Exploring metabolic vulnerabilities for personalized therapy in AML

Genetic abnormalities in hematopoietic stem or progenitor cells have been shown to contribute to the development of acute myeloid leukemia (AML), leading to uncontrolled growth of leukemic cells and suppression of normal hematopoiesis. Although standard induction therapy with chemotherapy and targeted therapies leads to high-rates of complete remissions in patients with AML, those remissions are often short-lived and prognosis, especially for older patients, remains poor.

The development of therapeutic interventions for AML which lead to sustained tumor control are needed. To this end, Lucille Stuani, from the Université de Toulouse 3 Paul Sabatier, Toulouse, FR, and colleagues analyzed potential metabolic targets which may exploit vulnerabilities of AML cells and thereby prevent survival and proliferation.

Major metabolic dysregulations in AML

Most cancer cells, regardless of cancer type, have altered metabolisms. One example is the increase in the uptake of glucose required to satisfy demands of the cell and a shift to aerobic glycolysis in cancer cells (also known as the Warburg effect).

Glucose metabolism

Substantial efforts have been made in the last few decades to investigate the role of glycolysis and oxidative phosphorylation in both healthy cells and cancer cells. Researchers have tried to understand the use of glucose in the cancer cells, and whether the inhibition of metabolic pathways for glucose could be exploited for the development of better therapies.

One in vivo study¹ found that deletions of pyruvate kinase isozymes M2 (PKM2) and lactate dehydrogenase A (LDHA), two glycolytic enzymes, preserve the functioning of healthy hematopoietic stem cells, but inhibit the initiation of leukemia. Anti-leukemic activity was also observed in vivo and in vitro through the inhibition of 6-phosphogluconate dehydrogenase (6PGD) and G6PD which could be further enhanced with cytarabine.² Also, sensitivity to inhibition of G6PD was found to be driven by mTORC1 activity, and inhibiting mTORC1 induces a change to oxidative metabolism. Consequently, the anti-leukemic effect of mTOR inhibitors (such as everolimus) can be further enhanced when combined with anti-glycolytic agents, demonstrating the strong link between mTOR activity and AML cell metabolism.

Amino acid metabolism

Studies have highlighted a major role for amino acids in the biology of leukemia.⁵ Glutamine, a non-essential amino acid, is a substrate in the tricarboxylic acid (TCA) cycle. As leukemic cells can become dependant on the action of glutamine for tumor growth⁵ and interfering with glutamine metabolism can abrogate tumor development in vivo, as demonstrated with the knockdown of the molecule that transports glutamine, SLC1A5. Also, the anti-tumor effect of L-asparaginase has been attributed to its activity on the enzyme glutaminase.

Lipid and sterol metabolism

The lipid biosynthesis molecular pathway is often reprogrammed in leukemic cells to increase the biomass of the cell, and the clinical benefit of targeting this pathway has been supported by numerous studies.⁴ The induction of apoptosis in AML cells has been observed through the inhibition of key lipogenic enzymes, such as stearoyl CoA desaturase 1 (SCD1) and fatty acid synthase (FASN). Furthermore, BaP, a combination of bezafilbrate and medroxyprogesterone acetate, was observed to inhibit SCD1, and redirected pyruvate utilization to the activation of pyruvate carboxylase (PC),  leading to AML growth arrest and differentiation. Another target has been the mevalonate pathway which can be inhibited by blocking 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) using statins. HMG-CoA is not only involved in cholesterol synthesis but also in the generation of ubiquinone which helps with electron transfer during oxidative phosphorylation.

Oxidative phosphorylation and mitochondrial metabolism

AML cells have a higher mitochondrial mass, along with an increased oxygen consumption rate in comparison to healthy progenitor cells. Studies have reported the amplification of mitochondrial DNA in AML cells,⁶ which correlates with higher cytoplasmic nucleoside kinase expression. However, the abundance of mitochondria in AML cells does not translate into enhanced respiration which hints to an impaired capability of AML cells to cope with oxidative stress.

Another target could be the mitochondrial enzyme dihydro-orotate dehydrogenase (DHODH), which mediates the conversion of dihydroorotate (DHO) to orotate, thereby provide electrons via ubiquinone to the process of oxidative phosphorylation in AML cells

Another relevant target is fatty acid oxidation (FAO), as increased FAO and carnitine palmitoyltransferase 1 (CPT1a) expression was correlated with a worse outcome in normal karyotype AML patients. The enzyme PHD3, prolyl-hydroxylase 3, is involved in FAO, where it activates acetyl-CoA carboxylase 2 (ACC2), and thereby inhibits FAO.

Targeting metabolic vulnerabilities in AML

Cancer cell metabolism can offer promising targets for therapeutic approaches. Numerous pathways that could potentially help in the development of novel therapeutic approaches have been outlined below.

Aerobic glycolysis

An anti-proliferative effect can be observed in AML cell lines when treated with 2-deoxyglucose (2-DG) to inhibit glycolysis and glycosylation of oncogenic proteins. However, clinical trials have not shown many successes. Identification of the glucose storage pathway or other glycolytic sources, such the biosynthetic enzymes GYS1/1 and GBE1, could lead to the observation of biomarker that could aid prognosis of AML.

