Targeting of vulnerabilities of drug-tolerant persisters identified through functional genetics delays tumor relapse

Drug-tolerant persisters (DTPs) are a rare subpopulation of cells within a tumor that can survive therapy through nongenetic adaptive mechanisms to develop relapse and repopulate the tumor following drug withdrawal. Using a cancer cell line with an engineered suicide switch to kill proliferating cells, we perform both genetic screens and compound screens to identify the inhibition of bromodomain and extraterminal domain (BET) proteins as a selective vulnerability of DTPs. BET inhibitors are especially detrimental to DTPs that have reentered the cell cycle (DTEPs) in a broad spectrum of cancer types. Mechanistically, BET inhibition induces lethal levels of ROS through the suppression of redox-regulating genes highly expressed in DTPs, including GPX2, ALDH3A1, and MGST1. In vivo BET inhibitor treatment delays tumor relapse in both melanoma and lung cancer. Our study suggests that combining standard of care therapy with BET inhibitors to eliminate residual persister cells is a promising therapeutic strategy.

Distinct spatiotemporal dynamics of CD8+ T cell-derived cytokines in the tumor microenvironment

Cells in the tumor microenvironment (TME) influence each other through secretion and sensing of soluble mediators, such as cytokines and chemokines. While signaling of interferon γ (IFNγ) and tumor necrosis factor α (TNFα) is integral to anti-tumor immune responses, our understanding of the spatiotemporal behavior of these cytokines is limited. Here, we describe a single cell transcriptome-based approach to infer which signal(s) an individual cell has received. We demonstrate that, contrary to expectations, CD8+ T cell-derived IFNγ is the dominant modifier of the TME relative to TNFα. Furthermore, we demonstrate that cell pools that show abundant IFNγ sensing are characterized by decreased expression of transforming growth factor β (TGFβ)-induced genes, consistent with IFNγ-mediated TME remodeling. Collectively, these data provide evidence that CD8+ T cell-secreted cytokines should be categorized into local and global tissue modifiers, and describe a broadly applicable approach to dissect cytokine and chemokine modulation of the TME.

CD4+ helper T cells endow cDC1 with cancer-impeding functions in the human tumor micro-environment

Despite their low abundance in the tumor microenvironment (TME), classical type 1 dendritic cells (cDC1) play a pivotal role in anti-cancer immunity, and their abundance positively correlates with patient survival. However, their interaction with CD4⁺ T-cells to potentially enable the cytotoxic T lymphocyte (CTL) response has not been elucidated. Here we show that contact with activated CD4⁺ T-cells enables human ex vivo cDC1, but no other DC types, to induce a CTL response to cell-associated tumor antigens. Single cell transcriptomics reveals that CD4⁺ T-cell help uniquely optimizes cDC1 in many functions that support antigen cross-presentation and T-cell priming, while these changes don't apply to other DC types. We robustly identify "helped" cDC1 in the TME of a multitude of human cancer types by the overlap in their transcriptomic signature with that of recently defined, tumor-infiltrating DC states that prove to be positively prognostic. As predicted from the functional effects of CD4⁺ T-cell help, the transcriptomic signature of "helped" cDC1 correlates with tumor infiltration by CTLs and Thelper(h)-1 cells, overall survival and response to PD-1-targeting immunotherapy. These findings reveal a critical role for CD4⁺ T-cell help in enabling cDC1 function in the TME and may establish the helped cDC1 transcriptomic signature as diagnostic marker in cancer.

CD26-negative and CD26-positive tissue-resident fibroblasts contribute to functionally distinct CAF subpopulations in breast cancer

Cancer-associated fibroblasts (CAFs) are abundantly present in the microenvironment of virtually all tumors and strongly impact tumor progression. Despite increasing insight into their function and heterogeneity, little is known regarding the origin of CAFs. Understanding the origin of CAF heterogeneity is needed to develop successful CAF-based targeted therapies. Through various transplantation studies in mice, we show that CAFs in both invasive lobular breast cancer and triple-negative breast cancer originate from mammary tissue-resident normal fibroblasts (NFs). Single-cell transcriptomics, in vivo and in vitro studies reveal the transition of CD26+ and CD26- NF populations into inflammatory CAFs (iCAFs) and myofibroblastic CAFs (myCAFs), respectively. Functional co-culture experiments show that CD26+ NFs transition into pro-tumorigenic iCAFs which recruit myeloid cells in a CXCL12-dependent manner and enhance tumor cell invasion via matrix-metalloproteinase (MMP) activity. Together, our data suggest that CD26+ and CD26- NFs transform into distinct CAF subpopulations in mouse models of breast cancer.

