The majority of lncRNAs are similar in origin to mRNAs and are transcribed by RNA polymerase II. They also have the same structure as mRNAs (13); however, lncRNAs also have certain characteristics, although these are not unique to lncRNAs. lncRNAs account for >60% of all ncRNAs [certain studies report rates of >70% (14) or 70–90% (15)], and most lncRNAs contain >2 exons (16). Some lncRNAs, owing to their reverse splicing or lack of a 5′ cap, are readily connected with their own 3′ poly-A tail, generating circRNA (17). Numerous lncRNA genes have miRNA sequences embedded in their exons or introns, making them a source of miRNAs (13). A number of chemical modification methods exist for lncRNAs, including adenosine methylation, cytosine modification, uridine isomerisation, guanosine methylation and ribose modification (18), resulting in more complex ncRNA biogenesis.

To date >900 lncRNAs have been identified (19), resulting in different methods of classifying lncRNAs (Fig. 1). Based on their location, they can be divided into nuclear, cytoplasmic and mitochondrial lncRNAs (20). According to their shape, lncRNAs can be divided into linear and circular lncRNAs (21). lncRNAs can also be categorised by source as follows: i) Sense transcript; ii) antisense transcript; iii) bidirectional transcript; iv) long intergenic ncRNAs (22); v) intron region; vi) promoter region (14); and vii) enhancer region (23). Finally, lncRNAs can be categorised by function as follows: i) Functional lncRNAs, ii) non-functional lncRNAs; and iii) lncRNAs for which transcription alone is sufficient to exert their effects and those that do not need to rely on transcripts to function (24). In general, the classification of lncRNAs is regarded as imperfect, as functional interactions and overlaps exist among different classifications of lncRNAs (5). In the future, a better lncRNA classification system will be required.

The function of lncRNAs varies depending on their spatial structure and location. lncRNAs have two functional forms including spliced and unspliced (25). The biological genetic information in mRNAs is mainly transmitted through ‘linear’ coding, while the biological information in lncRNAs mainly works through its different ‘spatial’ structural mechanisms (26). The main pathways through which lncRNAs function include epigenetic, transcriptional and post-transcriptional regulation (27). When lncRNAs perform their functions, they usually combine with DNA, RNAs and proteins to form the corresponding complexes. First, lncRNAs bind with DNA and play a role in decompressing chromatin (28). They can also affect chromatin accessibility (29), and the R-loop can make lncRNAs an ideal regulatory centre (30). Next, lncRNAs bind to RNAs. The lncRNAs regulate mRNA expression through competitive endogenous RNA mechanisms (31). Finally, lncRNAs bind with proteins to form complexes, regulating protein structure and function (32). Some lncRNAs directly bind to proteins, affecting cellular signal transduction and gene expression.

The functions of lncRNAs vary depending on their location (Fig. 2). The functions of lncRNAs in the nucleus include: i) Providing scaffolding for DNA (33), RNAs and proteins (34); ii) recruiting chromatin-remodelling complexes to remodel the chromatin (35); iii) preventing RNA polymerase II from attaching to the gene promoter and interfering with transcription (36); and iv) alternative splicing (37). The functions of lncRNAs in the cytoplasm include the following: i) Recruiting proteins that promote mRNA degradation and affect the stability of mRNAs (13); ii) sequestering specific miRNAs through the competing endogenous RNA (ceRNA) mechanism to regulate translation (38); iii) interfering with post-translational modifications; and iv) affecting the structure and function of proteins (39). Although the function of lncRNAs is understood to a certain extent, it cannot be inferred from its primary structure, unlike mRNAs. Therefore, further in-depth exploration of lncRNA functions is required.

The intervertebral disc (IVD) is the largest non-blood supply tissue in the human body (40). The IVD is composed of three parts: i) The cartilage endplate (CEP) on the upper and lower sides; ii) the surrounding annulus fibrosus (AF); and iii) the nucleus pulposus (NP) in the middle (41). The CEP is ~1-mm thick and contains a number of micropores for material exchange in the NP, but contains no nerve tissue. The AF is divided into outer and inner rings, and the front and sides are twice as thick as the back. Nerve endings are only distributed in the outer ring (42). The AF is very strong and plays a major role in supporting and stabilising the spine (42). The NP is a gel-like tissue with a high water content, containing collagen, proteoglycans, NP cells and water (43). The composition and structure change with age, accounting for 50–60% of the IVD cross-section (44). The toughness and elasticity of the IVD enable it to perform important physiological functions, such as supporting the spine, buffering pressure and providing spine mobility (45).

