The impact of RNA binding proteins and the associated long non-coding RNAs in the TCA cycle on cancer pathogenesis

ABSTRACT

The tricarboxylic acid (TCA) cycle is a central route for generating cellular energy and precursors for biosynthetic pathways. Emerging evidences have shown that the aberrations of metabolic enzymes which affect the integrity of TCA cycle are implicated in various tumour pathological processes. Interestingly, several TCA enzymes exhibit the characteristics of RNA binding properties, and their long non-coding RNA (lncRNA) partners play critical regulatory roles in regulating the function of TCA cycle and tumour progression. In this review, we will discuss the functional roles of RNA binding proteins and their lncRNA partners in TCA cycle, with emphasis placed on the cancer progression. A further understanding of RNA binding proteins and their lncRNA partners in TCA cycle, as well as their molecular mechanisms in oncogenesis, will aid in developing novel layers of metabolic targets for cancer therapy in the near future.

Abbreviations: CS: citrate synthase. AH: aconitase, including ACO1, and ACO2. IDH: isocitrate dehydrogenase, including IDH1, IDH2, and IDH3. KGDHC: α-ketoglutarate dehydrogenase complex, including OGDH, DLD, and DLST. SCS: succinyl-CoA synthase, including SUCLG1, SUCLG2, and SUCLA2. SDH: succinate dehydrogenase, including SDHA, SDHB, SDHC, and SDHD. FH: fumarate hydratase. MDH: malate dehydrogenase, including MDH1 and MDH2. PC: pyruvate carboxylase. ACLY: ATP Citrate Lyase. NIT: nitrilase. GAD: glutamate decarboxylase. ABAT: 4-aminobutyrate aminotransferase. ALDH5A1: aldehyde dehydrogenase 5 family member A1. ASS: argininosuccinate synthase. ASL: adenylosuccinate synthase. DDO: D-aspartate oxidase. GOT: glutamic-oxaloacetic transaminase. GLUD: glutamate dehydrogenase. HK: hexokinase. PK: pyruvate kinase. LDH: lactate dehydrogenase. PDK: pyruvate dehydrogenase kinase. PDH: pyruvate dehydrogenase complex. PHD: prolyl hydroxylase domain protein.

KEYWORDS:

• RNA binding protein
• long non-coding RNA
• TCA cycle
• TCA cycle enzymes
• tumour progression

1. Introduction

Aberrant metabolism is a major hallmark of cancer [1–3]. In the 1920s, Otto Warburg found that the glucose consumption of tumour tissues was remarkably higher than that of normal tissues [4]. The increased glucose uptake and preferential utilization of glucose through aerobic glycolysis have been validated across diverse tumor types have been linked to an unfavourable prognosis [5–7]. These early findings not only revealed the clinical relevance of aberrant tumour metabolism but also laid the groundwork for the application of tumour detection and imaging via positron emission tomography with radiolabeled glucose analogs [8]. Recently, new insights are emerging into the involvement of TCA cycle in tumour metabolic reprogramming and cancer progression [9]. It has been reported that tumour cells are capable of dissociating glycolysis from the TCA cycle, by utilizing alternative energy fuel sources (e.g. glutamine and fatty acids) in order to fulfill their elevated metabolic demands [6,10]. Importantly, extensive research has demonstrated that a variety of oncogenes and tumour suppressors play a vital role in regulating the uptake and breakdown of energy sources during the TCA cycle [10–15]. The mutations and/or dysregulations of multiple TCA cycle enzymes, including fumarate hydratase (FH), isocitrate dehydrogenase (IDH), ketoglutarate dehydrogenase complex (KGDHC), and succinate dehydrogenase (SDH) etc., are commonly harboured in human cancers [16–19]. Taken together, these studies have presented conclusive evidence to establish the significance of the TCA cycle in cancer metabolism and tumorigenesis.

Interestingly, several TCA cycle enzymes have the RNA binding capability, and are thus classified as RNA binding proteins (RBPs) [14,20]. RBPs have multiple roles in human cancer [21]. The malleability of their RNA-binding domains, in combination with their structural flexibility, allows RBPs to govern the metabolism of a broad spectrum of transcripts. Meanwhile, they establish highly dynamic interactions with long non-coding RNAs (lncRNAs), as well as with coding RNAs and other proteins [22]. LncRNAs constitute a large part of the human functional non-coding transcribed genome [23]. It is well recognized that these large classes of RNA transcripts have been implicated in many pathological processes, especially in cancer metabolism and development [12,14,24]. Cancer development has also been causally linked to disruptions in RBP-lncRNA network activity [14,25]. For instance, disrupting the interaction between lncRNA GAS5 and malate dehydrogenase 2 (MDH2) could affect FH – MDH2–citrate synthase (CS) complex formation and breast cancer tumorigenesis [14]. Blocking lncRNA IDH1-AS1 dramatically reduced IDH1’s enzymatic activities and inhibited Hela cell proliferation [25]. Hence, unravelling the complex web of interactions between RBPs and their associated lncRNAs in the TCA cycle may enhance our comprehension of tumour biology and reveal novel targets for cancer treatment. We will depict an overview of the roles played by RBPs and their lncRNA partners in TCA cycle and tumour progression.

