Novel Intergenically Spliced Chimera, NFATC3-PLA2G15, Is Associated with Aggressive T-ALL Biology and Outcome

Gene fusion is a frequent hallmark of leukemia and can arise due to a variety of structural chromosomal rearrangements, including translocation (e.g., BCR-ABL1), inversion (e.g., CBFβ-MYH11), and interstitial deletion (e.g., FIP1L1-PDGFRA; ref. 1). Fusion products are often critical mediators of leukemogenesis and therefore represent attractive therapeutic targets, as typified by the founder example of BCR-ABL kinase inhibition with imatinib (2).

The advent of high-throughput RNA sequencing (RNA-seq) has provided novel insights into the transcriptional landscapes of normal and malignant cells. It is now clear that expression of fusion mRNAs in the absence of structural rearrangements is more common than previously recognized. In particular, transcriptional read-through of a single mRNA between contiguous loci, also known as cis-splicing of adjacent genes (cis-SAGe), has been estimated to occur at 4% to 5% of the human genome (3, 4). Expression of the resultant intergenically spliced chimeric mRNAs (ISC) is frequent in normal cells (5, 6). Mounting evidence suggests that multiple cancers demonstrate aberrant ISC expression and that experimental inhibition of specific fusion transcripts can be toxic for malignant cells (7, 8).

To investigate whether cis-SAGe generates biologically important fusions in T-acute lymphoblastic leukemia (T-ALL), we performed RNA-seq of 12 diagnostic leukemic samples. We detected a high frequency of T-ALL–associated ISCs and notably found that expression of the NFATC3-PLA2G15 chimera correlated with aggressive disease biology.

Paired-end stranded RNA-seq (2 × 50 bp) of the initial series of 12 samples was performed with poly(A)-enriched RNAs using the SOLiD HQ5500XL system (Life Technologies). Mapping, coverage, and fusion discovery were determined using LifeScope (Life Technologies) using default parameters, with reference to version hg19 of the human genome. Selection of fusion transcripts required a total of 3 reads spanning two distinct gene transcripts (including at least 1 paired-end read and 1 split read). Fusions detected in normal thymic RNA-seq samples were removed. Fusions that involved genes located within 30 kb of each other with the same transcriptional orientation were defined as candidate ISCs. Sequencing of the second series of 12 samples was performed with poly(A)-enriched RNAs using the Illumina platform (paired-end 2 × 75 bp) and mapped with the TopHat-fusion algorithm. Sequencing depth in this series ranged between 30 and 70 millions of reads.

Extra information regarding the choice of filter criteria and parameters for detection of ISCs is included in the Supplementary Methods. Details of sequencing read depth are shown in Supplementary Table S1.

RT-PCR was performed using two fusion-specific primer sets:

• Set 1: 5′: CAACCATTGGTCTGCAGGAC; 3′: GGTGTGGGGACGCCAGTAC;

• Set 2: 5′: CAGGGGGGTCTTTCTGCAC; 3′: GGTGTCTGCACGAACACCTTC.

Transcript expression was confirmed by direct sequencing of PCR products.

NFATC3-PLA2G15 levels were quantified by a fusion-specific TaqMan QPCR system:

• 5′: GAACCAGAAGATCGAGAGCCTAAC; 3′: TCCGGTTGTTGTCTCCATCA,

• probe: TTGCAACCATTGGTCTGCAGGACATC. Fusion transcript levels were normalized to ABL expression: 5′: TGGAGATAACACTCTAAGCATAACTAAAGGT;

• 3′: GATGTAGTTGCTTGGGACCCA, probe: CCATTTTTGGTTTGGGCTTCACACCATT.

5′ RACE PCR was performed using the SMARTer RACE cDNA Amplification Kit (Clontech), following the manufacturer's instructions. A schematic representation of primer positions is shown in Supplementary Fig. S1. The sequences were:

• Primer NFATC3-PLA2G15: GGGATCCGGTTGTTGTCTCCATCATCTA,

• Primer NFATC3 WT: AGGCTGAAGCTGAGGAGATGGTGGCC.

