lncRNAs: a new generation of targets and biomarkers in thyroid cancer.

Introduction
lncRNAs are non-translated transcripts longer than 200 bp (or 500 bp according to a recent consensus statement paper), encoded by intergenic regions or overlapping, partially or entirely, with protein-coding genes; however, despite the proximity, their transcriptional regulation often remains independent of nearby protein-coding genes (1). lncRNAs are involved in a wide range of physiological and pathological processes by regulating gene expression through in cis or in trans mechanisms: they can act at the epigenetic, transcriptional, post-transcriptional, translational, and post-translational levels. These molecules can be classified by their mode of action into the following:Decoys: sponge RNAs or miRNAs preventing their binding to targets.

Guides: direct proteins to specific genome regions, through sequence complementarity, to regulate transcription or chromatin structure.

Scaffolds: promote protein–protein, protein–DNA, or protein–RNA interactions through binding by base-pairing of their primary sequence and/or by three-dimensional structures (2) (Fig. 1).

Unequivocally, most of the lncRNAs have been identified through a precise and straightforward experimental scheme as miRNA sponges, while there is still limited information on lncRNAs acting as guides or scaffolds due to the lack of standardized experimental tools. This implies that the unexplored aspects of these molecules could conceal significant information that can be utilized to better address these neoplasms.
Due to their diverse roles, lncRNAs are crucial in numerous biological processes, and their dysregulation significantly contributes to disease pathogenesis. Many are associated with the onset and progression of cancer through altered expression and/or mutations in their genes, functioning as oncogenes or tumor suppressors.
Thyroid cancer (TC) is the most common endocrine tumor and the seventh most prevalent malignancy worldwide, with 821,214 cases diagnosed and 47,507 deaths annually (0.44% of all cancers) (WHO, Feb 2025). TC incidence has risen globally, although mortality has slightly declined. This observation led to the hypothesis that the higher incidence may be due to advancements in detection methodologies (3). Over 90% of diagnosed TCs are well-differentiated papillary, follicular, or oncocytic and generally have favorable outcomes (4). TC cells can accumulate genetic mutations over time, resulting in rarer and more aggressive subtypes, including poorly differentiated and anaplastic thyroid carcinomas (5). Depending on the histotype and aggressiveness of the tumor, therapeutic approaches can vary, from surgery to radiotherapy to targeted therapy. This underscores the importance of identifying molecular targets to enhance our ability to combat tumor cells and utilize them as prognostic markers. Indeed, expanding our understanding of tumor development is foundational for improving treatments and follow-up care. This review outlines current knowledge on lncRNAs in TC, distinguishing between upregulated (oncogenes) (Table 1) and downregulated (tumor suppressors) transcripts (Table 2), also highlighting their potential application as therapeutic, prognostic, or diagnostic factors.

Upregulated lncRNAs

HOTAIR
Homeobox transcript antisense RNA (HOTAIR) is transcribed from the antisense strand of the HOXC cluster locus, serving as a scaffold for two key histone modification complexes. It binds to the polycomb repressive complex 2 (PRC2) to epigenetically repress transcription from the HOXD locus, and it interacts with the lysine-specific demethylase 1 (LSD1)-CoREST complex to remove histone markers linked to gene activation (6). Recently, HOTAIR has been implicated in TC with two independent studies highlighting its upregulation in TC samples (7, 8). Zhang et al. later showed HOTAIR is also present in the plasma of TC patients, and, through in vitro analyses, they demonstrated that it acts as an oncogene, controlling cancer proliferation and invasion, linking its expression to the progression and prognosis of thyroid carcinoma (7). Furthermore, HOTAIR overexpression in the lymph node metastasis of papillary thyroid cancer (PTC) suggests its role in epithelial-to-mesenchymal transition (EMT) via the Wnt/β-catenin pathway, by inducing SNAIl and ZEB1, which enhance migration. Silencing HOTAIR downregulates β-catenin and enhances Wnt inhibitor 1 (WIF1), reducing invasiveness (9). Furthermore, it sponges tumor suppressive miRNAs, such as miR-1, thus enhancing cyclin D2 (CCND-2) and contributing to tumor progression (8). Other target miRNAs are miR-17-5p and miR-761, enhancing cell viability, migration, and invasion (10). While miR-17-5p exact mechanism is still unknown, miR-761 binds protein phosphatase 2A (PP2A)-specific methyl esterase (PPME1). This protein, a molecular target in various tumors, is positively correlated with Erk signaling (11): HOTAIR upregulation limits miR-791 availability, reinforcing PPME1’s promotion of the oncogenic Erk pathway (12). In addition, HOTAIR baits miR-488-5p, boosting nucleoporin 205 (NUP205) and the apoptotic protein BCL-2, thereby aiding tumor growth (13). Under hypoxia, a crucial tumor promoting condition (14), HOTAIR recruits RELA, a transcription factor from the NFκB/RelA family, which in turn activates miR-181a. The latter is a component of cancer-secreted exosomes, recently identified as an oncomiR in PTC (15). Presumably, miR-181a guides hypoxia-induced angiogenesis by diminishing GATA6 levels in target cells (16). Overall, HOTAIR’s upregulation is tied to multiple oncogenic pathways, highlighting its potential as a diagnostic and therapeutic target in TC.

