Ribosomal Protein S6: A Potential Therapeutic Target against Cancer?

Ribosomal protein S6 (RPS6) is a component of the 40S small ribosomal subunit and participates in the control of mRNA translation. Additionally, phospho (p)-RPS6 has been recognized as a surrogate marker for the activated PI3K/AKT/mTORC1 pathway, which occurs in many cancer types. However, downstream mechanisms regulated by RPS6 or p-RPS remains elusive, and the therapeutic implication of RPS6 is underappreciated despite an approximately half a century history of research on this protein. In addition, substantial evidence from RPS6 knockdown experiments suggests the potential role of RPS6 in maintaining cancer cell proliferation. This motivates us to investigate the current knowledge of RPS6 functions in cancer. In this review article, we reviewed the current information about the transcriptional regulation, upstream regulators, and extra-ribosomal roles of RPS6, with a focus on its involvement in cancer. We also discussed the therapeutic potential of RPS6 in cancer.


Background
Eukaryotic ribosome (80S) consists of the 60S large and 40S small ribosomal subunits, which contain four ribosomal RNAs (rRNAs; 5S, 5.8S, and 28S rRNAs in the large subunit and 18S rRNA in the small subunit) and 79 ribosomal proteins (RPs) [1,2]. RPs are RNAbinding proteins and are primarily involved in the regulation of protein translation [3]. RPs also function in ribosome biogenesis in the nucleolus. They act as RNA chaperones to ensure the stabilization and correct folding of rRNAs to assemble ribosomal subunits. Dysregulation of ribosome biogenesis may result in tumorigenesis [4]. RPs also have multiple extra-ribosomal functions, including regulation of cell-cycle arrest, cell proliferation, cell migration and invasion, apoptosis, DNA damage repair, malignant transformation, and tumorigenesis via both p53/mouse double-minute 2 homolog (MDM2)-dependent and p53independent mechanisms [3]. For example, ribosomal protein L5 (RPL5) [5], RPL11 [6,7], and RPL23 [8,9] exert anticancer effects through the activation of the tumor suppressor p53 by specifically binding to its inhibitor MDM2.
Ribosomal protein S6 (RPS6), also known as phosphoprotein NP33 [10], is a component of the 40S small ribosomal subunit as a ribosomal RNA-binding protein [11,12]. RPS6 is evolutionarily conserved across eukaryotes from yeast to vertebrates [12] and plays important roles in ribosome biogenesis, protein translation, cell proliferation (increase in cell number), cell growth (increase in cell mass/volume), DNA repair, apoptosis, cell differentiation, and glucose metabolism in both normal and cancer cells [13][14][15]. RPS6 is the first ribosomal protein identified to be phosphorylated by protein kinases [16] and is one of the only two RPs known to be phosphorylated [17].
RPS6 is used as a readout of the mechanistic/mammalian target of rapamycin complex 1 (mTORC1) activity in many diseases, including cancers, since alleviated activity of mTORC1, an upstream regulator of RPS6, is commonly found in many types of cancer cells [2,[18][19][20]. However, considering the almost five-decade-long history of research on RPS6, the functions and therapeutic implications of this protein and its downstream regulatory pathways have not been well appreciated yet. Recently, we found that dual inhibition of epidermal growth factor receptor (EGFR) and MET kinase activities resulted in the synergistic anti-proliferation effect through downregulation of RPS6 in triple-negative breast cancer (TNBC) cells [21]. In addition, knockdown of RPS6 was enough to suppress the proliferation of TNBC cells in vitro. These results motivate us to investigate previous studies on the therapeutic potential of RPS6 targeting. In this review article, we analyzed the roles of RPS6 in tumorigenesis and evaluated the potential of using it as an anticancer therapeutic target.

