LncRNA-Dependent Mechanisms of Transforming Growth Factor-β: From Tissue Fibrosis to Cancer Progression

Transforming growth factor-β (TGF-β) is a crucial pathogenic mediator of inflammatory diseases. In tissue fibrosis, TGF-β regulates the pathogenic activity of infiltrated immunocytes and promotes extracellular matrix production via de novo myofibroblast generation and kidney cell activation. In cancer, TGF-β promotes cancer invasion and metastasis by enhancing the stemness and epithelial mesenchymal transition of cancer cells. However, TGF-β is highly pleiotropic in both tissue fibrosis and cancers, and thus, direct targeting of TGF-β may also block its protective anti-inflammatory and tumor-suppressive effects, resulting in undesirable outcomes. Increasing evidence suggests the involvement of long non-coding RNAs (lncRNAs) in TGF-β-driven tissue fibrosis and cancer progression with a high cell-type and disease specificity, serving as an ideal target for therapeutic development. In this review, the mechanism and translational potential of TGF-β-associated lncRNAs in tissue fibrosis and cancer will be discussed.


Introduction
Long non-coding RNAs (lncRNAs) are transcripts with lengths of over 200 nucleotides that together with short microRNAs (miRNAs), small interfering RNAs (siRNAs), small nucleolar RNAs (snoRNAs), small nuclear RNAs (snRNAs), and PIWI-interacting RNAs (piRNAs) constitute a spectrum of non-coding RNA molecules (ncRNAs) characterized by their gene-regulating functions [1][2][3]. Of these, lncRNAs and miRNAs are two major classes of ncRNAs that participate in the pathogenesis of cancer and fibrotic diseases, as dysregulation of lncRNAs and miRNAs interferes with the control of crucial biological processes, including cell proliferation, apoptosis, and extracellular matrix homeostasis [4,5].
Since their discovery with high-throughput RNA sequencing, various biological functions of lncRNAs have been revealed, suggesting that these RNAs may be a missing piece of a complex gene-regulatory mechanism that supports higher life forms and not junk transcripts left over from evolution [6,7]. Emerging evidence indicates that lncRNAs are important for fine-tuning transcription, contributing to the regulatory network of spatial and temporal gene expression with high specificity. Thus, lncRNAs may represent an optimal target for diagnosing and treating diseases.
Transforming growth factor-β (TGF-β) is a crucial cytokine that drives the pathogenesis and development of chronic inflammatory diseases, particularly chronic kidney disease (CKD) and cancer [8][9][10][11]. TGF-β signaling is highly activated in experimental models and in patient biopsies of CKD associated with enhanced production and deposition of extracellular matrix (ECM) components like collagens and fibronectins, contributing to the disruption of tissue structure and eventually leading to end-stage renal disease (ESRD) with complete loss of function [12,13]. However, fibrotic response in a resolvable manner is an essential mechanism for repairing initial tissue injury. Therapy directly targeting TGF-β may lead to side effects in vital organs [14]. This scenario also applies to cancer in that pleiotropic cytokine TGF-β initially suppresses proliferation and induces apoptosis of cancer cells but promotes immunosuppression, angiogenesis, and cancer cell growth via stromal cells in established tumors [15,16]. Therefore, identifying pathogenic downstream mediators of TGF-β is essential for developing a specific strategy that suppresses the pathogenic activity while preserving the therapeutic and physiological activities of TGF-β. TGF-β induced lncRNAs with high spatial and temporal specificity in the pathogenesis of inflammatory disease, serving as an ideal target for developing targeted TGF-β signaling therapy [1,17,18]. In this review, the mechanism of TGF-β-induced lncRNAs in the pathogenesis of renal fibrosis and cancer will be discussed.

