The dysfunction of cells and their induced diseases may result from the dysregulated progression of expression of massive RNAs, including mRNAs and noncoding RNAs (ncRNAs) (1). In addition, up to 98% of the human genome can be transcribed into RNAs, most of which may represent RNAs that do not encode proteins (2), indicating that these transcribed RNAs are regarded as ncRNAs. ncRNAs longer than 200 nucleotides are called long ncRNAs (lncRNAs) (3), which were widely considered to be just transcriptional noise previously (4-6). With the emergence of sensitive and high-throughput sequencing technology, numerous lncRNAs have been detected and recognized. According to the NONCODE (v5.0) database, 172,216 lncRNA transcripts have been identified in the human genome (7). Currently, research has found that an increasing number of lncRNAs have regulatory functions and participate in various biological processes through different mechanisms, such as pluripotency, differentiation, proliferation, or cell survival (8). For example, lincRNA regulator of reprogramming (lincRNA-RoR) promotes the survival of liver cancer cells during hypoxic stress by regulating the hypoxic signaling pathway, or plays a role in pluripotent or embryonic stem cells by blocking the activation of cell stress pathways (9,10).

LncRNA can bind to proteins, RNA, and DNA, thereby controlling gene expression levels in epigenetics, transcriptional regulation, and other mechanisms through genetic imprinting, chromatin remodeling, cell cycle regulation, splicing regulation, mRNA degradation, and translation regulation (11,12). Scientists are increasingly paying attention to the functional annotation of lncRNAs, especially their role in diseases (13). Liver cancer, including hepatocellular carcinoma (HCC), has become an important cancer in both developing and developed countries. Genetic research has shown that ncRNAs are involved in the initiation, regulation, and progression of HCC (14). LncRNA is receiving increasing attention in the pathogenesis of liver diseases and has potential diagnostic, prognostic, and therapeutic significance (15). The expression of highly upregulated liver cancer (HULC) lncRNA in plasma is known to be a novel mRNA-like lncRNA biomarker for the diagnosis of HCC (16). LncRNA participates in the pathogenesis of HCC by regulating cancer-related signaling pathways such as the MAPK signaling pathway (17).

RNA binding proteins (RBPs) control almost all aspects of RNA fate, including transcription, splicing, modification, cellular transport, translation, and degradation (18,19). Therefore, changes in the expression level, structure, localization, or function of RBPs may have far-reaching effects on various RNAs (19). RBPs interact with RNAs through the RNA binding domain (RBD), which include typical RNA recognition motifs, K homology domains, Piwi/Argonaute/Zwille domains, DEAD/DEAH helicases, and zinc finger domains (20-22). Previous studies have demonstrated that similar to lncRNAs, numerous RBPs are altered in cancer (23,24). In the field of RNA research, the search for the role of RNAs and RBPs in cancer has always been a focus of research (12). The traditional view suggests that RBPs regulate RNAs, but it has been proven that RNAs can also regulate the function of RBPs. In addition, there are also numerous reports suggesting that lncRNAs can also influence the functions of RBPs (25,26). Thus, it is a crucial issue to elucidate the underlying mechanism of how lncRNAs and RBPs interact and influence the progression of liver cancer.

A previous study summarized the interaction between lncRNAs and RBPs in liver cancer, but it was conducted ten years ago, and did not clearly elucidate the exact interaction mechanism between lncRNAs and RBPs in liver cancer (27). Due to the lack of information on lncRNA-RBP interactions, it hinders a better understanding of the mechanisms of lncRNA RBP interaction networks in the occurrence and development of liver cancer. With the deepening of research, more and more RBDs in RBPs have been discovered, providing a theoretical basis for the development of more drugs targeting RBDs in proteins. In the present review, the interaction between lncRNAs and RBPs specifically in liver cancer disease was examined. Notably, how lncRNA-RBP pairs affect the development or progression of liver cancer, as well as their downstream targets, was elucidated, and the mechanism of abnormal interaction networks in the occurrence and development of liver cancer was investigated.