ABHD11-AS1
Located on chromosome 7, ABHD11-AS1 regulates cell proliferation, migration, and invasion in several cancers, and recently, it has been linked to TC. In PTC, its serum levels correlate with tumor size, stage, and lymph node metastasis, suggesting a link to poor prognosis (17). Moreover, in vitro and in vivo ABHD11-AS1 knockdown inhibits cell proliferation and promotes apoptosis while also restraining tumor growth and metastasis. At the molecular level, the cancer-related protein STAT3 was found to be responsible for the transcriptional activation of ABHD11-AS1 in PTC; conversely, the lncRNA sponges miR-1301-3p, which causes upregulation of STAT3 mRNA levels. This mechanism defines a positive feedback loop involving ABHD11-AS1 and STAT3 that activates the downstream tumor-promoting PI3K/AKT pathway, thereby elucidating the oncogenic function of this lncRNA (18). Furthermore, a recent study highlighted a link between ABHD11-AS1 and EPS15L1, a substrate of EGFR tyrosine kinase activity involved in the regulation of cell proliferation, differentiation, growth, and survival. It was shown that overexpression of ABHD11-AS1 enhances EGFR, STAT3, p-STAT3, and EPS15L1, although the specific miRNA regulating this entire process remains unidentified (17). Finally, it is known that ABHD11-AS1 overexpression in advanced-stage PTC causes a miR-199a-5p-mediated increase in SLC1A5/ASCT2 levels. This molecule is a Na+-dependent neutral amino acid transporter that functions as an oncogene in many human cancers, primarily by transporting glutamine to support tumor growth (19). SLC1A5 is expressed exclusively by BRAF p.V600E mutated tumor cells in the thyroid (20). These data indicate that ABHD11-AS1 is a tumor promoter in TC, and it could be employed as a diagnostic and prognostic factor; moreover, its interplay with the PI3K/AKT pathway and glutamine transport paves the way for considering it as a therapeutic target.

ZFAS1
ZFAS1 (ZNFX1-antisense-RNA1) is transcribed antisense to the protein-coding ZNFX1 gene and accommodates three small nucleolar RNAs. It has been depicted as tumor suppressor in some tumors, while it acts as an oncogene in others, such as colorectal and gastric tumors (21, 22). Recently, ZFAS1 has been found overexpressed in human TC tissues, and its upregulation occurs from the early stages of neoplastic transformation, correlating with TNM stage, lymph node metastasis, and recurrence. In human anaplastic TC (ATC) cells, ZFAS1 knockdown decreased proliferation and cell cycle arrest: bioinformatic analyses predicted that it may be part of a competing endogenous RNA (ceRNA) circuitry involving miR-150-5p and miR-590-3p, along with more than one hundred mRNAs associated with DNA replication, ribosome function, transcription, translation, ubiquitin-mediated proteolysis, and sister chromatid separation. Despite the need for further validations, these findings suggest that ZFAS1 may serve as a promising biomarker and prognostic factor for TC (23). In PTC, it stimulates proliferation while inhibiting apoptosis, promoting tumor growth in vivo. Also here, it sponges the tumor suppressor miR-590-3p, thereby upregulating HMGA2, a transcriptional regulatory factor well-recognized as a biomarker for TC (24). ZFAS1 is a direct p53 target: in its wild-type form, TP53 suppresses ZFAS1 via miR-135b-5p and miR-193a-3p, whereas mutated TP53 fails to do so, contributing to oncogenesis (25). Doubtlessly, this lncRNA can be considered pivotal for TC development and evolution, although plenty of information still needs to be uncovered. Indeed, the ZFAS1/miR-590-3p/HMGA2 axis or the interaction with p53 is probably only a small piece in the complex mechanisms linking ZFAS1 to TC.