TGF-β1 Signaling Pathways
The TGF-β superfamily is a group of cytokines with shared properties in biosynthesis, signal transduction, and other functions. It consists of four major subfamilies and a group encompassing various divergent members. TGF-β1-3 are the three distinct isoforms; TGF-β1 lacks a TATAA box in its promoter, in contrast to TGF-β2 and TGF-β3 with TATAA boxes [19]. TGF-β1 is well recognized as a key driver of fibrosis and cancer progression, while the roles of TGF-β2 and TGF-β3 are still largely unclear [20]. TGF-β1 can be directly activated by reactive oxygen species, pH, and proteases in various contexts such as tissue injury, stress, viral infection, carcinogenesis, tissue fibrosis, and inflammation [21,22]. TGF-β1 peptide is expressed and secreted as a nonactive complex with latent TGF-β binding protein that is cleaved to release active TGF-β1 [17] for TGF-β1 receptor type II binding, triggering downstream signaling via TGF-β1 receptor type I kinase (TβRI) [23]. Smad proteins are key players in the canonical pathway, while non-Smad signalings are noncanonical pathways. Smad2/3 activated by TβRI kinase forms a heterotrimeric complex with Smad4, then translocates into the nucleus to bind with target genes to initiate transcription [8,17,24]. Simultaneously, TGF-β1 induces the expression of a Smad ubiquitin regulatory factor (Smurf) that degrades Smad7, an inhibitory Smad that competes with Smad3 and Smad2 for binding to TβRI to reinforce TGF-β1/Smad signaling [25].

TGF-β1 Signaling in Kidney Diseases
TGF-β1 is primarily involved in a dynamic pathophysiological process that leads to renal fibrosis. It is significantly upregulated in the injured kidney as a primary step in tissue scarring [8,26,27] and the progressive forms of kidney disease [28]. The latest studies have revealed the diverse roles of TGF-β1, for instance as a major inducer of macrophage polarization [23], myofibroblast differentiation, and accumulation in the fibrotic kidney, which is primarily reduced by conditional deletion of TGFβ receptor 2 (Tgfbr2) [29][30][31].
Smad3 is the canonical downstream of TGF-β1 signaling, serving as a key mediator of kidney fibrosis by promoting myofibroblast accumulation and fibrogenic molecule production in multiple experimental renal diseases, which is dramatically suppressed by Smad3 deletion [42][43][44][45]. Increasing evidence suggests that macrophages are a key player in the Smad3-dependent fibrogenic progression, particularly via the direct mechanism of MMT [46,47]. In a chimeric study with Smad3 −/− and Smad3 +/+ GFP + bone marrow transplanted into irradiated mice with unilateral ureteral obstruction (UUO), Smad3-deleted macrophage (Smad3 −/− GFP + F4/80 + ) failed to undergo MMT to generate myofibroblasts (GFP + α-SMA + ) for collagen-I deposition in the fibrotic kidney, which is in contrast to the profound MMT activity of Smad3 wildtype macrophage (Smad3 +/+ GFP + F4/80 + ) [48]. These findings demonstrate that Smad3 is the key regulator of MMT [48]. However, targeting Smad3 may cause dysregulation of the immune system, contributing to the development of autoimmune disease [49]. Thus, Smad3 direct downstream targets responsible for pathogenic processes including MMT were identified for developing an antifibrotic strategy with minimal side effects. Single-cell RNA seq resolved the cell-cell transcriptome of MMT, revealing proto-oncogene tyrosine-protein kinase Src-and neural transcription factor Pou4f1-centric regulatory gene networks driving MMT in the injured kidney in vivo and TGF-β1-induced bone marrow-derived macrophages in vitro [30,41]. Further molecular study reveals that Src and Pou4f1 are the direct targets of Smad3, where Src inhibition and macrophage-specific Pou4f1 silencing mimicked the protective effect of Smad3 −/− in MMT suppression and associated myofibroblast generation and collagen-I production, representing a precision strategy for targeting MMT [30,41]. Owing to lncRNA's temporal and spatial specificity, Smad3-dependent LncRNA-regulating MMT may be identified to further enhance the precision of targeted MMT antifibrotic therapy.