Through the application of high throughput sequencing technologies, the discovery and functional studies of lncRNAs have been expanded with blowout style (6). Numerous lncRNAs have been identified and verified in association of cancer development (28). As ncRNAs, lncRNAs in most cases interact with other molecules to modulate the tumorigenesis or progression of cancers, indicating the essential functions of lncRNA partners. RBPs are the most important type of lncRNA partners and can form RNA-protein (RNP) complexes with lncRNAs (29). Canonical RBPs have a wide range of substrates, including various types of RNAs, such as those associated with the heterogeneous nuclear ribonucleoprotein (hnRNP) family. In terms of functionality, RBPs can regulate the localization and function of lncRNAs (30). Conversely, lncRNAs can also alter the localization, modification, or even structure of associated RBPs to modulate their functions. In addition, the most important functions of RNP complexes are to regulate the downstream targets and influence the development of cancers (31). Thus, it is essential to identify the interactions between lncRNAs and RBPs to investigate their functions in liver cancer.

Transcriptional regulation establishes and maintains specific cell states and its misregulation can cause a broad range of diseases (53). In this regulatory area, a large proportion of lncRNAs localize in the nucleus and can regulate transcriptional programs (54). Mechanistically, lncRNAs can regulate gene expression at the transcriptional and post-transcriptional levels by targeting local or distant genes (6). LncRNAs not only bind to transcription factors (TFs) or other protein complexes, but also recruit them to promoters. In this case, TFs are specific RBPs that interact with lncRNAs. The common mechanism for transcriptional regulation by lncRNAs is to recruit or block components of chromatin or histone-modifying complexes (20,55). At the transcriptional level, a novel regulatory pattern has been widely observed in which lncRNAs bind to histone-modifying complexes and recruit these complexes to specific locations. In this section, the lncRNA-RBP pairs controlling transcription in liver cancer are summarized. LncRNA heart and neural crest derivatives expressed 2-antisense RNA 1 (HAND2-AS1) was shown to recruit the INO80 chromatin remodeling complex to activate bone morphogenetic protein (BMP) receptor type 1A (BMPR1A) transcription and BMP signaling transduction, thus promoting stem cell self-renewal of liver cancer. In addition, BMPR1A was found to promote liver cancer stem cell (CSC) self-renewal through BMP signaling (56). BMP signaling has been shown to involve the regulation of stem cells, including hair follicle stem cells, blood stem cells, and pluripotent stem cells (57-59). These findings also indicate that INO80 is a novel RBP that interacts with HAND2-AS1. A previous study demonstrated that hnRNPK, an RBP from the hnRNP family, interacts with small nucleolar RNA SNORD126, an orphan C/D box snoRNA. The formed RNP complex was shown to regulate fibroblast growth factor receptor 2 (FGFR2) expression at the transcriptional level, then activate the PI3K-AKT pathway, and ultimately promote HCC development (60). In addition, hnRNPs can interact with lncRNAs and inhibit gene expression. For example, in macrophages, lncRNA lnc13 was shown to form a complex with the p42 subtype of hnRNPD, which can promote histone deacetylase 1 (HDAC1) binding to the promoter regulated by lnc13 and ultimately inhibit the expression of inflammatory factors, indicating that hnRNPD and lnc13 play an important role in recruiting HDAC1 to promoter regions (61).

Another important lncRNA, HOX transcript antisense RNA (HOTAIR) was shown to reduce the recruitment of the cAMP response element-binding protein (CREB), p300, RNA polII onto the SET domain-containing 2 (SETD2) promoter region that inhibits SETD2 expression and its phosphorylation, and finally promote human liver cancer stem cell malignant growth (62). Notably, two lncRNAs, including HULC and MALAT1, can cooperate to aggravate liver cancer stem cell growth by increasing telomeric reapeat binding factor 2 expression via transcriptional regulation (63). The highly expressed LINC00324 in liver cancer has been shown to regulate Fas ligand (FASL) expression by interacting with Spi-1 proto-oncogene (PU.1) (64). By recruiting the nucleosome remodeling and deacetylase (NuRD) complex, long non-coding RNA histone deacetylase 2 (lncHDAC2) was demonstrated to localize on the promoter of promoter of patched 1 (PTCH1) to inhibit its expression, leading to activation of Hedgehog signaling (65). LncRNA HOTTIP was shown to function in a cis manner to directly control the homeobox A locus gene expression by interacting with the WD repeat domain 5/mixed-lineage leukemia complex, thereby promoting HCC progression and predicting patient outcome (66). A recent study demonstrated the molecular and cellular functions of lncRNA linc01134. Linc01134 was shown to interact with RBP IGF2BP1 to promote the stability of Yin Yang-1 (YY1) expression, and YY1 also promoted the transcription of linc01134, forming a positive feedback loop to dictate the progression of HCC. In addition, linc01134 was demonstrated to act as a miR-324-5p sponge to exert its function (67).