2. Overview of TCA cycle in cancer

The typical hallmark of cancer is the uncontrolled growth and division of tumour cells, which requires a boost in both energy generation and macromolecular synthesis. In response to heightened metabolic demands, cancerous cells often undergo aerobic glycolysis, also referred to as the Warburg effect, as a means of generating energy, building blocks, and NADPH to support their survival and rapid proliferation while maintaining redox homoeostasis [26]. Until recently, the role of the TCA cycle in cancer metabolism and tumorigenesis has been largely disregarded, probably due to the prevailing assumption that impaired mitochondrial activity is a ubiquitous hallmark of the Warburg effect and cancer progression. Based on ¹³C-labelled metabolomic analysis, recent research has challenged the notion that cancers have impaired mitochondrial activity, revealing that mitochondrial function is often fully intact in these diseases. More importantly, cancer cells are capable of decoupling glycolysis from the TCA cycle, by utilizing alternative energy sources (e.g. asparagine, glutamine and fatty acids) to satisfy their increased metabolic demands [6,10,27,28] (Figure 1). Despite exhibiting a heightened demand for glucose, cancer cells differ from normal cells in how they metabolize glucose. While most cellular glucose in normal cells enters the TCA cycle as pyruvate, cancer cells often favour lactate production over acetyl-CoA generation even under normoxic condition [29–32] (Figure 1).

Figure 1. Biochemical reactions that drive the TCA cycle. By integrating the KEGG metabolic pathways (https://www.genome.jp/kegg/pathway.html#metabolism, hsa00010, hsa00020, hsa00071, and hsa00250), the biochemical reactions in TCA cycle were summarized. The TCA cycle is comprised of 8 steps, three of which are irreversible, including CS-driven the generation of citrate from oxaloacetate and acetyl-CoA; IDH-driven the conversion of isocitrate to α-KG; and KGDHC-driven the formation of succinyl-CoA from α-KG (hsa00020). Acetyl-CoA derived from pyruvate, which is the end product of glycolysis, is the typical input for the TCA cycle (hsa00010). In addition, outside sources, including the production of acetyl-CoA from β-oxidation of fatty acids (hsa00071), the production of fumarate, succinate, and α-ketoglutarate from protein catabolism (hsa00250), and the production of oxaloacetate from pyruvate (hsa00010), can also provide the intermediates in the TCA cycle.

Hypoxia is a hallmark of tumour microenvironment present in the majority of solid tumours [33,34]. Tumour cells have developed various adaptive mechanisms under hypoxic conditions to sustain their growth and outcompete normal cells. A crucial player in the hypoxia response of tumours is hypoxia-inducible factor alpha (HIFα) [33]. Under low-oxygen conditions, HIFα proteins accumulate and migrate to the nucleus where they bind to the hypoxia response element (HRE) located in the 5’ regulatory regions of their downstream target genes, thus activating a diverse array of gene transcription events, such as the transcriptional regulation of glucose transporters (SLC2A1 and SLC2A3), glycolytic enzymes (e.g. hexokinase (HK), lactate dehydrogenase (LDH), and pyruvate kinase (PK)), and pyruvate dehydrogenase kinase (PDK) [30] (Figure 2). These enzymes, on the one hand, promote glycolysis and lactate production; on the other hand, suppress the activity of the pyruvate dehydrogenase complex (PDH), which is responsible for the conversion of pyruvate to acetyl-CoA. Hence, the activation of HIF coordinates a metabolic programme that enhances the breakdown of glucose via aerobic glycolysis, resulting in a diversion of glucose away from the TCA cycle (Figure 2). As a means of compensating for the reduced glucose supply to the TCA cycle, tumour cells activated by HIF often augment their utilization of glutamine, which primarily feeds the TCA cycle through α-ketoglutarate (α-KG), thereby facilitating reductive carboxylation [35,36].

Figure 2. Interplay among TCA cycle, glycolysis, and tumorigenic signalling. Oncogenic mutations of IDH lead to the production of 2-HG. Deficiencies of SDH and FH in cancer cells elevate succinate and fumarate, respectively. Accumulation of 2-HG, succinate, and fumarate inhibits the degradation of HIF-1α, and results in HIF-1α’s nuclear translocation, which brings about the interaction of HIF-1α and the hypoxia response element (HRE) and thus promotes a plethora of tumour-related genes. The downstream genes of HIF-1α include glycolytic enzymes (e.g. HK, PK, LDH) and PDK, which promote the catabolism of glucose through aerobic glycolysis and shift glucose away from the TCA cycle. In addition, the accumulation of fumarate also stabilizes NRF2 through either altering cellular redox state via succination of GSH or a direct succination of its negative regulator KEAP1. Also, NRF2 can translocate to the nucleus and function as a transcriptional factor through interacting with the antioxidant response element (ARE) upon the promoter of multiple tumour-related genes.