3.1. TGF-β1-Associated lncRNAs in Kidney Diseases 3.1.1. lncRNAs in TGF-β1 Induced EMT Studies in recent decades revealed that lncRNA is one of the pathogenic downstream regulators of TGF-β1 signaling in inflammatory diseases [4,18] (Figure 1). The majority of lncRNAs exert their biological effects by altering transcriptional or posttranscriptional processes such as transcription factor recruitment, RNA maturation, protein synthesis, and transport. These lncRNAs are also capable of changing chromatin structure via polycomb repressive complex 2 (PRC2) and repressing miRNAs via complementary binding (sponging) [50,51]. For instance, lncRNAs and protein-coding genes were induced via similar mechanisms that shared histone-modification profiles and exon/intron architecture, but lncRNAs are predominantly localized in the nucleus and expressed at a lower level, although with higher tissue specificity compared with coding genes [52]. Most novel lncRNAs were discovered in primates with high-throughput RNA sequencing, revealing disease-associated TGF-β1-dependent lncRNAs that were significantly upregulated in experimental disease conditions in vivo and in vitro [53] (Table 1). EMT is a crucial pathogenic process in kidney fibrosis. Cell-cell connections between tubular epithelial cells were progressively lost and ECM molecules were actively produced by EMT-derived cells to transform nephrons into functionless scar tissue [39]. Numerous studies suggested that TGF-β1-dependent lncRNAs regulate EMT. LncRNAs have been proposed to modulate gene expression indirectly as competing endogenous RNAs (ceRNAs), where they compete with a network of mRNAs and circular RNAs (circRNAs) to bind to microRNAs [54]. LncRNAs acting as ceRNAs may represent significant modulation of the canonical TGF-β pathway. PVT1 (plasmacytoma variant translocation 1) is the first lncRNA identified to be associated with diabetic nephropathy, where single-nucleotide polymorphisms (SNPs) significantly associated with end-stage renal disease (ESRD) of type 2 diabetes were located in PVT1 [55]. Further study reveals that high glucose levels induced PVT1 to stimulate TGF-β1, PAI-1, and FN1 expression, which is further amplified by PVT1-derived miR-1207-5p to accelerate ECM accumulation in the diseased glomeruli of diabetic nephropathy (DN) [56]. lncRNA-MGC (megacluster) is a host of 40 miRNAs upregulated in the TGF-β1-treated mesangial cells in vitro and the glomeruli of diabetic mice in vivo via transcription factor CHOP [57]. Inhibiting MGC by antisense oligos GapmeRs in the diabetic kidneys of streptozotocin-injected mice effectively suppressed a cluster of miRNAs and profibrotic gene expression (Col1a2, Col4a1), contributing to the dramatic reduction of PAS-positive areas, glomerular basement membrane (GBM) thickness, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive cells [57]. These results suggest a role of miRNAs in the regulation of MGC-driven DN progression.
in the TGF-β1-treated mesangial cells in vitro and the glomeruli of diabetic mice in vivo via transcription factor CHOP [57]. Inhibiting MGC by antisense oligos GapmeRs in the diabetic kidneys of streptozotocin-injected mice effectively suppressed a cluster of miR-NAs and profibrotic gene expression (Col1a2, Col4a1), contributing to the dramatic reduction of PAS-positive areas, glomerular basement membrane (GBM) thickness, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive cells [57]. These results suggest a role of miRNAs in the regulation of MGC-driven DN progression.

