An amino acid trade-off between Acute Lymphoblastic Leukemia blasts and bone marrow mesenchymal stromal cells: a possible mechanism for asparaginase resistance

Abstract

Acute Lymphoblastic Leukemia (ALL) is due to the neoplastic transformation of lymphoid progenitor cells, which start to proliferate in uncontrolled way, to accumulate in bone marrow (BM) and blood, and to invade other organs such as lymph nodes, spleen, liver and brain. The immature neoplastic cells are called blasts. ALL is the most common pediatric cancer and represents 25% of all oncologic diagnoses and 80% of all leukemia cases in children. A peculiar metabolic feature of ALL blasts is the low expression of Asparagine Synthetase (ASNS), the enzyme that synthetizes asparagine (Asn) from aspartate (Asp) and glutamine (Gln). Therefore, ALL blasts are strictly dependent on extracellular Asn. For this reason, L-Asparaginase (ASNase) is used in ALL therapy. providing the first (and, until now, the sole) example of anti-cancer treatment based on a tumor-specific metabolic vulnerability. ASNase hydrolyzes plasma Asn and, with a lower affinity, Gln, thus causing a severe nutritional stress in ALL blasts and, eventually, their death. The ASNases used in clinical medicine derive from Escherichia coli and Erwinia chrysanthemi. Both produce a severe and long lasting Asn depletion, while the Erwinia enzyme has a higher glutaminolytic activity than the E. coli counterpart. E. coli ASNase and its pegylated form are used in the induction and re-induction phases of ALL therapy, whereas the other is adopted in case of hypersensitivity due to immune reaction. Although ASNase exploitation led to a substantial improvement of ALL prognosis, leukemic cells may become resistant to the enzyme in a minority of cases. ASNS induction has been invoked as the adaptive mechanism underlying resistance to ASNase, but the expression levels of the enzyme do not correlate with sensitivity to ASNase in vivo. For this reason, other mechanisms must support ALL blast survival during ASNase treatment. Among these, it has been proposed that tumor microenvironment could play an important role in this drug resistance. Several years ago, it has been demonstrated that a line of immortalized human BM-derived Mesenchymal Stromal Cells (MSCs) protects leukemic blasts from ASNase through Asn synthesis and secretion. More recently, it has been also claimed that adipocytes, another component of the BM stroma derived from MSCs, reduce murine ALL blast sensitivity to ASNase through the release of Gln. These data suggest that MSCs could have an important role in the nutritional support of ALL blasts during the treatment with ASNase. Therefore, the aim of this study is the elucidation of the mechanisms underlying the protection of ALL blasts by MSCs during ASNase treatment. Two original approaches have been adopted: a) evaluating the response of MSCs to ASNase, and b) verifying if MSCs obtained from ALL patients or healthy donors equally support ASNase-treated ALL cells. Primary MSC, obtained from either ALL patients (ALL-MSCs) or healthy donors (HD-MSCs) were cultured in standard growth medium. In the experiments aimed to evaluate MSC-dependent protection, primary MSCs were cultured in the presence of the BCP-ALL cell line RS4;11 or primary blasts. Gln and Asn depletion was obtained incubating cells with ASNase from Erwinia chrysanthemi or in Gln- Asn-free medium. Both ALL-MSCs and HD-MSCs were poorly sensitive to treatment with ASNase, which caused a drastic fall of intracellular Gln and Asn. Initially, MSCs ceased to proliferate, but rapidly activated an amino acid stress response, indicated by the phosphorylation of the α subunit of the human initiation factor 2 (eIF2α), and induced ASNS and the pro-apoptotic factor CHOP. However, after a few days, they markedly increased the expression of Glutamine Synthetase (GS), which is the enzyme able to synthetize Gln from glutamate and ammonium, stabilized intracellular Gln levels and rescued proliferation. Under these conditions, ASNS protein and Asn levels did not show consistent changes, whereas CHOP expression was lowered at control levels. GS inhibition, mediated by the irreversible inhibitor Methionine-L-sulfoximine (MSO), or