AFAP1-AS1 and LINC00514
Actin filament-associated protein 1-antisense RNA 1 (AFAP1-AS1) is a tumor promoter in many malignancies, such as esophageal adenocarcinoma and colorectal cancers (26, 27). Recent findings indicate that AFAP1-AS1 is overexpressed in ATC, where it sponges miR-155-5p, derepressing ETS1, a transcription factor that regulates many cancer-related genes, including RAS, MET, and also ERK phosphorylation (28). In differentiated TC, AFAP1-AS1 expression is also elevated, with lower levels of this lncRNA correlating to better survival rates, thus establishing a new prognostic marker. In vitro analyses of human differentiated TC cell lines revealed that AFAP1-AS1 knockdown inhibited tumor growth, promoted apoptosis, and hindered migration through EMT, although no molecular mechanisms were described (29). A separate study showed that AFAP1-AS1 overexpression in TC tissues sequesters miR-204-3p, which normally regulates dual specificity phosphatase 4 (DUSP4) mRNA levels, thereby leading to its enhancement (30). DUSP4, a member of the mitogen-activated protein kinase phosphatase 2 (MKP2) family, plays a crucial role in dephosphorylating and inactivating MAPKs, thereby fine-tuning proliferation (31). Moreover, in PTC, DUSP4 overexpression correlates with BRAF p.V600E mutation, highlighting a potential biomarker for tumor aggressiveness associated with cancer progression and metastasis formation (32). Notably, miR-204-3p is part of another ceRNA network involving LINC00514 and CDC23, which enhances the latter, thereby driving PTC progression (33).

DOCK9-AS2
DOCK9 antisense RNA2 (DOCK9-AS2) is an exosomal lncRNA. Exosomes are small vesicles allowing cell to exchange molecules (34). Cancer cells exploit this communication mechanism to impact their environment; more specifically, cancer stem cells (CSCs) and non-CSCs use exosomes to affect each another, establishing a dynamic communication network (35). The interesting aspect of DOCK9-AS2 is that, as an exosomal lncRNA, it can be delivered, exerting oncogenic functions in target cells. DOCK9-AS2 is upregulated in PTC tissues and cell lines and has been detected in the exosomes of PTC patients. This is particularly relevant since detection of circulating lncRNAs, such as HOTAIR and DOCK9-AS2, opens avenues for non-invasive diagnostics and prognostic tools, considering also that in TC fine-needle aspiration cytology can sometimes yield inconclusive results. At the molecular level, it has been demonstrated that DOCK9-AS2 enhances the Wnt/β-catenin pathway in two independent ways: in the nucleus, it recruits the tumor promoter transcriptional factor SP1 to promote CTNNB1 transcription; in the cytoplasm, it sponges miR-1972 to stabilize CTNNB1 mRNA (36, 37, 38). This evidence locates DOCK9-AS2 as the first known lncRNA used as a signal by TC cells. Moreover, its implication in β-catenin regulation involves it in cancer progression and invasiveness.

NPSR1-AS1
Neuropeptide S receptor 1 antisense RNA 1 (NPSR1-AS1), antisense to the NPSR1 gene, is a rare case of scaffold lncRNA promoting proliferation, migration, and invasion in TC. Its silencing reduced these malignant behaviors and inhibited the EMT by increasing E-cadherin and decreasing N-cadherin and Vimentin levels (39). Mechanistically, NPSR1-AS1 binds to the RNA-binding protein ELAVL1, stabilizing the mRNA of its nearby gene NPSR1. Elevated NPSR1 then activates the MAPK signaling pathway, contributing to TC progression. Thus, NPSR1-AS1 enhances NPSR1 expression and MAPK pathway activation via ELAVL1 interaction (39).

TNRC6-AS1
TNRC6-AS1 is a lncRNA transcribed from the reverse strand of trinucleotide repeat-containing 6C (TNRC6C). A recent study reported the oncogenic nature of TNRC6-AS1 in PTC, linking its overexpression to increased proliferation, migration, and invasion rates. Interestingly, it is inversely correlated with the coding gene TNRC6C, which is repressed in thyroid tumor samples. Further analyses revealed that TNRC6-AS1 affects the protein-coding gene, and they both influence iodine metabolism genes, such as TSH-R, SLC5A5, TPO, and SLC26A4, as the restoration of physiological expression levels of the two resulted in the rescue of the mentioned iodine-related genes. These findings may offer insights to improve response to radiotherapy, given that the iodine genes are essential for its outcomes and functioning (40). Furthermore, in PTC, TNRC6-AS1 guides DNA methyltransferases on the CpG islands of the target gene serine/threonine-protein kinase 4 (STK4) promoter (41). Hypermethylation of CpG islands leads to silencing of the downstream gene, which can serve as a mechanism for the knockdown of tumor suppressor genes during cancer development and progression (42). STK4 downmodulation has been linked to the nuclear translocation of Yes-associated protein (YAP), promoting cell proliferation while inhibiting apoptosis and autophagy (41). YAP is a component of the Hippo pathway, a crucial signaling network for tissue development and regeneration, which is often dysregulated in cancer (43).