Smad3-Dependent lncRNAs in Kidney Diseases
Smad3 plays an important role in TGF-β-driven renal inflammation and fibrosis, but potential side effects in vital organs limit its therapeutic application [70]. Therefore, Lan's group further identify several Smad3 downstream profibrotic lncRNAs as therapeutic targets against renal fibrosis via RNA sequencing [71] (Table 2). Smad3-WT-specific upregulated lncRNAs were extracted from the Smad3-dependent transcriptomes of both UUO and anti-GBM kidneys, eventually revealing 21 potential fibrogenic lncRNAs suppressed by Smad3 deletion [71]. The lncRNAs GAS5 [72], LRNA9884 [73], and lnc-TSI [74] are Smad3 direct targets, regulated by its direct binding to the regulatory sequence of lncRNAs as detected by ChIP-PCR assay, and its influence on the expression levels of lncRNAs was further confirmed by luciferase reporter assay. Smad3 transcriptionally regulates LRNA9884 in the advanced glycation end product-stimulated embryonic fibroblasts (MEFs) and kidney of diabetic mice (db/db) via direct binding on LRNA9884 promoter, promoting MCP-1-mediated renal inflammation via direct binding to its promoter [73]. Moreover, LRNA9884 is also involved in the pathogenesis of acute kidney injury (AKI), which is highly expressed in the tubular epithelial cells of AKI kidneys, promoting IL-1β-induced inflammatory cytokine production (MCP-1, TNF-α, and IL-6) via transcriptional regulation of macrophage migration inhibitory factor (MIF) to trigger MIF/NF-κB pathway [75]. Therefore, LRNA9884 inhibition might be a potential therapy for DN and AKI.  LncRNA Ptprd-IR (np_4334, intron of protein tyrosine phosphatase receptor delta), is one of the 21 Smad3-WT-specific upregulated lncRNAs under fibrotic conditions [71]. Smad3 transcriptionally regulates Ptprd-IR expression in TGF-β1-stimulated mouse renal tubular epithelial cells (mTECs) and UUO kidneys via a conserved Smad3 binding site on Ptprd-IR's promoter [76]. Interestingly, Ptprd-IR promotes TGF-β1-and IL-1β-mediated activation of the NF-κB pathway, resulting in pro-inflammatory cytokine production and renal inflammation in UUO kidneys in vivo and mTEC in vitro, while it has no effect on TGF-β1-induced renal fibrosis [76].
LncRNA Erbb4-IR (intron of Erb-B2 receptor tyrosine kinase 4) is expressed in diabetic kidneys of db/db mice, and AGEs stimulated MEFs via a Smad3-dependent mechanism [71,77]. Erbb4-IR promotes fibrotic progression of a diabetic kidney in vivo, and Col-I/IV expression in AGE-stimulated mouse mesangial cells and tubular epithelial cells in vitro, via sponging renal protective miR-29b through direct binding to its 3' untranslated region (UTR) [80]. In a UUO kidney and its in vitro model, Smad3 regulates Erbb4-IR to promote Col-I and α-SMA expression via suppressing Smad7, a suppressor of TGFβ/Smad3 signaling by direct interaction with the 3' UTR of Smad7. Therefore, renal Erbb4-IR silencing effectively restored Smad7 expression against UUO-induced fibrotic progression [80]. This also applied to the pathogenesis of ischemia-reperfusion-induced AKI, where TGF-β/Smad3 signaling was further amplified by Erbb4-IR [78].

TGF-β1 Signaling in Tumor Progression
The role of TGF-β1 is also pleotropic in the tumor microenvironment (TME), simultaneously regulating both pro-and anticancer processes during tumor progression. TGF-β1 signaling activation has been associated with metastasis and poorer prognosis due to the induction of EMT and drug resistance [82,83], but disruption of TGF-β1 signaling also leads to poor prognosis and accelerated tumor progression as TGF-β1-induced cancer cell apoptosis was relieved [84]. EMT is induced in epithelial tumor cells with prolonged exposure to TGF-β1, phenotypic changes including the loss of cell-cell adhesion between epithelial cancer cells to allow their migration and invasion, and the acquisition of protumoral cancer-associated fibroblast phenotype [22,85]. Thus, EMT is a critical step in cancer metastasis and a critical contributor to patients' poor prognosis. At the molecular level, EMT is featured by the downregulation of E-cadherin and upregulation of the EMT markers N-cadherin, Vimentin, α-SMA, and EMT-associated transcription factor SNAI1/2 and ZEB1 [86]. Moreover, TGF-β1 promotes drug resistance by upregulating slow-cycling cancer stem cells (CSCs) that escape from chemotherapy and contribute to the recurrence [83,87]. Interestingly, TGF-β also induces the apoptosis of cancer cells by inhibiting the cell cycle via expression of cyclin-dependent kinases including p15 Ink4b , p21 Cip1 , and p57 Kip2 and pro-apoptotic factor Bim [88,89]. Therefore, targeting the pathogenic downstream effectors of TGF-β1 would selectively suppress its pro-tumoral effects without interfering with the anticancer effect of TGF-β1; LncRNA with high tissue and disease specificity represents an ideal TGF-β1-targeting strategy for anticancer therapy (Figure 2). β1-stimulated renal tubular epithelial NRK-52E cells [64]. Their pathogenic role in renal fibrosis was confirmed by gene silencing, where TGF-β-induced expression of profibrotic molecules Col1a1 and Col3a1 was regulated by TCONS_00088786, while Ctgf and Fn1 were controlled by TCONS_01496394 [64]. In addition, Smad3-dependent antifibrotic lncRNA GAS5 (growth arrest-specific 5) was largely suppressed in a UUO kidney and TGF-β1-induced mTEC, contributing to the Col-I and Fn expression. Further mechanistic study reveals that GAS5 interacted with miR-142-5p, which binds to the 3′UTR of Smad3 to suppress TGF-β1-induced apoptosis and Col-I/Fn expression of mTEC [72].