LncRNAs not only interact with TFs, but also recruit them to promoter regions to regulate the transcription process. At the transcriptional level, a new pattern has been widely observed in which lncRNAs bind to histone-modifying complexes and recruit histone-modifying complexes to specific locations. Enhancer of zeste 2 (EZH2) is the subunit of polycomb repressive complex 2 (PRC2), which is involved in maintaining the transcriptional repressive state of genes by modifying histones. In liver cancer, DDX11-AS1 was shown to interact with EZH2 and DNA methyltransferase 1, and to suppress large tumor sppressor kinase 2 (LATS2) expression through a histone modification mechanism, indicating DDX11-AS1 may be a novel oncogene in hepatocarcinogenesis by repressing LATS2(68). With regard to lncRNA SNHG9, it was also shown to interact with EZH2 and recruit EZH2 to the promoter region of phosphatase and tensin homolog (PTEN), thereby inhibiting transcription of PTEN by increasing histone H3 lysine 27 trimethylation levels in the promoter region. As PTEN is a canonical tumor suppressor, these findings indicated that SNHG9 promotes liver CSC self-renewal and tumor formation, and exacerbates HCC progression by inhibiting PTEN (69).

RNA modifications, including N6-methyladenosine (m⁶A) and 5-methylcytosine (m⁵C), have been shown to significantly influence the functions of lncRNAs in cancer (70). It has been reported that m⁶A modification plays an important role in liver carcinoma and lipid metabolism (71). The m⁶A reader has been shown to play a crucial role in the stability of lncRNAs. Expression of lncRNA MEG3 was downregulated in HCC tissues, which was probably induced by N⁶-methyladenosine modification mediated by m⁶A writer METTL3. In addition, overexpression of MEG3 was shown to inhibit the proliferation, migration and invasion of HCC cells through miR-544b/BTG anti-proliferation factor 2 signaling (51). It was shown that METTL3 mediates the modification of lncRNA nucleolar protein interacting with the FHA domain of MKI67 (NIFK)-AS1 and increases its expression. Highly expressed NIFK-AS1 was demonstrated to promote the proliferation, colony formation, migration, and invasion of HCC cells through a ceRNA network involving an NIFK-AS1/miR-637/AKT1 axis (72). Peng et al explored the role of lipopolysaccharide (LPS) in immune regulation of HCC, and detected the participation of lncRNA m⁶A modification. LPS treatment increased the expression of methyltransferase-like protein 14 (METTL14), which promotes m⁶A modification of miR-155 host gene (MIR155HG). The expression of MIR155HG was then stabilized by m⁶A ‘reader’ protein ELAVL1. Finally, MIR155HG functioned as a ceRNA to modulate the expression of programmed death-ligan 1 by the miR-223/signal transducer and activator of transcription (STAT)1 axis (73). METTL14 was also shown to mediate m⁶A modification of lncRNA Rho GTPase acitvating protein 5 (ARHGAP5)-AS1, which is significantly increased in HCC by m⁶A reader IGF2BP2. The high expression of ARHGAP5-AS1 was demonstrated to interact with cold shock domain containing E1 (CSDE1) and attenuate the interaction between CSDE1 and tripartite motif containing 28, preventing CSDE1 degradation via the proteasome, finally contributing to HCC prognosis (74). Compared with m⁶A, m⁵C modification is less known and reported. A previous study showed that the expression of lncRNA H19 is also regulated by m⁵C modification. RNA methyltransferase NOP2/Sun RNA methyltransferase 2 increased the m⁵C level of H19, thus increasing its stability. The m⁵C-modified H19 lncRNA could be specifically bound by G3BP stress granule assembly factor 1, leading to MYC accumulation and finally associated with poor differentiation of HCC (75). In summary, these findings demonstrated that RBP-mediated RNA modification can profoundly regulate the expression and function of lncRNAs, which can modulate the function of downstream targets and affect the progression of liver cancer.