In addition to the HIF-regulated TCA cycle, mutations in the TCA cycle enzymes also affect the accumulation of HIF. For example, mutated IDH develops a neomorphic catalytic activity, enabling it to convert α-KG into the 2-hydroxyglutarate (2-HG) using NADPH as a cofactor [37]. The accumulation of 2-HG inhibits the prolyl hydroxylase domain proteins (PHDs), which play critical roles in tumour metabolic adaptation, angiogenesis, proliferation, and metastasis via negatively regulating HIF-1α [38]. However, pseudohypoxia, a state characterized by constitutive activation of HIF-1α under normoxic conditions, can manifest in certain types of cancer such as those with IDH mutations [39]. Given the structural resemblance of succinate and fumarate to α-KG, their accumulation can trigger a pseudohypoxic reaction in tumours harbouring SDH and FH mutations [40–42]. Thus, the establishment of a pseudohypoxic state driven by HIF, accompanied by tumour behavioural alterations, appears to be a common characteristic of various TCA cycle enzyme deficiencies and metabolic alterations in cancer (Figure 2). Apart from affecting HIF-driven alterations, fumarate accumulation may also contribute to tumorigenesis by supporting the stabilization of the transcription factor nuclear factor erythroid 2-like 2 (NFE2L2, or NRF2), a key regulator of the cellular antioxidant defence system. Genetic targeting of NRF2 can govern a cellular detoxification and antioxidant defence programme and impair the K-Ras-induced lung and pancreatic tumorigenesis [43,44] (Figure 2). Interestingly, inhibition of NRF2 can inactivate HIF-dependent lncRNA NLUCAT1, leading to higher reactive oxygen species (ROS), increased susceptibility to cisplatin-induced apoptosis, and lower proliferation rates of lung adenocarcinoma cells [45]. In this case, the tumour metabolic lncRNA regulator NLUCAT1 is regulated by both NRF2 and HIFα. In fact, as master transcriptional regulators of energy metabolism in cancer, the activities of NRF2 and HIFα are closely correlated with the abundance of lncRNAs [46–48]. Additionally, it has been widely demonstrated that lncRNAs also play critical roles in regulating metabolic processes, including TCA cycle [14,20,25,49–51].

3. Roles of lncRNAs in regulating TCA cycle

LncRNAs play regulatory roles at a plethora of cellular processes, most of which require the interaction with RBPs [52]. RBPs are well known for their key roles in post-transcriptional events. During the structural and functional diversity, RBPs are essential for regulating RNA modification, localization, splicing, transport, translation, degradation, and stability [21,53]. Dysfunction of RBPs can lead to perturbations in RNA metabolism, resulting in wide alterations in the cell’s transcriptome and proteome. These changes can have significant implications for cell behaviour, including growth, proliferation, invasion, and apoptosis. The majority of RBPs that have been investigated in relation to cancer are considered canonical RBPs, which possess canonical RNA-binding domains (RBDs) such as RNA-recognition motifs (RRMs) [54]. In addition to canonical RNA-binding proteins, many RBPs without canonical RBDs but endowed with RNA-binding activity and have well- defined cellular functions were also identified, including some TCA enzymes [55].

TCA cycle occurs in the mitochondrial matrix and involves a sequence of biochemical reactions that are catalysed by several enzymes, including ATP Citrate Lyase (ACLY), CS, aconitase (AH), IDH, KGDHC, succinyl-CoA synthase (SCS), SDH, FH, and MDH (Figure 1). An early example that highlighted the role of a metabolic enzyme as an RBP in post-transcriptional gene regulation was the discovery of ACO1, also referred to as iron regulatory protein 1 (IRP1). IRP1 could interact with mRNA to regulate the iron content in cells [56]. Other enzymes of the TCA cycle, such as CS, MDH2, SUCLG1, and IDH, were consistently identified in RNA – protein interactome studies [57,58]. In addition to binding with exons, introns, and untranslated regions (UTR) of mRNA and regulating its processing, RBPs can also bind to lncRNAs and form a complex or interplay with each other to function in a variety of tumour-related biological processes. Current research indicates that TCA cycle enzymes interact with and are regulated by lncRNAs in different types of cancers, leading to specific metabolic changes that are closely related to tumorigenesis (Table 1 and Figure 3). In-depth understanding of the complex interplay between TCA cycle enzymes and lncRNAs will help provide a more comprehensive picture of how tumours undergo metabolic reprogramming.

Figure 3. Mechanisms of lncRNAs in regulating TCA cycle enzymes. TINCR binds to ACLY and protects it from ubiquitin degradation to maintain total cellular acetyl-CoA levels. LINC00477 interacts with ACO1 and suppresses the conversion ability from citrate to isocitrate by ACO1. IDH1-AS1 enhances IDH1 enzymatic activity through promoting its homodimerization, leading to increased production of α-KG. GAS5 inhibits FH – MDH2–CS complex formation by increasing SIRT3-mediated MDH2 deacetylation. AC020978 binds to MDH2 and protects it from ubiquitin degradation.

Table 1. The role of TCA cycle enzymes and their associated lncRNAs in tumour progression.