Array competitive genomic hybridization (CGH) was performed using the Affymetrix Genome-Wide Human SNP Array 6.0, using leukemic DNA extracted from 115 T-ALL samples at diagnosis. CGH data were analyzed using Chromosome Analysis Suite software (Affymetrix).

The pMSCV-IRES-GFP (pMIG)-HA-NFATC3 vector was generated by insertion of a Topo HA-NFATC3 fragment [obtained following PCR amplification from the pOTB7-NFATC3 plasmid (Biovalley)] between the XhoI and EcoRI restriction sites of the pMIG multiple cloning site. The pMIG-HA-NFATC3-PLA2G15 vector was generated following PCR amplification of a 1.3-kb segment of NFATC3-PLA2G15 cDNA from a human leukemic sample. The amplified fragment was then cloned into the pMIG-NFATC3 vector backbone, using a naturally occurring MfeI site in the NFATC3 cDNA.

Cells were attached to slides using poly-l-lysine (0.01%), fixed with formaldehyde (3.5%), and permeabilized with Triton X-100 (1%). Slides were incubated with an anti-NFATC3 (R&D Systems, MAB5834) primary antibody and a goat anti-mouse 555 Alexa red (Life #A21422) secondary antibody. Images were acquired on a Carl Zeiss LSM 700 confocal microscope with Zen 2011 software and processed using ImageJ software (NIH).

The pΔODOLO-NFAT/luc vector contains three copies of the distal NFAT binding site in the IL2 gene promoter, upstream of a Drosophila Adh promoter that drives luciferase expression. The pΔODOLO-Luc vector is the identical vector lacking the NFAT-binding sites, providing a control for the basal activity of the Adh promoter. 293T cells were transfected with NFATC3 expression vectors and luciferase vectors using Lipofectamine 2000 Reagent (Life Technologies). The total amount of DNA transfected per experiment was kept constant through the addition of empty pMIG vector, where appropriate. Luciferase activity was measured in triplicate 48 hours after transfection, using the Dual Luciferase Reporter Assay System (Promega). Values for pΔODOLO-NFAT/luc activity were corrected for both measured pΔODOLO-Luc activity and NFATC3 protein expression, quantified using the BioRad ChemiDoc XRS+ machine with Image Lab software.

Gene set enrichment analysis (GSEA) was performed using a set of genes described to be regulated by calcineurin/NFAT in normal peripheral lymphocytes (http://www.ncbi.nlm.nih.gov/biosystems/137993 and Supplementary Table S2). GSEA was run using signal-to-noise for the ranking gene metric and 1,000 permutations. The analysis was performed using RNA-seq data from 20 T-ALL samples, which were defined as being NFATC3-PLA2G15–high or low according to the results of NFATC3-PLA2G15 qRT-PCR (see above). The 10 NFATC3-PLA2G15–high cases all had expression levels in the highest quartile of results for T-ALL samples, whereas the 10 NFATC3-PLA2G15–low cases all had expression levels in the lowest two quartiles.

NSG mice were maintained under specific pathogen-free conditions in the animal facilities of the Institut Curie (Orsay, France). All experimental procedures were performed in strict accordance with the recommendations of the European Commission (Directive 2010/63/UE) and French National Committee (87/848; authorization APAFiS #7393-20161028104744-v1). Following injection of patient T-ALL blasts obtained at leukemia diagnosis, mice were followed for tumor engraftment by regular flow cytometry analysis of peripheral blood using hCD45 and hCD7 (eBioscience) as markers for human leukemic cells. Mice were euthanized when terminally ill, as evidenced by either severe dyspnea or weakness caused by leukemic dissemination in the thymus or vital organs (bone marrow, lung, and liver), respectively.

Statistical analyses and survival curves for patient-derived murine xenografts and for human T-ALL patients treated during the GRAALL-2003 and -2005 studies were calculated using Prism 5 (GraphPad). Kaplan–Meier survival curves were compared using the log-rank (Mantel–Cox) test. The GRAALL-2003 and GRAALL-2005 studies were registered at http://www.clinicaltrials.gov as #NCT00222027 and #NCT00327678, respectively. Informed consent was obtained from all patients at trial entry. Both studies were conducted in accordance with the Declaration of Helsinki and approved by local and multicenter research ethical committees. The sole criteria for inclusion in the current project were a diagnosis of T-ALL and the availability of diagnostic material for measurement of NFATC3-PLA2G15 expression.

