Exploiting genetic and environmental vulnerabilities to target DHODH in CNS tumors

Abstract

This review explores innovative therapeutic strategies for treating central nervous system (CNS) tumors by targeting their unique metabolic dependencies. This approach marks a significant departure from traditional cytotoxic treatments, focusing instead on the metabolic vulnerabilities created by the tumor’s microenvironment and genetic profile. A key area of interest is the de novo pyrimidine synthesis pathway, which is crucial for DNA and RNA synthesis, DNA repair, and protein glycosylation. We highlight the potential of dihydroorotate dehydrogenase (DHODH) inhibitors, which have shown promising anti-tumor activity in preclinical models. The blood–brain barrier, while a challenge for drug delivery, may enhance the efficacy of these inhibitors by maintaining a unique metabolic environment in the brain. Specific brain tumors, such as glioblastoma multiforme, MYC-amplified medulloblastoma, and IDH mutant gliomas, exhibit heightened sensitivity to DHODH inhibition. We suggest that the unique metabolic environment of the brain could make DHODH a more effective therapeutic target for brain tumors compared to other cancer types. Despite the speculative nature of these findings, the compelling preclinical data warrant further investigation into brain-penetrant DHODH inhibitors for CNS malignancies.

Key Points.

• EGFR signaling mediates resistance to DHODH inhibitors in glioblastoma.

• Pyrimidine salvage and Clusterin upregulation are key resistance mechanisms.

• Combination therapy targeting EGFR and DHODH overcomes resistance.

Recent advances in cancer research have unveiled a promising new approach to treating central nervous system (CNS) tumors by targeting metabolic vulnerabilities that are created by the tumor’s microenvironment and specific genetic profile.^1–4 This strategy represents a significant shift from conventional cytotoxic treatments and has the potential to be a highly targeted and effective therapeutic approach. De novo pyrimidine synthesis has emerged as one such critical metabolic dependency.

The de novo pyrimidine synthesis pathway utilizes glutamine as a nitrogen source to generate dihydroorotate (Figure 1A).⁵ Dihydroorotate is then converted to orotate by the mitochondrial redox enzyme dihydroorotate dehydrogenase (DHODH), and orotate is then converted to the precursor pyrimidine, uridine monophosphate (UMP), from which all other pyrimidines can be derived. The only other known cellular source of pyrimidines is the pyrimidine salvage pathway (Figure 1B).^6–8 The pyrimidine salvage pathway consists of a network of metabolite transporters and pyrimidine-modifying enzymes that take up and convert nucleosides and nucleotides into the pyrimidine precursors UMP and cytidine monophosphate (CMP). Uridine-cytidine kinase 2 (UCK2), which phosphorylates uridine and cytidine to generate UMP and CMP, respectively, is the principal and rate-limiting enzyme in the pyrimidine salvage pathway, and its activity is thought to be a critical determinant of the pyrimidine salvage capacity of cells.⁹ Under physiologic conditions, extracellular pyrimidines are limiting, and pyrimidine salvage activity is insufficient to meet cellular pyrimidine requirements. Under these conditions, de novo synthesis is the principal mechanism by which cells maintain pyrimidine homeostasis. Consequently, the de novo pyrimidine synthesis pathway plays a critical role in promoting the function of a wide range of cellular processes, including DNA and RNA synthesis, DNA repair, and protein glycosylation.¹⁰

Figure 1.

Mechanisms of pyrimidine acquisition in cells. (A) Schematic of the de novo pyrimidine synthesis pathway that converts glutamine into the precursor pyrimidine UMP. (B) Schematic of the pyrimidine salvage pathway that scavenges extracellular pyrimidine nucleosides and nucleotides to produce the precursor pyrimidines UMP and CMP. CAD, trifunctional carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, dihydroorotase; DHODH, dihydroorotate dehydrogenase; UMPS, uridine monophosphate synthetase; UPP, uridine phosphorylase; CDA, cytidine deaminase; UPRT, uracil phosphoribosyltransferase; UCK2, uridine-cytidine kinase; DPYD, dihydropyrimidine dehydrogenase.

