Fatty acids are essential nutrients, either obtained externally from daily meals or derived internally from de novo synthesis and breakdown of cellular triacylglycerol (TAG) and phospholipid. As important building blocks of the body, they also participate in energy metabolism and cellular signaling pathways to maintain physiological functions. However, deregulated fatty acid metabolism favoring excess lipid biosynthesis and deposition eventually predisposes the body to metabolic disorders and carcinogenesis.

Fatty acids have different turnovers in the body (Fig. 1). They can break down through a series of mitochondrial β-oxidation processes into acetyl-CoA, which then enters the tricarboxylic acid cycle to aid ATP generation. Alternatively, fatty acids can be incorporated into TAG, phospholipids or cholesterol esters. It is noteworthy that the two distinct pathways require a common initial step known as fatty acid activation by acyl-CoA synthetase (ACS) (1,2). Long-chain ASCs (ACSLs) are essential enzymes for the activation of the most abundant long-chain fatty acids (12–20 carbons) (2,3). By contrast, fatty acids with <6 carbons (e.g., acetate, propionate and butyrate) are activated by short-chain ACSs (ACSS) and fatty acids with 6–10 carbons are activated by medium-chain ACSs, while very long-chain fatty acids (>20 carbons) are activated by very long-chain ACSs (1–3).

Five isoforms of ACSLs, namely ACSL1, ACSL3, ACSL4, ACSL5 and ACSL6, are present in mammals, and they have overlapping but specific roles in the activation of long-chain fatty acids. Individual ACSL isoforms may have preferred substrates for activation, dependent on the chain length and saturation status of fatty acids (1,2). For example, ACSL3 and ACSL4 can activate polyunsaturated fatty acids (PUFA), but ACSL3 prefers oleic acid, whereas ACSL4 favors arachidonic acid (AA) and adrenic acid (1,2).

Recent pathological studies have found the abnormal expression of ACSLs in cancer tissues in comparison to neighboring non-cancerous tissues (2,3), as discussed below. Furthermore, their oncogenic roles and certain involved molecular mechanisms have been unveiled. The present review thus focuses on ACSLs, evaluating the recent evidence of the pathological functions of ACSLs in carcinogenesis and cancer development, and discussing the chemotherapeutic potential of targeting ACSLs.

ACSL-mediated fatty acid activation contributes to the two distinct pathways of anabolic lipid biosynthesis and catabolic fatty acid oxidation. The manner in which cells adapt to channel fatty acids into the distinct anabolic or catabolic pathways is currently poorly understood. Certain early hypotheses suggested that the subcellular localization of ACSLs and specified fatty acid transportation system may contribute to channeling fatty acids into different turnovers (1,2,4). Fig. 2 depicts a mitochondrion, where ACSL1, ACSL4 and ACSL5 localize and support fatty acid synthesis and β-oxidation, and a peroxisome, where ACSL1 and ACSL4 are involved in alkyl lipid biosynthesis and β-oxidation. ACSL1, ACSL3 and ACSL4 have also been found to reside in the endoplasmic reticulum (ER), facilitating glycerolipid synthesis and ω-oxidation (a minor pathway for medium-chain fatty acids in the normal physiological condition, but an alternative pathway when β-oxidation is defective). In addition, ACSL3 in lipid droplets may aid neutral lipid synthesis and lipid droplet formation (5).

Tissue-specific loss-of-function studies are particularly useful to unveil the physiological function of individual ACSL isoforms. The majority of studies have focused on ACSL1, the founding member of the ACSL family. Liver-specific knockout of ACSL1 reduced TAG synthesis and fatty acid oxidation (6). In contrast, adipose-specific (7), skeletal muscle-specific (8) or heart-specific (9) knockout of ACSL1 exhibited common phenotypes of decreased fatty acid oxidation alone. In addition, transgenic mice with all-tissue ACSL5-knockout exhibited decreased adiposity and increased energy expenditure, indicating its major role in fatty acid biosynthesis and deposition (10). Thus, the specific role of ACSL5 may be different from that of ACSL1. However, for the remaining ACSL isoforms, there are no such loss-of-function mouse models available.

Taken together, these results suggest that ACSLs are important for fatty acid metabolism. Whether individual ACSL isoforms channel fatty acids toward anabolic metabolism or catabolic metabolism may be dependent on their subcellular localization and interaction with specific fatty acid transportation systems. It is also possible that certain compositions of fatty acids could affect the expression and localization of ACSL isoforms in physiological and pathological conditions.