Drug-induced epigenomic plasticity reprograms circadian rhythm regulation to drive prostate cancer towards androgen-independence

In prostate cancer, androgen receptor (AR)-targeting agents are very effective in various disease stages. However, therapy resistance inevitably occurs and little is known about how tumor cells adapt to bypass AR suppression. Here, we performed integrative multi-omics analyses on tissues isolated before and after 3 months of AR-targeting enzalutamide monotherapy from high-risk prostate cancer patients enrolled in a neoadjuvant clinical trial. Transcriptomic analyses demonstrated that AR inhibition drove tumors towards a neuroendocrine-like disease state. Additionally, epigenomic profiling revealed massive enzalutamide-induced reprogramming of pioneer factor FOXA1 - from inactive chromatin sites towards active cis-regulatory elements that dictate pro-survival signals. Notably, treatment-induced FOXA1 sites were enriched for circadian clock component ARNTL. Post-treatment ARNTL levels associated with poor outcome, and ARNTL knockout strongly decreased prostate cancer cell growth. Our data highlight a remarkable cistromic plasticity of FOXA1 following AR-targeted therapy, and revealed an acquired dependency on circadian regulator ARNTL, a novel candidate therapeutic target.

Clinically applicable CD34+-derived blood dendritic cell subsets exhibit key subset-specific features and potently boost anti-tumor T and NK cell responses

Allogeneic stem cell transplantation (alloSCT), following induction chemotherapy, can be curative for hemato-oncology patients due to powerful graft-versus-tumor immunity. However, disease recurrence remains the major cause of treatment failure, emphasizing the need for potent adjuvant immunotherapy. In this regard, dendritic cell (DC) vaccination is highly attractive, as DCs are the key orchestrators of innate and adaptive immunity. Natural DC subsets are postulated to be more powerful compared with monocyte-derived DCs, due to their unique functional properties and cross-talk capacity. Yet, obtaining sufficient numbers of natural DCs, particularly type 1 conventional DCs (cDC1s), is challenging due to low frequencies in human blood. We developed a clinically applicable culture protocol using donor-derived G-CSF mobilized CD34⁺ hematopoietic progenitor cells (HPCs) for simultaneous generation of high numbers of cDC1s, cDC2s and plasmacytoid DCs (pDCs). Transcriptomic analyses demonstrated that these ex vivo-generated DCs highly resemble their in vivo blood counterparts. In more detail, we demonstrated that the CD141⁺CLEG9A⁺ cDC1 subset exhibited key features of in vivo cDC1s, reflected by high expression of co-stimulatory molecules and release of IL-12p70 and TNF-α. Furthermore, cDC1s efficiently primed alloreactive T cells, potently cross-presented long-peptides and boosted expansion of minor histocompatibility antigen-experienced T cells. Moreover, they strongly enhanced NK cell activation, degranulation and anti-leukemic reactivity. Together, we developed a robust culture protocol to generate highly functional blood DC subsets for in vivo application as tailored adjuvant immunotherapy to boost innate and adaptive anti-tumor immunity in alloSCT patients.

Flagellin/TLR5 Stimulate Myeloid Progenitors to Enter Lung Tissue and to Locally Differentiate Into Macrophages

Toll-like receptor 5 (TLR5) is the receptor of bacterial Flagellin. Reportedly, TLR5 engagement helps to combat infections, especially at mucosal sites, by evoking responses from epithelial cells and immune cells. Here we report that TLR5 is expressed on a previously defined bipotent progenitor of macrophages (MΦs) and osteoclasts (OCs) that resides in the mouse bone marrow (BM) and circulates at low frequency in the blood. In vitro, Flagellin promoted the generation of MΦs, but not OCs from this progenitor. In vivo, MΦ/OC progenitors were recruited from the blood into the lung upon intranasal inoculation of Flagellin, where they rapidly differentiated into MΦs. Recruitment of the MΦ/OC progenitors into the lung was likely promoted by the CCL2/CCR2 axis, since the progenitors expressed CCR2 and type 2 alveolar epithelial cells (AECs) produced CCL2 upon stimulation by Flagellin. Moreover, CCR2 blockade reduced migration of the MΦ/OC progenitors toward lung lavage fluid (LLF) from Flagellin-inoculated mice. Our study points to a novel role of the Flagellin/TLR5 axis in recruiting circulating MΦ/OC progenitors into infected tissue and stimulating these progenitors to locally differentiate into MΦs. The progenitor pathway to produce MΦs may act, next to monocyte recruitment, to fortify host protection against bacterial infection at mucosal sites.