By virtue of the fact that tumor cells are highly transcriptionally active and highly proliferative, they require particularly high levels of pyrimidines. This suggests that a therapeutic window exists in which inhibiting de novo pyrimidine synthesis can target tumor cells without significantly affecting surrounding normal cells. Efforts to target de novo pyrimidine synthesis have focused on inhibiting DHODH, the rate-limiting step in the de novo pathway, and a number of highly specific DHODH inhibitors have been developed that can rapidly suppress UMP levels in cells.^11–13 These inhibitors have shown potent anti-tumor activity in preclinical models of numerous tumor types, including acute myeloid leukemia and B-cell lymphoma,^13–15IDH mutant glioma,¹ diffuse midline glioma,²MYC-amplified medulloblastoma,³ and glioblastoma multiforme.⁴ However, several early-phase clinical trials that tested the anti-tumor activity of DHODH inhibitors in patients with hematologic malignancies failed to show significant clinical efficacy. While the reasons for this remain to be fully elucidated, emerging evidence suggests that DHODH inhibition may also exert immunomodulatory effects, potentially influencing the tumor microenvironment and anti-tumor immune responses. Further investigation of these immunomodulatory effects, particularly in the context of CNS tumors, is warranted. Furthermore, it is hypothesized that inhibition of DHODH results in a significant increase in circulating levels of pyrimidine salvage substrates, which allows tumor cells to bypass their requirement for DHODH. Indeed, high-dose uridine supplementation has been shown to completely rescue the cytotoxic and anti-proliferative effects of DHODH inhibitors in vitro.^1,2,4,13,15 This begs the question of whether certain DHODH-dependent cancers, particularly those within the CNS, might be less capable of utilizing pyrimidine salvage pathways. The blood–brain barrier presents a formidable obstacle to effectively treating brain tumors due to the poor brain penetrance of many anti-cancer drugs,^16,17 may offer a unique advantage in this context. While direct clinical evidence supporting this hypothesis is currently lacking, preclinical data and theoretical arguments suggest that the blood–brain barrier, by regulating the brain’s metabolic environment, could limit the availability of key pyrimidine salvage substrates. Specifically, the restricted transport of both glutamine and uridine across the blood–brain barrier may hinder the ability of tumor cells to bypass DHODH inhibition through salvage pathways, thereby enhancing the efficacy of DHODH inhibitors in the CNS. This restricted access to salvage pathways could create a therapeutic window where DHODH inhibition effectively targets tumor cells while minimizing systemic toxicity.

There are several mechanisms that could potentially make glial brain tumors particularly sensitive to inhibition of de novo pyrimidine synthesis:

Glutamine Metabolism in the CNS

Glutamine cannot freely diffuse across the blood-brain barrier and has to be actively transported by metabolite transporters.¹⁸ Moreover, neuronal cells, while efficiently recycling neurotransmitters, require a constant supply of glutamine to replenish the neurotransmitter pool and maintain the glutamate-glutamine cycle(Figure 2A).¹⁹ This continuous demand for glutamine in the neuronal compartment, coupled with its restricted transport across the blood–brain barrier, contributes to significantly lower glutamine concentrations in brain interstitial fluid (~80 μM) when compared with blood plasma (~600 μM).^18,20 These brain-specific features of glutamine metabolism could limit glutamine availability to glial cells such that their capacity to synthesize pyrimidines is lower than that of peripheral tissues.

Figure 2.

(A) In neuronal cells, glutamine is required for the production of both the excitatory neurotransmitter gamma-aminobutyric acid (GABA) and the inhibitory neurotransmitter glutamate. Glial cells take up both GABA and glutamate and convert them back to glutamine, which can then be either pumped out of the cell to replenish glutamine stores in the extracellular space or utilized to meet the metabolic requirements of the glial cells. (B) Uridine transport across membranes is mediated by the SLC29A family of equilibrative nucleoside transporters and the SLC28A family of concentrative nucleoside transporters.

Uridine Homeostasis

Uridine levels in normal human plasma are estimated to be in the low micromolar range.²¹ In cancer patients being treated with a DHODH inhibitor, the plasma concentrations of uridine and other nucleotides and nucleosides are likely significantly higher due the release of intracellular metabolites by dying tumor cells. It is generally assumed that these high plasma levels of salvage substrates are responsible for the failure of DHODH inhibitors to show clinical activity by allowing surviving tumor cells to bypass their requirement for DHODH activity. As is the case for glutamine, uridine cannot freely diffuse across the blood-brain barrier and has to be actively transported.^22,23 Two classes of nucleoside transporters mediate uridine uptake: low-affinity equilibrative nucleoside transporters (ie, the SLC29A family of transporters), which require high levels of plasma uridine to have activity, and higher-affinity concentrative nucleoside transporters (ie, the SLC28A family of transporters), which can actively transport uridine across the blood-brain barrier at physiologic plasma uridine levels. Due to the relatively low efficiency of all of these transporters (Figure 2B), combined with the restricted transport across the blood–brain barrier, uridine levels are significantly lower in brain interstitial fluid (< 1 μM) than in plasma (2.5–10 μM).^24,25 This limited uridine availability in the brain restricts the ability of glial tumors to utilize the pyrimidine salvage pathway to bypass DHODH inhibition effectively. By making it more difficult for tumor cells to acquire sufficient pyrimidines through salvage when DHODH is inhibited, this increases their reliance on de novo synthesis and enhances their sensitivity to DHODH inhibitors.

Most tumor cells, if treated with high enough concentrations of DHODH inhibitor, will display some degree of cell cycle arrest and cytotoxicity. However, there are specific cancers, including several subtypes of brain tumors, that appear to be exquisitely sensitive to the inhibition of de novo pyrimidine synthesis.