TGF-β1 Signaling in Tumor Progression
The role of TGF-β1 is also pleotropic in the tumor microenvironment (TME), simultaneously regulating both pro-and anticancer processes during tumor progression. TGF-β1 signaling activation has been associated with metastasis and poorer prognosis due to the induction of EMT and drug resistance [82,83], but disruption of TGF-β1 signaling also leads to poor prognosis and accelerated tumor progression as TGF-β1-induced cancer cell apoptosis was relieved [84]. EMT is induced in epithelial tumor cells with prolonged exposure to TGF-β1, phenotypic changes including the loss of cell-cell adhesion between epithelial cancer cells to allow their migration and invasion, and the acquisition of protumoral cancer-associated fibroblast phenotype [22,85]. Thus, EMT is a critical step in cancer metastasis and a critical contributor to patients' poor prognosis. At the molecular level, EMT is featured by the downregulation of E-cadherin and upregulation of the EMT markers N-cadherin, Vimentin, α-SMA, and EMT-associated transcription factor SNAI1/2 and ZEB1 [86]. Moreover, TGF-β1 promotes drug resistance by upregulating slow-cycling cancer stem cells (CSCs) that escape from chemotherapy and contribute to the recurrence [83,87]. Interestingly, TGF-β also induces the apoptosis of cancer cells by inhibiting the cell cycle via expression of cyclin-dependent kinases including p15 Ink4b , p21 Cip1 , and p57 Kip2 and pro-apoptotic factor Bim [88,89]. Therefore, targeting the pathogenic downstream effectors of TGF-β1 would selectively suppress its pro-tumoral effects without interfering with the anticancer effect of TGF-β1; LncRNA with high tissue and disease specificity represents an ideal TGF-β1-targeting strategy for anticancer therapy (Figure 2).

lncRNAs in TGF-β1 Induced EMT
Studies in the past decade have identified a number of TGF-β1-induced lncRNAs in different types of cancer (Table 3). Most TGF-β1-induced lncRNAs were found to be involved in the regulation of EMT, drug resistance, apoptosis, and the proliferation of cancer cells. TGF-β1 induces the LncRNA MALAT-1 (metastasis-associated lung adenocarcinoma transcript 1) in several types of cancer to promote EMT via different downstream mechanisms [90][91][92][93][94][95][96][97]. In metastatic bladder cancer, MALAT1 directly interacts with suz12 (a component of histone-modifying complex-PRC2) to regulate the expression of EMTassociated genes (E-/N-cadherin, fibronectin, MMP9), contributing to the migration and metastasis of cancer cells in vitro and in vivo [95]. One study of aggressive renal cell carcinoma further reveals that MALAT1 enhances the levels of Ezh2 and H3K27me3 on E-cadherin promoter region, epigenetically suppressing E-cadherin to promote EMT [93]. In addition to EMT, MALAT1 promotes metastasis by enhancing angiogenic hepatocellular carcinoma cells; VEGF-A expression and associated angiogenesis in vivo were regulated by MALAT1 in a miR-140-dependent manner [90]. Moreover, MALAT1 also controls the bioavailability of TGF-β1 in the extracellular space by activating the transcription of latent TGF-β binding protein 3 (LTBP3) via recruiting Sp1 to its promoter [96].     The lncRNA HOTAIR interacts with PRC2 complex to alter the chromatin state for the metastatic phenotype of breast cancer cells [135]. HOTAIR promotes PRC2 occupancy on the promoters of 854 genes detected by genome-wide promoter array assay on ChIP enriched with EZH2, H3K27me3, and SUZ12 target genes in HOTAIR-overexpressing cells [135]. Among the PRC2-associated genes, 35 genes are associated with the stemness of cancer cells, contributing to the TGF-β1-induced upregulation of CD133 + /CD44 + CSC populations [134]. In renal carcinoma, HOTAIR regulates the proliferation and invasiveness