Enzyme	Associated lncRNA(s)	Roles of the associated lncRNA(s)	References
ACLY	TINCR	TINCR binds with ACLY and protects it from ubiquitin degradation to maintain total cellular acetyl-CoA levels and thus promotes nasopharyngeal carcinoma progression and chemoresistance.	[20]
ACO1	LINC00477	Isoform 1 of LINC00477 interacts with ACO1 and suppresses the conversion ability from citrate to isocitrate by ACO1 and gastric cancer cell growth.	[49]
IDH1	IDH1-AS1	IDH1-AS1 interacts with IDH1 and promotes IDH1’s enzymatic activities, which inhibits Hela cell proliferation.	[25]
MDH2	GAS5 AC020978	GAS5 binds with MDH2, and modulates mitochondrial tricarboxylic acid flux by disrupting metabolic enzyme tandem association of FH, MDH and CS in breast cancer; AC020978 binds with MDH2 and facilitate non-small cell lung cancer progression by activating MDH2-directed AKT pathway.	[14,50]

3.1. ACLY

ACLY is an enzyme located in the cytosol that facilitates the conversion of citrate and coenzyme A (CoA) to oxaloacetate (OAA) and acetyl-CoA in an Mg-ATP-dependent manner. This enzyme plays an essential role in cancer metabolism by potentially restricting cytosolic citrate levels, thereby regulating glycolysis and signalling pathways to promote cancer progression [59,60]. Its overexpression was correlated with poor prognosis and differentiation status of acute myeloid leukaemia and lung adenocarcinoma [27,61]. Human ACLY is a tetramer with a molecular weight of 480 kD and is composed of 1101 amino acids. It consists of a C-terminal citryl-CoA lyase (CCL) domain and an N-terminal citryl-CoA synthetase (CCS) module that is responsible for its catalytic activity [62]. Zheng and co-workers found that the N-terminal of ACLY could interact with lncRNAs [20]. By performing RNA pull-down assay followed by mass spectrometry (MS), Zheng et al. first identified the interaction between lncRNA TINCR and ACLY. Then, followed by the ACLY RNA immunoprecipitation (RIP) assay and the immunofluorescence (IF) of ACLY combined with TINCR fluorescence in situ hybridization (FISH) staining, Zheng et al. validated the endogenous colocalization and interaction of TINCR and ACLY. Next, Zheng et al. generated HA-tagged deletion mutants of ACLY to pinpoint the specific domains of the enzyme that interact with TINCR. Finally, RNA pull-down assays indicated that the N-terminal region of ACLY, which encompasses the ATP-binding, citrate-binding, and CoA-binding domains, was essential for the binding of the lncRNA TINCR. Mechanistically, TINCR binds to ACLY and protects it from ubiquitin degradation to maintain total cellular acetyl-CoA levels, which regulates a set of tumour-related events, including regulation of the PADI1–MAPK – MMP2/9 axis, histone H3K27 acetylation, and de novo lipid biosynthesis. Silencing TINCR inhibits nasopharyngeal carcinoma progression and cisplatin resistance [20] (Figure 3). In addition to TINCR, lncRNA FLJ22763 also regulates the mRNA and protein levels of ACLY. Ectopic expression of FLJ22763 suppresses the expression of ACLY, the biological malignant behaviour of gastric cancer cells and the xenograft tumour growth [63]. However, it is unclear whether FLJ22763 can bind with ACLY, and whether the negative association between ACLY and FLJ22763 is due to their binding affinity.

3.2. ACO1

ACO1 is a protein located in the cytosol, which acts as a crucial enzyme in the TCA cycle and binds to mRNA to regulate intracellular iron concentrations [56]. Under high iron levels, ACO1 interacts with a 4Fe-4S cluster and transforms into an active cytosolic aconitase enzyme [64]. In contrast, when iron levels are insufficient, ACO1 binds to conserved stem-loop structures known as iron-responsive elements (IREs) in the UTRs of mRNAs responsible for iron metabolism. The impact of the IRE on regulation is determined by its position and context within the mRNA: ACO1-IRE located in the 5’ UTR of mRNA represses translation, while ACO1-IRE located in the 3’ end can indirectly activate translation via suppressing mRNA degradation [64]. The dual roles of ACO1 in regulating TCA cycle and iron homoeostasis are closely related to tumorigenesis [65,66]. Interestingly, in addition to binding with mRNAs, ACO1 also interacts with a lncRNA variant of LINC00477, as predicted by RNA-binding Protein Database (RBPDB) and validated by RNA pull-down and RIP assays (Figure 3). Gain-of-function and loss-of-function assays of the LINC00477 variant showed that it could suppress the functions of ACO1, including ferritin mRNA translation for cellular iron and glycol metabolism, conversion ability from citrate to isocitrate, and gastric cancer cell proliferation and migration [49].

