Lin28 Regulates Cancer Cell Stemness for Tumour Progression

Simple Summary Cancer stem cells (CSCs) are a small population of tumour cells bearing stemness characteristics. They have been deemed as a root of cancer development and a promising drug target for anticancer therapy. Herein, we discuss the roles of an RNA-binding protein, Lin28, in regulating cancer cell stemness to drive tumour progression. Lin28 acts on various types of RNAs in cancer cells and regulates these RNA functions to control the expression of oncogenes. In these ways, Lin28 promotes cancer cell survival, growth, and invasion. We also discuss recent efforts in developing Lin28 inhibitors targeting CSCs in tumours. Abstract Tumours develop therapy resistance through complex mechanisms, one of which is that cancer stem cell (CSC) populations within the tumours present self-renewable capability and phenotypical plasticity to endure therapy-induced stress conditions and allow tumour progression to the therapy-resistant state. Developing therapeutic strategies to cope with CSCs requires a thorough understanding of the critical drivers and molecular mechanisms underlying the aforementioned processes. One such hub regulator of stemness is Lin28, an RNA-binding protein. Lin28 blocks the synthesis of let-7, a tumour-suppressor microRNA, and acts as a global regulator of cell differentiation and proliferation. Lin28also targets messenger RNAs and regulates protein translation. In this review, we explain the role of the Lin28/let-7 axis in establishing stemness, epithelial-to-mesenchymal transition, and glucose metabolism reprogramming. We also highlight the role of Lin28 in therapy-resistant prostate cancer progression and discuss the emergence of Lin28-targeted therapeutics and screening methods.


Introduction
Cancers such as prostate cancer (PC) contain heterogeneous cell populations and, therefore, respond to anticancer treatments differently [1]. This heterogeneity is reflected not only in genomics and transcriptomics but also RNA processing re-programming that can give rise to tumour cells with various co-existing phenotypes and histological presentations [2][3][4]. When a targeted therapy is applied to a tumour, it will be effective in suppressing some populations of tumour cells and, at the same time, may provide an opportunity to allow other cell populations to thrive and eventually develop into therapy-resistant tumours. One good example is how androgen receptor (AR)-targeted therapy in metastatic PC can result in short-term tumour suppression effects but also induces castrate-resistant PC (CRPC) in the long term that presents an AR-indifferent phenotype containing weakly expressed AR and AR-regulated PSA, cancer stem cell (CSC) markers (e.g., CD44, CD133, BMI1, EZH2), and even AR negative neuroendocrine PC (NEPC) markers (e.g., CHGA, SYP) [5][6][7]. In the absence of Lin28B, pri-let-7s are cleaved into prelet-7 miRNAs and then converted to mature let-7 with the aid of Dicer in both cytoplasm and nucleus. The stages at which Lin28A and/or B inhibit let-7 maturation are marked. The lower right square shows the structure of pri-let-7. The upper right square demonstrates the site where Dicer cleaves pre-let-7. The left square shows the bipartite binding of Lin28A/B to pri/pre-let-7 members. Created with BioRender.com.

Lin28 Regulates let-7 Maturation
As shown in Figure 1, two paralogs of Lin28 bind to let-7 to prevent it from being further processed [39,40]. The two conserved domains, CSD and ZKD, have a high affinity to let-7 precursors at the terminal loop region of preE [30]. Specifically, ZKD binds to the 3' end of the preE bulge containing the GGAG motif, while CSD preferentially binds to the preE loop with the NGNGAYNNN motif [30]. Mayr et al. [41] showed that CSD binds to microRNA first and then induces a conformational change in the preE, making it available for ZKD binding. Although CSD has a higher affinity to miRNA, ZKD binding to the sequence is required and sufficient to recruit TUTase enzymes and induce oligouridylation [42]. In the cytosol, the Lin28B CSD recognizes the preE loop of pre-let-7 while allowing ZKD to bind with the conserved motif preE bulge within let-7 (GGAG) [22,32]. ZKD binding to miRNA blocks the Dicer cleavage site, which is proximal to the GGAG-conserved preE bulge, and prevents pre-let-7 maturation [22]. It is worth mentioning that some members of the let-7 family can escape Lin28-mediated inhibition [43]. For instance, Triboulet et al. [43] confirmed that human let-7a-3 and its murine ortholog, mouse let-7c-2, bypass Lin28A-mediated repression. However, these pathways must be less common because most of the studies confirm that let-7 levels are tightly regulated by Lin28A and Lin28B. In the absence of Lin28B, pri-let-7s are cleaved into pre-let-7 miRNAs and then converted to mature let-7 with the aid of Dicer in both cytoplasm and nucleus. The stages at which Lin28A and/or B inhibit let-7 maturation are marked. The lower right square shows the structure of pri-let-7. The upper right square demonstrates the site where Dicer cleaves pre-let-7. The left square shows the bipartite binding of Lin28A/B to pri/pre-let-7 members. Created with BioRender.com.