Our RNA-seq analysis pipeline is depicted in Fig. 1A, and notably excluded fusions that were also detected in normal thymic RNA sequenced in parallel. Strikingly, we found that 55 of the 140 total candidate fusions involved genes located within 30 kb of each other, in the same transcriptional orientation. This distance is consistent with that previously observed for cis-SAGe (9), suggesting that ISCs are a common event in T-ALL. In total, putative ISCs were detected in 10 of 12 samples, with a median of 4 (range, 0–15) per patient. Full details of the candidate ISCs detected in this study are shown in Supplementary Table S3.

We noted that several of the ISC gene partners have important roles in normal and leukemic T-cell development. We decided to perform further analysis on the candidate ISC NFATC3-PLA2G15, which was detected in 2 of 12 T-ALLs (Fig. 1B). Nuclear factor of activated T cell (NFAT) proteins are critical regulators of normal thymopoiesis and mature T-cell function (10), and murine Nfatc3 has specific roles in Th cell differentiation from naïve to effector states (11), and in double positive (CD4⁺CD8⁺) to single positive (CD4⁺/CD8⁻ or CD4⁻/CD8⁺) thymocyte transition (12). We have also previously shown that the calcineurin/NFAT pathway activation is essential for T-ALL leukemia-initiating cell function (13). PLA2G15 (Phospholipase A2 Group XV) is located 16 kb downstream of NFATC3 and encodes a lysosomal enzyme with both phospholipase and transacylase activities (14–16). We initially confirmed the presence of NFATC3-PLA2G15 mRNA in leukemic cells by RT-PCR and direct sequencing (Fig. 1C). We additionally verified that the same NFATC3-PLA2G15 fusion transcript was detectable in an independent RNA-seq series of T-ALL samples that were analyzed using a different system (Supplementary Fig. S1). As expression in patient samples appeared variable, we designed a fusion-specific qRT-PCR, to quantify the levels of NFATC3-PLA2G15 transcription more precisely. We found that primary T-ALLs exhibited a wide range of NFATC3-PLA2G15 expression, with low levels being found in the majority. Of note, NFATC3-PLA2G15 levels in normal tissue samples were consistently very low or undetectable (Fig. 1D). The results of 5′ RACE PCR of leukemic cDNA were consistent with initiation of fusion transcription in the first exon of NFATC3 (Supplementary Fig. S2). We also performed array CGH of 115 diagnostic T-ALL samples and found no evidence of microdeletions that would result in NFATC3-PLA2G15 transcript expression (Supplementary Fig. S3), providing strong evidence that NFATC3-PLA2G15 is a true ISC that is generated by cis-SAGe.

We next tested the activity of the NFATC3-PLA2G15 fusion. Immunofluorescent staining showed that NFATC3-PLA2G15 could localize to the nucleus, suggesting that the fusion might have the capacity to affect NFAT target gene transcription (Fig. 2A). The results of luciferase assays showed that the fusion had NFAT reporter activity in vitro, but that this was considerably lower than the WT protein (Fig. 2B; Supplementary Fig. S4). We then performed GSEA of expression data from primary T-ALL samples, using a set of genes that are regulated by calcineurin/NFAT in normal human lymphocytes (Supplementary Table S2). This analysis revealed that T-ALL cases with high NFATC3-PLA2G15 levels had generally lower expression of canonical NFAT target genes than NFATC3-PLA2G15–low cases (Fig. 2C).