Downregulated lncRNAs

GAS8-AS1
GAS8-AS1 is an antisense lncRNA transcribed on the opposite strand of the GAS8 gene, with its transcription starting from a promoter region that remains to be fully characterized. This lncRNA was initially mapped on a region on chromosome 16q that is frequently deleted in prostate and breast cancers (44). The first evidence depicting GAS8-AS1 as a tumor suppressor emerged from a whole-exome sequencing analysis conducted by Pan et al. (45) on a Chinese cohort of 91 paired PTC and normal tissues, in which this lncRNA was the second most frequently mutated gene with a significantly low expression in PTC samples (45). Upon the ectopic expression of wild-type GAS8-AS1, Pan and colleagues discovered that the proliferation of PTC and MTC cell lines was significantly hindered compared to controls, while the opposite occurred when the lncRNA was targeted by siRNAs (45). These findings were independently reproduced in hepatocellular carcinoma (46) and colorectal cancer (47) cell lines. Chen et al. found that GAS8-AS1 indirectly regulates the stability of cyclin G2 mRNA (CCNG2) by sponging miR-135b-5p, a well-known oncomiR in many cancers, including PTC and MTC (48). This likely leads to an accumulation of cyclin G2 that delays cell cycle progression, which was measured in both cultured cells and tumor masses from nude mice grafted with GAS8-AS1-silenced cells (49). In addition, GAS8-AS1 can promote uncontrolled proliferation and invasiveness in ATC (29, 50). Although the precise mechanism is unclear, both Zhao et al. (47) and Zha et al. (51) suggest that it can regulate transcription by recruiting epigenetic regulators. Indeed, this lncRNA can guide the epigenetic writer complex MLL1/WDR5 to trimethylate H3K4s on the GAS8 gene promoter, facilitating the transcription of this tumor suppressor gene, which in turn regulates microtubule sliding during mitotic spindle assembly (46). Furthermore, GAS8-AS1 also plays a role in the activation of autophagy, which is frequently hijacked by aggressive tumors as a form of metabolic adaptation to the new necessities of metastasizing cancer. Stress-related p38 activation of ATF2, a transcription factor commonly stimulated during autophagy, promotes the expression of GAS8-AS1, which prompts two effects. On the one hand, it hinders cell cycle by sponging miR-135b-5p, leading to the accumulation of cyclin G2. On the other hand, GAS8-AS1 sponges miR-187-3p and miR-1343-3p, which downregulate ATG5 and ATG7, respectively, necessary for expanding the newly formed autophagosome (52). As Qin et al. demonstrated, the loss of GAS8-AS1 in PTC removes one of the safeguard mechanisms against tumor progression while effectively rescuing its expression and sensitizes the tumor cells to autophagy, simultaneously decreasing the proliferation rate (53). Collectively, these findings suggest that alterations in GAS8-AS1 function likely occur earlier in cancer onset, probably while the tumor remains well-differentiated rather than in aggressive tumors, as genetic hits to this tumor suppressor may help remove obstacles that advanced-stage tumors consider already surpassed.

PTCSC3
PTCSC3 is a tumor suppressor lncRNA, identified in thyroid after careful investigation of the surrounding locus in linkage with rs944289, a tag SNP on chromosome 14q13.3. Indeed, the systematic characterization of this lncRNA was conducted by Jendrzejewski et al., who were the first to elucidate the biological significance of rs944289 as occurring in a proximal enhancer of a 1154 bp long non-coding RNA gene thereafter named papillary thyroid carcinoma susceptibility candidate 3 (54). Shortly after, Fan et al. studied the effects of reintroducing PTCSC3 in thyroid carcinoma cell lines, where its expression is decreased. They reported the inhibition of cell growth upon PTCSC3 recovery and identified oncomiR-574-5p as an interactor of this lncRNA (55). PTCSC3 restoration negatively regulates the expression of the EF-hand calcium-binding protein S100A4, thereby limiting tumor invasiveness and reducing angiogenesis and extracellular matrix remodeling (56). PTCSC3 also interferes with the dysregulated oncogenic Wnt/β-catenin pathway, which provides the survival advantage often observed in PTC (57). At least two research groups have linked tumor suppressor PTCSC3 deregulation to the Wnt/β-catenin pathway, although parts of this complex regulatory process remain unknown. Xia et al. described the epistatic control of PTCSC3 on the expression of Frizzled co-receptor LRP6 and scaffolder protein Axin in glioma cells (58). They showed that PTCSC3 overexpression reduces LRP6 abundance while simultaneously exerting a significant positive regulation on both Axin transcript and protein. Consequently, the Wnt pathway becomes strongly desensitized by PTCSC3 rescue, as the receptor is less prone to activation; meanwhile, the increased amount of Axin in the cytoplasm effectively targets β-catenin for degradation, reducing the expression of target genes c-MYC and CCND1. Axin also seizes YAP/TAZ in the cytoplasm, thereby lowering the expression of EMT genes SNAI1 and ZEB while promoting the expression of CDH1 in the process (58, 59). Later, Wang et al. demonstrated how PTCSC3 stabilizes the transcript of suppressor of cancer cell invasion, SCAI, from miR-574-5p and explained how this pioneering transcription factor can recruit the SWI/SNF complex to suppress gene expression from both pathways (60, 61). The same conclusions, albeit in cervical carcinoma, have been independently reached by both Zhang et al. and Tong et al. (62, 63), further confirming this mechanism. Even more interesting is PTCSC3 ability to directly modulate the half-life of STAT3 protein and its response to cytotoxic drugs. Indeed, Wang et al. discovered the presence of STAT3 protein in PTCSC3 pulldown, and upon overexpression, the protein abundance of STAT3 decreased to 1/5 of baseline levels. This, in turn, triggered a downregulation of the downstream INO80 gene, increasing susceptibility to doxorubicin. ATC cells in which PTCSC3 expression was restored expressed less P-glycoprotein (MDR1) and were more susceptible to doxorubicin treatment (64). Although the understanding of this lncRNA is far from complete, the evidence gathered thus far depicts PTCSC3 as a fundamental ceRNA and epigenetic regulator that successfully restrains thyroid carcinoma from progressing into a more aggressive form.