3.3. IDH1

IDH1 is a cytosolic enzyme that converts isocitrate to α-KG in TCA cycle [67,68]. The mutations in IDH1 are prevalent in various types of cancer [69]. Mutant IDH1 reveals altered enzymatic activity leading to the synthesis of 2-HG, which acts as an oncometabolite that has pleotropic effects on tumorigenesis [18]. The enzymatic activity of IDH1 is also regulated by lncRNA IDH1-AS1. By performing RIP and pull-down assays, Xiang et al. found that endogenous IDH1-AS1 was coprecipitated with IDH1 but not the mutant form of IDH1 (R132H) in Hela cells. To identify the WT IDH1 domain that physically interacts with IDH1-AS, Xiang et al. also used a domain mapping strategy of RIP assay, and finally revealed an intermediate part of IDH1 (103–286 aa), in which a helix structure formed by aa 271–286 was identified, which could bind with IDH1-AS1. Deletion of the residues 271–286 attenuated the interaction between IDH1 and IDH1-AS1, and inhibited the enzymatic activity of IDH1. This work eventually discovered that lncRNA IDH1-AS1 could bind and regulate IDH1 enzymatic activity by promoting its homodimerization, resulting in the elevated production of α-KG, downregulated expression of HIF-1a, and a decrease in glycolysis [25] (Figure 3). Besides, IDH1-AS1 has been found to regulate IDH1 in glioma. Wang et al. showed that IDH1-AS1 could regulate glioblastoma cell viability via IDH1. Although Wang et al. demonstrated the effect of IDH1-AS on IDH1 homodimerization and enzymatic activity in U251 cells, the interaction between IDH1-AS and IDH1 has not been validated in glioblastoma cells [70].

3.4. MDH2

MDH2 is a mitochondrial enzyme essential for the conversion of malate to oxaloacetate [71]. MDH2 has the potential to be a viable target for cancer treatment, owing to its effects on drug sensitivity and ATP production [72]. For example, MDH2 is highly expressed in doxorubicin-resistant prostate cancer cells and uterine cancer cells. Overexpression of MDH2 enhances the P-glycoprotein’s ability to eliminate chemotherapeutic drugs, which can be a contributing factor to the shorter periods of relapse-free survival observed in patients following chemotherapy [73]. MDH2 may also play a role in conferring resistance to docetaxel in prostate cancer cells, possibly through the JNK pathway [74]. In recent years, the involvement of lncRNA in MDH2-directed regulations has been revealed. Xu et al. demonstrated that lncRNA AC020978 interacts with MDH2 and activates the AKT pathway by stabilizing MDH2, thus affecting the progression and metastasis of non-small cell lung cancer (NSCLC) [50] (Figure 3). The interaction between AC020978 and MDH2 was originally found by a biotin labelled AC020978 pull-down assay followed by MS analysis, which was further validated by MDH2 RIP assay and AC020978-FISH/MDH2-IF colocalization assay. Xu et al. investigated the binding sites of AC020978 and MDH2 by conducting a series of pull-down assays using truncated mutants of AC020978. The results identified the MDH2-specific binding sequence as being located within a 300-428nt long region in AC020978. Meanwhile, a set of truncated MDH2-based RIP assays were also performed. The findings demonstrated that the segment of MDH2 consisting of amino acids 121–260 was capable of binding with AC020978. In addition to AC020978, lncRNA GAS5 was also reported as an MDH2-interacted lncRNA, which was also originally identified by an RNA pull-down assay followed by MS analysis [14]. After that, detection of RNA-protein interactions using cell lysates or recombinant MDH2 confirmed the direct interaction between GAS5 and MDH2. The MDH2-binding motif of GAS5 was also revealed (GAS5-Loop 2; 269–465 nucleotides) in this work. Functionally, GAS5 disrupts the metabolic enzyme tandem association of FH, MDH, and CS to modulate mitochondrial tricarboxylic acid flux via binding with MDH2 (Figure 3). GAS5 expression shows a negative correlation with the levels of its associated mitochondrial metabolic enzymes, which improves the overall survival of patients with breast cancer [14].

Protein interaction is a fundamental mechanism through which lncRNAs exert their biological functions. Thus, the identification of binding partners of lncRNAs is crucial for unravelling their biological functions and deciphering the underlying molecular mechanisms. In summary, the aforementioned lncRNA-binding TCA enzymes (ACLY, ACO1, IDH1, and MDH2) were all first identified based on RNA pull-down and MS analysis. As a matter of fact, more than a thousand novel RNA binding proteins lacking known RNA-binding domains have been characterized by MS analysis in large-scale studies [75–86]. Several comprehensive databases documenting lncRNA-protein interactions have emerged in recent years, listing over 47 million entries of RNA interactions with DNAs, proteins, and other RNAs [87–89]. In addition, a number of computational tools have been established to predict potential binding partners or even the binding sites [90–92]. The advancements in technologies for assessing the interaction between lncRNA and RBPs have accelerated the progress in understanding the role of lncRNAs in the TCA cycle. Targeted analysis of TCA cycle enzymes associated with a specific lncRNA can be highly beneficial in uncovering the unknown functions of the lncRNAs under investigation. However, the biological functions of many novel lncRNA-TCA cycle enzyme partners discovered by large-scale studies or bioinformatics prediction remain elusive, and in-depth studies are warranted to gain a deeper understanding of the dynamics of these interactions and their roles in the development and progression of various tumours.