The Role of Lin28 in Cancers
The roles of Lin28A and Lin28B in cancer development and progression have been elucidated over the last decade. Numerous studies have found that overexpression of Lin28A/B leads to reduced levels of the let-7 members that stimulate tumour cells to gain stemness and invasiveness. Not surprisingly, clinical studies have confirmed that Lin28 overexpression correlates with higher chances of metastasis and overall poor survival rates, and we have summarized the main findings on Lin28 in various cancer types in Table 1. The broad range of cancer types and cancer cell lines that rely on Lin28/let-7 regulation indicates a central role for the pathway in cancer biology. Indeed, a comprehensive analysis of Lin28A and Lin28B expression over a panel of 527 human cancer cell lines revealed that the paralogs are overexpressed in roughly 15% of cancer types [25]. The plethora of Lin28A/B functions is explained by the abundance of downstream genes, including direct targets of let-7, Sox2, HMGA2, and c-Myc [17]. Furthermore, various studies have shown that therapeutic inhibition, gene knockout or RNA silencing of Lin28 reverse the stemness and invasive phenotypes ( Table 1), highlighting that aberrant Lin28 expression is functional in the control of tumour cell aggressiveness. Additionally, overexpression of Lin28B in transgenic murine models was sufficient to induce hepatoblastoma and hepatocellular carcinoma, while deletion or silencing of Lin28B gene prolonged survival [46]. These studies collectively demonstrate the key importance of Lin28A/B in the development and progression of cancers. (1) Lin28B ↑ in patients leads to short progression-free and low survival rates (2) Lin28B blocks let-7a maturation and positively correlates with pri/pre-let-7a (3) A positive correlation between Lin28B and insulin-like growth factor II (IGF-II) may contribute to the adverse effect in ovarian cancer [64] Ovarian primitive germ cell tumours (GCTs) Ovarian cancer tumour tissues Both (1) Lin28A ↑ significantly contributes to primitive ovarian GCTs (primary and metastatic dysgerminomas, gonadoblastoma, Yolk Sac tumour, and embryonal carcinoma), immature teratomas, and neuroepithelial tissues (2) Lin28A can be used as a diagnostic marker for GCTs [26]  (1) Lin28B ↑ pre-metastatic genes in lung tissues, facilitating neutrophil accumulation, N2 conversion, and the development of the immunesuppressive pre-metastatic niche (2) Let-7s exosomes ↓ in pre-metastatic niches ↑ IL-6 and IL-10, contributing to the incidence of lung metastasis and increased tumour size (3) Lin28B and let-7s are markers for poor survival and lung metastasis [66] Breast cancer