of cancer cells by promoting H3K27me3 in cell cycle-related genes' promoters (p53, p21, and p16) via EZH2 and by sponging miR-141 to regulate ZEB1 expression [131,132]. Interestingly, lncRNA MEG3 brings the GA-rich distal regulatory element and PRC2 into close proximity via two interacting sequences encoded to target the GA-rich region and the PRC2 complex, thus modulating the transcription of the TGF-β pathway genes [119]. Further analysis of lung cancer cell lines reveals that MEG3 epigenetically suppressed E-cadherin and miRNA-200s via recruitment of JARID2 and EZH2 and associated histone H3 methylation [117]. The lncRNA ATB promotes the EMT of several types of cancer cells [4,124,126,127] via sponging miR-200s, an EMT suppressor inhibiting ZEB1/2 expression via direct binding to their 3 UTR, which is detected between ATB and miR-200a/b/c in an RNA immunoprecipitation (RIP) experiment [4,138,139]. The ATB/miR-200s/ZEB1/2 axis was demonstrated in the EMT of hepatocellular carcinoma and gastric and breast cancer cells [4,125,126]. Moreover, in HCC cells, ATB also interacts with the IL-11 transcript to extend its half-life, facilitating the synthesis and secretion of IL-11 into the supernatant as the autocrine signal for STAT3 activation [4]. This IL-11/STAT3 axis is essential for the metastatic colonization of the lung and liver [4].
LncRNAs regulate miRNAs targeting key activators of EMT to drive the phenotyping changes in cancer cells. ZEB1/2 are transcriptional repressors of E-cadherin, which is essential for the initiation of EMT [139,140]. ZEB1 is regulated by LINC00673 and LINC00273 via sponging miR-150-5p and miR-200a-3p, respectively, to reduce the binding of miRNAs on the 3 UTR of ZEB1 [103,120]. Meanwhile, ZEB2 and Notch-1 are regulated by XIST via miR-367/miR-141 and miR-137, respectively, in NSCLC cells, contributing to increased pulmonary metastasis nodules in vivo and cell migration in vitro [113,141]. Interestingly, the transcription of ZEB2 is also regulated by its natural antisense transcript ZEB2 NAT, which overlaps the splicing site at 5 UTR of ZEB2 as a potential internal ribosome entry site (IRES) sequence to promote ZEB2 translation, accounting for the EMT of colon adenocarcinomas and urinary bladder cancer cells [128,129]. In addition to NAT, the long non-coding transcript of MMP2 (LncRNA-MMP2-2) is also induced by TGF-β1 and contained in exosomes to promote host gene MMP2 expression for enhancing cancer cell migratory ability [114]. Likewise, the lncRNA miR-155 host gene (MIR155HG) promotes EMT via derived miR-155-5p, which targets the 3' UTR of SOX10 to regulate its transcription. Moreover, zinc finger protein SNAI1/2 is a powerful initiator of EMT induced by TGF-β1associated lncRNAs for suppressing E-cadherin expression [107,108,129]. In ovarian cancer cells, SNAI2 expression is upregulated by PTAF via sponging miR-25, which suppresses SNAI2 expression via direct binding to its 3 UTR [115]. In lung and pancreatic cancer cells, SNAI1/2 is upregulated by MEG8 via suppressing miR-34a and miR-203 through histone H3 methylation of their regulatory regions [116]. SNAI1 expression is also induced in esophageal squamous carcinoma cells by lncRNA SPRY4-IT1, driving the EMT of thyroid cancer cells and contributing to metastasis and poor patient prognosis [136,137]. Another EMT-regulating transcription factor Twist is upregulated in colorectal and pancreatic cancer cells by TGF-β1-induced lncRNA TUG1, suppressing E-cadherin expression to promote the migration capacities of these cancer cells [104,105]. Serine/arginine-rich splicing factor 6 (SRSF6) is highly expressed in the cancer tissues of patients with metastatic tumors and is induced by TGF-β1 via suppressing SRSF6 repressing lncRNA LINC01133 [123].

