Integrative multiomics enhancer activity profiling identifies therapeutic vulnerabilities in cholangiocarcinoma of different etiologies

OBJECTIVES

Cholangiocarcinoma (CCA) is a heterogeneous malignancy with high mortality and dismal prognosis, and an urgent clinical need for new therapies. Knowledge of the CCA epigenome is largely limited to aberrant DNA methylation. Dysregulation of enhancer activities has been identified to affect carcinogenesis and leveraged for new therapies but is uninvestigated in CCA. Our aim is to identify potential therapeutic targets in different subtypes of CCA through enhancer profiling.

DESIGN

Integrative multiomics enhancer activity profiling of diverse CCA was performed. A panel of diverse CCA cell lines, patient-derived and cell line-derived xenografts were used to study identified enriched pathways and vulnerabilities. NanoString, multiplex immunohistochemistry staining and single-cell spatial transcriptomics were used to explore the immunogenicity of diverse CCA.

RESULTS

We identified three distinct groups, associated with different etiologies and unique pathways. Drug inhibitors of identified pathways reduced tumour growth in in vitro and in vivo models. The first group (ESTRO), with mostly fluke-positive CCAs, displayed activation in estrogen signalling and were sensitive to MTOR inhibitors. Another group (OXPHO), with mostly BAP1 and IDH-mutant CCAs, displayed activated oxidative phosphorylation pathways, and were sensitive to oxidative phosphorylation inhibitors. Immune-related pathways were activated in the final group (IMMUN), made up of an immunogenic CCA subtype and CCA with aristolochic acid (AA) mutational signatures. Intratumour differences in AA mutation load were correlated to intratumour variation of different immune cell populations.

CONCLUSION

Our study elucidates the mechanisms underlying enhancer dysregulation and deepens understanding of different tumourigenesis processes in distinct CCA subtypes, with potential significant therapeutics and clinical benefits.

PBRM1-deficient PBAF complexes target aberrant genomic loci to activate the NF-κB pathway in clear cell renal cell carcinoma

PBRM1 encodes an accessory subunit of the PBAF SWI/SNF chromatin remodeller, and the inactivation of PBRM1 is a frequent event in kidney cancer. However, the impact of PBRM1 loss on chromatin remodelling is not well examined. Here we show that, in VHL-deficient renal tumours, PBRM1 deficiency results in ectopic PBAF complexes that localize to de novo genomic loci, activating the pro-tumourigenic NF-κB pathway. PBRM1-deficient PBAF complexes retain the association between SMARCA4 and ARID2, but have loosely tethered BRD7. The PBAF complexes redistribute from promoter proximal regions to distal enhancers containing NF-κB motifs, heightening NF-κB activity in PBRM1-deficient models and clinical samples. The ATPase function of SMARCA4 maintains chromatin occupancy of pre-existing and newly acquired RELA specific to PBRM1 loss, activating downstream target gene expression. Proteasome inhibitor bortezomib abrogates RELA occupancy, suppresses NF-κB activation and delays growth of PBRM1-deficient tumours. In conclusion, PBRM1 safeguards the chromatin by repressing aberrant liberation of pro-tumourigenic NF-κB target genes by residual PBRM1-deficient PBAF complexes.

Measuring kinetics and metastatic propensity of CTCs by blood exchange between mice

Existing preclinical methods for acquiring dissemination kinetics of rare circulating tumor cells (CTCs) en route to forming metastases have not been capable of providing a direct measure of CTC intravasation rate and subsequent half-life in the circulation. Here, we demonstrate an approach for measuring endogenous CTC kinetics by continuously exchanging CTC-containing blood over several hours between un-anesthetized, tumor-bearing mice and healthy, tumor-free counterparts. By tracking CTC transfer rates, we extrapolated half-life times in the circulation of between 40 and 260 s and intravasation rates between 60 and 107,000 CTCs/hour in mouse models of small-cell lung cancer (SCLC), pancreatic ductal adenocarcinoma (PDAC), and non-small cell lung cancer (NSCLC). Additionally, direct transfer of only 1-2% of daily-shed CTCs using our blood-exchange technique from late-stage, SCLC-bearing mice generated macrometastases in healthy recipient mice. We envision that our technique will help further elucidate the role of CTCs and the rate-limiting steps in metastasis.