Glutaminolysis inhibition and amino acid depletion

Inhibiting glutaminase using CB-839 has reduced mitochondrial activity in AML cells, suggesting that glutamine exerts control on mitochondrial oxidative metabolism in AML. Clinical trials are currently in progress to assess the benefit of CB-839 in patients.⁷

L-asparaginase has been observed to lead to the inhibition of mTORC1 by transforming glutamine into glutamate, and has exhibited anti-leukemic activity.³ Although it is currently one of the standard drugs for treating patients with acute lymphoblastic leukemia (ALL), its use in patient with AML was of limited success due to reduced sensitivity of AML cells to L-asparaginase potentially due to changes in the AML tumor microenvironment.

Using a mycoplasma-derived enzyme of arginine deiminase bound to polyethylene glycol (ADI-PEG20) leads to degradation of arginine and has shown to reduce tumor burden in AML in synergy with cytarabine. ADI-PEG20 as a selective inhibitor of AML cell growth is currently under investigation in a phase II study

Inhibition of the mevalonate pathway

The anti-leukemic effect of statins such as HMG-CoA inhibitors, were found to be additive with conventional chemotherapies in primary AML samples. Phase I and II trials⁸ combining these drugs have observed an encouraging response rate of 75% in patients with relapsing AML. Interestingly, these results could not be repeated in newly diagnosed patients with AML hinting for fundamental changes in metabolism between untreated and pre-treated AML cells.

OxPHOS, BCL2 and mitochondrial dependencies

As AML cells are increasingly dependant on mitochondria, they become vulnerable to mitochondria metabolism. Inhibiting mitochondrial translation and mitochondrial protease ClpP have both been studied as methods of disabling mitochondrial function.⁶ Other approaches to inhibit the electron transport chain (ETC) complex I have explored metformin (used for treatment of type 2 diabetes) and 2’,3’-dideoxycytidine used for treating AIDS,⁶ and despite not being investigated yet, could be used in combination with current AML drugs.

The inhibition of BCL2 with ABT-199 (venetoclax) leads to the impairment of mitochondrial respiration and selectively targets reactive oxygen species (ROS)-low leukemia stem cells (LSCs) that are unable to switch to glycolysis to maintain energy production.

Metabolic stratification to decipher specific vulnerabilities and develop more efficient therapies in patient genetic subgroups

Isocitrate dehydrogenase mutations (IDH)

Around 20% of all patients with AML have IDH1 (cytosolic) or IDH2 (mitochondrial),⁹ reinforcing that metabolic investigations should be conducted for the condition. The metabolic pathways of IDH mutations are involved in TCA cycle, oxidative phosphorylation and lipid biosynthesis, among others. Therefore, combining IDH inhibitors with other drugs interfering with mitochondrial activity could show great promise. 

FMS-like tyrosine kinase 3 (FLT3) mutations

FLT3 mutations, including FLT3-internal tandem duplication (ITD), can be observed in 30% of all patients with AML, and usually display an enhanced rate of relapse, and poor prognosis. New combinations of drugs are needed for this subgroup of patients with AML as, despite good responses to initial responses to monotherapy being promising,¹⁰ enhanced disease-free survival was not observed. Interestingly, FLT3-ITD seems to be associated with higher glycolytic activity favoring ATP transfer from OxPHOS which provides mitochondrial protection against mitochondrial death pathways. Therefore, a combination of glycolytic inhibitors with FLT3-ITD inhibitors merits further investigation.

Other mutational and cytogenetic subgroups

For patients with AML in other subgroups such as those with p53, RAS or CEBPa mutations, or for those with other karyotypes, metabolic dysregulation and the biology behind it is largely unknown. One study identified that the creatinine kinase pathway in the EVI1 subgroups of patients with AML represses the myeloid differentiation regulator RUNX1, promoting the expression of creatine kinase mitochondrial 1 (CKM1) . Inactivating CKM1 abrogates ATP production, and therefore mitochondrial function, and prolonged the survival of mice engrafted with EVI1-espressing cells. Thus, the numerous metabolic pathways in the various subgroups in AML could be investigated, and may lead to the development of novel therapies to achieve good clinical outcomes.

Current limitations in cancer metabolism studies and metabolism-based therapeutic strategies

Increasing concerns on cancer metabolism have arisen over the last decade. The main focuses are the reproducibility of published data, differences in efficacy in vitro and in vivo studies, and high attrition rates for cancer drugs.

Many studies have tried to develop cell culture medium that have nutrient levels close to those found in human serum. One study found that the growth of cancer cells in HPLM containing human plasma levels of uric acid led to the inhibition of de novo pyrimidine synthesis. Despite large-scale RNAi and CRISPR being powerful tools used to identify metabolic genes essential for the proliferation of AML cells, the medium they are grown in can hugely affect the reproducibility of the results.

Conclusion

Myeloid leukemias have been described as metabolic disorders, and should be regarded as such for metabolic-based, personalized treatments. However, AML cells often display complex metabolic capabilities that may limit the efficacy of drugs, and induce metabolic adaptations within the leukemic cells and the tumor environment which leads to resistance.

However, as most of the metabolic pathways described in AML cells also occur in healthy cells, deciding on the best therapy options can be difficult. Further in situ research into specific metabolic pathways in the subgroups in AML is required to decide on when to utilize or eliminate certain metabolites through therapy.