Although cancer genomics is the central topic in tumour research, the correlation between tumour and metabolism needs to be evaluated in detail. One of the hallmark events during tumorigenesis is the evolution of metabolic patterns that sustain tumour proliferation and survival (1). Metabolic reprogramming is an essential feature of malignant tumours to provide enough energy and essential bioactive substances for tumour growth (2). Therefore, some scholars have proposed that tumour is not only a genetic disease but also a metabolic disease (3). This hypothesis originated from the discovery of the Warburg effect or aerobic glycolysis (4). Subsequently, various anti-metabolic drugs have been used in anti-tumour therapy.

One-carbon metabolism refers to the process in which some amino acids form a group containing a carbon atom (also called a one-carbon unit), which participates in the biosynthesis after being transferred. One-carbon unit cannot exist in the free form and is bound to the N5 and N10 positions of tetrahydrofolate (THF), which is a key coenzyme that undergoes transformations to enable the transfer and participation of one-carbon units in metabolism (5). Serine is a non-essential amino acid, which is the third major contributor to cancer cell metabolism after glucose and glutamine. The amino acid participates in the formation of nucleic acids, lipids, amino acids and coenzymes and is also required for cell proliferation (6). The one-carbon units derived from serine contribute to the synthesis of purines and thymidine, methylation reactions, and the production of NADPH involved in anti-oxidant defences in cells (5).

Here, we reviewed the role of serine-associated one-carbon metabolism in different cancers and addressed the possibility of targeting this metabolic pathway as an anti-tumour therapeutic strategy.

The intracellular one-carbon units are mainly derived from the catabolism of amino acids, such as glycine, histidine, serine, tryptophan, and methionine. These units are produced by pathways involved in serine/glycine metabolism, glycine cleavage system, and other amino acid metabolism (4). Serine is the major contributor of one-carbon units in cancer metabolism. On the one hand, cancer cells can take up exogenous serine quickly; on the other hand, cancer cells can synthesise serine (7). Consequently, cancer cells ensure optimal serine supply and enough one-carbon units to meet their requirements.

In recent years, over-activated serine/glycine metabolic pathways have been implicated in the development of cancer (8). The serine de novo synthesis pathway becomes the main source of serine in the absence of serine or during rapid cell proliferation. Cancer cells obtain most of their energy needs through aerobic glycolysis, and the anabolism of serine is an important branch of glycolysis (9). The process consists of three enzymatic reactions. First, 3-phosphoglycerate (the intermediate metabolite of glycolysis) is converted to 3-phosphohydroxypyruvate (pPYR) by 3-phosphoglycerate dehydrogenase (PHGDH) with the simultaneous conversion of oxidised nicotinamide adenine dinucleotide (NAD+) to reduced nicotinamide adenine dinucleotide (NADH) (10). Next, phosphoserine aminotransferase 1 (PSAT1) catalyses the conversion of pPYR to phosphoserine (pSER) and α-ketoglutarate. Finally, pSER is dephosphorylated by phosphoserine phosphorylase (PSPH) to form serine (11). Serine is converted to glycine by serine hydroxymethyltransferase (SHMT)1 and SHMT2 in the cytoplasm and mitochondria, respectively. SHMT transfers methyl group from serine to THF and participates in the folic acid cycle (12). SHMT2 catalyses the production of the one-carbon unit from serine in the rapidly proliferating cells, and this unit enters the cytoplasm in the form of formic acid to participate in nucleotide synthesis. In SHMT2-deficient cell lines, SHMT1 catalyses serine catabolism to fulfil the requirements of one-carbon groups for thymidine and pyrimidine synthesis (13).

Cancer cells require more serine because of their higher proliferative capacity than normal cells. Moreover, their genes associated with serine metabolism also show changes corresponding to metabolic changes (7).

PHGDH is a key enzyme required at the first step of the serine de novo synthesis pathway. Possemato (14) reported that the PHGDH gene is in a genomic region of recurrent copy number gain in breast cancer, and the PHGDH protein levels are elevated in 70% of estrogen receptor-negative breast cancers. Breast cancer cells with a high PHGDH expression have increased serine synthesis flux. Song et al. suggested that the expression of PHGDH is increased in pancreatic cancer tissues compared with the adjacent normal tissues. Increased expression of PHGDH is associated with tumour size, lymph node metastasis, and TNM stage of patients with pancreatic cancer. Therefore, PHGDH could serve as an important prognostic indicator and therapeutic target for pancreatic cancer (15).