The subcellular localization of molecules can indicate their potential functions and influences on biological processes. Increasing evidence suggests that RBPs can affect the cellular localization of lncRNAs, and vice versa. The functions of lncRNAs are tightly associated with their subcellular localizations, and have been previously summarized (54). It has been shown that RBP HuR can affect the stability and localization of its target RNA (76). The oncoprotein Yes associated protein (YAP) was shown to upregulate the expression of MALAT1 at the transcriptional and post-transcriptional levels, while serine and arginine rich splicing factor 1 (SRSF1) played the opposite role. SRSF1 inhibited YAP activity and downregulated MALAT1 expression by preventing co-occupation of T-cell factor/β-catenin on YAP and MALAT1 promoters. Conversely, overexpression of YAP impaired the nuclear preservation of SRSF1 and itself by interacting with angiokinetin. This effect eliminated the inhibitory effect of SRSF1 on MALAT1 in the nucleus. In addition, the high expression of YAP was consistent with the low nuclear accumulation of SRSF1 in human liver cancer tissue. Thus, RBP SRSF1 was demonstrated to exert important regulation on MALAT1 expression via protein-protein interaction, but not by directly interacting with MALAT1 RNA (77).

In most cases, H19 is a carcinogenic lncRNA. However, H19 can suppress HCC progression metastasis and the expression of markers of EMT. Mechanistically, H19 has been shown to be associated with the protein complex hnRNPU/PCAF/RNA polII, and to transport this complex into nucleus, thus activating the miR-200 family by increasing histone acetylation. Ultimately, the epigenetic activation of the miR-200 family was demonstrated to contribute to H19-mediated suppression of HCC metastasis (52). Concurrently, it has also been observed that lncRNA H19 can accelerate the occurrence of liver cancer. miR-675 embedded in the first exon of H19 can promote H19 transcription through histone modification and subsequent early growth response 1 transcription in an indirect manner. Finally, H19 may promote tumorigenesis by activating PKM2 to alter the expression and function of oncogenes with an as-yet unknown mechanism in human liver cancer (78). These findings suggest that lncRNAs can have various functions under specific conditions, which is probably induced by their interacting partners.

As one of the first identified cancer-related lncRNAs, HOTAIR has been shown to be aberrantly expressed in numerous tumors, including liver cancer, and is closely related to the occurrence and development of tumors (79). HOTAIR is reported to influence the localization of its interacting proteins. SETD2 is a lysine methyltransferase that catalyzes histone H3 lysine 36 trimethylation and has been revealed to play important roles in the regulation of transcriptional elongation, RNA splicing, and DNA damage repair (80). LncRNA HOTAIR was shown to reduce the reappearance of CREB, p300, and RNA polII in the SETD2 promoter region, thereby inhibiting SETD2 expression and its phosphorylation. Therefore, HOTAIR greatly impaired the DNA damage repair ability and increased microsatellite instability of liver stem cells, ultimately promoting their malignant growth (62). A similar mechanism was also observed for lncRNA LEF1-AS1, which recruits TF CCAAT/enhancer-binding protein (CEBP)β to the promoter region of cell division cycle associated 7 (CDCA7) to promote CDCA7 expression, therefore upregulating EZH2 expression and promoting proliferation and invasion of HCC cells (81).

LncRNA CEBPA-divergent transcript (DT), a divergent transcript of CEBPA, was revealed to have the ability to promote growth, migration and invasion of hepatoma cells. CEBPA-DT could bind to hnRNPC, facilitating cytoplasmic translocation of hnRNPC, and enhancing the interaction between hnRNPC and discoidin domain receptor 2 mRNA. In addition, CEBPA-DT could also facilitate nuclear translocation of β-catenin, which promotes Snail family transcriptional repressor 1 expression and finally induces the EMT process (82).