The Role of Lin28 in Regulating the CSC Phenotype
Early evidence of the key role of the Lin28/let-7 axis in the development of the CSC phenotype was presented by Yu et al. [71] in a study on breast cancers. They postulated that low levels of cellular let-7 result in upregulation of HMGA2, a suppressor of differentiation, and H-RAS, an enhancer of self-renewal ability [71]. They showed that an artificial upregulation of let-7 through lentiviral infection reduced the regenerative abilities of undifferentiated cells and, subsequently, prevented tumour growth and metastasis [71]. Similar findings were reported for pancreatic cancer, where low levels of let-7 were found to be responsible for the upregulation of pluripotency genes, including Sox2, Sox9, Oct4, Nanog, and c-Myc [72]. Further evidence supporting the role of Lin28/let-7 in the development of treatment resistance through the CSC phenotype was demonstrated in breast and gastric cancer [71,73]. Overall, the Lin28/let-7 axis plays an important role in cancer stemness, and an understanding of mechanisms underlying Lin28-mediated CSC initiation might illuminate some important targets for metastatic and therapy-resistant cancer treatment.
Several known let-7-regulated signalling pathways are implicated in CSC development, including the Wnt/β-catenin, NOTCH/hedgehog, MAPK/ERK, PI3K/AKT, and STAT3/NFκB/cytokines pathways [74]. A significant amount of research on breast cancer progression linked Lin28/let-7 with the development of CSC phenotypes through the aforementioned signalling pathways. For example, breast cancer has been shown to respond to tamoxifen treatment and radiotherapy more effectively as a result of let-7c/d-induced suppression of the Wnt pathway, which stimulates CSC renewal and proliferative abilities [75,76]. Similarly, the CSC phenotype in breast cancer cells was suppressed after the attenuation of Lin28A expression caused let-7b-dependent Wnt pathway inhibition [77]. Besides breast cancer, the Wnt signalling suppressed by let-7 was shown to inhibit EMT and self-renewal phenotypes in CSCs of lung and esophageal cancers, as well as in hepatocellular carcinoma [77][78][79]. The importance of the Lin28/let-7 axis on Wnt signalling was also highlighted in a study of Wnt-activated esophageal squamous cell carcinoma (ESCC) clinical tissues. A clear correlation between metastasis, migration, and invasion with the let-7 members was observed [57].
The NOTCH/hedgehog pathway has also been indirectly linked to the stemness phenotype and metastatic abilities of cancer cells. For instance, sonic hedgehog protein, along with CSC cell surface markers, was found to be upregulated in highly metastatic pancreatic cancer [80]. Such upregulation was accompanied by low levels of let-7 members, which were shown to directly affect sonic hedgehog expression [81]. In non-small lung cancer cells (NSCLC), upregulation of let-7c, along with miR-200b, coupled with inhibition of the Hh signalling pathway resulted in a reduction of CSC markers and EMT phenotype, as well as re-sensitization to chemotherapy [82].
Aberrant regulation of the STAT3/NFkB pathway, which promotes mesenchymal and CSC phenotypes, is also related to the Lin28/let-7 axis. For example, in NSCLC, NFkB binds to Lin28B chromatin, leading to overexpression of Lin28B and subsequent suppression of let-7 [83]. Similarly, STAT3 directly stimulates Lin28A/B expression by binding to its promoter in breast cancer cells [84]. In both studies, NFkB-and STAT3mediated activations of Lin28 enhanced EMT and CSC phenotypes. STAT3/NFkB pathway is stimulated by pro-inflammatory cytokines produced by M1 macrophages, thus mediating an immune-dependent activation of the Lin28/let-7 axis [85]. Other studies also reported that IL6 cytokines create a positive feedback loop in Lin28/let-7 axis activation through stimulation of NFkB, leading to the transformation of breast epithelial cells into CSCs. Overall, the STAT3/NFkB pathway has been found to deliver immune signals down to the Lin28/let-7 axis.

The Role of Lin28 in EMT and Metastasis
EMT, the initiation step of invasion and metastasis, is defined by the destabilization of the adherent and tight junctions that maintain the integrity and apical-basal polarity of epithelia [86,87]. The transformation of the epithelial cells into mesenchymal cells is a characteristic change when tumour cells gain an invasive phenotype and obtain motility to progress into more malignant tumours [86,87]. The epithelial cell adherent junction consists of cadherin adhesion receptors, such as N cadherin and E cadherin, which are responsible for calcium-dependent adhesion and maintenance of epithelial cell plasticity, respectively [87]. During EMT, cells lose E cadherin but gain N cadherin expression. EMT is also accompanied by several other phenotypic changes in epithelial cells, including cytoskeleton dynamics, which allow cancer cells to gain high mobility [87].