lncRNAs Associated with TGF-β1-Induced Drug Resistance
Cancer stem cells (CSCs) are a subpopulation with self-renewal and high tumorigenic capacity that contribute to TGF-β1-induced drug resistance [83]. LncRNA UCA1 promotes the stemness and proliferation of cancer cells for escaping from anticancer treatment, where stemness regulators Nanog, ALDH1, and HXK2 expression were regulated by the UCA1/miR-1, miR-203/Slug axis [99,100]. UCA1 also promotes PD-L1 expression in anaplastic thyroid carcinoma cells to suppress CD8+ T-cell-mediated cell cytotoxicity via sponging miR-148a to prevent its suppression on PD-L1 expression [98]. miRNA-145 plays a suppressive role in CSC formation, which is inhibited by LET and MACC1-AS1 against the gemcitabine resistance of bladder cancer cells and the 5-FU and oxaliplatin resistance of gastric cancer cells, respectively [110,122,142]. In breast cancer cells, the lncRNA Has2as mediates TGF-β-induced stemness via regulating a panel of stemnessassociated transcription factors (Pou5f1, Sox2, Nanog, Zfp42). In glioma cells, lncRNA LINC00115 promotes the sphere formation of stem-like cancer cells in vitro and tumor growth in vivo via sponging miR-200s to reduce their binding on the 5 UTR of ZNF596, a downstream stemness regulator of glioma cells [109]. The lncRNAs CASC11 and HAND2-AS1 were upregulated in small cell lung cancer (SCLC) and non-small cell lung cancer cells (NSCLC), respectively, enhancing TGF-β1 expression to promote a CDD133+ CSC population of SCLC cells [106,107]. In hepatocellular carcinoma cells, TGF-β1 induced the enrichment of lincRNA-ROR (linc-ROR) in extracellular vesicles, promoting the formation of spheroid CD133 + stem-like cells against sorafenib-, doxorubicin-, and camptothecininduced cancer cell death [130]. Another lncRNA, H19, promotes CSCs and their formation of spheroid HCC cells, which is regulated by the noncanonical TGF-β/TGFBR2/SOX2 and PI3K/AKT/miR-675 axis [101,102].

Therapeutic Strategies Targeting lncRNAs
Due to the dual roles of TGF-β1 in physiological and pathological contexts, targeting TGF-β1 is not an optimal therapeutic strategy as there is evidence that TGF-β1 deficiency might impair host immunity and cause autoimmune diseases [26]. Although many studies have demonstrated that blocking TGF-β1 protects against progressive renal fibrosis and cancer, others have also highlighted the potential consequences of TGF-β1 inhibition, for instance lethal inflammation observed in TGF-β1-deficient mice at 3 weeks of age [155]. Currently, TGF-β1 inhibitors have not been approved for cancer or fibrosis therapy due to reported cytotoxicity in recent clinical trials. A case in point is fresolimumab, a human anti-TGF-β1 monoclonal antibody (mAb), was found to have no significant effect on proteinuria, eGFR, or serum creatinine in focal and segmental glomerulosclerosis (FSGS) [156]. Furthermore, side effects including pustular rash, herpes zoster, bleeding, skin lesions, and cancer have also been observed after anti-TGF-β therapies. Of note, fresolimumab was found to be involved in the development of cutaneous lesions, and results showed that keratoacanthomas were the most common cutaneous neoplasms observed as adverse events in therapies targeting TGF-β [157]. Therefore, targeting the downstream effector of TGF-β may represent a therapeutic strategy specifically against the pathogenic effects of TGF-β without major disturbances to the immune system. Much insight has been gained into how lncRNAs regulate processes such as fibrosis, tumorigenesis, and ECM accumulation, where they can act via binding to Smad proteins, serving as miRNA sponges or interacting with other signaling pathways. These lncRNAs can be used for diagnosis and targeted when their pathogenic mechanisms are elucidated. Findings from preclinical studies have shown the potential of targeting TGF-β-associated lncRNAs for treating kidney diseases and cancers.