Accumulating findings have demonstrated that almost all ACSL members are deregulated in clinical cancer and that a number are associated with poor patient survival. The expressional changes of ASCLs in different types of cancer are summarized in Fig. 3, with the asterisk marks indicating statistically significant associations with patient prognosis. Upregulation of ACSL1 was found in multiple types of cancer, including colon (11,12), breast (13) and liver (14,15) cancer, and myeloma (16), while downregulation was found in lung squamous cell carcinoma (17). More importantly, ACSL1 overexpression was associated with a poor clinical outcome in colon cancer patients (18). Like those of ACSL1, the majority of studies of ACLS4 favor an oncogenic role. ACSL4 has been shown to be overexpressed in multiple cancer types, including colon (11,19), breast (20), liver (21–23) and prostate (24) cancer, while another study has shown its downregulation in gastric cancer (25). In terms of patient survival outcome, ACSL4 overexpression predicted poorer patient survival in stage II colon cancer [together with upregulated stearoyl-CoA desaturase (SCD)] (12) and in liver cancer (together with downregulated growth arrest and DNA damage inducible β) (21).

By contrast, ACSL5 appears to exhibit the opposite function in cancer. ACSL5 was downregulated in colon (26) and breast (27) cancer, where ACSL1 and ACSL4 were upregulated. ACSL5 was also decreased in bladder cancer (28). More importantly, lower expression of ACSL5 in breast cancer was associated with a worse prognosis (27). An exception was reported in glioma, where ACSL5 was upregulated (29). One recent study also demonstrated that fibroblast growth factor receptor 2 (FGFR2)-ACSL5 chimera RNA rendered clinical gastric cancer cells resistant to treatment with FGFR inhibitors (30).

The role of ACSL3 in cancer is comparatively complex. ACSL3 was found to be overexpressed in lung cancer (31), prostate cancer (32) and estrogen receptor-negative breast cancer (33). Furthermore, its upregulation predicted an unfavorable prognosis in prostate cancer (32). However, opposing results were also determined in studies conducted in prostate cancer and breast cancer. The expression of ACSL3 was decreased in high-grade and metastatic prostate cancer (34). Homozygous deletion of ACSL3 was found in breast cancer following chemotherapy, and was associated with increased risks of recurrence and distant metastasis (35). These contradicting results may suggest that the roles of ACSL3 vary among the different stages of cancer. Lastly, few studies have reported on ACSL6 in cancer. Only one case report found a t(5;12)(q23-31;p13)/ETV6-ACSL6 gene fusion in chronic leukemia, which rendered cancer cells resistant to treatment using a tyrosine kinase inhibitor (36).

A recent bioinformatics study (37) utilizing online databases Oncomine (https://www.oncomine.org/resource/login.html) and PrognoScan (http://www.abren.net/PrognoScan/) stated certain novel associations between ACSL expression and cancer survival outcomes. On the one hand, which was in line with previous findings (12,18,27), the study showed that the overexpression of either ACSL1 or ACSL4 was associated with a poorer prognosis in patients with colon cancer, and that ACSL5 downregulation was associated with poor survival in breast cancer patients. On the other hand, the study suggested certain novel links. High ACSL1 expression was associated with worse survival in lung cancer patients and high ACSL3 was associated with worse survival in patients with melanoma. However, there are also opposing results (37) linking higher ACSL expression with improved patient survival in multiple types of cancer, including breast, ovarian, brain and lung cancer, and acute myeloid leukemia. The dry-lab bioinformatic approach enabled fast analyses and novel associative findings, but the obtained results may require further validations using datasets from the other databases or sources.

Taken together, the results from the majority of studies supported the fact that ACSL1 and ACSL4 serve oncogenic roles in the majority of cancer types, whereas ACSL5 may suppress tumor development. However, the function of ACSL3 in cancer may be dependent on the stages of prostate and breast cancer. The aforementioned ACSL-knockout transgenic mice models, alone or cross-bred with other transgenic mice with known oncogenes [e.g., KRAS proto-oncogene, GTPase and HBx (hepatitis B virus X-interacting protein), etc.], will be useful to shed light on the cause-effect function of ACSLs in carcinogenesis and cancer development.