RMST
Rhabdomyosarcoma 2 associated transcript, or RMST, was first discovered in malignant soft tissue tumors with various degrees of musculoskeletal differentiation. Its sequence was cleverly reconstructed using library screening and RACE, after which a northern blot revealed a 1.25 kb long transcript mapping to chromosome 12q21 (65). The work of Uhde et al. and Chordoff et al. established RMST function in the brain, with the former group observing RMST expression extending from the dorsal midline to the most anterior tip of the telencephalon in the developing mouse brain, ultimately restricted to midline dopaminergic neurons in the adult mouse brain (66). Meanwhile, the latter group described the overall high conservation among orthologs from humans to frogs (67). Yakushina et al., by examining eight microarray datasets and two validation RNA-seq datasets, demonstrated that RMST downregulation is a characteristic feature of ATC, linking dysregulated RMST to cancer progression and aggressiveness (68). De Martino et al. first systematically characterized RMST in thyroid carcinoma, showing its tumor-suppressive role in aggressive ATC and suggesting that its high expression in thyroid helps maintain thyroid cell differentiation (69). The authors showed that restoring RMST in ATC cells where it is absent leads to a sharp decrease in stemness markers, including NANOG, OCT4, and SOX2, resulting in a reduced proliferation capacity and increased apoptotic events, clearly suggested by the delayed decline in cell number and shrinkage of tumoral thyrospheres. Furthermore, they demonstrated how RMST can repress the mesenchymal markers SNAI1, SLUG, and TWIST1, significantly restraining the migration and invasion capacity of ATC cells. The evidence gathered aligns well with the proposed and verified mechanism of action of RMST in physiological development and pathological events, reaffirming its role as a key tumor suppressor in thyroid tissue (70, 71, 72).

PAR5
PAR5 is one of the transcripts characterized by Sutcliffe et al. originated from the Prader-Willi/Angelman region (PAR) on chromosome 15q11-13, whose alterations are responsible for the imprinting disorders Prader-Willi and Angelman syndromes. PAR5 was initially identified through the northern blot as an ∼12 kb RNA enriched in skeletal muscle and brain (73). PAR5 was first studied in high-grade glioma and glioblastoma by Zhang et al., where its downregulation might help stratify patients with worse prognoses (74, 75). PAR5 function is closely linked to that of EZH2. The interaction between lncRNAs and EZH2 is commonly observed, as this protein regularly binds MALAT1 (76), XIST (77), and HOTAIR (6), although the dynamics by which PAR5 modifies the PRC2 complex through its association with EZH2 depends on the circumstances. According to Wang et al., PAR5 directly interacts with EZH2 and SUZ12 in a manner that helps stabilize the PRC2 complex, resulting in a bona fide increased methylation activity on its targets, although they did not evaluate regulatory methylation on gene loci per se (78). This exploration, conducted in glioma-cultured cells, also revealed that restoring PAR5 significantly reduced EGFR, VEGF-A, and cyclin A levels while decreasing AKT phosphorylation, highlighting this lncRNA tumor suppressor’s ability to direct methylation suppression on these oncogenes. The function observed by Wang and colleagues matches with the role of a trans-acting lncRNA already established in other instances (6, 77, 79), making PAR5 appear mostly as a scaffolder/guide RNA of the PRC2 complex, although this interaction can be largely dependent on tissue-specific context. Indeed, on one pivotal work regarding the PAR5/EZH2 axis in thyroid oncology, Pellecchia et al. proposed a different mechanism for this lncRNA–protein complex. They demonstrated that PAR5 downregulation is exclusive to ATC, and upon restoring PAR5 in ATC cultured cells, it negatively impacts EZH2 stability through interaction while also significantly reducing PRC2-mediated H3K27 trimethylation around the CDH1 promoter, consequently affecting proliferation and invasiveness (80).