4. Alterations of TCA cycle enzymes are closely related to tumorigenesis

Numerous cancers harbour alterations that affect the integrity of TCA cycle have been found [93–112]. For instance, the mutations of IDH1/2 have been commonly found in secondary glioblastoma and low-grade glioma, and relatively rare in angioimmunoblastic T-cell lymphomas, acute myeloid leukaemia (AML), prostate, colorectal, and thyroid cancer [105,113–116]. Patients with IDH1/2 mutations often have better outcomes than those carrying wild-type IDH genes [18]. The mutations of SDH, which regulate the inactivation of protein function in tumour cells and thus affect angiogenesis, proliferation, and cell survival, have been widely identified in hereditary paraganglioma, pheochromocytoma, renal tumour, gastrointestinal stromal tumour, thyroid tumour, testicular seminoma, and neuroblastoma [106,117–123]. Another example is the mutations of FH, which often occur in bladder, breast, testicular, renal cancer, as well as cutaneous and uterine leiomyomas, can alter the morphological features of tumour cells, affect tumour cell migratory potential, and vary the ability to respond to oxidative stress and DNA damage [107,124–128].

In addition, the dysregulation or activity changes of TCA cycle enzymes (e.g. IDH1/2, CS, and KGDHC) have also been revealed in several studies. For example, it has been demonstrated that the overexpression of IDH1 and IDH2 in glioblastoma cells could inhibit the growth and migration of cancer cells [129]. CS plays a pivotal role in the TCA cycle by catalysing a rate-limiting step. The overexpression or increased enzymatic activity of CS has been found in renal, ovarian, and pancreatic cancer, affecting the proliferation, migration, and invasion of tumour cells [108,130,131]. KGDHC is also a TCA cycle rate-limiting enzyme, comprised of three components: dihydrolipoamide dehydrogenase (DLD), dihydrolipoamide S-succinyltransferase (DLST), and α-KG dehydrogenase (OGDH). The downregulation of OGDH, which regulates tumour cell cycle, migration, and viability, has been widely identified in colorectum, breast, oesophageal, lung, pancreatic, and cervical cancer [109,132–134]. Besides, lower expression of OGDHL, one of the rate-limiting components of OGDH complex, has also been observed in breast, cervical, lung, hepatocellular carcinoma, oesophageal, and pancreatic cancers, which is closely associated with advanced tumour stage [110,135]. DLST has been found to be upregulated in T-cell acute lymphoblastic leukaemia (T-ALL) and neuroblastoma. The elevated expression of DLST is an indicator of poor treatment response and a more aggressive disease course [111,136]. Inhibition of DLD has been found to decrease both ferrous iron accumulation and lipid peroxidation, resulting in the inhibition of ferroptosis in head and neck cancer [137]. Unlike the well-explicit enzymes with certain oncogenic or suppressor roles, MDH2 has been shown to exhibit bifunctional roles. A previous study highlighted the downregulated expression of MDH2 in pheochromocytoma/paraganglioma caused by a germline mutation, and another study demonstrated the overexpression of MDH2 in clinical prostate cancer specimens [74,112]. The overexpression of MDH2 is considered an important factor for chemotherapy resistance in cancer. However, the in vitro experiments in prostate and doxorubicin-resistant uterine cancer cells also showed that MDH2 silencing could impair the viability of tumour cells [74,138].

Despite the prevalence of TCA cycle enzyme mutations and dysregulation in various malignancies, their actual roles in cancer development are still unclear, leaving many unanswered questions. Gaining a comprehensive understanding of the reasons behind the increased prevalence of specific mutations or dysregulation in certain cancers, compared to others, is crucial for researchers. This may be partly attributed the fact that different types of cancer have distinct underlying drivers and/or specific requirements on the path towards malignant transformation. Considering the expression specificity of lncRNAs in different types of cancers, lncRNAs may play a regulatory role in modulating the specificity of TCA cycle enzymes. Notably, cell death and inhibition of tumorigenesis may result from a significant level of mitochondrial dysfunction, whereas a moderate level of dysfunction could lead to an increase in mitochondrial ROS production and redox rebalancing, ultimately promoting tumour cell growth and invasiveness [139]. Therefore, gaining a comprehensive comprehension of the functions of TCA cycle enzymes and their regulators in cancer is vital for devising efficacious therapeutic approaches to combat various forms of cancer in the future.

5. Targeting the TCA cycle enzymes and the associated lncRNAs

The TCA cycle plays a crucial role as the central hub for catabolic and anabolic reactions in tumour cells, and the dysregulation or mutation of enzymes involved in this pathway can significantly impact its modulation and functioning. Therefore, the TCA cycle enzymes and the regulators that cause alterations in these enzymes, such as lncRNAs, can serve as promising targets for the development of new cancer therapies [140].

5.1. Targeting TCA cycle enzymes

In the case of IDH-mutated glioma cells, the administration of mutant IDH (mIDH) inhibitors resulted in suppressed growth and stimulated differentiation, whereas primary human AML cells with IDH mutation exhibited cell differentiation upon treatment [141,142]. These preclinical data revealing the targeted treatment of IDH mutations sheds light on advancing the clinical development of mIDH inhibitors. AG-221 was the first molecule to undergo clinical trials in late 2013. It is a selective, small molecule inhibitor of mutant IDH2 (Table 2). In 2014, AG-221 received orphan drug status for AML from the Food and Drug Administration (FDA), and later in 2019, the European Medicines Agency (EMA) also granted it a similar designation. Then, the mutant IDH1 and IDH2 inhibitor AG-881, which could penetrate the blood-brain barrier, entered the clinical trial (Table 2). Several studies have explored the potential of AG-881 in the treatment of glioblastoma and haematopoietic malignancy, leading up to a Phase III clinical trial that commenced in early 2020 to evaluate its efficacy in glioblastoma patients (Table 2).