Histone methyltransferase DOT1L controls state-specific identity during B cell differentiation

Differentiation of naïve peripheral B cells into terminally differentiated plasma cells is characterized by epigenetic alterations, yet the epigenetic mechanisms that control B‐cell fate remain unclear. Here, we identified a role for the histone H3K79 methyltransferase DOT1L in controlling B‐cell differentiation. Mouse B cells lacking Dot1L failed to establish germinal centers (GC) and normal humoral immune responses in vivo. In vitro, activated B cells in which Dot1L was deleted showed aberrant differentiation and prematurely acquired plasma cell characteristics. Similar results were obtained when DOT1L was chemically inhibited in mature B cells in vitro. Mechanistically, combined epigenomics and transcriptomics analysis revealed that DOT1L promotes expression of a pro‐proliferative, pro‐GC program. In addition, DOT1L indirectly supports the repression of an anti‐proliferative plasma cell differentiation program by maintaining the repression of Polycomb Repressor Complex 2 (PRC2) targets. Our findings show that DOT1L is a key modulator of the core transcriptional and epigenetic landscape in B cells, establishing an epigenetic barrier that warrants B‐cell naivety and GC B‐cell differentiation.

Integrin α3β1 in hair bulge stem cells modulates CCN2 expression and promotes skin tumorigenesis

Although hair bulge stem cells are not the cancer cells-of-origin, they contribute to two-stage DMBA/TPA skin carcinogenesis in an α3β1-dependent manner.

Epidermal-specific deletion of integrin α3β1 almost completely prevents the formation of papillomas during 7,12-Dimethylbenz[a]anthracene/12-O-tetradecanoylphorbol-13-acetate (DMBA/TPA) two-stage skin carcinogenesis. This dramatic decrease in tumorigenesis was thought to be due to an egress and premature differentiation of α3β1-depleted hair bulge (HB) stem cells (SCs), previously considered to be the cancer cells-of-origin in the DMBA/TPA model. Using a reporter mouse line with inducible deletion of α3β1 in HBs, we show that HB SCs remain confined to their niche regardless of the presence of α3β1 and are largely absent from skin tumors. However, tumor formation was significantly decreased in mice deficient for α3β1 in HB SCs. RNA sequencing of HB SCs isolated from short-term DMBA/TPA–treated skin showed α3β1-dependent expression of the matricellular protein connective tissue growth factor (CCN2), which was confirmed in vitro, where CCN2 promoted colony formation and 3D growth of transformed keratinocytes. Together, these findings show that HBs contribute to skin tumorigenesis in an α3β1-dependent manner and suggest a role of HB SCs in creating a permissive environment for tumor growth through the modulation of CCN2 secretion.

Androgen modulation of XBP1 is functionally driving part of the AR transcriptional program

Prostate cancer development and progression is largely dependent on androgen receptor (AR) signaling. AR is a hormone-dependent transcription factor, which binds to thousands of sites throughout the human genome to regulate expression of directly responsive genes, including pro-survival genes that enable tumor cells to cope with increased cellular stress. ERN1 and XBP1 - two key players of the unfolded protein response (UPR) - are among such stress-associated genes. Here, we show that XBP1 levels in primary prostate cancer are associated with biochemical recurrence in five independent cohorts. Patients who received AR-targeted therapies had significantly lower XBP1 expression, whereas expression of the active form of XBP1 (XBP1s) was elevated. In vitro results show that AR-induced ERN1 expression led to increased XBP1s mRNA and protein levels. Furthermore, ChIP-seq analysis revealed that XBP1s binds enhancers upon stress stimuli regulating genes involved in UPR processes, eIF2 signaling and protein ubiquitination. We further demonstrate genomic overlap of AR- and XBP1s-binding sites, suggesting genomic conversion of the two signaling cascades. Transcriptomic effects of XBP1 were further studied by knockdown experiments, which lead to decreased expression of androgen-responsive genes and UPR genes. These results suggest a two-step mechanism of gene regulation, which involves androgen-induced expression of ERN1, thereby enhancing XBP1 splicing and transcriptional activity. This signaling cascade may prepare the cells for the increased protein folding, mRNA decay and translation that accompanies AR-regulated tumor cell proliferation.