Keap1 mutation renders lung adenocarcinomas dependent on Slc33a1

Approximately 20-30% of human lung adenocarcinomas (LUAD) harbor loss-of-function (LOF) mutations in Kelch-like ECH Associated-Protein 1 (KEAP1), which lead to hyperactivation of the nuclear factor, erythroid 2-like 2 (NRF2) antioxidant pathway and correlate with poor prognosis1-3. We previously showed that Keap1 mutation accelerates KRAS-driven LUAD and produces a marked dependency on glutaminolysis4. To extend the investigation of genetic dependencies in the context of Keap1 mutation, we performed a druggable genome CRISPR-Cas9 screen in Keap1-mutant cells. This analysis uncovered a profound Keap1 mutant-specific dependency on solute carrier family 33 member 1 (Slc33a1), an endomembrane-associated protein with roles in autophagy regulation5, as well as a series of functionally-related genes implicated in the unfolded protein response. Targeted genetic and biochemical experiments using mouse and human Keap1-mutant tumor lines, as well as preclinical genetically-engineered mouse models (GEMMs) of LUAD, validate Slc33a1 as a robust Keap1-mutant-specific dependency. Furthermore, unbiased genome-wide CRISPR screening identified additional genes related to Slc33a1 dependency. Overall, our study provides a strong rationale for stratification of patients harboring KEAP1-mutant or NRF2-hyperactivated tumors as likely responders to targeted SLC33A1 inhibition and underscores the value of integrating functional genetic approaches with GEMMs to identify and validate genotype-specific therapeutic targets.

CRISPR-mediated modeling and functional validation of candidate tumor suppressor genes in small cell lung cancer

Small cell lung cancer (SCLC) is a highly aggressive subtype of lung cancer that remains among the most lethal of solid tumor malignancies. Recent genomic sequencing studies have identified many recurrently mutated genes in human SCLC tumors. However, the functional roles of most of these genes remain to be validated. Here, we have adapted the CRISPR-Cas9 system to a well-established murine model of SCLC to rapidly model loss-of-function mutations in candidate genes identified from SCLC sequencing studies. We show that loss of the gene p107 significantly accelerates tumor progression. Notably, compared with loss of the closely related gene p130, loss of p107 results in fewer but larger tumors as well as earlier metastatic spread. In addition, we observe differences in proliferation and apoptosis as well as altered distribution of initiated tumors in the lung, resulting from loss of p107 or p130 Collectively, these data demonstrate the feasibility of using the CRISPR-Cas9 system to model loss of candidate tumor suppressor genes in SCLC, and we anticipate that this approach will facilitate efforts to investigate mechanisms driving tumor progression in this deadly disease.

Identification of DHODH as a therapeutic target in small cell lung cancer

Small cell lung cancer (SCLC) is an aggressive lung cancer subtype with extremely poor prognosis. No targetable genetic driver events have been identified, and the treatment landscape for this disease has remained nearly unchanged for over 30 years. Here, we have taken a CRISPR-based screening approach to identify genetic vulnerabilities in SCLC that may serve as potential therapeutic targets. We used a single-guide RNA (sgRNA) library targeting ~5000 genes deemed to encode "druggable" proteins to perform loss-of-function genetic screens in a panel of cell lines derived from autochthonous genetically engineered mouse models (GEMMs) of SCLC, lung adenocarcinoma (LUAD), and pancreatic ductal adenocarcinoma (PDAC). Cross-cancer analyses allowed us to identify SCLC-selective vulnerabilities. In particular, we observed enhanced sensitivity of SCLC cells toward disruption of the pyrimidine biosynthesis pathway. Pharmacological inhibition of dihydroorotate dehydrogenase (DHODH), a key enzyme in this pathway, reduced the viability of SCLC cells in vitro and strongly suppressed SCLC tumor growth in human patient-derived xenograft (PDX) models and in an autochthonous mouse model. These results indicate that DHODH inhibition may be an approach to treat SCLC.