Table 2. Selected clinical trials of small molecules targeting the TCA cycle enzymes for tumour.

Target	Drug	Phase	Condition(s)	Identification No.
IDH2	AG-221	I/II	Hematologic neoplasms	NCT01915498
		I/II	Solid tumour, glioma, angioimmunoblastic T-cell lymphoma, intrahepatic cholangiocarcinoma, chondrosarcoma	NCT02273739
IDH1/2	AG-881	I	Advanced haematologic malignancies, includes AML and myelodysplastic syndromes	NCT02492737
		I	Advanced solid tumours including glioma, cholangiocarcinoma, chondrosarcoma	NCT02481154
		III	Grade 2 glioma, residual glioma, recurrent glioma	NCT04164901
KGDHC	CPI-613	I	Advanced or metastatic cancer, metastatic cancer, lymphoma, solid tumours, advanced malignancies	NCT00741403
		I	(Advanced or metastatic) CRC	NCT02232152
		I/II	Cancer, pancreatic cancer	NCT00907166
		II	Cancer	NCT01832857
		III	Relapsed/refractory acute myeloid leukaemia	NCT03504410
		III	Pancreatic cancer metastatic	NCT03504423

Besides targeting mIDH, recent studies suggest that KGDHC represents a significant vulnerability in multiple cancers and holds potential as an effective therapeutic target. Anderson and co-workers utilized a MYC-driven model of T-ALL to demonstrate that the heterozygous inactivation of DLST led to a significant delay in tumour onset, without affecting the normal development of animals [136]. Moreover, their study also revealed that the inactivation of DLST in T-ALL cells resulted in the disruption of the TCA cycle, deceleration of cell growth, and induction of apoptosis [136]. In a recent study conducted by Ilic and colleagues, it was found that cancer cells with oncogenic PI3K mutations rely on all three components of KGDHC, particularly OGDH, to facilitate their proliferation [143]. The results of the study provide support for the idea of targeting KGDHC as a potential approach for cancer therapy. Later, CPI-613, a multipurpose metabolic inhibitor (PDH and KGDHC) and α-lipoic acid (LA) derivative, was established. It is designed to target the dysregulated TCA cycle of tumour cells by inhibiting the enzymes PDH and KGDHC [144,145]. Clinical trials of CPI-613 were launched in the beginning of 2008, with a focus on advanced, metastatic, recurrent, or refractory diseases (Table 2). Thereafter, several Phase II and III trials were conducted to evaluate its safety and efficacy (Table 2). The FDA and EMA granted orphan drug status to CPI-613 for treating pancreatic cancer and AML [146].

5.2. Targeting lncRNAs

LncRNAs exhibit tissue-, disease-, or developmental-specific expression patterns, which make them a promising candidate for use as a diagnostic biomarker and therapeutic target [147]. The ability of lncRNA to exhibit specific expression patterns has been exploited to selectively target cancer cells, while sparing healthy cells, resulting in a potential strategy for the development of cancer therapies with fewer side effects. For instance, BC-819 (DTA-H19), a double-stranded DNA plasmid that contains the gene for diphtheria toxin-A under the influence of the H19 regulatory sequence, has been employed to target H19 overexpressing cancer cells. By intratumorally injecting BC-819, the expression of diphtheria toxin is induced, which effectively reduces the size of various tumour types [148]. Aside from their potential for selectively eradicating cancer cells, the properties of lncRNAs, such as their expression patterns, functions, and structural characteristics, can also be investigated as potential tools for cancer diagnosis and treatment. For example, the expression of ACLY-binding lncRNA TINCR has been found to be associated with cancers, including prostate, gastric, hepatocellular carcinoma, oesophageal, lung, colon, breast, bladder, and oral squamous cell carcinoma [149]. The diagnostic potential of TINCR was demonstrated in oesophageal squamous cell carcinoma and hepatocellular carcinoma, where TINCR levels were overexpressed at around 14-fold and 8.7-fold, respectively. Additionally, the significant correlations among TINCR dysregulation and clinicopathological characteristics, such as gender, age, histological grade, Ki-67 level, lymph node metastasis, tumour node metastasis stage, and tumour size in bladder and breast cancer were also validated [150,151]. However, the dysregulation patterns of TINCR can be varied across different cancer types. Due to the pleiotropic effects of TINCR dysregulation, efforts to target TINCR may lead to toxicity in both healthy cells and cancer cells. Thus, further investigations are needed to evaluate the potential of TINCR as an effective therapeutic target. Another example is MDH2-binding lncRNA GAS5. GAS5 is a lncRNA that exhibits diverse functions in the regulation of gene expression. Previous research has found that this lncRNA plays a role in the induction and suppression of cell apoptosis and tumorigenesis, respectively. Early investigations conducted in glioblastoma and NSCLC have suggested that the plasma levels of GAS5 may serve as a valuable diagnostic and prognostic biomarker for these diseases [152]. Given the non-invasive nature of liquid biopsy compared to traditional biopsy methods, it is anticipated that the promising findings of these studies will be readily translated into clinical practice in the foreseeable future.

