A metabolic link between the urea cycle and cancer cell proliferation

ABSTRACT

Clinical observations in citrullinemia type I, an inborn error of metabolism, led us to explore the benefits of somatic ASS1 silencing in cancer. We found that downregulation of ASS1 results in preferential utilization of its substrate, aspartate, for pyrimidine synthesis to support cell proliferation. Reducing aspartate availability for pyrimidine synthesis restricted cancerous proliferation.

The urea cycle, an important metabolic pathway in the liver, is essential to convert excess nitrogen from ammonia and aspartate into urea. Urea cycle disorders (UCDs) are rare inborn errors of metabolism (IEMs) caused by deficiency of the enzymes or transporters involved in ureagenesis. Citrullinemia, a classic UCD, occurs in two forms: citrullinemia type I (CTLN1) caused by dysfunction of the cytosolic enzyme argininosuccinate synthase 1 (ASS1), which utilizes aspartate and citrulline as substrates to generate argininosuccinate, and citrullinemia type II (CTLN II) resulting from deficiency of the aspartate transporter, citrin (SLC25A13), which transports mitochondrial aspartate into the cytosol. The argininosuccinate generated by ASS1 is further cleaved by another urea cycle enzyme, argininosuccinate lyase (ASL), to generate arginine and fumarate (Fig. 1). In both types of citrullinemia, there are increased levels of citrulline (hence the name) and of ammonia, due to the lack of nitrogen flux through the urea cycle.^1,2 If untreated, the hyperammonemia can be life threatening or lead to devastating neurologic consequences. Thus, we were intrigued as to why ASS1, a gene required for a crucial pathway like ureagenesis, is silenced in multiple types of cancer.

Figure 1.

ASS1 levels regulate aspartate availability for pyrimidine synthesis. Schematic illustration of the metabolic flux of aspartate for the synthesis of either pyrimidines or arginine. In ASS1 deficiency (CTLN I), as well as in cancers with ASS1 deficiency, there is a potential diversion of the aspartate toward pyrimidine. ASL, argininosuccinate lyase; ASS1, argininosuccinate synthase; CAD, carbamoyl-phosphate synthase 2, aspartate transcarbamylase, and dihydroorotase; NOS, nitric oxide synthase.

It has been reported by others that many cancers exhibit somatic silencing of ASS1 in the tumor tissues, and that loss of ASS1 correlates with poor outcome.³ One consequence of ASS1 downregulation in cancer cells that has been well studied is arginine auxotrophy, i.e., the inability of cancer cells to synthesize arginine and their dependence on exogenous arginine. In fact, this led to the development of arginine catabolizing drugs as therapeutic agents. This strategy can be beneficial, but typically only for a limited time as the cancer cells can develop drug-resistance by re-expressing ASS1.³ Since the enzyme directly responsible for arginine synthesis is ASL (Fig. 1) and there are cancers in which both ASL and ASS1 are silenced, we hypothesized that ASS1 silencing in cancer offers an additional metabolic advantage independent of its role in arginine synthesis.⁴ Thus, we focused our studies on aspartate, the substrate for ASS1.

Using computational modeling, we found that ASS1 silencing is predicted to enhance the flux of aspartate toward pyrimidine synthesis^5,6 by a reaction catalyzed by the cytosolic enzyme complex carbamoyl-phosphate synthase 2, aspartate transcarbamylase, and dihydroorotase (CAD). To confirm these results, we looked toward evidence from humans with the IEMs CTLN I and CTLN II. Patients with IEM offer a unique opportunity to dissect the consequences of single gene perturbation in contrast to the multiple genetic changes that occur in cancer.⁷ Specifically, patients with citrullinemia allowed us to test our hypothesis that ASS1 deficiency in CTLN I causes increased aspartate flux toward pyrimidine synthesis; by comparing the phenotype of CTLN I to that of CTLN II, we were able to focus on the flux of mitochondria-derived aspartate. Encouragingly, we found that urinary levels of orotate, a product reflecting the synthetic activity of CAD, were significantly elevated in humans with CTLN I compared to the normative values from the control population and those with CTLN II. Corroboratively, growth restriction has been reported in humans with CTLN II² but no growth aberrancies have been reported in CTLN I. Using fibroblasts from these patients, we further showed that ASS1-deficient CTLN I cells exhibited increased utilization of aspartate for pyrimidine synthesis.⁸ Importantly, using mouse intestines we showed that ASS1 downregulation is a biological response during proliferation, even in non-cancerous states.⁸

We then assessed whether the increase in pyrimidine synthesis was the metabolic advantage conferred by silencing of ASS1 in cancer cells. Using metabolic modeling and data analysis of publicly available databases involving patients with cancer and cancer cell lines, we confirmed that ASS1 downregulation in cancer concurs with CAD upregulation and increases the growth rate. We then conducted detailed studies in two types of cancer, melanoma and osteosarcoma, in which clinical outcomes in patients have been shown to depend on levels of ASS1.^9,10 We modulated ASS1 levels in these cancer cell lines and found that ASS1 levels inversely alter the proliferation rate and pyrimidine synthesis by changing aspartate availability for pyrimidine synthesis.⁹ These results were confirmed in vivo using xenografts in an immunodeficient mouse model, where melanoma cells transduced with shASS1 demonstrated increased pyrimidine synthesis and gave rise to an increased tumor burden.

In parallel, analysis of The Cancer Genome Atlas (TCGA) database showed a significant upregulation of citrin in cancer. To test whether cancers with ASS1 deficiency depend on citrin for transport of mitochondrial aspartate, we transfected cells with ASS1 downregulation with citrin-siRNA and found that loss of citrin restricted the high proliferation rate associated with ASS1 downregulation. Thus, cells with ASS1 downregulation are addicted to mitochondria-derived aspartate to support their enhanced proliferation. Importantly, we found this to be clinically relevant as patients who had elevation of citrin together with downregulation of ASS1 in their tumors had the worst prognosis.

Finally, we showed that interventions in the flux of aspartate toward pyrimidine synthesis pathway by blocking citrin, CAD, or pyrimidine synthesis have additive benefits to arginine-depleting agents in restricting the proliferation of ASS1-deficient cancers. Hence, such combination therapy might address the resistance that develops following usage of these drugs and could provide the rationale behind these treatment modalities.

In summary, our study provides a direct metabolic link between urea cycle components and cancer as part of the metabolic rewiring that accompanies carcinogenesis and underscores the importance of understanding metabolic fluxes in IEMs as a means to study more common disorders like cancer.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

SCSN is supported by Baylor College of Medicine IDDRC Grant (1 U54 HD083092) and by the Doris Duke Charitable Foundation (DDCF 2013095). AE is incumbent of the Leah Omenn Career Development Chair and is supported by research grants from the European research program (CIG618113, ERC614204), the Israel Science Foundation (1343/13; 1952/13), and from the Minerva grant award (711730). AE received additional support from the Adelis Foundation, the Henry S. and Anne S. Reich Research Fund, the Dukler Fund for Cancer Research, the Paul Sparr Foundation, the Saul and Theresa Esman Foundation, from Joseph Piko Baruch, and from the estate of Fannie Sherr.