Loss of p53 triggers WNT-dependent systemic inflammation to drive breast cancer metastasis

Cancer-associated systemic inflammation is strongly linked to poor disease outcome in patients with cancer^1,2. For most human epithelial tumour types, high systemic neutrophil-to-lymphocyte ratios are associated with poor overall survival³, and experimental studies have demonstrated a causal relationship between neutrophils and metastasis^4,5. However, the cancer-cell-intrinsic mechanisms that dictate the substantial heterogeneity in systemic neutrophilic inflammation between tumour-bearing hosts are largely unresolved. Here, using a panel of 16 distinct genetically engineered mouse models for breast cancer, we uncover a role for cancer-cell-intrinsic p53 as a key regulator of pro-metastatic neutrophils. Mechanistically, loss of p53 in cancer cells induced the secretion of WNT ligands that stimulate tumour-associated macrophages to produce IL-1β, thus driving systemic inflammation. Pharmacological and genetic blockade of WNT secretion in p53-null cancer cells reverses macrophage production of IL-1β and subsequent neutrophilic inflammation, resulting in reduced metastasis formation. Collectively, we demonstrate a mechanistic link between the loss of p53 in cancer cells, secretion of WNT ligands and systemic neutrophilia that potentiates metastatic progression. These insights illustrate the importance of the genetic makeup of breast tumours in dictating pro-metastatic systemic inflammation, and set the stage for personalized immune intervention strategies for patients with cancer.

CDK4/6 inhibitors target SMARCA4-determined cyclin D1 deficiency in hypercalcemic small cell carcinoma of the ovary

Inactivating mutations in SMARCA4 (BRG1), a key SWI/SNF chromatin remodelling gene, underlie small cell carcinoma of the ovary, hypercalcemic type (SCCOHT). To reveal its druggable vulnerabilities, we perform kinase-focused RNAi screens and uncover that SMARCA4-deficient SCCOHT cells are highly sensitive to the inhibition of cyclin-dependent kinase 4/6 (CDK4/6). SMARCA4 loss causes profound downregulation of cyclin D1, which limits CDK4/6 kinase activity in SCCOHT cells and leads to in vitro and in vivo susceptibility to CDK4/6 inhibitors. SCCOHT patient tumors are deficient in cyclin D1 yet retain the retinoblastoma-proficient/p16INK4a-deficient profile associated with positive responses to CDK4/6 inhibitors. Thus, our findings indicate that CDK4/6 inhibitors, approved for a breast cancer subtype addicted to CDK4/6 activation, could be repurposed to treat SCCOHT. Moreover, our study suggests a novel paradigm whereby critically low oncogene levels, caused by loss of a driver tumor suppressor, may also be exploited therapeutically.

XenofilteR: computational deconvolution of mouse and human reads in tumor xenograft sequence data

BACKGROUND

Mouse xenografts from (patient-derived) tumors (PDX) or tumor cell lines are widely used as models to study various biological and preclinical aspects of cancer. However, analyses of their RNA and DNA profiles are challenging, because they comprise reads not only from the grafted human cancer but also from the murine host. The reads of murine origin result in false positives in mutation analysis of DNA samples and obscure gene expression levels when sequencing RNA. However, currently available algorithms are limited and improvements in accuracy and ease of use are necessary.

RESULTS

We developed the R-package XenofilteR, which separates mouse from human sequence reads based on the edit-distance between a sequence read and reference genome. To assess the accuracy of XenofilteR, we generated sequence data by in silico mixing of mouse and human DNA sequence data. These analyses revealed that XenofilteR removes > 99.9% of sequence reads of mouse origin while retaining human sequences. This allowed for mutation analysis of xenograft samples with accurate variant allele frequencies, and retrieved all non-synonymous somatic tumor mutations.

CONCLUSIONS

XenofilteR accurately dissects RNA and DNA sequences from mouse and human origin, thereby outperforming currently available tools. XenofilteR is open source and available at https://github.com/PeeperLab/XenofilteR .