Although targeting lncRNAs has been the subject of extensive research over the past decade, most of the studies on lncRNA targeting are still in the preclinical stage [153]. However, the expression of lncRNAs can be effectively repressed using small molecule inhibitors, antisense oligonucleotides, or RNA interference (RNAi) technology. Furthermore, gene therapy strategies can be employed to introduce beneficial tumour suppressor lncRNAs into cells for potential therapeutic benefit. In the future, lncRNAs are anticipated to expand the range of potential targets for RNAi and CRISPR-Cas genome editing technologies. In addition, to devise therapeutic approaches that target RBPs, researchers have screened RNA interference-based oligonucleotides and small molecule inhibitors that are designed to alter the levels of target RNAs and disrupt RBP-RNA interactions [154]. The utilization of high-throughput techniques to identify RNA binding partners and the integration of multiple datasets can aid researchers in developing new prognostic biomarkers or therapeutic targets for various human cancers.

6. Conclusions and prospects

To date, as the repertoire of lncRNAs continues to expand, the process of comprehending their mechanisms of action is an ongoing endeavour, and we are still learning how to use them as weapons against human cancer. However, numerous pieces of evidence have demonstrated the power of lncRNAs in a variety of cancers, and the new advancement in computational algorithms, together with sophisticated massive sequencing and innovative techniques (e.g. CRISPR-Cas9), the prospect of discovering effective cures for cancer is within reach. And now, what has become apparent, even with the incomplete understanding of the molecular processes through which lncRNAs contribute to the specific tumour-related cellular activities, is that lncRNAs participate extensively in tumour-related regulatory networks, including cellular metabolism reprogramming.

Cellular metabolism reprogramming is a hallmark of cancer initiation and progression [1–3]. Over the past decade, there has been an increasing recognition of metabolic phenotypes that distinguish cancer cells from adjacent normal tissues. When these phenotypes represent actionable therapeutic vulnerabilities, they hold significant potential for the development of novel cancer treatments. The TCA cycle is one of the crucial metabolic pathways that enable various cancers to utilize fatty acids, amino acids, and glucose to efficiently fulfill the cell’s bioenergetic, biosynthetic, and redox balance requirements [140,155]. In addition to investigating the function and regulation of TCA cycle components, there is also growing interest in developing therapies that target these components.

RBPs and their binding lncRNAs are recently identified as novel layers that cause direct drastic changes in TCA cycle enzymes and metabolites, and lead to distinct metabolic, genetic, and epigenetic alterations that are associated with the development and progression of tumours. Due to the functional versatility of lncRNAs, they played key roles in expanding the scope of RNA binding TCA cycle enzymes regulatory components in TCA cycle and tumour progression [22,156]. This represents a pivotal factor in the clinical development of metabolic inhibitors for oncology. Due to the compensatory pathways that cancer cells can employ to evade treatment and the context-dependent nature of their metabolic properties, developing effective therapies targeting cancer metabolism remains a significant challenge [157–161]. Thus, a better understanding of the expression and function of the dysregulated pathways directed by TCA cycle enzymes and their binding lncRNAs in each cancer type may shed new light on the development of novel therapeutic strategy for cancer treatment and accelerating our race towards a healthier world.

In this review, we have provided an overview of the interplay between TCA cycle enzymes and their binding lncRNAs in tumorigenesis. The interactions between lncRNAs and TCA cycle enzymes have emerged as critical regulators in both tumour metabolism and tumorigenesis, and become the potential targets for anti-tumour treatment. However, we have limited our discussions to the interactions that have been experimentally validated. As a matter of fact, other TCA cycle enzymes, such as CS, SUCLG1, MDH3, etc., have also been identified as RNA binding proteins in the RNA interactome studies [57,58]. Nevertheless, the role of these TCA cycle enzymes in cancer, and the relevance of their RNA binding activities, are still elusive. Incorporating multiple RNA-binding, tumour dysregulation, and mutation data sources in the future will enhance our understanding of the relationships between TCA cycle enzymes and cancer.

Acknowledgments

We thank Dr. Rongkang Hu from Anhui Normal University for his constructive suggestions. We also thank the platforms and resources provided by the Outstanding Innovative Research Team for Molecular Enzymology and Detection in Anhui Provincial Universities, the Key Laboratory of Biomedicine in Gene Diseases and Health of Anhui Higher Education Institutes, and Anhui Provincial Key Laboratory of the Conservation and Exploitation of Biological Resources. The authors would also like to express their gratitude to EditSprings (https://www.editsprings.cn) for the expert linguistic services provided.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work is supported by the National Natural Science Foundation of China (31970598 to X.W. and 32071270 to G.Z.), the Major Science and Technology Projects in Anhui Province (202003a06020009 to G.Z.), the University Natural Science Research Project of Anhui Province (2022AH050211 to T.S.) and the start-up funds from Anhui Normal University (762189 to T.S.).