Lin28 Promotes EMT and Metastasis in let-7-Dependent Manner
Lin28 is one of the well-studied EMT-inducing factors [19]. Clinical evidence has shown that Lin28A and B are correlated with cancer cell invasion and migration. Early research first marked low levels of let-7 family members as a prerequisite for metastasis and one of the key drivers of mesenchymal phenotype switch in epithelial cells [88][89][90][91]. The discussed mechanisms involved loss of let-7-regulated suppression and subsequent upregulation of genes, such as E-cadherin transcriptional repressors ZEB1 and SIP1 in a breast cancer model, GAB2 and FN1 in mammary carcinoma, and developmental genes Twist and Snail in an oral squamous cell carcinoma (OSCC) model. Importantly, the key role of let-7 in EMT was further proven by reversing the EMT process through the restoration of let-7 levels [90,92]. These findings suggest that Lin28A and B are both direct negative regulators of let-7 in initiating EMT.
The direct role of Lin28A/B in the development of EMT through let-7 regulation has also been experimentally demonstrated by a series of studies. Lin28A/B was shown to be significantly upregulated in mesenchymal type cells (MDA-MB-231 and SK-3rd) compared to epithelial type cells (MCF-7 and BT-474) [20]. Furthermore, Lin28A/B was found to be responsible for higher colony formation rates and migratory abilities in breast cancer cells. In triple-negative breast cancer (TNCB) cells, overexpression of Lin28A/B lowers the level of E cadherin while increasing N cadherin, thereby evidently contributing to EMT development and conferring cancer stem cell characteristics [19]. Similarly, overexpression of Lin28B in TNCB was found to promote lung and breast cancer metastasis through the development of an immune-suppressive pre-metastatic niche marked by neutrophil recruitment and subsequent conversion of N2 phenotype [66]. In particular, Lin28B was shown to increase IL-6 and IL-10 cytokine production to polarize neutrophils into metastatic-promoting N2 phenotype-inhibiting T cells, which normally prevent malignant cell migration [66]. These studies collectively demonstrate that Lin28 initiates EMT and metastasis in breast cancer models not only by causing cell-cell adhesion loss but also by acting as an immune-suppressive agent.
Another study emphasized the role of Lin28 in EMT and metastasis by discovering Lin28 as a target of a known metastasis suppressor in PC, called Raf kinase inhibitory protein (RKIP) [67]. The study showed that RKIP indirectly suppresses Lin28A/B through mitogen-activated protein kinases (MAPK) and Myc [67]. The reduction of Lin28A/B consequently restores let-7 levels, which in turn block the expression of a metastasisactivating gene HMGA2 [67]. The dysregulation of RKIP in breast cancer cells was linked to breast cancer development and invasion through the reversion of the aforementioned cascade involving HMGA2-let-7-Lin28A/B-Myc-MAPK-RKIP [67]. Importantly, while most of the studies implicated the role of Lin28 in cancers by showing the downstream effects of Lin28, this study proved its significance by showing that Lin28 is a direct upstream regulator of RKIP.

Lin28 Promotes EMT and Metastasis in let-7-Independent Manner
Lin28 can also regulate cell invasion and migration independently of let-7. For example, Lin28A/B was found to act as a post-translational driver of metastatic phenotype via direct enhancement of the translation of MYCN-induced transcripts in MYCN-amplified neuroblastoma [69]. In this study, ZKD of Lin28B bound the RPS29 protein, which was shown to correlate with low patent survival, cause cell cycle arrest, and induce cancer cell morphological differentiation. Importantly, the authors showed that the loss of Lin28B reduced the metastatic potential of the tumour cells, leading to a complete reversal of stage 4 to stage 1 of neuroblastoma. This let-7-independent function of Lin28 demonstrates that Lin28 has a broad range of effectors that are highly clinically relevant.