In addition, hnRNPK has been identified as a driving factor for RNA nuclear retention. MALAT1 lacks short scattered nuclear elements (SINEs), which can easily translocate to the cytoplasm and bind more strongly to TAR DNA protein 43 (TDP-43). TDP-43 can even be guided to the cytoplasm by MALAT1, leading to a decrease in the TDP-43 expression. Notably, the SINEs of MALTA1 can regulate the interaction between hnRNPK and MALAT1, which increases the nuclear retention of MALAT1. This may occur through direct interaction between SINEs and KH RNA binding domain containing, signal transduction associated 1, containing the KH RNA binding domain, as well as transformer-2α homolog protein, which binds to hnRNPK (83).

LncRNAs can also sponge miRNA, acting as ceRNAs, to modulate expression of downstream genes, the regulatory mechanism of which is more abundant and deeper explored in cancers (84). This process is predominantly regulated by RBPs such as the miRNA partner Argonaute 2 (AGO2). AGO2 contains four functional core domains and has catalytic activity; its interaction with miRNAs guides the complex to the target RNA body, and can induce gene silencing at the post-transcriptional level (85). For example, lncRNA proliferating cell nuclear antigen pseudogene 1 (PCNAP1) was shown to contribute to hepatitis B virus (HBV) replication by attracting miR-154, thus increasing the expression level of PCNA, finally accelerating the growth of HCC (86). HOXD-AS1 was demonstrated to also act as a ceRNA to attract miR-130a-3p and upregulate expression of SRY-box containing transcription factor 4, thus activating the expression of EZH2 and matrix metalloproteinase (MMP)2 and ultimately facilitating HCC metastasis (87). LncRNA BCAR4 has also been shown to act as a ceRNA to promote liver cancer progression by sponging miR-1261 and upregulating anaphase promoting complex subunit 11 expression (88). As a well-known lncRNA, MALAT1 has been demonstrated to promote liver cancer progression, and one of the underlying mechanisms was reported to be that MALAT1 sponges miR-204 to release the expression and function of sirtuin 1 (SIRT1), which can significantly promote HCC migration and invasion (89). LncRNA growth arrest-specific transcript 5 (GAS5) was shown to be downregulated in the NK cells of patients with liver cancer. By exploring the underlying mechanism, it was found that GAS5 serves as a sponge of miR-544, and GAS5 overexpression was demonstrated to increase RUNX3 expression, interferon-gamma (IFN-γ) secretion, and NK cell cytotoxicity, indicating the killing effect of GAS5 on cancer cells (48). LncRNA HOTAIR can also act as a ceRNA to promote growth of liver cancer by targeting miR-217(90). In addition, lncRNAs have been shown to not only sponge miRNA, but also RBPs and act as ceRNAs. Specifically, lncRNAs have been demonstrated to serve as ‘bait’ to prevent protein binding to target RNA. These lncRNAs can bind to RBPs and regulate the post-transcriptional fate of target mRNA (91). Furthermore, circular RNAs (circRNAs), a specific type of lncRNAs, have also been shown to be crucial ceRNAs that sponge miRNAs (92,93). Previous studies have described the functions and mechanisms of circRNAs in liver cancer (94,95). For example, circDLC1 can attract RBP HuR to subsequently reduce the interaction between HuR and MMP1 mRNAs, thus inhibiting the expression of MMP1 and ultimately contributing to inhibition of HCC progression (96). With regard to clinical application, circRNAs exhibit potential as novel biomarkers and therapeutic targets in HCC (97).

