The Pro-Tumorigenic Vascular Niche Sustains the T-Cell Acute Lymphoblastic Leukemia Phenotype and Fosters Resistance to Therapy

Introduction. T-cell acute lymphoblastic leukemia (T-ALL) is a genetically heterogeneous malignancy associated with a high risk of treatment failure. Efforts to improve outcomes have focused on underlying genetic defects. However, new evidence suggests that the microenvironment can foster drug resistance/relapses. Identification of factors that contribute to microenvironment-mediated chemo-refractoriness remains an important challenge. Here, we sought to construct an in vitro platform to dissect tumor-host interactions and to optimize drug treatments using Patient-Derived Tumor Xenograft models (PDTX) of high risk adult T-ALL and engineered human endothelial cells.

Methods. T-ALL PDTX were established and serially passaged in NSG mice. Engraftment was monitored by flow cytometry of peripheral blood and/or MRI. Mice were sacrificed and leukemic cells were harvested from the spleen/bone marrow. To determine the ex vivo growing conditions, we first cultured a panel of 8 "bona fide" T-ALL cell lines and 11 PDTX cells alone in complete RPMI 20% FCS supplemented with IL2, IL12, IL15 and IL7; or co-cultured with human E4-ORF1 endothelial cells (ECs) without ILs in complete RPMI 20% FCS or serum/cytokine-free media. CDK4/6, MEK, PI3K and JAK inhibitors were used at 0.1 and 1 µM alone and in combination. Cell titer glo, cell titer blue, Annexin-V and S-cell cycle analysis were used as readouts. Total RNA from cells before and after co-culture was extracted for paired-end RNA sequencing on an Illumina HiSeq2500.

Results. To study the supporting role of ECs, we first co-cultured ECs with T-ALL cell lines in vitro (serum/cytokine free co-culture) and showed that ECs could reproducibly sustain the viability of 3/8 cell lines (Loucy, KOPTK1, P12 Ichikawa) serum/cytokine-free media. A partial rescue was seen with 3 additional lines (HPB-ALL, CCRF-CEM, CUTLL1), while 2 (KE37, DND41) underwent massive cell death. We next tested whether either ILs or CXCL12 could provide anti-apoptotic signals and demonstrated that KOPTK1 and Loucy were only partially rescued by IL15 or CXCL12. Conversely, IL7, although capable of inducing a robust upregulation of pSTAT5, had no effect (CCRF-CEM and CUTLL1). We then characterized 11 PDTX from 15 high-risk adult T-ALL patients. All PDTX were serially propagated and caused T-ALL in subsequent NSG mice (massive spleen and bone marrow infiltration with extensive paravertebral mass associated with paralysis and multi-organ involvement). Genomic analysis (RNA-seq) demonstrated a high concordance between primary (pre-implant) and PDTX samples. All of them were extensively studied ex vivo, demonstratingthat T-ALL PDTX cells could only survive in ILs supplemented media, even better if enriched of growth factors and supplements for the expansion of human hematopoietic cells. However, when PDTX cells were treated with targeting compounds they all underwent massive apoptosis. Conversely, individual PDTX T-ALL could be selectively rescued by ECs, allowing the construction of individual drug response profile. To extend these data, 7 PDX T-ALL samples were screened against a 430-targeted compound library in supplemented RPMI or Stem Span media. Results indicated differential cell killing and gain (NFKB, BTK) and loss (TP-53, IGF-1R) of targets.

Conclusions. These data clearly demonstrate a key role of aberrantly activated vascular niche in T-ALL cell maintenance and drug resistance. We envisage that drug screening of EC+T-ALL will lead to the identification of actionable targets in each individual patient. Our report supports the potential for future personalized curative strategies aimed at targeting both tumor cells and host tissue supporting niche elements disrupting pro-tumorigenic signals within leukemia cell niches.