Glioblastoma (GBM)

We recently reported that the orally bioavailable, blood–brain-barrier-penetrant DHODH inhibitor BAY2402234 significantly impedes glioblastoma stem cell (GSC) proliferation and induces GSC cell cycle arrest, DNA damage, and apoptosis.⁴ The anti-tumor effects of BAY2402234 vary based on genetic context, with MYC-amplified GSCs showing particularly high sensitivity and EGFR-amplified GSCs showing significant de novo resistance. Myc gene amplification is found in ~2% of glioblastoma cases (Putative copy-number calls on 563 cases, TCGA-GBM, determined using GISTIC 2.0). Interestingly, expression of the constitutively active EGFR variant EGFRvIII is sufficient to confer resistance to DHODH inhibitor-sensitive GSCs. This finding reinforces the long-standing observation that metabolism and signaling are intrinsically linked.²⁶ Furthermore, unpublished data from our lab suggests that combined targeting of pyrimidine synthesis and cytokine signaling may be an effective strategy against EGFR-amplified GBM (Bar et al., in preparation).

MYC-amplified Medulloblastoma

Using a genome-wide CRISPR-Cas9 screening approach combined with metabolomic profiling, Gwynne, et al identified DHODH as a specific dependency in MYC-amplified group 3 medulloblastomas (G3MB-MYC).³ They found that genetic knock-out of DHODH or chemical inhibition of DHODH with BAY2402234 similarly impairs G3MB-MYC cell proliferation and tumor sphere formation in vitro, and that BAY2402234 treatment abrogates G3MB-MYC engraftment in orthotopic mouse models. These effects are associated with alterations in the expression of specific MYC target genes involved in cell growth and proliferation, rather than a global decrease in all MYC targets. Although the biological basis for the dependence of G3MB-MYC tumors on DHODH has not been definitively established, it is notable that MYC protein stability is dependent on post-translational UDP-glycosylation, and it is hypothesized that inhibition of de novo pyrimidine synthesis impairs MYC stabilization and function.^27,28

IDH Mutant Glioma

Using a chemical synthetic lethal screening approach, Shi, et al identified drugs that target de novo pyrimidine synthesis as selective inhibitors of IDH1 mutant glioma cells.¹ They found that human oligodendroglioma (HOG) cells expressing mutant IDH1 are significantly more sensitive to BAY2402234 than isogenic HOG cells expressing wild-type IDH1, and BAY2402234 treatment of mice bearing IDH1-mutant glioma xenografts results in significantly prolonged survival. They hypothesize that the product of the mutant IDH enzyme, the oncometabolite (R)-2-hydroxyglutarate (2-HG), can induce replicative stress and inhibit DNA repair, and that the resulting replication-associated DNA damage, even in the context of a low overall proliferation rate as seen in low-grade gliomas, increases the reliance of IDH mutant tumor cells on de novo pyrimidine synthesis for DNA repair processes.

Diffuse Midline Glioma (DMG)

Using a genome-wide CRISPR-Cas9 screening approach, Pal et al identified DHODH as a specific dependency in DMG, a highly aggressive and treatment-refractory pediatric cancer.² They found that DMG cells exhibit significantly greater sensitivity to DHODH inhibition than adult glioblastoma cells and normal astrocytes. They attribute this increased sensitivity, at least in part, to the fact that DMG cells express high levels of the catabolic enzyme dihydropyrimidine dehydrogenase (DPYD). DPYD converts uracil to alanine and thereby downregulates the pyrimidine salvage capacity of the cells (Figure 1A). This observation suggests that high expression of wild-type DPYD and/or high DPYD activity, as measured by pharmacogenomic (PGx) testing, could be used as a predictive biomarker for response to DHODH inhibitors.

In conclusion, preclinical studies have identified DHODH as a critical dependency in several subtypes of brain tumors, including glioblastoma, MYC-amplified medulloblastoma, IDH mutant glioma, and diffuse midline glioma. These studies demonstrate that DHODH inhibition can effectively suppress tumor cell proliferation, induce cell death, and prolong survival in preclinical models. We suggest that the unique metabolic environment of the brain, characterized by restricted transport of glutamine and uridine across the blood–brain barrier, has the potential to enhance the efficacy of DHODH inhibitors in CNS tumors. While early clinical trials of DHODH inhibitors in hematologic malignancies have not shown significant efficacy, this may be due to the ability of peripheral tumors to bypass DHODH inhibition through robust pyrimidine salvage pathways. The limited availability of salvage substrates in the brain may create a more favorable context for DHODH inhibition in CNS tumors. Further investigation is needed to determine whether this preclinical promise translates to clinical benefit in patients with CNS malignancies. Clinical trials are warranted to evaluate the efficacy and safety of brain-penetrant DHODH inhibitors in these specific tumor subtypes.

Funding

This work has been supported by NIH R01 CA187780 (E.E.B) and R01 CA227640 (J.A.L).

Conflict of interest statement.Dr. Eli E. Bar declares no competing financial interests or other conflicts of interest related to this work. Dr. Julie A. Losman declares no competing financial interests or other conflicts of interest related to this work.