As the IVD has no blood vessels, its self-repair ability is poor, making it prone to degeneration (40). IDD initiates in NP cells (46), and this is hypothesised to be primarily due to the joint action of genetic and environmental factors. Smoking, physical activity and poor living habits are contributing factors to IDD, although genetic factors are the most important (47). It is estimated that >70% of cases are caused by genetics and the programmed cell death drives the occurrence of IDD (48). After the onset of IDD, the water content in the IVD decreases, and the extracellular matrix (ECM), including type II collagen and proteoglycans, degrades (49). After the CEP and AF are destroyed, the barrier between the IVD, and the circulatory and immune systems is broken, triggering an inflammatory reaction (50). Under the influence of these pathogenic factors and changes in the microenvironment, the proliferative ability of the NP cells decreases, causing cell apoptosis, autophagy and senescence, ultimately resulting in IDD (5). Surgical treatment of IDD also has challenges, including infection (51), recurrence (52) and adjacent spondylosis (53). Therefore, understanding the pathogenesis of IDD and exploring new therapeutic targets are crucial.

Compared with normal tissues, IDD exhibits numerous differentially expressed ncRNAs, including lncRNAs, which play an important role in the occurrence and development of IDD. Wan et al (54) used lncRNA-mRNA microarrays to study lncRNA expression in NP specimens from human degenerative NP tissue and normal groups. The results revealed that the expression levels of 67 lncRNAs increased and those of 49 decreased. In another study, the same lncRNA-mRNA microarray technology was used to analyse the lncRNA expression in NP specimens of human degenerative NP tissue and normal groups, and 135 significantly upregulated and 170 downregulated lncRNAs were found (55). A similar study found that 2,418 mRNAs and 528 lncRNAs were differentially expressed in the IDD group compared with the control group (56). Shi et al (57) recently constructed a ceRNA network containing 15 lncRNAs, 9 miRNAs and 103 mRNAs, and discovered a lncRNA/miRNA/mRNA axis, which affects the progression of IDD, namely, the small nucleolar RNA host gene 5/miR-299-5p/activating transcription factor 2 axis. However, further experimental verification is still needed. In another study, a ceRNA network was constructed using 15 lncRNAs and 21 miRNAs, and identified the two hub genes MAPK8 and CAPN1 in this regulatory network as key biomarkers of IDD (58). These studies demonstrate that there are obvious differences in lncRNA expression in IDD and that lncRNAs do not function independently but form an extensive ceRNA network system with other ncRNAs.

As aforementioned, IDD originates primarily from NP cells. As early as 2006, researchers found that transplantation of human NP cells into degenerated IVDs in rabbits could improve the degree of disc degeneration (59). A subsequent similar study showed that transplantation of NP stem cells into degenerated IVDs can more effectively improve IDD than transplantation of NP cells (60). These studies indicate that NP cells play a crucial role in the occurrence and development of IDD. The state of NP cells is influenced by the ECM, inflammation, apoptosis, necrosis and ageing factors. Under a series of negative influences, this ultimately leads to dysregulation of the NP cell phenotype, thereby participating in the occurrence and development of IDD (61). Jiang et al (62) reported that miR-365 can alleviate the development of IDD by regulating the synthesis and degradation of the ECM in NP cells. Propionibacterium acnes has been shown to induce the apoptosis of NP cells through the TLR2/JNK pathway, which can alter the process of IDD (63). Inflammation in NP cells also affects IDD, whereby extracellular lactate can promote activation of the NLR family pyrin domain containing 3 (NLRP3) inflammasome, increasing inflammatory levels in the NP cells and thereby promoting IDD (64). Experimental results from Du et al (65) suggested that activation of cannabinoid receptor type 2 can delay the ageing of NP cells, restore the balance of ECM metabolism and attenuate IDD. Overall, changes in the ECM, cell death, cell inflammation and cell ageing are all part of the complex mechanisms underlying the occurrence and development of IDD. They interact with each other, collectively influencing the process of IDD.

The ECM, type II collagen and proteoglycan (aggrecan) are the main components of the NP. The ECM is produced by NP cells, and the balance between ECM synthesis and degradation contributes to the biomechanical balance and structural stability of the IVD, which is also considered a key indicator for evaluating NP cell function (66). Subsequent studies indicate that lncRNAs can affect the synthesis and degradation of the ECM through a ceRNA mechanism and can directly regulate hub genes or signalling pathways. Gao et al (67) studied degenerated human tissues and cells, and found that prostate androgen-regulated transcript 1 (PART1) was upregulated in degenerated NP tissues. By contrast, PART1 knockdown led to increased ECM synthesis, reduced degradation and enhanced proliferation of NP cells. Growth arrest specific 5 has been shown to be expressed at abnormally high levels in human degenerative NP tissues (68). After downregulation, angiopoietin 2 has been shown to be inhibited in a miR-17-3p-dependent manner, promoting ECM remodelling, inhibiting NP cell apoptosis and improving IDD (68). A recent study showed that the transcription factor FOXO3 may enhance the competitive binding of HOXA transcript at the distal tip (HOTTIP) and miR-615-3p by activating HOTTIP transcription, increasing the expression of the target gene collagen type II α1 of miR-615-3p, inducing NP cell proliferation, and reducing apoptosis to prevent ECM degradation (69). Krüppel-like factor 3 antisense RNA 1 (KLF3-AS1), like HOTTIP, was also shown to be expressed at low levels in degenerative NP tissues. Overexpression of KLF3-AS1 can increase NP cell viability, prevent cell apoptosis and increase ECM synthesis (70).