Moreover, numerous lncRNAs involved in liver cancer were reported to not be associated with RBPs. Knockdown of ACTA2-AS1 has been shown to promote liver cancer cell proliferation, migration and invasion, however the underlying mechanism has not been clearly elucidated (98). High expression of SNHG4 was shown to indicate a poor overall survival of patients with liver cancer, while the underlying mechanisms remain unclear (99). Ding et al demonstrated that lncRNA HOTAIR inhibits the mRNA and protein levels of RNA binding motif protein (RBM38) in HCC cells, thus promoting cell migration and invasion; whereas, how HOTAIR and RBM38 interact or cooperate to promote HCC progression has yet to be determined (100). Another well-known lncRNA, H19, can be detected in the exosomes released by CD90⁺ liver cancer cells, and was reported to mediate the phenotype of endothelial cells (101). LncRNA deleted in lymphocytic leukemia 2 (DLEU2) has been shown to interact with EZH2. DLEU2 knockdown inhibited the proliferation of HCC cells, while EZH2 knockdown attenuated this inhibition. However, how DLEU2 and EZH2 interact and their molecular functions remain unclear (102).

The function of proteins is usually regulated by introducing chemical modifications after translation. Different types of post-translational modifications (PTMs) can affect multiple functions of proteins, thereby influencing cellular function and ultimately leading to the occurrence and development of cancer (103,104). To date, it has been found that some lncRNAs can regulate PTMs of their binding proteins through mechanisms such as phosphorylation, ubiquitination, and acetylation, thereby regulating protein degradation or production, and affecting protein expression levels and activity (105). p53 determines cell fate by regulating the transcription ability of multiple target genes, and its PTMs play a crucial role in its activity. However, how the PTMs of p53 are dysregulated remains a question worth exploring. MALAT1, a transcript associated with lung adenocarcinoma metastasis, has been shown to be involved in the PTMs of p53 and to affect the occurrence and development of tumors. According to previous studies, MALAT1 is widely present in various cancers, including HCC (106), and is associated with the proliferation and metastasis of cancer cells. MALAT1 was shown to reduce the acetylation of p53 by competing with the interaction of SIRT1 and deleted in breast cancer 1, thereby releasing SIRT1 and thus regulating the activity of SIRT1(107). Similarly, APAL [also referred to as a Polo-like kinase 1 (PLK1)-related lncRNA] was shown to promote Aurora kinase A (AURKA)-mediated PLK1 phosphorylation by increasing the interaction between PLK1 and AURKA (108). In addition, lncRNA mesenchymal stem cell upregulation factor (lncRNA-MUF), was demonstrated to specifically interact with glycogen synthase kinase-3 β (GSK-3β) and annexin A2 (ANXA2) through different but partially overlapping sites, thereby regulating the progression of HCC. In addition, the increased expression of lncRNA-MUF promoted the interaction between ANXA2 and GSK-3β. Conversely, depletion of lncRNA MUF inhibited these binding effects (109).

In a recent study, lncRNA tetraspanin 12 (TSPAN12) was shown to be stabilized and partially increased by METTL3-mediated m⁶A modification in HCC. Subsequently, TSPAN12 directly interacted with eukaryotic translation inititation factor 3 subunit I (EIF3I) and sentrin-specific protease 1 (SENP1), and also enhancde EIF3I-SENP1 interaction. This interaction was shown to inhibit the SUMOylation of EIF3I, ultimately activating the Wnt/β-catenin signaling pathway and stimulating EMT and metastasis in HCC (110). By competing with HOTAIR to bind to EZH2, lncRNA-p21 not only broke down the PRC2 complex to release EZH2, but also enhanced the interaction between EZH2 and STAT3 to trigger STAT3 methylation (111). It is worth noting that the interaction between GAS5 and YAP was shown to enhance YAP phosphorylation, promoting its ubiquitination and degradation, thereby inhibiting YAP signaling. Therefore, lncRNA-GAS5 was demonstrated to inhibit the progression of colorectal cancer by dysregulating YAP in vivo and in vitro (112). Previous research has shown that lncRNAs can serve as regulatory factors for ubiquitin-mediated protein hydrolysis, promoting the degradation of ubiquitin-binding proteins. HOTAIR was shown to interact with E3 ubiquitin ligase and its substrates, thereby enhancing the ubiquitination and degradation of Ataxin-1 and Snuportin-1(113). Although the aforementioned studies were not completely derived from liver cancer research, the findings on these important lncRNAs and RBPs indicate that these regulatory mechanisms may also occur and play important roles in liver cancer, which should be further explored in the future.