KLHL14-AS1
KLHL14-AS1 is an intriguing thyroid-associated lncRNA, not only because it plays a tumor suppressor role but also for its specific expression in thyroid since the very early stages of thyrocyte specification during development. The first evidence of Klhl14-AS1 emerged from its detection in a microarray of laser micro-dissected E10.5 developing mouse thyroid primordia by Fagman et al. (81). They discovered that Klhl14-AS1 was the most enriched transcript (still under the RIKEN provisional name 4930426D05Rik) in the thyroid bud compared to the whole embryo, alongside Bcl2 and Pax8. Afterward, Credendino et al. investigated the expression profile of Klhl14-AS1 in adult mouse, showing that it is expressed in various tissues and confirming that mouse Klhl14-AS1 is, in fact, highly conserved among mammals, to the extent that the entire locus and its vicinity are syntenic between humans and mice (82). Subsequent analysis by the same group revealed the role of mouse and human KLHL14-AS1 as a tumor suppressor in TC (83). They demonstrated that downregulation of KLHL14-AS1 increases cell viability and proliferation, and through RNA pulldown assays, they showed that it serves as a decoy for miR-182-5p and miR-20a-5p, which target the mRNA of two thyroid differentiation regulators, PAX8 and BCL2. Interestingly, the interaction between KLHL14-AS1 and these miRNAs is conserved across humans and rodents, emphasizing its significance in thyroid physiology and pathology.

BGLT3
Beta globin locus transcript 3 (BGLT3) stands as an outsider in relation to the thyroid, as it subjected to the same transcriptional regulation as the β-globin-like locus on chromosome 11p15.4. In thyroid, HBB, the β-globin gene, is expressed in very low amounts and does not produce a significant quantity of protein, while BGLT3 expression and function in erythroid cells is closely associated with the BGLT3 and HBG2 γ-globin genes, which are generally repressed in favor of the HBB gene upon the fetal-to-adult hemoglobin switch. Indeed, HBG1/HBG2 gene expression is nearly zero, meaning that also BGLT3 should go undetected. Intriguingly, BGLT3 expression is regulated in normal thyroid by HIF1α, which is positively modulated by the PI3K and MAPK pathways (84). In the early phases of thyroid oncogenesis, BRAF p.V600E or RET/PTC oncogenes drive the expression of HIF1α (85), although the concomitant overexpression of c-MYC, belonging to the same signaling pathway, quickly overturns HIF1α regulation by suppressing BGLT3 transcription at its locus (86). The current literature provides only one instance that explores BGLT3 as a tumor suppressor in TC, namely the study by Zhao et al. In addition to demonstrating the repressive effect of c-MYC on the BGLT3 promoter, they observed a pronounced attenuation of cellular proliferation, growth, and migration of PTC cells in which they restored BGLT3 expression, attributing these effects to the increased stability of PTEN. They initially co-precipitated PTEN via RNA pulldown of BGLT3, proving the interaction between them. Subsequently, they reported PTEN reduced ubiquitylation, attributed to the interaction with this lncRNA–protein complex and OTUD3 deubiquitylase (87). However, presenting OTUD3 to PTEN is not the sole relevant function of BGLT3 in thyroid carcinogenesis as this lncRNA is also involved in DNA repair. Hu et al. found that silencing BGLT3 caused severe chromosomal instability, with an increase in chromosome breaks, radial figures, and dicentric chromosomes (88). Interestingly, BGLT3 is also a scaffold for the major protein involved in homologous recombination as the lncRNA binds to the C-terminal BCRT domain of BARD1, which then interacts with BRCA1, facilitating the recruitment of RAD51 to initiate strand recognition. Essentially, BGLT3 acts as a tether that stabilizes the BRCA1/BARD1 complex to damaged DNA. In addition, BGLT3 interacts with PARP1, directly bringing the lncRNA to sites of damage. Although this role of BGLT3 has not been directly studied in TC, it holds significance because poorly differentiated thyroid carcinoma and ATC – the more aggressive and undifferentiated forms of TC – often exhibit complex genomes and chromosomal instability due to impaired DNA repair mechanisms (89).

ASMTL-AS1
Acetylserotonin O-methyltransferase-like antisense RNA 1, or ASMTL-AS1, is a novel lncRNA that may play a role as either an oncogene or a tumor suppressor, depending on the cancer type. As a tumor suppressor, it has been shown to sponge miR-1228-3p, increase SOX17 mRNAs, and suppress β-catenin expression in triple‐negative breast cancer (90). In PTC, ASMTL-AS1 is frequently reported to be highly downregulated, which generally worsens the prognosis. Feng et al. found that ASMTL-AS1 primarily resides in the cytoplasm of non-tumorigenic thyroid cells, where it can increase FOXO1 expression by acting as a sponge for miR-93-3p and miR-660, thereby inhibiting their repressive activity on FOXO1 mRNAs (91). The ectopic expression of FOXO1 in PTC cells in which ASMTL-AS1 is knocked down blocked the enhanced glycolysis induced by ASMTL-AS1 loss, showing that FOXO1 mediates the inhibitory effects of the lncRNA on PTC cell proliferation. Interestingly, in PTC cells, ASMTL-AS1 is transcriptionally activated by FOXO1 itself, creating a positive feedback regulatory loop that is crucial for suppressing PTC glycolysis and growth. Lower levels of ASMTL-AS1 in PTC are associated with overexpression of miR-93-3p and miR-660, alongside reduced FOXO1 levels, leading to an increased glycolysis rate that provides sufficient energy for the uncontrolled growth of cancer cells.