Shang et al (71) showed that docosahexaenoic acid, which plays a protective role in numerous chronic diseases, can alleviate excessive ECM degradation by reducing nuclear paraspeckle assembly transcript 1 overexpression in degenerative NP tissues. This previous study also revealed that ZNFX1 antisense RNA 1 (ZFAS1) was upregulated in degenerative NP tissues and NP cells treated with IL-1β, and that ZFAS1 competitively bound to miR-4711-5p through a ceRNA mechanism, thereby affecting the expression of adaptor-associated kinase 1, which is the target of miR-4711-5p. Silencing ZFAS1 may inhibit IDD progression by reducing NP apoptosis and ECM degradation (72). Similar to most lncRNAs, long intergenic nonprotein coding RNA 284 was revealed to be upregulated in IDD tissues and IL-1β-induced NP cells, and it activated the Wnt/β-catenin pathway by sponging miR-205-3p, thereby inhibiting NP cell proliferation and ECM synthesis (73). Zhan et al (74) showed that Hox transcript antisense intergenic RNA (HOTAIR) expression was positively associated with the IDD grade. Oe-HOTAIR enhanced NP cell autophagy, and promoted NP cell ECM degradation, apoptosis and senescence (74). A subsequent study found that HOTAIR regulates ECM synthesis and degradation in NP cells (75). In addition, inhibiting HOTAIR in in vivo animal models can effectively reduce IDD symptoms (75). In general, lncRNAs, miRNAs and circRNAs form an intricate regulatory network that jointly affects the synthesis and degradation of the ECM in IDD and IDD progression (Table I and Fig. 3).

Owing to different damage factors, NP cells die through a number of the same mechanisms as other cells, such as apoptosis, pyroptosis, autophagy, ferroptosis and necrosis (76). Currently, the most studied process is NP cell apoptosis. As aforementioned, PART1 not only affects the synthesis and degradation of the ECM through the miR-93/matrix metalloproteinase-2 signalling axis, but increased PART1 expression in degenerated tissues can also promote the apoptosis of NP cells and promote IDD (67). Similarly, HOTTIP, while preventing ECM degradation through a ceRNA mechanism, can also inhibit the apoptosis of NP cells and promote proliferation, improving the progression of IDD (69). In degenerated NP cells, small nucleolar RNA host gene 6 (SNHG6) expression has been reported to be upregulated and SNHG6 can induce NP cell apoptosis by targeting miR-101-3p (77). Chen et al (70) illustrated the possible impact of KLF3-AS1 on IDD. Oe-KLF3-AS1 enhanced NP cell viability and prevented apoptosis. The experimental results confirmed that KLF3-AS1 overexpression improved degenerative changes in NP cells through the miR-10a-3p/zinc finger and BTB domain-containing protein 20 (ZBTB20) axis. Yu and Li (78) detected low expression of membrane associated guanylate kinase, WW and PDZ domain containing 2 (MAGI2)-AS3 and IL-10, and high expression of miR-374b-5p in lipopolysaccharide (LPS)-induced NP cells. Notably, miR-374b-5p is a target of MAGI2-AS3 and IL-10. MAGI2-AS3 can increase the expression of IL-10 by competing with miR-374b-5p, thereby reducing the inflammatory response and cell apoptosis. Thus, the regulation of NP cell apoptosis by lncRNAs is mainly based on the ceRNA mechanism.

LncRNAs regulate the apoptosis of NP cells and affect pyroptosis and autophagy. For instance, MIR155 host gene (MIR155HG) has been shown to be upregulated in degenerated human NP tissues, and further experiments revealed that MIR155HG sponges miR-223-3p and promotes NLRP3 expression, thereby inducing pyroptosis in NP cells (79). Furthermore, platelet-rich plasma-derived extracellular vesicles may inhibit tert-butyl hydroperoxide-induced NP cell injury by upregulating metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) expression; MALAT1 regulates the miR-217/silent information regulator sirtuin 1 (SIRT1) signalling axis, and its overexpression can alleviate NP cell pyroptosis (80). As aforementioned, HOTAIR can affect the synthesis and decomposition of the ECM in NP cells and the apoptosis of NP cells. Simultaneously, HOTAIR can also affect autophagy and alter the progression of IDD (75). As the first lncRNA discovered (81), H19 imprinted maternally expressed transcript (H19) is highly conserved and widely expressed in mammals. Sun et al (82) found that H19 was highly expressed in degenerated NP tissues, and experimentally verified that H19 can promote autophagy and apoptosis in NP cells through the miR-139-3p/C-X-C motif chemokine receptor type 4/nuclear factor-κB (NF-κB) signalling axis, thereby affecting IDD (Table I and Fig. 3).