Based on the current literature research, there are several lncRNA and RBP databases that have summarized the identified lncRNAs and RBPs, their functions, subcellular localizations, association with diseases, and even their interaction pairs. LncBook2.0 (https://ngdc.cncb.ac.cn/lncbook/) is a curated database that integrates human lncRNAs with multi-omics annotations (114). In this version, LncBook2.0 identified the lncRNA-protein interaction pairs using the RNA binding sites of 356 RBPs from HepG2 and K562 cells, which are obtained by the eCLIP-seq experiments (115). Specifically, HepG2 is a type of liver cancer cell line, indicating that the identified lncRNA-RBP interaction pairs from LncBook2.0 can provide an abundant resource for the functional interactions between lncRNAs and RBPs in liver cancer cells. Typically, specific motifs exist within the lncRNA binding site of protein regulatory factors. High throughput CLIP-seq technology has provided comprehensive RBP datasets from various cell types. The recently established database CLIPdb collected 395 public CLIP-seq datasets for 111 RBPs from four organisms, providing a valuable resource to search and visualize the interactions between RBPs and RNAs, including lncRNAs (116).

In addition, the experimentally validated lncRNAs and circRNAs associated with human cancers have been collected and categorized in Lnc2Cancer 3.0 database, which includes 9,254 lncRNA-cancer associations, with 2,659 lncRNAs and 216 cancer subtypes. This database provides an abundant resource for the identification of lncRNAs associated with liver cancer. There were 117 entries related with liver cancer when the database was searched (117). However, this database does not provide detailed functional information for these validated lncRNAs. Finally, a review article has been published that listed the RNA-protein interaction database resources, which can help scientists quickly search for lncRNAs and proteins of interest (118).

In addition to database resources, the rapidly developing technologies and methods have also aided in identifying the lncRNA-RBP complexes in various diseases. Specifically, next generation sequencing technology has been shown to facilitate the identification of novel lncRNAs, as well as the interaction sites between RBPs and RNAs. The aforementioned CLIP and RNA immunoprecipitation methods can identify the sites bound by a specific RNA-binding protein on endogenous RNAs (119). In addition, RNA pulldown, RNA antisense purification, and chromatin isolation by RNA purification methods can help scientists identify interacting proteins given a lncRNA of interest (120-122). In addition to experimental methods to capture lncRNA-RBP pairs, in silico inference methods have also been developed for the prediction of these RNP complexes. With the increased accuracy of predictive algorithms, inference methods can provide highly confident prediction results, which accelerates the discovery of functional RNP complexes (123).

The identification of disease-driving lncRNA-RBP interactions in liver cancer has opened unprecedented opportunities for therapeutic intervention. Unlike traditional protein-targeting drugs, RNA-based therapeutics offer the unique advantage of targeting ‘undruggable’ molecules by sequence-specific recognition. Subsequently, the current therapeutic strategies targeting lncRNA-RBP networks, including ASOs, small molecule inhibitors, and clustered regularly interspaced short palindromic repeats (CRISPR)-based approaches, are presented alongside the critical challenges facing clinical translation.

The convergence of advanced delivery technologies, structural biology, and artificial intelligence-driven drug design promises to overcome current limitations. Novel approaches including antibody-oligonucleotide conjugates enable cell-specific targeting beyond hepatocytes, while exosome-based delivery systems offer biocompatible, immunologically inert alternatives to synthetic nanoparticles. As our understanding of lncRNA-RBP structural dynamics expands, the development of orally bioavailable small molecule modulators may finally unlock the full therapeutic potential of these critical regulatory networks in liver cancer.