Abstract A24: Using single-cell RNA-seq approaches to decipher heterogeneity in autochthonous mouse models of small cell lung cancer

Small cell lung cancer (SCLC) is one of the most aggressive cancer types, and patients in clinic usually (>60% cases) present with metastasis. Current therapies for SCLC have not changed from 1980, and they include a combination therapy of cisplatin and etoposide. Most patients relapse after initial response, and the 5-year survival for extensive stage SCLC is 2%. One possible explanation for the failure of conventional and targeted therapies in cancers is the cellular heterogeneity that exists within tumors. Thus, understanding phenotypic heterogeneity at the single-cell level can be leveraged to predict mechanisms of resistance, which enables the design of effective combination therapies. In this project, we used as a model system an autochthonous mouse model of human SCLC, in which we deleted p53, RB, and p130. We collected multiple primary tumors, circulating tumor cells (CTC), and lymph node (proximal site) and liver (distant site) metastases from the same mouse, and across different mice. These samples were collected at the stage that should correspond to the limited/extensive stage of human SCLC, which is most commonly seen in clinic. Then we used single-cell RNA-seq (sc-RNAseq) methods to define and understand transcriptional heterogeneity in these samples. Furthermore, we aimed to understand the evolutionary relationships between primary tumors, CTCs, lymph node and liver metastases using different computational approaches such as diffusion maps. Further defining and functionally annotating transcriptional heterogeneity will help us better understand the disease, and find the new Achilles heel for targeting.

Note: This abstract was not presented at the conference.

Citation Format: Nemanja Despot Marjanovic, Sheng Rong Ng, Aviv Regev, Tyler Jacks. Using single-cell RNA-seq approaches to decipher heterogeneity in autochthonous mouse models of small cell lung cancer [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A24.

Abstract A12: CRISPR-mediated modeling and functional validation of candidate tumor-suppressor genes in small cell lung cancer

Small cell lung cancer (SCLC) is a highly aggressive subtype of lung cancer that remains among the most lethal of solid tumor malignancies. Recent genomic sequencing studies have identified a large number of recurrently mutated genes in human SCLC tumors. However, apart from a few notable examples, the functional roles of most of these genes remain to be validated. The CRISPR-Cas9 system has been utilized in genetically engineered mouse models of cancer to generate somatic genetic alterations in tumor cells in vivo. This has greatly accelerated efforts to study the roles of candidate genes of interest in various cancers. In this study, we adapted the CRISPR-Cas9 system to murine models of SCLC (mSCLC) to enable rapid modeling and functional validation of candidate tumor-suppressor genes in SCLC. We generated a Cre-activated allele of Cas9 and crossed it into both the Trp53flox/flox; Rb1flox/flox and the Trp53flox/flox; Rb1flox/flox; p130flox/flox models of mSCLC. By infecting these animals with adenoviral vectors coexpressing Cre recombinase and a target sgRNA, we can initiate SCLC tumors harboring loss-of-function mutations in candidate genes of interest. To validate this system, we infected animals with adenoviral vectors expressing sgRNAs targeting either p107 (Rbl1) or p130 (Rbl2), two genes that are recurrently mutated in a subset of human SCLC tumors. SCLC tumors that developed in these animals harbored loss-of-function mutations at the respective sgRNA target sites in the genome. Loss of p107 or p130 expression in these tumors was confirmed by immunohistochemical staining. As expected, CRISPR-mediated loss of p130 in Trp53flox/flox; Rb1flox/flox animals resulted in acceleration of tumor progression compared with animals infected with a control sgRNA, consistent with the phenotype previously observed in Trp53flox/flox; Rb1flox/flox; p130flox/flox animals. Loss of p107 in Trp53flox/flox; Rb1flox/flox animals also resulted in acceleration of tumor progression, as seen by an increase in tumor burden and a decrease in overall survival compared with control animals. Such an acceleration was not observed with loss of p107 in Trp53flox/flox; Rb1flox/flox; p130flox/flox animals, suggesting potential functional redundancy between p107 and p130 in SCLC. In summary, we have demonstrated the feasibility of using the CRISPR-Cas9 system to rapidly model loss-of-function mutations in candidate tumor-suppressor genes in mSCLC, and have used this approach to validate the role of p107 as a tumor suppressor in SCLC. Such an approach will greatly accelerate efforts to investigate the molecular mechanisms of tumor progression in SCLC.