CATIP-AS1
lncRNA CATIP-AS1, or ciliogenesis associated TTC17 interacting protein antisense RNA 1, is significantly downregulated in both PTC and ATC, with its reduced expression correlating with a poor patient prognosis. According to Qi et al., CATIP-AS1 functions by directly sponging miR-515-5p in TC cells, thereby inhibiting its regulatory effect on the target mRNA SMAD4, a crucial mediator in the TGF-β signaling pathway (92). The upregulation of SMAD4 induced by overexpressed CATIP-AS1 results in increased levels of epithelial markers, such as E-cadherin and ZO-1, while simultaneously reducing the expression of mesenchymal markers, such as N-cadherin and vimentin, thus inhibiting EMT. Consequently, the overexpression of CATIP-AS1 has been shown to inhibit cell proliferation and migration while being associated with an increased rate of apoptosis, identifying CATIP-AS1 not only as a tumor suppressor but also as a potential therapeutic target for managing aggressive TCs.

MPPED2-AS1
Metallophosphoesterase domain-containing 2 antisense RNA 1, or MPPED2-AS1 for short, previously known as RP5-1024C24.1, is a lncRNA located on chromosome 11p14.1 in an antisense orientation relative to the MPPED2-AS1 gene, of which it is a natural antisense. MPPED2 encodes a metallophosphodiesterase protein with tumor suppressor functions in various cancers, including cervical cancer, neuroblastoma, glioblastoma, and oral squamous cell carcinoma. Pellecchia et al. showed that MPPED2-AS1 and MPPED2 are both significantly downregulated in benign and malignant thyroid neoplasms, including well-differentiated and undifferentiated thyroid carcinomas, proving to be strongly positively correlated with each other (93). MPPED2-AS1 positively regulates MPPED2 expression by inhibiting DNA methyltransferase 1 (DNMT1), leading to decreased hypermethylation and inactivation of the MPPED2 promoter. Overexpression of MPPED2-AS1 in thyroid carcinoma cells has been shown to reduce cell proliferation and migration, partly by upregulating PTEN and decreasing AKT phosphorylation. Similar inhibitory effects on cell growth and migration were also observed with MPPED2 overexpression, suggesting that the tumor suppressive role of MPPED2-AS1 is mediated through MPPED2 induction. These findings indicate that MPPED2-AS1, along with MPPED2, is a critical tumor suppressor in thyroid carcinogenesis.

GAS5
The growth arrest specific 5 gene, or GAS5, is transcribed into two major transcripts, named GAS5a and GAS5b, whose alternative splicing gives rise to two, among the at least 29 others, splice variants that regulate cell growth and proliferation in two opposite ways (94). Nevertheless, GAS5 appears to be mostly downregulated (95), classifying it as a tumor suppressor in most tumors. A substantial amount of literature on GAS5 demonstrates that it primarily acts as a negative regulator of the PI3K/AKT/mTOR pathway, thereby limiting the ability of the cell to proliferate (96). This is appealing to the TC discourse, as PI3K/AKT/mTOR is greatly active in RAS-driven follicular TC (FTC) and ATC. Indeed, a prospective study by Guo et al. found that GAS5 was significantly downregulated in TC, with no difference between histological types, compared to benign thyroid tissues (97). Moreover, lower GAS5 expression correlates not only with poorer prognosis but also with an increased expression of CDK6, which is particularly overexpressed in FTC and can indirectly reinforce PI3K/AKT signaling. Indeed, it was Liu and colleagues who explained this peculiar behavior by showing how GAS5, acting as a sponge for miR-221-3p, preserves the levels of CDKN2B and its encoded protein, p15INK4b, which directly inhibits CDK6 by preventing its binding with D-type cyclins, thus hampering cell cycle progression (98). Moreover, GAS5 can exert indirect control over PI3K/AKT signaling by sponging miR-196a-5p and miR-182-5p, effectively to sustain the abundance of FOXO1 and FOXO3, two transcription factors that act redundantly and synergistically to arrest cell growth (96). Finally, GAS5 influence over PI3K/AKT culminates in preserving the level of PTEN, a tumor suppressor itself that switches off the PI3K downstream signal and restricts cell proliferation (99). Recently, a preliminary study on PTC cell lines confirmed that ectopic expression of GAS5 can positively regulate genes usually induced by IFN-α/IFN-β response, as IFI44, which is part of the negative feedback loop that interferes with JAK/STAT pro-survival signaling (100). Considering its intensive control over PI3K/AKT/mTOR and JAK/STAT pathways, GAS5 is not only a good example of tumor suppressor lncRNA in thyroid but also a candidate for future RNA-based drug to take advantage of.

