Targeting leukemia with a metabolic genotoxin

In this issue of Blood, Vekariya et al¹ find increased DNA damage in acute myeloid leukemia (AML) arising from an elevated burden of a metabolic genotoxin, formaldehyde. To protect itself, AML cells rely on DNA repair mediated by DNA polymerase θ (POLQ), which can be targeted as a novel therapeutic strategy against AML (see figure).

In 1946, Goodman et al² showed that the nitrogen mustard compound used in poison gas was an effective treatment for blood cancers. Nitrogen mustard is a potent genotoxin through alkylation of DNA, thus inhibiting DNA replication and limiting the growth of highly proliferative cells. Almost 80 years later, genotoxic chemotherapy remains the backbone of cancer treatment, proving that DNA damage is a reliable mechanism to kill cancer cells. However, the adverse effects are considerable because of the lack of specificity against normal proliferative tissue. AML therapy is no exception, with genotoxic daunorubicin and cytarabine remaining front-line therapy for many patients. To develop more cancer-specific genotoxic treatments, we must better understand how DNA repair operates in cancer cells and target cancer cells’ vulnerability to specific types of DNA damage.

Both normal and cancer cells rely on complex and diverse DNA repair pathways to protect their DNA against the constant attack by reactive chemicals found in the environment and within the body.³ Of these reactive genotoxins, the most fundamental are oxygen and water, essential molecules for life. Another class of reactive genotoxins produced in the body are aldehydes, which can attack DNA to produce a range of toxic adducts and cross-links. Mammals produce toxic levels of a particular aldehyde called formaldehyde⁴ and rely on detoxification by acetaldehyde dehydrogenase 2 (ALDH2) and alcohol dehydrogenase 5 (ADH5) enzymes. Furthermore, DNA repair mechanisms such as the Fanconi anemia (FA) pathway are necessary to prevent accumulation of toxic formaldehyde-DNA lesions.⁵ In humans with an inborn deficiency in ALDH2 and ADH5, the increased formaldehyde-DNA damage in blood cells results in bone marrow failure and increased frequency of leukemia. Given the ubiquitous nature of aldehydes in the body, important questions are whether these metabolic genotoxins and their respective DNA repair pathways are altered in cancer cells, and whether this could provide more specific anti-cancer treatments.

Vekariya et al provide proof in principle that metabolism-derived genotoxins, such as formaldehyde, can be therapeutically exploited. An important discovery is that leukemic cells harbor elevated levels of formaldehyde compared with normal blood cells. The excess formaldehyde burden in leukemic cells originates from their increased serine/1-carbon utilization. Restriction of 1-carbon metabolism decreased the formaldehyde level in leukemic cells. How do leukemic cells protect themselves from the cytotoxic formaldehyde? The authors find a key DNA repair enzyme, POLQ, is upregulated in leukemic cells to repair formaldehyde-derived DNA-protein adducts through microhomology-mediated end joining. Most strikingly, genetic and pharmacologic inhibition of POLQ resulted in cytotoxicity only in leukemic cells, while sparing normal blood stem cells. This finding was validated in xenograft models, where treatment of leukemia-engrafted mice with the POLQ inhibitor novobiocin and the FLT3 inhibitor quizartinib resulted in synergistic restriction of leukemic growth, translating to significantly prolonged survival.

This study advances the field through several discoveries. Although the 1-carbon cycle, along with other metabolic pathways, is able to produce formaldehyde,⁵ it was not clear which pathway is the most significant contributor to total cellular formaldehyde. Vekariya et al now present compelling evidence supporting 1-carbon metabolism as a meaningful source of formaldehyde when upregulated in leukemic cells. This finding could potentially apply to many other cancer types. This study also shows POLQ to be important in cellular defense against endogenous formaldehyde-DNA lesions and adds to several mammalian DNA repair pathways known to protect against aldehyde-DNA damage, including the FA pathway, homologous recombination (HR), nonhomologous end joining, nucleotide excision repair, and SprT-like N-terminal domain nucleases.⁵ Previous studies have demonstrated synthetic lethality in cancers deficient in both POLQ and HR,^6,7 thus raising the possibility that formaldehyde could be the common genotoxic threat defended by multiple DNA repair pathways, and a prevalent genotoxin in many human cancers. The selective vulnerability of AML cells to POLQ inhibition while sparing healthy blood cells is of particular interest. Several factors are likely to contribute to the AML-specific cytotoxicity. First, the elevated formaldehyde in AML cells compared with normal cells restricts the excess burden of aldehyde toxicity to leukemic cells. Second, in support of a previous study,⁸ this study shows that in AML, the formaldehyde catabolism enzyme ALDH2 is downregulated. The reduced detoxification of formaldehyde likely contributes to the elevated formaldehyde in leukemic cells. Third, upregulation of POLQ in AML compared with normal blood cells indicates a functional requirement for POLQ to limit the cytotoxicity of formaldehyde genotoxicity.

This study provides needed data to address several important future questions. If formaldehyde is an Achilles heel in AML, can we further synergize leukemic cells to formaldehyde toxicity by targeting POLQ in combination with other formaldehyde protective pathways (eg, ALDH2 and ADH5 detoxification enzymes) and the FA DNA repair pathway? In addition, there are ≈500 million people worldwide, predominantly in East Asia, who carry a natural polymorphism in the ALDH2 gene that inactivates enzymatic activity.⁹ Could AMLs arising in these individuals be more sensitive to aldehyde-directed therapy, or conversely be more susceptible to toxic adverse effects because of the loss of formaldehyde catabolism in normal cells? Precision medicine could play an important factor in these treatment decisions. Extending beyond formaldehyde, are there other metabolic genotoxins that can be therapeutically harnessed for leukemia therapy, such as increased levels of malondialdehydes and fatty aldehydes reported in AML?^8,10 To fully explore this question, we will need to deploy sensitive assays to measure these reactive chemicals in patient samples, and test inhibition of protective pathways respective to these other aldehydes. Overall, the findings by Vekariya et al bring into focus a fundamental question regarding the physiological sources of DNA damage in cancer cells, and how such insight can translate into novel therapeutic strategies.

Conflict-of-interest disclosure: The author declares no competing financial interests.