PSAT1 is a key enzyme in serine biosynthesis, and its overexpression is associated with colon, non-small cell lung, and breast cancers (16). Liao et al. found that high PSAT1 expression is associated with a poor prognosis of nasopharyngeal cancer and is an indicator of an advanced cancer stage, indicating that PSAT1 is a potential prognostic biomarker (16). PSAT1 is an independent prognostic factor for colorectal cancer. It is overexpressed in the colorectal cancer tissues compared with the normal tissues, and its overexpression is associated with a response to chemotherapy using irinotecan, 5-fluorouracil, and leucovorin and shorter survival time (17).

PSPH is a key enzyme catalysing the last step of the serine synthesis pathway. Sun (9) found that aberrant expression of PSPH is highly correlated with mortality in patients with hepatocellular carcinoma. Li (18) described that most colorectal cancer cell lines overexpress PSPH, which supports cancer growth during 5-FU treatment. Therefore, PSPH can be targeted for increasing the anti-cancer efficacy of conventional therapy in patients with colorectal cancer.

SHMT is an important enzyme in the one-carbon unit cycle associated with serine. It also participates in the rapid proliferation of various cancers; therefore, SHMT is an actively studied target for anti-cancer drug development. Sun (19) reported that stromal SHMT1 expression was highest in anaplastic carcinomas, tumoural SHMT1 was higher in poorly differentiated carcinomas and papillary thyroid carcinomas, and stromal SHMT1 expression was associated with short disease-free survival in follicular variant papillary thyroid carcinomas. Dual SHMT1/2 knockout blocked HCT-116 colon cancer xenograft formation, suggesting that both SHMT1 and 2 may be involved in the association between serine metabolic reprogramming and lymphoma. However, the role of SHMT2 in mitochondria may be more crucial (12).

Approximately 40%–80% of patients with cancer suffer from malnutrition. Cancer cells can obtain energy to support their survival and growth even in nutrient-deficient conditions (20). Both exogenous and endogenous serine support the development of cancer (2, 7, 21, 22). Cancer cells rely on exogenous serine to maintain their high proliferation rate. Therefore, serine deprivation has anti-cancer effects in patients with different genetic backgrounds and cancer types.

p53 mutations make colon cancer cells sensitive to serine deprivation because they cannot induce the necessary G1 blockade in vivo. The cells lacking p53 cannot respond to serine depletion, leading to oxidative stress. Such cells cannot guide the residual serine to produce glutathione for nucleotide synthesis. Additionally, these mutations can result in decreased viability and severely impaired cancer cell proliferation, indicating that serine deprivation may have a potential role in the treatment of p53-deficient cancers (23).

Serine deprivation can also effectively inhibit spontaneous cancer growth driven by Apc loss (intestinal cancer) or Myc activation (lymphoma) (21). Pancreatic ductal adenocarcinoma (PDAC) mice on the serine-free diet showed a 50% reduction in cancer growth. The researchers also used an FDA-approved drug LOXO-101 to prevent axons from entering PDAC. The drug blocked the activation of the receptor protein on the surface of neurons that interacted with nerve growth factor (also known as TRK-A), thereby inhibiting cancer innervation. The use of this drug alone could not slow down the growth of PDAC cancers in mice. However, a combination of this drug with the serine-free diet reduced the growth rate of PDAC by 50% compared with the diet alone. This suggests that nerves are necessary to support the growth of PDAC cells in serine-deprived cancer regions. TRK inhibitors have been approved for the treatment of TRK fusion cancers, and approximately 40% of patients with PDAC who cannot produce serine after surgery may be treated with a combination of TRK inhibitors and a low-serine diet to reduce cancer recurrence. However, this strategy has not been tested in clinical trials (24).

Serine deprivation can also enhance the effect of cancer therapy through the mitochondrial complex I inhibitor metformin (or phenformin). However, in vivo experiments showed that neither phenformin alone nor serine deprivation could inhibit the growth of allogeneic colon adenocarcinoma in mice, but their combination produced a marked inhibitory effect (25). Naama Kanarek concluded that a serine-free diet is a promising approach to inhibit cancer progression, which should be tested in combination with other cancer treatment methods in patients with diverse genetic backgrounds. However, experiments conducted with cell lines showed that serine depletion alone was more effective than simultaneous deprivation of serine and glycine. Therefore, future research should consider serine deprivation alone, rather than the combination of serine and glycine deprivation (26). Therefore, serine (but not glycine) deprivation supports the reduction of one-carbon metabolism, which then inhibits cancer cell proliferation.

The serine synthesis pathway is catalysed by several enzymes, such as PHGDH, SHMT1/2, and MTHFD1/2, and their suppression may decrease cancer cell growth and survival. However, drugs targeting these enzymes are still in the initial stages of development, and the identification of appropriate inhibitors will be crucial for cancer treatment (27).