Abstract 1041: XenofilteR: Computational dissection of mouse and human reads in PDX and xenograft sequence data

Mouse xenografts from (patient-derived) tumors (PDX) or tumor cell lines are widely used as models to study various biological and preclinical aspects of cancer. However, analysis of their RNA and DNA profiles is challenging, because they comprise reads not only from the grafted human cancer but also from the murine host. The reads of murine origin can result both in the generation of false positives in mutation analysis of DNA samples and obscure gene expression levels when sequencing RNA. Therefore, we developed the open-source R-package XenofilteR, which separates mouse from human sequence reads based on the number of discordant base pairs between each read and the reference genomes. To assess the accuracy of XenofilteR, we generated sequence data by in silico mixing of mouse and human whole genome and whole exome DNA sequence data. This analysis revealed that XenofilteR removes >99.9% of sequence reads of mouse origin while retaining sequence reads of human origin. The filtering allowed for mutation analysis of PDX samples with accurate variant allele frequencies, and retrieved all non-synonymous somatic mutations present in the original tumor. These findings were further validated in breast cancer and melanoma PDX samples, confirming the retrieval of accurate variant allele frequencies and somatic mutations. In conclusion, XenofilteR accurately dissects sequence reads from mouse and human origin in PDX sequence data, thereby outperforming currently available tools.

Citation Format: Oscar Krijgsman, Roelof JC Kluin, Kristel Kemper, Thomas Kuilman, Julian R. de Ruiter, Vivek Iyer, Josep V. Forment, Paulien Cornelissen-Steijger, Iris de Rink, Petra ter Brugge, Ji-Ying Song, Sjoerd Klarenbeek, Ultan McDermott, Jos Jonkers, Arno Velds, David J. Adams, Daniel S. Peeper. XenofilteR: Computational dissection of mouse and human reads in PDX and xenograft sequence data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1041.

Knockout of the Arp2/3 complex in epidermis causes a psoriasis-like disease hallmarked by hyperactivation of transcription factor Nrf2

The Arp2/3 complex assembles branched actin filaments, which are key to many cellular processes, but its organismal roles remain poorly understood. Here, we employed conditional Arpc4 knockout mice to study the function of the Arp2/3 complex in the epidermis. We found that depletion of the Arp2/3 complex by knockout of Arpc4 results in skin abnormalities at birth that evolve into a severe psoriasis-like disease hallmarked by hyperactivation of transcription factor Nrf2. Knockout of Arpc4 in cultured keratinocytes was sufficient to induce nuclear accumulation of Nrf2, upregulation of Nrf2 target genes and decreased filamentous actin levels. Furthermore, pharmacological inhibition of the Arp2/3 complex unmasked the role of branched actin filaments in Nrf2 regulation. Consistent with this, we revealed that Nrf2 associates with the actin cytoskeleton in cells and binds to filamentous actin in vitro Finally, we discovered that Arpc4 is downregulated in both human and mouse psoriatic epidermis. Thus, the Arp2/3 complex affects keratinocyte shape and transcriptome through an actin-based cell-autonomous mechanism that influences epidermal morphogenesis and homeostasis.

Abstract 5409: Assessment of the MammaPrint 70-gene profile using RNA sequencing technology

Introduction: Improvements in RNA processing have enabled microarray diagnostics for formalin-fixed, paraffin-embedded (FFPE) tissue. Recently, MammaPrint, a prognostic 70-gene profile for early-stage breast cancer, was successfully translated to FFPE tissue showing to be substantially equivalent to fresh tissue. In recent years, RNA-sequencing (RNA-Seq) became the standard method for transcriptome analysis, because of its low background signal and its ability of quantifying a large dynamic range of expression levels. Here we report a preliminary analysis of the FFPE MammaPrint 70-gene profile using RNA-Seq technology and the comparison with the MammaPrint® microarray diagnostic test in a series of FFPE samples.

Methods: RNA-Seq was carried out using a strand-specific RNA library preparation followed by target enrichment of the coding region of the human transcriptome without relying on the presence of poly-A tail. RNA sequencing libraries were prepared starting from a minimal amount of 20 ng of total RNA based on the DV200 metric assessment. The library pools were single-end sequenced on the Illumina HiSeq 2500 instrument at the length of 65bp. The resulting sequences were mapped to the human reference genome (build 38) using TopHat v2.1. Tophat was guided by using a transcriptome index from Ensembl (version 77). The HTSeq-count tool was used to generate the total number of uniquely mapped reads for each gene. Gene expressions were normalized with Count Per Million (CPM) normalization and log2 transformed afterwards. Microarray data of the sample were available for analysis comparison.

