Free Asparagine or Die: Cancer Cells Require Proteasomal Protein Breakdown to Survive Asparagine Depletion

Cancer cells have altered metabolic fluxes compared with normal tissues, prompting efforts to develop metabolic interventions to selectively kill cancer cells. Indeed, inhibitors of nucleotide synthesis have proven efficacy in conventional chemotherapy regimens, and new metabolic inhibitors are currently being evaluated for cancer therapy in clinical trials (1). Cancer metabolism is also targeted by modifying metabolite availability by administration of purified enzymes that systemically deplete metabolites that are essential for cancer cell viability (2). One such enzyme is asparaginase, a purified bacterial enzyme that depletes systemic asparagine levels. The introduction of asparaginase treatment has helped dramatically improve cure rates for pediatric acute lymphoblastic leukemia (ALL), and is now part of the standard therapy protocol (3). The ALL-specific sensitivity appears to be a consequence of an endogenous downregulation of the de novo asparagine synthesis enzyme asparagine synthetase. In the absence of asparagine synthesis, ALL cells depend on environmental asparagine, which is eliminated by asparaginase treatment.

The therapeutic efficacy of asparaginase in ALL has prompted efforts to understand its mechanism and to expand its usage in other cancers. Unfortunately, thus far, asparaginase has had clinical success only in ALL, despite numerous clinical trials in other neoplasms. In 2019, Hinze and colleagues (4) used a CRISPR screen in asparaginase-resistant ALL cells to reveal that negative regulators of WNT signaling are required for resistance to asparaginase treatment. They showed that WNT signaling activation, through inhibition of GSK3α, slowed proteasomal turnover of proteins, a phenomenon known as WNT/STOP (ref. 5; Fig. 1). The authors concluded that rapid proteasomal turnover is important for asparaginase resistance because it liberates amino acids from proteins, thereby supporting the free intracellular asparagine pool when environmental asparagine is unavailable. Importantly, a small-molecule inhibitor of GSK3α was sufficient to confer asparaginase sensitivity to resistant ALL cells, resulting in impressive survival improvements when coadministered in mouse models. Collectively, these data indicated that WNT/STOP activation may expand the usage of asparaginase treatments beyond ALL.

In this issue of Cancer Discovery, Hinze and colleagues (6) advance this discovery by evaluating opportunities to likewise increase asparaginase sensitivity in colorectal cancer. The majority of colorectal cancers have mutations activating the WNT signaling pathway. Out of these, 10% to 15% have mutations causing inhibition of GSK3α, whereas the remainder have mutations downstream of GSK3α. Hinze and colleagues focus on three different WNT-activating lesions: a GSK3α inhibiting R-spondin 3 fusion and mutations in APC or β-catenin, which are both downstream of GSK3α. According to their previous findings, it is predicted that only the R-spondin fusion tumors have active WNT/STOP and the resulting asparaginase sensitivity. To test this, they generated KRAS/p53-mutant intestinal organoids with either of the three WNT-activating lesions. As predicted by their hypothesis, R-spondin fusion tumors were highly sensitive to asparaginase, whereas both APC and β-catenin mutants were unaffected. Notably, this relationship was upheld in vivo when these organoids were engrafted in mice: APC-mutant tumors were unresponsive to asparaginase treatment whereas R-spondin fusion tumors showed marked regression. This experiment is therefore consistent with the model from ALL cells, where GSK3α inhibition is sufficient to promote sensitivity to asparaginase treatment.

Next, Hinze and colleagues investigated whether GSK3α inhibition alone could induce asparaginase sensitivity in resistant colorectal cancer cells. Both small-molecule and shRNA-mediated inhibition of GSK3α activity was sufficient to induce asparaginase sensitivity in colorectal cancer cells and in organoids. To test this in mice, they use a combination of a GSK3α inhibitor, BRD0705, and asparaginase on an impressive array of colorectal cancer models, covering tumors derived from mouse organoids and patient-derived xenografts, localized tumor growth, and metastasis models. In every model, they observed striking responses to combination therapy. Notably, tumors were completely refractory to individual treatments, indicating that this synergistic interaction is necessary for antitumor efficacy. Collectively, these experiments provide compelling evidence that GSK3α inhibition may be the key to unlocking the therapeutic potential of asparaginase treatment for solid tumors in humans, in colorectal cancer and beyond.

Mouse experiments using asparaginase and BRD0705 indicated that the combination treatment is well tolerated with no overt toxicities. Consistent with this result, Hinze and colleagues observe that cells from normal human intestinal epithelium are unaffected by combination treatment, raising the question of how normal cells retain asparagine levels in these conditions. The authors hypothesize that autophagy may act as an alternative protein recycling pathway available to normal cells but not colorectal cancer cells. Measurement of autophagy in organoid systems showed that introduction of KRAS and TP53 mutations sensitized cells to the combination therapy and blocked asparaginase-induced autophagy. Thus, it appears that the cancer-specific inability to activate autophagy is what makes colorectal cancer cells uniquely sensitive to the combination therapy. Curiously, even in the absence GSK3α inhibition, untransformed cells were sensitized to asparaginase treatment when autophagy was impaired, suggesting that normal cells do not have access to the same proteasomal response as colorectal cancer cells. How the balance of protein turnover capabilities between autophagy and the proteasome is regulated by oncogenic mutations will be of interest for further study.

Cells have three pathways by which they can acquire free asparagine: (i) environmental uptake, (ii) de novo synthesis, and (iii) protein recycling (Fig. 1). In the case of ALL, natural impairments in synthesis render them inherently vulnerable to loss of environmental uptake. Synthesis of asparagine can also be reduced by limiting its precursor aspartate through mitochondrial inhibition (refs. 7, 8; Fig. 1), and recent work has shown that mitochondrial inhibitors can sensitize solid tumors to impairments in environmental asparagine uptake (9). Hinze and colleagues add to this matrix by showing that losing protein recycling is also sufficient to sensitize cells to the loss of environmental uptake. Collectively, these results suggest that tumors can tolerate the loss of only one asparagine acquisition pathway, but not two, and support further approaches to test combination strategies to limit asparagine for cancer therapy.

Among the asparagine acquisition pathways, protein recycling is unique because it does not result in a net gain of asparagine. However, Hinze and colleagues clearly show that the release of protein-derived asparagine is important for survival upon asparaginase challenge, but specifically why this asparagine is needed to support cell viability is yet to be shown. One possibility is that, under asparagine-limiting conditions, protein-bound asparagine must be available for quick redistribution into time-sensitive protein synthesis. Alternatively, if asparaginase-treated cells were able to take up extracellular proteins, breakdown of imported proteins may be an important source of asparagine and net contribute to the amino acid balance of the cell. Regardless, the metabolic mechanism by which protein turnover supports viability will be an important future direction for better understanding this therapeutic approach.

One exciting implication of the mechanism described by Hinze and colleagues is that the synergy observed between amino acid limitation and treatments that impair proteasome function may not necessarily be limited to a WNT/STOP and asparaginase combination. Numerous treatment modalities that intend to deprive cancer cells of specific amino acids are being evaluated at all stages of clinical development, any of which could be synergistic with treatments that activate WNT/STOP. In addition, several proteasome inhibitors are already in clinical use in the treatment of multiple myeloma. It is thus conceivable that diverse approaches to proteasome disruption may be efficacious in combination with treatments that disrupt cancer cell amino acid availability. The remarkable effects of GSK3α inhibition and asparaginase combination therapy provide motivation to investigate the many possible combinations that target amino acid availability and protein recycling pathways to generate new cancer treatments.

No potential conflicts of interest were disclosed.