Transcriptional Regulatory Landscape of TCF3-PBX1-Positive Leukemia and Novel Targeted Treatments

Introduction

The oncogenic TCF3-PBX1 (TP; also known as E2A-PBX1) fusion gene results from a translocation between chromosomes 1 and 19 in pre-B-cell acute leukemia (pre-B-ALL). Both TCF3 and PBX1 function as transcription factors (TF), and their fusion generates a unique transcriptional landscape in t(1;19) leukemias which differs from other pre-B-ALL subtypes. Here we explored the transcriptional regulatory landscape of t(1;19)-leukemia genome-wide and sought novel targeted therapy options.

Materials & Methods

We modeled the oncogenicity of the fusion protein in leukemic cells by either expressing (Nalm6-TP, induced for 16 hours) or silencing (697-shTP, down to appr. 40 %) the fusion. Proliferation, apoptosis and cell cycle were studied using fluorometric reactions and flow cytometry. Patient samples (TP, n=4; other subtypes, n=18) and various cell line models were subjected to global nuclear run-on sequencing (GRO-seq), which provides a genome-wide map of nascent (primary) transcription. All transcripts that were altered after overexpression or silencing of the fusion-TF in cell models were inspected from the patient GRO-seq samples. TCF3-PBX1-regulated genomic regions were studied for enrichment of TF binding motifs and altered signaling pathways. Mature RNA levels and potential novel long non-coding RNAs were further validated by qPCR. Gene expression differences between t(1;19) and other subtypes were compared using a curated microarray data set containing 1304 pre-B-ALL samples retrieved from 15 different data sets from the Gene Expression Omnibus. ARACNE, a network inference algorithm, and GRO-seq were used to identify TFs correlating strongly with the t(1;19) subtype and to find novel drug targets. Drugs were tested in cell culture using t(1;19)-positive cell line models and patient samples either alone or in combination with known leukemia therapies.

Results

GRO-seq analysis allowed elucidation of the regulatory landscape downstream of the TCF3-PBX1 fusion protein. Directly regulated enhancer RNAs were matched to the cis regulated genes, and vice versa, to clarify enhancer-gene relations. As an example, correspondingly regulated enhancer regions were located for WNT16 and ANKS1B, two genes that are known targets of TCF3-PBX1, and that were found consistently upregulated in the studied sample sets. EBF3, a tumor suppressor gene, was one of the top hits in the network inference analysis and was also found consistently regulated by TCF3-PBX1. One of the identified druggable target was RORB which was directly upregulated by TCF3-PBX1. An inhibitor targeting RORB decreased viability of TCF3-PBX1-positive cell lines and cells from a t(1;19) patient. The effect was especially prominent when the inhibitor was combined with a low dose (1 nM) of vincristine, yielding a marked synergistic effect.

Conclusions

Our results provide the first genome-wide transcriptional regulatory landscape of TCF3-PBX1 leukemia. We also identified novel putative druggable targets and a potential inhibitor for this leukemia subtype.