Results: On average, we obtained 22 million reads assigned to gene per sample (min=15M, max=28M). The number of reads assigned to genes vary from 61% to 70% of the total number of reads. Between 80% and 90% of the reads assigned to genes mapped to protein coding genes which is comparable to fresh frozen material. The 70-gene signature was successfully mapped to the RNA-Seq genes. A median raw read-count of 384 was observed for the 70-gene profile among the samples. Importantly, we observed a high concordance (R2 Pearson correlation=0.97) between the MammaPrint index calculated using the RNA-Seq data and the correspondent Microarray MammaPrint index. Additionally, the BluePrint profile, a microarray diagnostic test for breast cancer molecular subtyping, was successfully translated to the RNA-Seq platform. As with the MammaPrint profile, BluePrint showed high concordance between the two technologies with high correlation values for each of the subtypes (Luminal R2 Pearson correlation=0.98, Basal R2 Pearson correlation=0.97, HER2 R2 Pearson correlation=0.77). Conclusions: Next Generation RNA-sequencing is a feasible technology to assess diagnostic signatures, such as the 70 gene MammaPrint and BluePrint profiles.

Citation Format: Lorenza Mittempergher, Jacob B. Spangler, Mireille H. Snel, Leonie J. Delahaye, Iris de Rink, Sun Tian, Annuska M. Glas, Rene Bernards. Assessment of the MammaPrint 70-gene profile using RNA sequencing technology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5409. doi:10.1158/1538-7445.AM2017-5409

Glycerophosphodiesterase GDE2 Promotes Neuroblastoma Differentiation through Glypican Release and Is a Marker of Clinical Outcome

Neuroblastoma is a pediatric embryonal malignancy characterized by impaired neuronal differentiation. A better understanding of neuroblastoma differentiation is essential for developing new therapeutic approaches. GDE2 (encoded by GDPD5) is a six-transmembrane-domain glycerophosphodiesterase that promotes embryonic neurogenesis. We find that high GDPD5 expression is strongly associated with favorable outcome in neuroblastoma. GDE2 induces differentiation of neuroblastoma cells, suppresses cell motility, and opposes RhoA-driven neurite retraction. GDE2 alters the Rac-RhoA activity balance and the expression of multiple differentiation-associated genes. Mechanistically, GDE2 acts by cleaving (in cis) and releasing glycosylphosphatidylinositol-anchored glypican-6, a putative co-receptor. A single point mutation in the ectodomain abolishes GDE2 function. Our results reveal GDE2 as a cell-autonomous inducer of neuroblastoma differentiation with prognostic significance and potential therapeutic value.

REV7 counteracts DNA double-strand break resection and affects PARP inhibition

Error-free repair of DNA double-strand breaks (DSBs) is achieved by homologous recombination (HR), and BRCA1 is an important factor for this repair pathway. In the absence of BRCA1-mediated HR, the administration of PARP inhibitors induces synthetic lethality of tumour cells of patients with breast or ovarian cancers. Despite the benefit of this tailored therapy, drug resistance can occur by HR restoration. Genetic reversion of BRCA1-inactivating mutations can be the underlying mechanism of drug resistance, but this does not explain resistance in all cases. In particular, little is known about BRCA1-independent restoration of HR. Here we show that loss of REV7 (also known as MAD2L2) in mouse and human cell lines re-establishes CTIP-dependent end resection of DSBs in BRCA1-deficient cells, leading to HR restoration and PARP inhibitor resistance, which is reversed by ATM kinase inhibition. REV7 is recruited to DSBs in a manner dependent on the H2AX-MDC1-RNF8-RNF168-53BP1 chromatin pathway, and seems to block HR and promote end joining in addition to its regulatory role in DNA damage tolerance. Finally, we establish that REV7 blocks DSB resection to promote non-homologous end-joining during immunoglobulin class switch recombination. Our results reveal an unexpected crucial function of REV7 downstream of 53BP1 in coordinating pathological DSB repair pathway choices in BRCA1-deficient cells.