Inflammation can accelerate the process of IDD and cause damage and dehydration of the IVD tissue, resulting in loss of its original elasticity and function. Therefore, preventing and controlling inflammation is crucial for treating IDD (83). LncRNAs are important molecules that regulate inflammatory responses. Jiang et al (84) found that IDD tissues showed decreased expression of family with sequence similarity 83 member H (FAM83H)-AS1 compared with normal tissues, and Oe-FAM83H-AS1 could promote NP-cell proliferation, and reduce inflammatory responses and ECM degradation. In addition, miR-22-3p mediated the effect of FAM83H-AS1 on degenerated NP tissue (84). In IL-1β-stimulated NP cells, the expression of HLA complex group 18 (HCG18) and follistatin-like protein 1 (FSTL1) has been shown to be increased. Subsequent overexpression experiments demonstrated the impact of the HCG18/miR-495-3p/FSTL1 signalling axis on NP tissue inflammation and apoptosis (85). Zhang et al (86) showed that melatonin can reduce inflammation and apoptosis in NP cells while upregulating maternally expressed gene 3 (MEG3) expression, whereas MEG3 expression was shown to be downregulated in the degenerated NP. The study revealed that melatonin can reduce NP cell inflammation and apoptosis through the MEG3/miR-15a-5p/peroxisome proliferator-activated receptor γ coactivator 1α/SIRT1 pathway. Similar to most lncRNAs described previously, MAGI2-AS3 also affects IDD progression through a ceRNA mechanism. MAGI2-AS3 has been reported to be poorly expressed in degenerated NP cells, whereas oe-MAGI2-AS3 can downregulate miR-374b-5p and upregulate IL-10 expression to improve inflammation and ECM degradation in NP cells (78). Exosomes derived from bone marrow mesenchymal stem cells (BMSC-Exos) have therapeutic effects on IDD. In co-culture experiments, BMSC-Exos delivered colon adenocarcinoma hypermethylated (CAHM) to inhibit the polarisation of M1 macrophages, thereby reducing the apoptosis of degenerated NP cells, ECM degradation and the expression of inflammatory factors, and improving IDD (87). In addition to the aforementioned lncRNAs, numerous lncRNAs discovered in previous studies also have important effects on the occurrence and development of IDD (56), requiring further verification experiments (Table I and Fig. 3).

Senescence of NP cells plays an important role in the development of IDD. During ageing, the regenerative capacity of NP cells weakens, apoptosis and inflammation increase, and cell metabolism becomes abnormal (65), ultimately promoting the occurrence of IDD under the combined action of various factors. In vivo and in vitro experiments have shown that the expression levels of adipogenesis regulatory factor (ADIRF)-AS1 and SERPINA1 are downregulated in high-grade degenerated NP tissues and that both have binding sites for miR-214-3p. Subsequent experiments revealed that ADIRF-AS1 can bind to miR-214-3p, thereby increasing SERPINA1 expression and ultimately delaying NP cell ageing and apoptosis (88). As aforementioned, HOTAIR can promote the apoptosis and ECM degradation of NP cells, and accelerate the ageing of NP cells, which can comprehensively affect the progress of IDD in these numerous aspects (74). Furthermore, transient receptor potential canonical 7 RNA 1-AS1 adsorbs miR-4769-5p through the classical ceRNA mechanism, inhibiting HPN and regulating ageing, vitality and ECM synthesis in NP cells (89). The methylation level of lncRNA activated by DNA damage (NORAD) has been shown to be significantly increased in ageing NP cells, and the expression of WTAP was revealed to be increased in the degenerated NP, significantly promoting the m6 modification of NORAD; the lack of NORAD can promote cellular senescence by affecting the expression of Pumilio-homology domain (90).

In general, research on the mechanisms of lncRNAs in IDD has made great progress since the discovery of the first lncRNAs, H19, and the mechanisms of numerous lncRNAs in IDD have been clarified step-by-step. The same lncRNAs can control IDD through different signalling channels. Different lncRNAs can also function through the same signalling pathway. The impact of lncRNAs on NP cells has also impacted a number of aspects, such as apoptosis, ECM degradation and synthesis, inflammatory response, autophagy and ageing. lncRNAs form a large and complex regulatory network with miRNAs, circRNAs and proteins. To date, most studies have focused on the role of lncRNAs in nucleus pulposus, and there has been little research on CEP and AF. Future research should endeavour to further the understanding of CEP and AF (Table I and Fig. 3).