Smad3-dependent lncRNAs with therapeutic potential in renal diseases have been identified in previous studies. Inhibition of Erbb4-IR alleviated renal fibrosis in fibrotic UUO and DN models [79,80,158]. Inhibition of Arid2, LncRNA_5318, and LRNA9884 also suppressed renal inflammation in UUO and diabetic models [73,75,81,159]. For antineoplastic therapy, the modulation of lncRNAs not only regulates the TME but also participates in combination with first-line therapy. Gemcitabine and cisplatin are standard therapy for advanced/metastatic carcinoma, where LINC01714 dramatically enhanced the gemcitabine sensitivity of cholangiocarcinoma cells [160]. LncRNA can be targeted by antisense-based strategies or by shRNAs, consisting of siRNAs and modified antisense oligonucleotides (ASOs) [161]. The ASO-based technologies including novel chemical modifications were optimized with multiple preclinical trials, and the efficiency of cellular uptake and the expression levels of targeted ncRNAs have largely improved [162]. The inhibition efficiency and toxicity were major concerns of directly administering lncRNA-targeting agents via tissue or tail vein injection, where toxicity is observed in a dose-dependent manner, i.e., off-target effects through nonspecific binding to similar nucleotide sequences [163][164][165]. However, repeated high dosages of siRNAs and gapmers are required for effective lncRNA inhibition in vivo. Therefore, post-delivery monitoring and optimizing effective concentration of lncRNAs therapeutics are critical for translational application. Novel noninvasive ultrasound microbubble-assisted (USMB) delivery largely reduced the concentrations of lncRNA-targeting agents in nontargeted tissue [166], contributing to the safety and effectiveness in preclinical studies [73,80,167]. Thus, USMB represents a realistic approach to translating lncRNA-targeted therapeutics with added value in postdelivery monitoring and assessment with its imaging function. Moreover, among the FDA-(Food and Drug Administration) and EMA-(European Medicines Agency) approved ASO-based therapies targeting mRNA expression in the liver, most were administered subcutaneously (mipomersen, inotersen, givosiran, volanesorsen, inclisiran, and lumasiran) [5], suggesting the potential for developing subcutaneously delivered lncRNA-targeted ASOs for kidney fibrosis and cancer.
Collectively, lncRNA-targeted therapy might represent an effective strategy for fibrosis and cancer due to its superior tissue and disease specificity. However, the translation of lncRNA-targeted therapy was limited by the dose-dependent toxicities associated with the delivery of lncRNA therapeutics and the lack of conservation among species; that is, human lncRNAs may lack mouse homologs for preclinical study, and mouse lncRNAs may lack human homologs for therapeutic development [168]. Therefore, an effective approach in identifying homologues among species or using humanized mouse models may facilitate the translation of experimental findings into preclinical settings for lncRNAbased therapeutics development [169]. In addition, TGF-β-dependent lncRNAs are disease specific, and in circulation, these lncRNAs are biomarkers of associated diseases. The FDA approved the use of the lncRNA PCA3 in urine as a biomarker for detecting prostate cancer with high sensitivity [170]. As the circulating levels of the TGF-β-associated lncRNAs UCA1 [171], H19 [172], and MALAT1 [173] are associated with disease progression, the potential of these lncRNAs for use as biomarkers for diagnosis could be further explored.

Conclusions
The translational development of therapeutics targeting the TGF-β1 signaling pathway has been largely hindered by its key regulatory roles in multiple physiological processes. In recent decades, the dissection of TGF-β1 signaling pathways has revealed numerous precise therapeutic targets, including lncRNAs for inflammatory diseases. Emerging evidence shows that lncRNAs are specific pathogenic mediators of TGF-β1, regulating a particular function of TGF-β1 during inflammatory disease progression that can be targeted to develop effective gene-based therapies. With the advancement of RNA sequencing at single-cell resolution and bioinformatic analysis, a more in-depth regulatory mechanism of lncRNAs in inflammatory diseases will be discovered. Disease-and cell-type-specific lncRNAs will be identified for the development of precision therapies against tissue inflammation and cancers.

Conflicts of Interest:
The authors declare no conflict of interest.