Abstract 2767: Identification of kinase fusion genes in bladder cancer through kinome-centered RNA sequencing

Urothelial carcinoma (UC) of the bladder is one of the most common cancers worldwide. The treatment of patients with metastatic or locally advanced UC consists of platinum-based chemotherapy. Unfortunately, a large number of patients will ultimately develop resistance. Therapeutic options at that point are very limited: the last FDA drug approval for the treatment of bladder cancer dates back more than 2 decades. Identification of activated signaling pathways in UC can provide new targets for treatment. Recent DNA and RNA sequencing projects in invasive UC have revealed somatic mutations in several cancer genes. Some of these mutated genes, such as FGFR3 and PIK3CA, could potentially guide therapy. However, our knowledge of gene rearrangements in bladder cancer remains limited. These fusion genes often involve multiple fusion partners, as was reported for FGFR3 fusions in bladder cancer, which represents a significant challenge for discovery and for subsequent diagnostic screening. Therefore, a global detection method is needed to fully understand the diversity of alterations driving this disease. We used a high-throughput platform to systematically profile kinase fusions through specific enrichment of kinase transcripts. Using this approach, we screened 80 muscle invasive UC specimens and identified a number of activating mutations and novel fusion transcripts. The fusion genes identified in this discovery set will be validated in a second cohort of UC specimens to determine their frequency. Functional validation will be presented for some of these fusion genes. These genetic alterations may provide new avenues for individualized molecular treatment of UC patients.

Citation Format: Floris Groenendijk, Iris de Rink, Laura Mertens, Annegien Broeks, Yann Neuzillet, Jeroen de Jong, Bas van Rhijn, Rene Bernards, Michiel van der Heijden. Identification of kinase fusion genes in bladder cancer through kinome-centered RNA sequencing. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2767. doi:10.1158/1538-7445.AM2014-2767

Abstract 3890: MicroRNA-551b is highly expressed in hematopoietic stem cells and expression in acute myeloid leukemia is associated with relapse and poor survival

Despite high remission rates after chemotherapy, only 30-40% of acute myeloid leukemia (AML) patients survive five years after diagnosis. The main cause for this treatment failure is insufficient eradication of a subpopulation of chemotherapy resistant leukemic cells with a stem cell-like character. These so-called leukemic stem cells (LSC) are thought to be responsible for relapse.

We hypothesized that success of novel anti-AML therapies relies on functional manipulation of genes including microRNAs, resulting in elimination of leukemic (stem) cells while sparing residual co-existing normal hematopoietic stem cells (HSC). We aimed at identification of microRNAs differentially expressed between HSC, LSC and the AML bulk obtained from the same AML bone marrow, taking into account the effects of the leukemic microenvironment.

To that end, we described immunophenotypic markers that can distinguish LSC from HSC. Moreover, we identified that HSC have higher aldehyde dehydrogenase activity than LSC (Schuurhuis et al. Plos One 2013). Comparing the microRNA expression profile of LSC with that of HSC showed that microRNA-551b (miR-551b) is highly expressed in residual HSC in the AML bone marrow. To determine whether miR-551b is a HSC specific microRNA we purified stem and progenitor cell subsets from normal bone marrow and showed that miR-551b is the highest expressed in the two most primitive CD34+CD38- populations i.e. CD90+CD45RA- HSC and CD90-CD45RA- multipotent progenitors.

To investigate whether the expression of miR-551b is of clinical importance in AML we determined its expression in AML bone marrow samples (n=154) and showed that high miR-551b is associated with lower complete remission (CR) rates after the first cycle of induction chemotherapy, shorter relapse free survival and a worse overall survival. In line with miR-551b being a stem cell miRNA, high expression in AML was associated with an undifferentiated morphology (FAB M0).

To shed more light on the functional role of miR-551b in AML we correlated the expression of miR-551b with overall gene expression in a large panel of AML patients. Many of the genes that highly correlated with miR-551b like; MLLT3, INPP4B, HTR1F, HOPX, PROM1 and others, are also present in published HSC signatures. In conclusion, miR-551b is specifically expressed in normal stem and multipotent progenitor cells and high expression in AML is associated with poor prognosis. Currently, our research focuses on the function of miR-551b in AML.