As a key achievement of this review, a clinically oriented summary of the lncRNA-RBP interactions catalogued in Table II is presented, stratified by therapeutic accessibility, biomarker potential, and druggability. This framework provides actionable guidance for prioritizing interactions in drug development pipelines. For example, the HULC-IGF2BP1 interaction can be treated as a diagnostic biomarker, as plasma HULC levels were shown to be associated with HCC diagnosis, while IGF2BP1 was demonstrated to destabilize HULC providing a druggable node (16,40). Another interaction pair is MALAT1-SRSF1, which can serve as a prognostic indicator for patients with HCC. The mutual inhibition of SRSF1 with YAP was shown to create a synthetic lethality opportunity by maintaining MALAT1 expression; thus, targeting SRSF1 or MALAT1 may represent a potential therapeutic strategy for HCC in future. Other than these two illustrated pairs, other lncRNA-RBP pairs can also be further investigated to identify their potential in clinical translation of HCC treatment.

The interactions between lncRNAs and RBPs are diverse and complex. At the transcriptional and/or post-transcriptional level, lncRNAs can recruit interacting proteins to specific sites in the genome to regulate cis or trans gene expression (12). LncRNAs can serve as scaffolds to promote the assembly of protein complexes. In addition, lncRNAs can also serve as bait to weaken the interaction between proteins and biological macromolecules (DNA, RNA, and proteins). As the topic of lncRNAs and RBPs has been widely discussed in previous studies (30,142,143), their interactions and biological outcomes in liver cancer were emphasized in the present review, especially in studies from the last 5 years. The identified lncRNA-RBP pairs were found to have profound effects on the tumorigenesis or progression of liver cancer, and most of these lncRNA-RBP interactions were shown to promote liver cancer development, indicating that these interactions can serve as potential therapeutic targets for liver cancer. However, it has also been reported that the functions of RBPs are diverse in both cancer and normal cellular activity, making them difficult or undruggable targets for liver cancer therapy (143). Additionally, lncRNAs were shown to exhibit more tissue-specific expression pattern and have less functions compared with proteins, making them desirable targets for liver cancer therapy, as was reported in a recent study (144). Thus, the identified lncRNA-RBP interactions in liver cancer can be further explored to identify the exact interacting sites or functional domains, which should be preferentially considered in drug design and testing.

Of note, some points and issues were not discussed in the present review. Numerous lncRNAs identified in recent years were shown to lack functional annotation and interacting partners, and many RBPs were demonstrated to play important roles in liver cancer development, but were not shown to interact with lncRNAs. These lncRNAs and RBPs were not considered, but their importance is not excluded. In addition, not all the lncRNA-RBP pairs were included, especially those involving lncRNAs, including circRNAs, as ceRNAs (94). Finally, some special interaction mechanisms were not fully described in the present review. In summary, the field of lncRNA-RBP interactions is growing rapidly and attracting more attention than before. It is predictable that the protein-lncRNA interaction network will provide essential resources for the understanding of lncRNA and RBP biological mechanisms, as well as the abundant molecular targets for the drug design and treatment of liver cancer in the near future.

Despite the rapid expansion of lncRNA-RBP interaction research in liver cancer, several fundamental limitations constrain clinical translation. First, the overwhelming majority of reported interactions rely exclusively on in vitro experiments using established cell lines, which fail to recapitulate the complex tumor microenvironment, heterogeneity, and stromal interactions characteristic of human HCC. Second, the lack of robust in vivo validation represents a significant translational bottleneck. Animal models, particularly patient-derived xenografts and genetically engineered mouse models that faithfully replicate human hepatocarcinogenesis, are essential for assessing pharmacokinetics, biodistribution, and off-target effects of RNA-targeted therapies. Third, clinical validation is conspicuously absent. Prospective cohort studies with well-annotated clinical parameters (etiology, cirrhosis status, treatment history) are urgently needed to validate the prognostic and therapeutic relevance of specific lncRNA-RBP pairs. Fourth, mechanistic studies often fail to establish causal relationships vs. correlative associations. Rigorous genetic rescue experiments, structure-function mapping of interaction domains, and competitive inhibition studies are required to distinguish direct regulatory interactions from indirect effects. Finally, the dynamic and context-dependent nature of lncRNA-RBP interactions presents therapeutic challenges. The same lncRNA (for example H19) can exert opposing effects depending on RBP availability, metabolic state, or cellular stress (145). This functional plasticity complicates target validation and necessitates sophisticated conditional knockout models to dissect context-specific functions.