Glutamine availability as a target for the control of Glutamine-Synthetase negative human cancers: the case of oligodendroglioma

The amino acid Glutamine (Gln) is a nutrient of fundamental importance for cell metabolism. It is involved in many metabolic pathways, such as the synthesis of non essential amino acids, nucleotides and hexosamines, the regulation of cell volume, the response to oxidative stress (through the maintenance of intracellular glutathione). Moreover, it refuels the Krebs cycle with carbon moieties (anaplerosis) and activates mTOR, possibly through the energization of leucine influx. It has been known since many years that several types of normal and cancer cells depend upon Gln availability to maintain adequate proliferative activity. Moreover, certain cancer cells require large amounts of Gln and undergo severe metabolic stress and apoptosis upon Gln restriction, a condition called “Gln addiction”. However, the molecular basis of Gln addiction and, more in general, the molecular and metabolic features that underlie the sensitivity to Gln depletion have not yet been defined. The impossibility to identify Gln-dependent tumors has hampered, until now, the possibility to exploit sensitivity to Gln depletion for therapeutic purposes. However, a drug that produces Gln depletion in plasma has an established clinical use for many years. This drug, L-asparaginase (ASNase), is one of the first-line agents for the therapy of Acute Lymphoblastic Leukemia, since it produces a marked depletion of plasma asparagine, which ALL blasts cannot synthetize. However, ASNase also produces a partial depletion of plasma Gln. This thesis aims at identifying determinants of Gln sensitivity in human cancer cells, using the sensitivity to ASNase as a device of potential translational interest. The cancer model adopted in this study is oligodendroglioma. Oligodendrogliomas are rare brain cancers, which are often characterized by lack of expression of Glutamine Synthetase (GS negativity). The hypothesis of the study is, therefore, that low GS expression may render oligodendroglioma cells particularly dependent upon external Gln. For verifying this hypothesis, the effects of the ASNase from Erwinia chrysanthemi, the most glutaminolytic variant in clinical use, were compared in two different human oligodendroglioma lines, HOG and Hs683, and in two glioblastoma cell lines, U87 and U373. Compared to glioblastoma cells, oligodendroglioma cells expressed far lower amount of the mRNA of GLUL, the gene which encodes for GS. Consistently, GS protein expression was readily detectable in U87 and in U373 cells, but not in HOG and Hs683 cells, remaining undetectable even after ASNase treatment, a condition known to increase GS abundance in other cell models. Under the same condition, GS expression was instead clearly increased in glioblastoma lines. Upon ASNase treatment, a marked depletion of cell Gln was detected in all the cell models. However, while HOG and Hs683 underwent an almost complete suppression of cell viability, glioblastoma lines were less sensitive. Moreover, the IC50 values for ASNase were lower in oligodendroglioma than in glioblastoma cells. The addition of the GS inhibitor methionine-L-sulfoximine (MSO) did not synergize ASNase effects in oligodendroglioma cells, while enhanced the effect of ASNase in U87 and U373 glioblastoma cells, confirming the absence of a functional GS in the former cell models. Moreover, HOG and Hs683 cells were more dependent than glioblastoma cells on the availability of extracellular Gln. These results consistently point to a relationship between lack of GS and dependence on extracellular Gln. Nevertheless, transfection experiments with GLUL in HOG and Hs683 cells, while produced a marked GS expression in transfected oligodendroglioma cells, failed to yield a significant protection from Gln deprivation. This result may suggest that the severe consequences of Gln depletion in oligodendroglioma cells are not the result of the sole poor GS expression. Definite conclusions would require the measurement of Gln content along with the determination of GS activity in GLUL-transfected cells. However, further support to the dependence of oligodendroglioma cells from extracellular Gln has come from studies on Gln transporters. Gln is transported into mammalian cells by several transport systems. Most of Gln influx is due to sodium-dependent transporters belonging to the SLC1 (ASCT transporters) and SLC38 (SNAT transporters) gene families. In particular, ASCT2, the product of the SLC1A5 gene, appears overexpressed in several tumors and in many models of Gln-addicted cancer cells in vitro. A preliminary characterization of Gln-transporter expression in oligodendroglioma cells indicated that these cells express members of both SLC1 and SLC38 families and, in particular, Hs683 cells showed high expression of the SNAT1 transporter. In the attempt to hinder the activity of these transporters and, hence, Gln transport, oligodendroglioma and glioblastoma cells were incubated in the presence of high concentrations of specific inhibitors. The results indicated that transport inhibition had larger inhibitory effects on the viability of oligodendroglioma cells, compared with glioblastoma cells. In particular, both HOG and Hs683 cells were very sensitive to the specific inhibitor of SNAT transporters 2-methylaminoisobutyric acid (MeAIB). These results suggest that oligodendroglioma cells depend upon this transporter for Gln fuelling. In the attempt to verify if the cytotoxic effects of Gln depletion were due to the inhibition of mTOR, the activity of the kinase was also studied in oligodendroglioma and glioblastoma cells. While ASNase caused a severe inhibition of the kinase activity in HOG cells (as well as in glioblastoma cells), mTOR activity was spared in Hs683 cells. However, both HOG and Hs683 cells exhibited a very low mTOR activity when incubated in amino acid-free saline solution, indicating that sensitivity to essential amino acids, such as leucine, is maintained in both models. Gln restitution to cells pre-incubated in amino acid free saline solution restored mTOR activity in HOG but not in Hs683 cells. Hs683 cells also exhibited enhanced sensitivity to the mTORC1 inhibitor rapamycin, thus showing a mTOR-dependent phenotype. This behavior is probably to attribute to a MTOR mutation, previously described in these cells, that constitutively increases the phosphorylation of the mTOR substrate S6K1. While these results exclude that mTOR inhibition plays a role in the cytotoxic effects of ASNase in oligodendroglioma cells, they indicate that Hs683 cells yield a model of Gln-independent mTOR and demonstrate that leucine and Gln have independent roles in the stimulation of mTOR. In summary, the results recounted in this thesis indicate that GS-negative oligodendroglioma cells are markedly dependent on extracellular Gln and that mTOR inhibition is not involved in the effect. Lack of significant GS expression may, therefore, constitute a marker of sensitivity to therapeutic approaches based on the reduced availability of the amino acid. This hypothesis awaits conclusive confirmation in vivo with models of oligodendroglioma and, possibly, other GS-negative tumors.