Citation Format: David C. de leeuw, Fedor Denkers, Peter Valk, Iris de Rink, Ron Kerkhoven, Gerrit Jan Schuurhuis, Gert J. Ossenkoppele, Linda Smit. MicroRNA-551b is highly expressed in hematopoietic stem cells and expression in acute myeloid leukemia is associated with relapse and poor survival. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3890. doi:10.1158/1538-7445.AM2014-3890

Identification of recurrent FGFR3 fusion genes in lung cancer through kinome-centred RNA sequencing

Oncogenic fusion genes that involve kinases have proven to be effective targets for therapy in a wide range of cancers. Unfortunately, the diagnostic approaches required to identify these events are struggling to keep pace with the diverse array of genetic alterations that occur in cancer. Diagnostic screening in solid tumours is particularly challenging, as many fusion genes occur with a low frequency. To overcome these limitations, we developed a capture enrichment strategy to enable high-throughput transcript sequencing of the human kinome. This approach provides a global overview of kinase fusion events, irrespective of the identity of the fusion partner. To demonstrate the utility of this system, we profiled 100 non-small cell lung cancers and identified numerous genetic alterations impacting fibroblast growth factor receptor 3 (FGFR3) in lung squamous cell carcinoma and a novel ALK fusion partner in lung adenocarcinoma.

High concordance for MammaPrint 70 genes by RNA next generation sequencing

3065

Background: The development of new biomarkers often requires fresh frozen (FF) samples. Recently we showed that microarray gene expression data generated from FFPE material are comparable to data extracted from the FF counterpart, including known signatures such as the 70-gene prognosis signature (Mittempergher L et al., 2011). As described by Luo et al (2010) RNA profiling using next generation sequencing (RNA-Seq) is now applicable to archival FFPE specimens. Methods: Technical performance and the comparison between the RNA-Seq 70-gene read-out and the MammaPrint test (Glas et al., 2006) is evaluated in a series of 15 patients (11/15 with matched FFPE/FF material). RNA-Seq was carried out using minor adjustments of the Illumina TruSeq RNA preparation method. RNA sequencing libraries were prepared starting from 100ng of total RNA. Next, the DSN (Duplex-Specific Nuclease) normalization process was used to remove ribosomal RNA and other abundant transcripts (Luo et al, 2010). The libraries were paired-end sequenced on the Illumina HiSeq 2000 instrument with multiplexing of 4 libraries per lane. The resulting sequences were mapped to the human reference genome (build 37) using TopHat 1.3.1(Trapnell et al., 2009). The HTSeq-count tool was used to generate the total number of uniquely mapped reads for each gene. Results: Between 14% and 45% of the total number of reads were assigned to protein-coding genes. The minimum coverage per 1000bp of CDS was 38 reads. The 70 MammaPrint genes were successfully mapped to the RNA-Seq transcripts. We calculated the Pearson correlation coefficient between the centroids of the original good prognosis template (van’t Veer et al., 2002) and the 70-gene read count determined by RNA-Seq of each sample. Predictions based on the 70-gene RNA-Seq data showed a high agreement with the actual MammaPrint test predictions (>90%), irrespective of whether the RNA-seq was performed on FF or FFPE tissue. Conclusions: New generation RNA-sequencing is a feasible technology to assess diagnostic signatures.

Abstract LB-386: The NuRD complex cooperates with DNMTs to maintain silencing of colorectal tumor suppressor genes

Many Tumor Suppressor Genes (TSGs) are silenced through synergistic layers of epigenetic regulation including abnormal DNA hypermethylation of promoter CpG islands, repressive chromatin modifications and enhanced nucleosome deposition over transcription start sites. Many of the protein complexes responsible for silencing of such TSGs remain to be identified. A subset of silenced TSGs controlling key regulatory signaling pathways in colorectal cancer cells can be partially reactivated by in cells disrupted for the DNA methyltransferase 1 and 3B (DNMT1 and 3B) or by DNMT inhibitors (DNMTi). Herein, we used proteomic and functional genomic approaches to identify additional proteins that cooperate with DNMTs in silencing these key silenced TSGs in colon cancer cells. We discovered that DNMTs and the core components of the NuRD nucleosome remodeling complex, chromo domain helicase DNA-binding protein 4 (CHD4), histone deacetylases 1 and 2 (HDAC1 and 2), occupy the promoters of several of these key hypermethylated TSGs and physically and functionally interact to maintain their silencing. Consistent with this, we find an inverse relationship between expression of HDAC1 and 2 and these TSGs in a large panel of primary colorectal tumors. We demonstrate that this DNMT-NuRD complex maintains the silencing of several negative regulators of the WNT signaling pathway. We find that depletion of CHD4 is synthetic lethal with DNMT inhibition in correlation with reactivation of TSGs, suggesting that their combined inhibition may be beneficial for the treatment of colon cancer.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr LB-386. doi:1538-7445.AM2012-LB-386