The Role of Lin28 in Glucose Metabolism
Lin28 plays a critical role in the Warburg effect, which is defined as the reprogramming of glucose metabolism to enhance tumourigenesis [93]. Cancer cells predominantly rely on aerobic glycolysis and utilize glucose as the energy source instead of more efficient default pathways, such as the citric acid cycle and oxidative phosphorylation in mitochondria [93]. The shifts in glucose metabolism provide additional plasticity for cancer cells and even enhance therapy resistance [94]. Cancer cells require higher energy demands, not only because of the increased replication and proliferation rates but also hostile tumour microenvironments, such as changes in pH. Thus the cancer cells' metabolic plasticity becomes critical for survival. This is even more true for cells that undergo EMT and metastasis, during which processes the cells experience higher oxidative stress and have to alter their metabolism to withstand the hurdles of metastatic spread [95].
Early evidence showed that the Lin28/let-7 axis regulates metabolism through pyruvate dehydrogenase kinase 1 (PDK1) in an oxygen-or hypoxia-inducible factor-1 (HIF-1)independent manner [96]. Lin28B was found to control the aerobic glycolysis metabolic state and lower the microenvironment pH through the Lin28B/Myc/miR-34a-5p pathway in breast cancer stem cells (BCSCs) [97]. Lin28B-positive MDA-MB-231 and H1299 cells used a significant amount of glucose for aerobic glycolysis and produced a high concentration of lactate, while Lin28B-negative MDA-MB-231 cells had significantly lower rates of glucose uptake and lactate secretion [97]. The high level of lactate in these BSCSs promoted a low pH microenvironment that enhanced stemness properties. Furthermore, low pH increased the abundance of cells expressing high levels of aldehyde dehydrogenase 1 (ALDH1), a CSC biomarker [97]. Myc and Lin28B proteins were also significantly overexpressed in spheroid cells and ALDH-overexpressed cells under a low pH environment [97]. Moreover, the biogenesis of miR-34a-4p, a Myc-suppressed miRNA, is downregulated by Lin28B [97]. Lin28B activates Myc expression, which, in turn, blocks miR-34a-5p to induce glycolysis and inhibits oxidative phosphorylation in BCSCs. Lin28B inhibition by a small molecule C902 reduced Myc expression and lactate secretion in BCSCs. This thereby reversed pH to close to neutral and restored miR-34a-5p levels, which led to reductions in tumour spheroid formation, BCSC migration capability, and tumour sizes. Moreover, the blockade of the Lin28/Myc/miR-34a-5p axis successfully hampered pulmonary metastasis of BCSC in in vitro models [97]. Based on these findings, this study concluded that BCSC relies on aerobic glycolysis activated through the Lin28B/Myc/miR-34a-5p pathway.
Similarly, another study determined the role of the Lin28B/let-7 pathway in regulating glycolytic metabolism, ATP production, and neurosphere formation in neuroblastoma (NB) [98]. The researchers proposed a pathway whereby ornithine decarboxylase (ODC) regulates polyamine production to induce hypusination of eIF-5A, which in turn regulates the Lin28B/let-7 pathway [98]. The study showed that ODC inhibition by difluoromethylornithine (DFMO) restores ATP levels, let-7 expression, and glycolytic metabolism in various degrees depending on the NB cell lines (BE(2)-C, SMS-KCNR, and CHLA90) [98]. Lin28B and MYCN protein expression and ATP levels significantly decreased in BE(2)-C and SMS-KCNR cell lines in response to DFMO treatment [98]. Furthermore, as let-7 expression increased upon DFMO treatment, there was a decrease in glycolytic metabolism in vivo [98]. Therefore, these researchers concluded that Lin28 significantly contributes to glucose metabolism reprogramming in NB cell lines.
Although one of the main known roles of Lin28 is to promote tumourigenesis and metabolic diseases, the reprogramming of glucose metabolism by Lin28A/B brings advantages to the resistance to diabetes, skeletal muscle maintenance, and acceleration of tissue repair [99][100][101]. For instance, overexpression of Lin28A and Lin28B leads to an insulin-sensitized state, which subsequently improves glucose tolerance in Lin28A/B genetically modified aging mice. It allowed the mice to gain resistance to diet-induced diabetes under high-fat diet conditions [99]. Another study showed that Lin28A knockout transgenic mice have impaired primordial germ cell development during embryogenesis and significantly lowered fertility in the adult state. [100]. Consistently, a deficiency in Lin28A lowers the ATP/AMP ratio and NADH/NAD ratio in embryonic growth and hinders the rate of tissue repair [101]. Overall, Lin28 acts as a double sword that reprograms glucose metabolism to induce the Warburg effect in cancers and acts as a factor necessary for healthy developmental processes.

Lin28 Targets Various Messenger RNA
Besides microRNAs, Lin28 also targets messenger RNAs in a cell type-dependent and region-specific manner [102,103]. Similar to conventional mRNA binding proteins, Lin28A/B binds an array of similar RNA sequence motifs [103]. Those motifs are characterized by uridine-rich fragments surrounded by guanosine residues. Lin28A/B increases the protein levels of cell cycle regulators predominantly by stabilizing their respective mRNAs through interaction with the exonic regions and 3' UTR of post-splicing mRNAs [103]. The conserved GGAGA sequence shared between most of the let-7 members was also found in many Lin28-targeted mRNAs [104]. By interacting with the RNA motifs in mRNAs, Lin28A/B serves as both positive and negative regulator of mRNA processing and translation. This was reported to occur through interactions with protein factors such as hnRNP F, TIA-1, FUS/TLS, and TDP-43 [104]. Interestingly, Lin28A/B autoregulates its own mRNA for translation [104]. These findings indicated that Lin28 has a broad range of functions other than controlling let-7 miRNA expression.
Moreover, Lin28 is a global regulator of different cellular compartments, including stress granules (SGs) and cytoplasmic processing bodies (P-bodies) [105]. The formation of SGs is a characteristic response of cancer cells when encountering therapy-induced stress. SGs contribute to cancer cell survival by reducing stress-related damage from oxidative, pH, and temperature fluctuations [105][106][107]. In SGs, Lin28A/B were found to bind T cell intracellular antigen 1 (TIA 1), an RNA-binding protein that stabilizes the mRNA and regulates splicing patterns [105,107]. Overexpression of TIA-1 induces SG assembly and alters gene expression and signal transduction patterns to promote cancer proliferation and metastasis [106,108]. These studies indicate that Lin28 may play a key role in regulating the survival response of tumour cells when encountering therapy-induced stress.