Citation Format: Sheng Rong Ng, William M. Rideout, III, Bethany L. Wagner, Tyler Jacks. CRISPR-mediated modeling and functional validation of candidate tumor-suppressor genes in small cell lung cancer [abstract]. In: Proceedings of the AACR Special Conference: Advances in Modeling Cancer in Mice: Technology, Biology, and Beyond; 2017 Sep 24-27; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(10 Suppl):Abstract nr A12.

Foxa2 and Cdx2 cooperate with Nkx2-1 to inhibit lung adenocarcinoma metastasis

Li et al. demonstrate that loss of Foxa2 and Cdx2 synergizes with loss of Nkx2-1 to fully activate the metastatic program in lung cancer. Silencing of these three transcription factors is sufficient to account for a significant fraction of the gene expression differences between the nonmetastatic and metastatic states in lung adenocarcinoma.

Despite the fact that the majority of lung cancer deaths are due to metastasis, the molecular mechanisms driving metastatic progression are poorly understood. Here, we present evidence that loss of Foxa2 and Cdx2 synergizes with loss of Nkx2-1 to fully activate the metastatic program. These three lineage-specific transcription factors are consistently down-regulated in metastatic cells compared with nonmetastatic cells. Knockdown of these three factors acts synergistically and is sufficient to promote the metastatic potential of nonmetastatic cells to that of naturally arising metastatic cells in vivo. Furthermore, silencing of these three transcription factors is sufficient to account for a significant fraction of the gene expression differences between the nonmetastatic and metastatic states in lung adenocarcinoma, including up-regulated expression of the invadopodia component Tks5_long, the embryonal proto-oncogene Hmga2, and the epithelial-to-mesenchymal mediator Snail. Finally, analyses of tumors from a genetically engineered mouse model and patients show that low expression of Nkx2-1, Foxa2, and Cdx2 strongly correlates with more advanced tumors and worse survival. Our findings reveal that a large part of the complex transcriptional network in metastasis can be controlled by a small number of regulatory nodes that function redundantly, and loss of multiple nodes is required to fully activate the metastatic program.

RPL39L is an example of a recently evolved ribosomal protein paralog that shows highly specific tissue expression patterns and is upregulated in ESCs and HCC tumors

Ribosomal proteins (RPs) have been shown to be able to impart selectivity on the translating ribosome implicating them in gene expression control. Many ribosomal proteins are highly conserved and recently a number of ribosomal protein paralogs have been described in mammals. We examined the expression pattern of RPs in differentiating mouse Embryonic Stem Cells (ESCs), paying particular attention to the RP paralogs. We find the RP paralog Rpl39l is highly expressed in ESC and its expression strongly correlates with hepatocellular carcinoma tumor (HCC) samples with high tumor grading and alpha-fetoprotein level giving it diagnostic potential. We further screen the expression pattern of all RPs and their paralogs across 22 different tissues. We find that the more recently evolved RP paralogs show a much greater level of tissue-specific expression. We propose that these RP paralogs evolved more recently to provide a greater level of gene expression control to higher eukaryotes.