GS silencing completely prevented the adaptation of ASNase-treated MSCs, pointing to the essential role of GS expression and/or activity. Erwinia ASNase caused a massive cell death in ALL cell cultures. However, when ALL-MSCs were co-cultured with ASNase-treated leukemic cells, the percentage of necrotic leukemia cells was reduced by more than 60%. The presence of MSO in the co-culture markedly hindered, but did not abolish the protective effect. Conversely, GS silencing in MSCs only marginally inhibited the protection on leukemia cells, demonstrating that the MSO inhibitory effect on protection was only partially due to the inhibition of GS activity in the MSCs and suggesting that most of the effect could derive from the suppression of GS activity in ALL cells themselves. These results are consistent with a working model in which ALL cells provide MSCs with Gln that is then used by the stromal cells to synthetize Asn to be exported. A support to this hypothesis came from the demonstration that MSCs are able to extrude Asn not only in Asn- and Gln-free cell culture medium but also in an advanced plasma-like medium either in the presence of physiological levels of Asn or in the absence of the amino acid, a condition that strictly mimics the BM situation during ASNase treatment. This finding prompted us to identify the transport routes involved in these metabolic exchanges. To this purpose, we have compared MSCs from healthy donors (HD-MSCs) with ALL-MSCs, demonstrating that the latter protected leukemic cells from ASNase more efficiently than the former. The expression of genes potentially involved in the protective effect was compared in two panels of primary MSCs from ALL patients and healthy donors, demonstrating that HD-MSCs expressed lower levels of the SNAT5 transporter than ALL-MSCs. SNAT5 is a bidirectional transporter of the System N family, able to mediate a symport of Na+ with Gln or Asn, either inward- or outward-directed, depending on the substrate gradients. Consistently with gene expression data, ALL-MSCs exhibited faster amino acid efflux. Further experiments demonstrated that ALL blasts induced SNAT5 expression in HD-MSCs but not in ALL-MSCs. These results suggest that the different expression of SNAT5 in MSCs derived from ALL patients is induced by leukemia blasts, and causes a faster efflux of Asn, thus potentially contributing to the enhanced protective effect on ASNase-treated ALL blasts. Further experiments are needed to definitely verify this hypothesis. While performing the research, we realized that the information available on the metabolism of MSCs themselves is very scarce. For this reason, we have preliminarily characterized MSCs metabolic features in Plasmax™, which is an advanced medium containing the physiological concentrations of an extended list of metabolites found in human plasma, and at low levels of oxygen, thus mimicking, as far as possible, BM conditions in vivo. The preliminary results obtained thus far indicate that all the MSC strains tested, from leukemia patients or healthy donors, proliferated at 1%O2, exhibited an aerobic glycolytic behavior and secreted citrate in the extracellular medium. Further experiments will ascertain if this metabolic behavior is modified by co-culturing MSCs with ALL blasts. In summary, Glutamine Synthetase and the transporter SNAT5 play key, but distinct, roles in the metabolic exchanges occurring in the ALL BM niche and, therefore, could be proposed as therapeutic targets to counteract ALL resistance to ASNase. These results allow to propose a working model of an amino acid trade-off between ASNS-negative Acute Lymphoblastic Leukemia blasts and mesenchymal stromal cells, in which not only MSCs actively sustain the metabolic needs of ASNS-deficient ALL cells but also ALL cells actively shape the metabolic landscape of the BM niche a) synthetizing and extruding Gln for Asn synthesis in MSCs and b) manipulating MSCs gene expression so as to enhance their capability to extrude Asn. This mechanism would be boosted upon the Gln and Asn depletion induced by the anti-leukemic drug L-Asparaginase, through GS-dependent MSC adaptation, but it may also operate under basal conditions, given the Asn auxotrophy of ALL blasts.