Cancer immunotherapy has made substantial progress in the treatment of various cancers over the past decade. A core phenomenon of malignancies is metabolic dysfunction, which reprograms metabolic homeostasis in cancers and stromal cells and affects metabolic modifications of specific proteins. These post-translational modifications, such as glycosylation and palmitoylation, alter the localisation, stability, and function of proteins. Most of these proteins are involved in the development and progression of cancers, as well as in acute or chronic inflammation. Therefore, metabolic modifications targeting immune checkpoints and inflammation offer an attractive therapeutic strategy for certain cancers (53). Palmitoylation is a process that covalently binds palmitic acid to protein residues in three different ways, namely S-palmitoylation, O-palmitoylation and N-palmitoylation. O-palmitoylation is linked to serine/threonine residues by oxygen ester bonds. In 2019, two research groups reported that S-palmitoylation maintained the stability of PD-L1 and suppressed T-cell cytotoxicity (54). However, additional research is required to explore the influence of serine and O-palmitoylation on PD-L1 in different cancers.

Ma et al. revealed that serine is an essential metabolite for the optimal expansion of T cells after antigen stimulation. Restriction of dietary serine damages the pathogen to drive the expansion of T cells in vivo, but does not affect the overall immune cell homeostasis. Mechanistically, serine provides glycine and one-carbon units for de novo nucleotide biosynthesis in proliferating T cells, and one-carbon units from formate can rescue T cells from serine depletion. The authors suggested that serine—as a critical immune metabolite —directly regulated adaptive immunity and the responsiveness and tolerance of cancer immunotherapy by controlling the proliferation of T cells (55).

Tregs are also affected by acid metabolism. Serine promotes the synthesis of glutathione and enters a carbon metabolic network, the key process of the Teff reaction. The specific deletion of the catalytic subunit of glutamate–cysteine ligase (Gclc) from mouse Tregs resulted in the loss of Treg-specific glutathione (56). Kurniawan et al. (56) found that glutathione-deficient Tregs showed increased serine metabolism and decreased FoxP3 expression. Additionally, Treg-specific Gclc-deficient mice showed an enhanced anti-cancer response. When mice were fed a serine-deficient diet to suppress serine supply, Gclc-deficient Tregs restored FoxP3 expression and immunosuppressive ability of mice (57).

Phosphatidylserine (PS) is mainly localised on the inner side of the phospholipid bilayer of the cell membrane in healthy cells. It migrates to the outer lobe during apoptosis. Wang et al. established PSout cancer models, in which cancer cells lacked the CDC50A component of PS turnover enzyme and continued to expose PS but still survived. The everted PS restricted the expression of MHC-I/II in tumour-associated macrophages, thereby inhibiting cancer antigen presentation and promoting the development of cancer. The PS receptor Tim-3 (but not Tim-4) mediates the recognition of PS in cancers. The research team also established PSin models when cancer cells lacked the PS-disrupting enzyme Xkr8 in vivo and cannot expose PS during normal apoptosis. The inversion of PS results in the accumulation of apoptotic cells and activates cGAS to produce cGAMP, thereby activating the type I interferon signalling pathway of immune cells in the cancer microenvironment. These immune cells, including tumour-associated macrophages and natural killer cells, work together to inhibit cancer growth. Simultaneously, the concentration of IL-10, a cytokine that suppresses anti-cancer immunity, was also reduced in the PSin models. Moreover, Xkr8 inactivation in combination with the anti-PD-1 regimen resulted in the complete elimination of cancers in mice. Although no small molecular drugs targeting Xkr8 are currently available, the silencing of Xkr8 in vivo with short hairpin RNA or siRNA resulted in externally blocked PS and powerful therapeutic oncostatic effects (58). Therefore, PS in cell membrane, which is synthesised from serine, may be targeted in cancer immunotherapy.

Serine-related one-carbon metabolism is important for the occurrence and development of cancers. Currently, researchers are actively exploring this metabolic pathway to develop clinically effective anti-tumour drugs, which can be used independently or in combination with immunotherapy. However, most of these drugs are in the research and development stage and have not yet entered the clinical trials. Therefore, their efficacy and safety in the real world need to be further confirmed.

JZhang wrote the ‘Introduction’, ‘Development of serine synthetic pathway’, and ‘Inhibitors for cancer therapy’ sections. JB wrote the ‘Serine deprivation for cancer therapy’ section. JZhao and HX designed the review and revised the manuscript. CG wrote the ‘Serine and one-carbon metabolism’ section. JW wrote the ‘Metabolic reprogramming of serine in cancers’ section. YC wrote the ‘Serine deprivation for cancer therapy’ section. All the authors approved the manuscript.