiviral agent in several animal models.

Development of Lactobacillus paracasei harboring nucleic acid-hydrolyzing 3D8 scFv as a preventive probiotic against murine norovirus infection

The protein 3D8 single-chain variable fragment (3D8 scFv) has potential anti-viral activity due to its ability to penetrate into cells and hydrolyze nucleic acids. Probiotic Lactobacillus paracasei engineered to secrete 3D8 scFv for oral administration was used to test the anti-viral effects of 3D8 scFv against gastrointestinal virus infections. We found that injection of 3D8 scFv into the intestinal lumen resulted in the penetration of 3D8 scFv into the intestinal villi and lamina propria. 3D8 scFv secreted from engineered L. paracasei retained its cell-penetrating and nucleic acid-hydrolyzing activities, which were previously shown with 3D8 scFv expressed in Escherichia coli. Pretreatment of RAW264.7 cells with 3D8 scFv purified from L. paracasei prevented apoptosis induction by murine norovirus infection and decreased messenger RNA (mRNA) expression of the viral capsid protein VP1. In a mouse model, oral administration of the engineered L. paracasei prior to murine norovirus infection reduced the expression level of mRNA encoding viral polymerase. Taken together, these results suggest that L. paracasei secreting 3D8 scFv provides a basis for the development of ingestible anti-viral probiotics active against gastrointestinal viral infection.