Lin28 and Prostate Cancer (PC)
Prostate cancer (PC) is one of the most common malignancies in men worldwide [109]. Even though PC generally has a relatively slow growth rate and low mortality rate, therapyresistant CRPC poses a significant threat and takes thousands of lives every year [110]. These tumours fail to respond to the first-line androgen deprivation therapies due to the loss of conventional dependency on the AR signalling for growth [111]. Many tumours progress into even more aggressive states by changing cancer cell lineage and transform into therapy-induced neuroendocrine prostate cancers (t-NEPCs) that have a median survival of less than one year [112]. Thus, there is a clear need for a comprehensive understanding of the biological mechanisms underlying PC development and its progression to develop effective treatments. Lin28 has been identified as one of the most promising candidates that regulate therapy-resistant tumour progression [21]. Herein, we discuss the role of Lin28 in the development and progression of PC.
Early studies linked the role of Lin28 to the development of CRPC through the observation that AR expression was regulated by the members of let-7 through c-Myc. High levels of Lin28A-suppressed let-7 in CRPC cells were found to indirectly stimulate AR expression [113]. A follow-up study confirmed that elevated levels of let-7c in PC xenografts significantly reduced tumour growth and invasion [114]. Furthermore, PC patient tissues were found to have a higher level of Lin28A protein than the benign tissues. Borrego-Diaz et al. [51] also reported that Lin28 is an important driver of PC, whereby Lin28B regulates c-Myc expression indirectly by binding to the intermediate c-Myc miRNA suppressor miR-212. Further mechanistic studies have shown that the LNCaP PC cell lines overexpressing Lin28A rendered these cells more resistant to antiandrogens, including enzalutamide, abiraterone, and bicalutamide [115]. The resistance was attributed to the Lin28-dependent stimulation of overexpression of the AR splice variant AR-V7. Lovnicki et al. [16] found that Lin28B, along with another core pluripotency stem cell gene, Sox2, are key drivers of t-NEPC [16]. Importantly, Lin28B induced a stem-like gene network and neuroendocrine biomarkers, including HMGA2, that stimulated the transition of AdPC to more aggressive t-NEPC. Furthermore, Lin28B stimulated EMT properties, which promote t-NEPC formation, by reducing N cadherin and increasing E cadherin. Similarly, another group illustrated that upregulation of Sox2 via Lin28B induces androgen ablation-induced neuroendocrine differentiation via neuroendocrine (NE) lineage in prostate luminal cells [116]. Overall, the Lin28/let-7 pathway regulates several downstream gene networks in CRPC progression.
The expression of Lin28 is regulated by oncogenic pathways such as ESE3/EHF [117]. In normal cells, ESE3/EHF binds Lin28 promoters and inhibits its transcription, which results in an upregulation of let-7 miRNA [118]. In this way, ESE3/EHF maintains a balance between self-renewal and differentiation of normal prostatic epithelium. Significant downregulation of ESE3/EHF was reported in PC [119,120]. It was estimated that approximately 50% of PCs acquired ETS dysregulation through gene rearrangement [117]. Lin28 silencing, both in vitro and in vivo, resulted in similar phenotypes in PC cells expressing high levels of ESE3/EHF. In particular, there was a decrease in PC cell spheroid formation and self-renewal abilities, indicating that Lin28 promotes tumourigenic and CSC-like phenotypes [52]. In addition, the effects of Lin28 silencing were tested on NOD scid gamma (NSG) mice, where tumourigenic ESE3KD-PrEC cells transfected with siRNA against Lin28 caused a reduction in xenograft growth in mice [52]. These findings suggest that ESE3/EHF is a key regulator of Lin28 expression in PCa.

