MYCN Is Required in Immature T-ALL and Bears Epigenetic Marks Associated with Cell Identity

T-cell acute lymphoblastic leukemia (T-ALL) accounts for 15% of pediatric and 25% of adult ALL cases and is associated with higher rate of relapse as compared to B-ALL. Although prognosis in T-ALL is excellent overall, outcomes for patients with intrinsic chemoresistance or those who relapse remains poor. Recent genome-wide studies have shed light into previously unknown oncogenes and signaling pathways; however, how different combinations of oncogenes cooperate remains unclear.

In hopes of generating more relevant experimental models, we developed an efficient and reproducible approach to generating T-ALL from primary human CD34+ cord blood (CB) cells by lentiviral transduction followed by culture first in vitro on OP9-DL1 feeders then injection into immunodeficient NSG recipient mice. This yielded aggressive disease within 12-20 weeks which was indistinguishable from spontaneously occurring human T-ALL as judged by whole transcriptomic analysis (Kusakabe et al., Nature Communications 2019). Unlike conventional mouse models of T-ALL, activated NOTCH1 alone was inefficient at transforming human cells whereas addition of LMO2, TAL1, and BMI1 (for a total of 4 delivered oncogenes, N+L+T+B) yielded high penetrance, short latency disease in vivo and showed marked growth advantage over NOTCH1-only transduced controls in vitro.

We now report whole transcriptomic analysis by RNA-seq of a collection of 75 synthetic CB leukemias generated with various combinations of N, L, T, and/or B oncogenes which yielded two major types of leukemia. One type resembled CD4- CD8- T cells, lacked TCRα/β rearrangements, and was significantly enriched for hematopoietic stem cell, inflammatory response, and MYC targets signatures, whereas the other corresponded to CD4+ CD8+ T cells, were TCRβ rearranged, and showed enrichment for Wnt/beta-catenin signalling and double positive thymocyte signatures. Clustering using a 2,824-gene list emphasizing normal T-cell development (Wong et al., Leukemia Research 2021) further characterized these two types as corresponding to earlier vs. later stages of early T-cell development. Since all synthetic leukemias carried activated NOTCH1 which drives MYC expression in T-ALL, it was not surprising that MYC transcripts were expressed consistently in all samples; however, we observed a striking elevation of MYCN expression specifically in early cluster samples as compared to late cluster samples (log2FC=3.7, padj<0.0001). Interestingly, ChIP-seq analysis of NLTB-transduced CB cells revealed enrichment for H3K4me3 marks broadly across the MYCN gene body, suggesting consistent expression across cells and thus providing a potentially critical function.

To assess for potential oncogene “addiction” to MYCN, select cell lines and primary synthetic leukemias were transduced with shRNAs or sgRNAs to knock-down (KD) or knock-out (KO) MYCN, respectively. By both approaches, bulk cell growth and/or clonogenic activity were reduced in KD/KO cells as compared to controls, with effects more pronounced in MYCN-high cells, supporting their functional dependence on MYCN. Furthermore, enforced expression of MYCN in combination with activated NOTCH1 in CD34+ CB cells resulted in increased clonogenic activity and differentiation delay in vitro but was insufficient to generate leukemia in vivo. Analysis of publicly available RNA- and ChIP-seq datasets revealed MYCN to be developmentally regulated in early T cells with highest expression in early T-cell progenitors (ETPs) and T-ALLs of less mature phenotype. Interestingly, T-ALL patients with high MYCN expression exhibited significantly shorter survival as compared to those with low MYCN (hazard ratio = 5.4, p=0.029), and was independent of designation as ETP-ALL in multivariate analyses (NCI TARGET dataset, n=265).

In summary, genetic modeling of T-ALL using primary human CB cells has revealed MYCN as a critical oncogenic effector associated with less mature phenotype tumors (e.g. ETP-ALL) and poor survival. Mechanistically, we identified broad H3K4me3 marks over the MYCN gene body suggesting that epigenetic regulation may be a key element in maintaining consistent MYCN expression across cells and defining their malignant identity. These results emphasize developmental stage-specific oncogene dependencies in T-ALL and could help motivate development of alternate therapeutic modalities.