Finally, we tested whether NFATC3-PLA2G15 transcription correlated with T-ALL biology in vivo. Strikingly, we found that higher NFATC3-PLA2G15 levels strongly predicted shorter time to leukemia development (Fig. 3A) and survival (Fig. 3B) in patient-derived T-ALL xenografts in immunodeficient mice (see Supplementary Fig. S5 for fusion transcript expression in individual samples used for xenografts). To estimate the clinical relevance of NFATC3-PLA2G15 expression, we analyzed the outcome of human T-ALL patients treated as part of the Francophone multinational GRAALL-2003 and -2005 studies. In line with the murine xenograft results, we found that patients with the highest quartile of NFATC3-PLA2G15 expression (Fig. 1D) had reduced overall and event-free survival compared with the rest of the patient cohort (Fig. 3C and D). As shown in Supplementary Table S4, NFATC3-PLA2G15–high patients did not differ from the NFATC3-PLA2G15–low group with regard to classical risk factors such as age, NOTCH1/FBXW7 mutations, and initial treatment response. The prognostic effect of NFATC3-PLA2G15 expression was however outweighed by our recently reported mutational classifier (see Supplementary Table S4), and the potential influence of NFATC3-PLA2G15 on patient outcome requires further examination in independent studies, ideally complemented by more extensive evaluation of fusion transcription by RNA-seq.

Further work is necessary to determine the mechanism by which the fusion may alter T-ALL biology. We found that, unlike constitutively nuclear mutants of NFAT (17), ectopic expression of NFATC3-PLA2G15 was insufficient to induce transformation of NIH 3T3 fibroblasts in vitro (data not shown). This finding is not unusual for confirmed T-ALL oncogenes. For example, most NOTCH1 gain-of-function mutants are not sufficient to induce leukemia in murine models (18). It is therefore likely that the oncogenic effects of the fusion might require additional cooperative events. It remains to be seen whether the altered NFAT activity detected in vitro contributes to reduced canonical NFAT target transcription in patient samples in vivo. The latter finding should be interpreted with caution, as we have previously found that the transcriptomic signature associated with calcineurin activity in murine T-ALL models is distinct from that seen in normal human T cells (13).

Interestingly, NFATC3-PLA2G15 fusions have also been described in an isolated case of acute myeloid leukemia (19) and in colorectal cancer, where experimental inhibition of the fusion transcript was reported to cause decreased proliferation and invasion of a cell line in vitro (20). Unfortunately, none of the T-ALL cell lines we tested had significant NFATC3-PLA2G15 expression (data not shown), so we were unable to investigate similar effects in a T-lymphoid context. Although the involved exons reported in the above cases differed from the T-ALL–associated chimera reported here, these data suggest that NFATC3-PLA2G15 ISCs might have important activities in other malignancies.

Finally, our discovery of frequent ISC expression in this series identifies a novel oncogenic mechanism in T-ALL and provides a rationale for further evaluation of this phenomenon in acute leukemia.

No potential conflicts of interest were disclosed.

Conception and design: J. Bond, J. Ghysdael, S. Spicuglia, V. Asnafi

Development of methodology: J. Bond, S. Spicuglia, V. Asnafi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Bond, C.T. Quang, G. Hypolite, J. Ghysdael, E. Macintyre, S. Spicuglia

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Bond, C.T. Quang, G. Hypolite, M. Belhocine, A. Bergon, J. Ghysdael, S. Spicuglia, V. Asnafi

Writing, review, and/or revision of the manuscript: J. Bond, C.T. Quang, M. Belhocine, J. Ghysdael, E. Macintyre, N. Boissel, S. Spicuglia, V. Asnafi

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G. Hypolite, G. Cordonnier, N. Boissel, V. Asnafi

Study supervision: J. Bond, V. Asnafi

The authors thank the IBiSA “Transcriptomics and Genomics Marseille-Luminy (TGML)” platform for sequencing of RNA samples. J. Bond was supported by a Kay Kendall Leukaemia Fund Intermediate Research Fellowship. The Necker laboratory (J. Bond, G. Hypolite, G. Cordonnier, E. Macintyre, and V. Asnafi) is supported by the Association Laurette Fugain and the INCa CARAMELE and 2015-PLBIO-06 Translational Research and PhD programs. J Bond was supported by the Kay Kendall Leukaemia Fund (KKL-699).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.