Clinical translation of lncRNAs in thyroid oncology
Treatments and diagnostic procedures based on noncoding RNAs have been gathering the attention of researchers for nearly three decades, resulting in observational and interventional cancer clinical trials alike increasing in number (101). To date, there are no open or completed interventional clinical trials based on noncoding RNAs conducted for TC. This is partly due to the fact that the management of most TCs, especially differentiated thyroid carcinomas, has the highest efficacy among solid tumors (SEER, 2025).
Nonetheless, aggressive and treatment-refractory subtypes, such as radioiodine-refractory and anaplastic thyroid carcinomas, could benefit from novel molecular targets. As highlighted in this review, restoration of physiological lncRNA expression can mitigate tumor aggressiveness by reducing proliferation and invasiveness or promoting apoptosis. One potential strategy involves combining the restoration of basal levels of lncRNAs implicated in cancer cell dedifferentiation (e.g. TNRC6-AS1 and Klhl14-AS) (40, 83) with radioactive iodine (131I) therapy. In this setting, the redifferentiation of cancer cells in both primary tumors and metastases could restore 131I uptake and treatment efficacy (102).
Another emerging approach could involve the rescue of lncRNAs regulating the production of proteins responsible for drug resistance and the drug itself, e.g. the case of PTCSC3 and MDR1 with doxorubicin susceptibility (64).
Current TC diagnosis and classification rely on established clinical workflows combining physical examination, serum TSH measurement, ultrasonographic risk stratification, and fine-needle aspiration biopsy (103). Nevertheless, the detectability of certain lncRNAs in patient serum (e.g. HOTAIR and DOCK9-AS2) suggests their potential as non-invasive biomarkers for post-treatment monitoring and disease surveillance (7, 36).

Conclusion and future perspectives
In this review, we outline lncRNAs whose roles in TC have been at least partially elucidated, highlighting the various pathological contexts in which they are involved. Indeed, lncRNAs are associated with many modes of action; however, a review of the available literature reveals that over 90% of characterized tumor-linked lncRNAs function as sponges. As reported in this review, numerous studies have demonstrated that the upregulation of a lncRNA is crucial for generating downstream imbalances in various targets, each acting as bait for one or more miRNAs that subsequently dysregulate additional mRNAs. This suggests that each sponge lncRNA may have a multimeric effect. Indeed, the complex network of intracellular RNAs complicates the understanding of independent interactions and characterizes the intricate molecular mechanisms underlying tumorigenesis over time. For instance, HOTAIR is a tumor promoter in TC, and independent studies have defined its involvement in malignancy, culminating in an understanding of the aberrant effects of its overexpression, which occur through sequestering miRNA-1, miRNA-17-5p, miR-761, and miR-488-5p, consequently impacting the post-transcriptional miRNA-driven regulation of many other genes (8). From another perspective, there is still a lack of information in the literature about lncRNAs functioning as scaffolds or guides in TC. Notably, it has been reported that TNRC6-AS1 acts as a tumor promoter by guiding DNA methyltransferase to the CpG island of the STK4 promoter (41), while DOCK9-AS2 directs the cancer-associated transcription factor SP1 to promote CTNNB1 transcription (36). However, the roles of these specific types of lncRNAs in TC remain largely unknown, leaving a substantial gap in our understanding.
This review aims to highlight how lncRNAs are involved in every cellular process and, therefore, contribute heterogeneously to many hallmarks of cancer cells. Nonetheless, almost all the knowledge we have up to date remains solely related to the sponging mechanism, indicating that a significant number of processes continue to be overlooked.
Considering the promising results obtained with the manipulation of many lncRNAs on the aggressiveness of cancer cells, and also the opportunity to exploit some of them as biomarkers, it is undeniable that deepening this field may represent an effective new approach to the most common endocrine malignancy.

Supplementary materials

Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding
ME research fellowship was funded by the project T3-AN-09 by the Italian Ministry of Health. GDV, RPC, and RM received no funding to declare.

Author contribution statement
RM: conceptualization (supporting). Writing – original draft (lead), and writing – review and editing (lead). ME: writing – original draft (supporting), and writing – review and editing (supporting). RPC: conceptualization (supporting), writing – review and editing (equal). GDV: conceptualization (lead), writing – review and editing (equal). All authors approved the final version of the manuscript.