Therapeutic Inhibition of Lin28
Despite a widespread view that RNA-binding proteins are undruggable, there is a growing body of research on targeting Lin28A and Lin28B with small molecules [121][122][123]. These small molecules bind to the surface of either CSD or ZKD of Lin28 and prevent Lin28/let-7 interaction from increasing let-7 biogenesis in cancer cells, resulting in suppression of cancer CSC and EMT phenotypes (Table 2). Notably, the studies outlined in this section employed novel high-throughput screening techniques, including fluorescence polarization (FP) and fluorescence resonance energy transfer (FRET), which are gaining popularity in screenings for RNA-binding protein inhibitors [124,125]. These novel technologies have enabled screenings of libraries containing up to 100,000 compounds ( Table 2). The basic principle of the two most commonly used techniques, FP and FRET, involves the use of fluorescently labeled macromolecules that excite light differently upon RNAprotein interactions [126]. In this way, an event of inhibition between Lin28 and let-7 RNA can be tracked.  (2) Toxic in Lin28 + mESCs and Lin28 -HeLa cells (Lin28B + ) (3) Direct binding: NMR spectroscopy [127] FP-based HTS of 15,000 natural product-inspired compounds (1) FP IC50 = 5.0 µM (2) Let-7 RT-qPCR: 2 times (let-7a/g) in JAR cells (Lin28A + , Lin28B + ) at 20 µM    The earliest implementation of the FRET assay for Lin28 inhibitor screening yielded compound 1632, which successfully restored let-7 levels in Huh7 cells and reduced "stemness" in mouse embryonic stem cells (mESC) [136]. Compound 1632 suppressed mRNA levels of known markers of pluripotency, including Pou5f1/Oct4, Rex1, and Stella, and increased differentiation. Furthermore, the authors demonstrated that 1632 effectively inhibited clonogenic growth in 22Rv1, PC3, DU145, and Huh7 cell lines and induced dosedependent reduction of tumour-sphere formation in 22Rv1 and Huh7 cells. Subsequent studies demonstrated that 1632 downregulates Lin28 mRNA levels, increases let-7 miRNA levels, and, importantly, suppresses cell surface expression of PD-L1, one of the key proteins involved in immune evasion of cancer cells [135]. 1632 was reported to block glycolytic, tumour sphere-forming, and migratory capacities in MDA-MB-231 and H1299 cell lines, as well as reduce the stemness phenotype [97]. The pharmacological effects of 1632 are also observed in vivo, as it inhibited tumour growth and metastasis in immuno- The earliest implementation of the FRET assay for Lin28 inhibitor screening yielded compound 1632, which successfully restored let-7 levels in Huh7 cells and reduced "stemness" in mouse embryonic stem cells (mESC) [136]. Compound 1632 suppressed mRNA levels of known markers of pluripotency, including Pou5f1/Oct4, Rex1, and Stella, and increased differentiation. Furthermore, the authors demonstrated that 1632 effectively inhibited clonogenic growth in 22Rv1, PC3, DU145, and Huh7 cell lines and induced dose-dependent reduction of tumour-sphere formation in 22Rv1 and Huh7 cells. Subse-quent studies demonstrated that 1632 downregulates Lin28 mRNA levels, increases let-7 miRNA levels, and, importantly, suppresses cell surface expression of PD-L1, one of the key proteins involved in immune evasion of cancer cells [135]. 1632 was reported to block glycolytic, tumour sphere-forming, and migratory capacities in MDA-MB-231 and H1299 cell lines, as well as reduce the stemness phenotype [97]. The pharmacological effects of 1632 are also observed in vivo, as it inhibited tumour growth and metastasis in immunodeficient orthotopic mouse models of MDA-MB-231 human breast cancer cells. Additionally, compound 1632 suppresses cell growth and clonogenicity in acute myeloid leukemia cell lines and tissue samples [62]. The mechanisms involved in the inhibition of the NFkB pathway induced leukemic stem cells and cell cycle arrest at the G1/S phase. Collectively, these studies demonstrated that compound 1632 is a promising drug candidate for Lin28 overexpressing tumours.
Other compounds identified by FRET also showed some promising activity in vitro. In particular, compound KCB3602 was found to increase let-7 levels in human choriocarcinoma JAR cells and, subsequently, downregulate gene products of the key drivers of CSC, including HMGA, IGF2, IMR1, Ras, and Lin28A [128]. Furthermore, KCB3602 effectively suppressed the sphere-like growth of cancer cells, a hallmark of CSCs. The authors confirmed the on-target effects using surface plasmon resonance (SPR) assays and showed that KCB3602 does not impair cell growth in a non-cancerous cell line. Another small molecule, benzopyranylpyrazole-based compound (1), was found to upregulate let-7 levels in JAR and PA-1 cell lines [129]. It also downregulated the expression of let-7 target genes, including HMGA2, c-Myc, and Ras. These compounds proved that the pharmacological inhibition of Lin28 suppresses CSC gene signatures. However, further investigations are necessary to determine if these inhibitory effects can be advanced to in vivo studies.