ACSLs serve essential roles in fatty acid activation and metabolism, and their expression is also readily controllable by intracellular fatty acids in physiological and pathological conditions. Physiologically, diet-derived fatty acids can activate several transcription factors, including peroxisome proliferator-activated receptors α and δ (PPARα and PPARδ), liver X receptors α and β (LXRα and LXRβ), cAMP response element-binding and specificity protein 1 (Sp1) (13,15,38,39). These transcription factors can then activate different ACSLs through direct binding to conserved responsive elements in the ACSL promoter regions. Pathologically, cancer cells could hijack these transcriptional mechanisms to upregulate ACSLs, as evidenced by the following three examples. Firstly, cancer cells can utilize an epigenetic approach to activate ACSL indirectly. PPARα increases ACSL1 transcription, but this action could be repressed by microRNA-9 (miR-9) in the normal liver. However, liver cancer overexpressed the long non-coding RNA (lncRNA) known as highly upregulated in liver cancer. This lncRNA could dampen miR-9 and relieve its suppression of PPARα expression and consequently turn on ACSL1 transcription (15). Secondly, oncogenes can activate ACSL transcription indirectly through modulation of the aforementioned transcription factor. In breast cancer cells, HBx was reported to act as a co-activator to sensitize Sp1-mediated transcription of ACSL1 (13). Thirdly, hepatitis B virus mutant large surface protein also increased ACSL3 expression through induction of ER stress, and this action could promote lipid biosynthesis and deposition (40). Glycogen synthase kinase 3β can also activate ACSL3 in an ER stress-dependent manner, but the exact mechanism remains unclear.

Apart from direct regulation of ACSL transcription, cancer cells can manipulate the mRNA stability of ACSLs by targeting their 3′-untranslated regions (3′UTR). Previous studies reported that miR-205 was downregulated in liver cancer (41) and lung squamous cell carcinoma (17) to augment ACSL1 mRNA stability, since miR-205 can directly bind to the 3′UTR of ACSL1 and facilitate its degradation. A recent study further showed that DNA polymorphism (rs8086, T/T genotype) in the 3′UTR of the ACSL1 gene enabled higher ACSL1 expression compared with the corresponding C/T or C/C genotype. It was suggested that the genotype difference in 3′UTR may affect the binding affinity of microRNA and the resultant ACSL1 mRNA stability. More importantly, those patients with the T/T genotype and colon cancer had a poor disease-free survival outcome (18). Note that, in liver cancer, HBx repressed miR-205 expression and consequently increased ACSL4 mRNA stability (41). It is thus rational to propose that HBx can activate ACSL1 using the same mechanism.

Alternatively, protein ubiquitination and the involved proteasome degradation pathway is another mechanism to regulate ACSL expression. Fatty acid AA treatment could enhance ACSL4 protein ubiquitination and shorten its protein half-life in HepG2 cells via a negative feedback loop (42), as AA is the preferred substrate of ACSL4. By contrast, in breast cancer, the hormone 17β-estradiol extended the half-life of the ACSL4 protein, and subsequently increased cellular uptake of AA and eicosapentaenoic acid (43). This 17β-estradiol-mediated ACSL4 upregulation was found to be essential for the invasiveness of estrogen receptor-expressing breast cancer cells.

ACSL-mediated lipid anabolism may promote cancer initiation. Silencing of ACSL3 markedly inhibited HCV secretion, suggesting that phospholipid generation by the ACSL pathway is involved in the replication and secretion of the hepatitis B virus and the hepatitis C virus (44). The latter two viruses are well-known etiologies of liver cancer. Note that cancer is characterized by 10 hallmarks, including self-sufficiency in growth signals, deregulated metabolism, tissue invasion and metastasis, and evasion of programmed cell death (45). Next, the effects of ACSLs on the relevant cancer hallmarks and the possible molecular mechanisms involved are summarized.

In light of the epidemic of overweight and obese individuals, excess fatty acid metabolism has increasingly been found to be associated with metabolic disorders and carcinogenesis. Based on the evidence examined thus far, ACSL1 and ACSL4 are overexpressed in the majority of cancer types and exhibit oncogenic activities. These ACSLs can promote unchecked cancer cell proliferation, switch on deregulated cancer metabolism, facilitate tumor invasion and metastasis, and evade programmed cell death. ACSL3 is relatively complex, with varying expression and function in the different stages of prostate and breast cancer. By contrast, ACSL5 is generally decreased in those cancer types where ACSL1 and ACSL4 are upregulated. It is thus imperative to clarify the molecular mechanisms involved in individual ACSL isoforms and design targeted therapies in a precision mode. Manipulation of ACSL activities by genetic or pharmacological approaches (e.g., triacsin C to inhibit ACSL1 and ACSL4, but spare ACSL5) is theoretically applicable, but the non-specific toxicity cannot be overlooked due to the importance of fatty acid activation in physiology. Notably, ACSL4 marks the sensitivity of ferroptosis, making it a desirable target in cancer. Future studies are necessary to deepen our understanding of ACSL-involved fatty acid metabolism and carcinogenesis.