Lin28 Inhibitor Screening: Fluorescence Polarization
Another commonly used technique is the FP assay, which was first used to screen for Lin28 inhibitors by Lightfoot and colleagues in 2016 [133]. The authors reported two small molecule inhibitors, SB/ZW/0065 and 6-hydroxy-DL-DOPA, had low micromolar activity in FP assays. Electrophoresis mobility shift assay (EMSA) was then used to confirm the disruption of Lin28/pre-let-7g complex. Furthermore, these molecules were shown to partially restore Dicer cleavage activity of pre-let-7g in vitro but failed to restore let-7g levels in P19 embryonal carcinoma cells. Later, Wang et al. [133] also employed FP screening of a large library and presented two small molecule inhibitors-namely, LI71 and TPEN-that targeted the CSD and ZKD domains of Lin28A and Lin28B, respectively. Importantly, the authors confirmed the direct binding of these molecules to their respective target domains-i.e., CSD and ZKD-with protein nuclear magnetic resonance (NMR) spectroscopy. LI71 treatment upregulated various members of the let-7 family in leukemia cell line K562 and mouse embryonic stem cells, while TPEN was generally toxic in vitro. Using LI71 as a scaffold, further modifications resulted in derivatives with stronger potency in FP and EMSA assays [131]. They performed structure-activity relationship (SAR) and docking-based binding mode analyses and proposed that the binding mode features are necessary to inhibit protein-RNA interaction at the Lin28 CSD interface. These features included a π-cation interaction with Lys102 and several hydrogen bonds with Lys78, Lys98, and Ala101. Interestingly, another study performed a similar series of experiments (i.e., FP HTS, SAR, docking) and presented a new class of trisubstituted pyrrolinone inhibitors, including the hit compound C902. This work highlighted the importance of Lys102 for the successful inhibition of Lin28 CSD [130].

Lin28 Inhibitor Screening: Other Methods
A few other studies have focused on the development of novel experimental highthroughput techniques in the context of Lin28/let-7 interaction inhibition. For instance, catalytic enzyme-linked click chemistry assay (cat-ELCCA) enabled the identification of two N,N -(1,2-phenylene)-dibenzenesulfonamide derivatives, CCG-233094 and CCG-234459 [134].
The compounds were shown to inhibit Dicer-mediated cleavage of pre-let-7d and ontarget effects confirmed by SPR. Another study focused on the development of a fluorescence intensity-based (FL) assay to screen for potential Lin28/pre-let-7 interaction disruptors [137]. The authors validated their method by identifying a novel scaffold, KCB170522, which was also active in EMSA and showed an expected increase in cellular let-7 levels. Besides small molecules, an epitope inhibitor named Nb-S2A4 was developed and shown to indirectly impact ZKD through recognition of N-terminal TUT4 and disruption of its association with Lin28B ZKD/pre-let-7 complex [132]. Nb-S2A4 prevents pre-let-7 degradation by allowing oligouridylation and monouridylation activities of TUT4 without stopping the formation of the Lin28B/pre-let-7 complex. In summary, several bioassays have been established to screen and validate Lin28 inhibitors that can block Lin28/let-7 interactions and suppress the CSC and EMT phenotypes of cancer cells. Although several Lin28 compounds had been discovered and developed, none have reached clinical trials, and more efforts in the field of research are needed.

Conclusions
Both Lin28A and Lin28B play important roles in cancer development and progression. As CSC promoters, they induce the stemness phenotype and drive the self-renewal and clonogenic capabilities of cancer cells. The Lin28-mediated stemness gene network confers therapy resistance in various cell and xenograft models. It promotes cancer cell migration and invasion, leading to cancer progression to more aggressive stages. Importantly, Lin28 overexpression is observed in a wide range of cancer types, highlighting that it could be used as a common prognostic marker and therapeutic target of cancer.
The development of Lin28 inhibitors as therapies is still at its early stage, as most studies focus on the screening techniques and aim to provide proof-of-concept for Lin28 inhibition. However, a few promising molecules have emerged and been shown to successfully restore let-7 levels and downregulate downstream oncogenes. Furthermore, some studies showed that the pharmacological effects of Lin28 modulation result in the suppression of tumour growth in animal models. Thus, the development of Lin28-targeted drugs presents an attractive strategy for cancer treatment, especially in highly aggressive metastatic forms. The rise of computational drug discovery should aid the development of such drugs. The largest performed screening for Lin28 inhibitors covered around 100,000 compounds, whereas the available chemical space counts in the billions nowadays. Thus, virtual screening coupled with novel in vitro screening techniques should make it possible to explore the vast chemical space and find safe and efficacious small molecule inhibitors of Lin28 protein.