Transcriptional Regulation of RPS6
The human RPS6 gene is located on chromosome 9p21, comprises six exons [22], and encodes a protein of 249 amino acids [23,24]. Contrary to rRNAs, which are transcribed by RNA polymerase I or III, RP genes are transcribed by RNA polymerase II [1]. Following transcription, the produced RP mRNAs are transported from the nucleoplasm to the cytoplasm for translation, after which the de novo formed RPs are translocated into the nucleus and concentrated in the nucleolus. Interestingly, the promoter regions of most RP genes are phylogenetically well conserved [1].
The promoters of the human RPS6 and mouse Rps6 genes share 80% sequence similarity [25], lack a consensus TATA box, and have high GpC content [22,25]. Consensus binding sites for specificity protein 1 (SP1) have been found in the upstream regions and within the first introns of both human RPS6 and mouse Rps6 genes [22,25]. The SP1 sites in the human RPS6 gene have been reported to minimally contribute to the transcription of the gene in COS-1 cells [26]. Additionally, the human RPS6 promoter contains four putative binding sites for GA-binding protein (GABP), a member of the ETS DNA-binding protein family. Potential binding sequences for CCAAT/enhancer binding protein (C/EBP) and the DNA-binding protein Ikaros have also been identified. In addition, the recombinant Yin-Yang 1 (YY1) protein has been shown to bind to the consensus binding site in the first intron of the human RPS6 gene. In fact, upstream GABP and SP1 sites and downstream YY1 sites are common features of several RP gene promoters [27]. However, the functional implications of GABP and YY1 have not been determined yet [26,28,29].
Chromatin immunoprecipitation experiments have shown that the MYC protein binds to the E-box motif in the TSC2 and RPS6 genes [30]. Interestingly, MYC differentially regulates the transcription of these genes. For example, the TSC2 protein level in the MYCdeficient rat fibroblast HO15 cells is increased, whereas that of RPS6 is reduced. MYC has been shown to repress the transcription of TSC2 and thereby activates the mTORC1/S6K pathway, but the involvement of MYC in the transcriptional regulation of the RPS6 gene has not been determined yet [31].
Zinc finger BED domain-containing protein 1 (ZBED1), an E3 SUMO-protein ligase, has been reported to bind to the upstream regions of RP genes, including RPS6 [32]. Depletion of ZBED1 reduces the transcription of the RPS6, RPSL10A, and RPL12 genes. The expression of the ZBED1 and RP genes is coordinately regulated via a cell-cycle-dependent manner. ZBED1 SUMOylates promoter-bound chromodomain-helicase-DNA-binding protein 3 (CHD3) to release it from the DNA and thereby increases the recruitment of RNA polymerase II to RP gene promoters [33].
Dual-specific tyrosine-phosphorylation-regulated kinase 1A (DYRK1A) has been reported to be recruited to the promoters of genes transcribed by RNA polymerase II [34]. In fact, DYRK1A is constitutively recruited to the promoters of cell-cycle-dependent genes, such as CDK12, RPS6, RPS11, and RPS12, and depletion of DYRK1A downregulates expression of these genes.
The RPS6 gene is transcribed in the mid-to-late G 1 phase, whereas rRNA genes are transcribed in the early G 1 phase concurrently with the nucleolar assembly [35]. In contrast to the RPS27a gene promoter, the human RPS6 gene promoter is not occupied by SP1 or the cAMP-responsive element-binding protein (CREB) [35]. Interestingly, E2F transcription factor 1 (E2F1) has been detected to bind to the RPS6 gene promoter [35]. Since E2F1 is a well-established transcription factor that regulates the cell-cycle-dependent transcription [36], it is highly plausible that E2F1 is a key transcriptional regulator of RPS6 during cell cycle progression. Additionally, the retinoblastoma-associated protein (RB) might contribute to the transcriptional upregulation of RPS6 by binding to E2F1 on the RPS6 promoter. RPS6-knockdown (KD) is known to downregulate phospho (p)-RB in various cancer cells [37][38][39][40]; however, it is elusive whether such a positive feedback loop between RPS6 and the p-RB level is involved in cancer progression. Furthermore, the precise roles and effects of E2F1 and RB on the transcription of RPS6 remain to be determined.

Post-Translational Modifications of RPS6
Post-translational modifications, including phosphorylation, acetylation, methylation, O-linked β-N-acetylglucosaminylation, and ubiquitinylation, have been described to participate in the regulation of the activity of RPs [41] (Figure 1). For example, ischemic preconditioning (IPC) induces S-nitrosylation of RPs, including RPS6, in perfused mouse hearts [42]. In addition, the R137 residue of RPS6 has been identified to be hydroxylated by lysine demethylase 8 (KDM8), an L-arginine (3R)-hydroxylase [43]. However, little is known about the regulation of the post-translational modifications of RPS6, except for phosphorylations, which we will discuss next.
The MAPK pathway consists of rat sarcoma (RAS)/v-raf-1 murine leukemia viral oncogene homolog (RAF)/MAPK-ERK kinase (MEK)/extracellular signal-regulated kinase (ERK) cascade. RSKs are downstream effectors of the MAPK pathway, which also regulates RPS6 activity through RPS6 phosphorylation [48]. Four RSK genes, namely RSK1-4, have been identified in mammals, and these four genes display high homology (78-90%), especially in two kinase domains [68]. In cells lacking S6K, phosphorylation of RPS6 at S235/236 is mediated by RSKs [69]. In resting cells, RSKs are mainly localized in the cytoplasm [70]. Upon stimulation of cells with a growth factor, RSK1 shows maximal kinase activity and transiently translocates to the plasma membrane and then to the nucleus [71]. Subcellular localization of RSK2 is differentially regulated. For example, upon mitogen stimulation of cells, RSK2 localizes to the nucleus; whereas oxidative stress induces RSK2 localization to the cytoplasmic stress granules to promote the survival of the cells [72].
Since S6K and RSK are the downstream effectors of the PI3K/AKT/mTORC1 and MAPK pathways, respectively, RPS6 constitutes a convergence node of these two signaling pathways ( Figure 2). In addition, synergistic crosstalk between the mTORC1 and MAPK signaling in the regulation of RPS6 phosphorylation has been reported (see below). ERK enhances S6K activity by phosphorylating it at T421/424 [73]. Additionally, ERK phosphorylates and thereby inhibits tuberous sclerosis 2 (TSC2), consequently leading to mTORC1 activation [74]. ERK also indirectly activates mTORC1 through RSK-mediated phosphorylation of the regulatory-associated protein of mTOR (RPTOR) [75]. RPS6 is also phosphorylated in hippocampal neurons by S6K, which is activated by cyclin-dependent kinase 5 (CDK5) in these cells [76]. Knocking-in the gene of a non-phosphorylatable RPS6 (rpS6 P−/− ) mutant protein resulted in small cell size, hypoinsulinemia, decreased β cell size, and muscle weakness in mice [77,78]. However, the exact roles of RPS6 phosphorylations still remain enigmatic.

Upstream Effectors of RPS6 Phosphorylation
Various factors regulating the phosphorylation of RPS6 have been identified, including extracellular proteins, kinases and phosphatases, membrane receptors, transcription factors, and adaptor proteins. For example, a genome-scale siRNA screening against 21,121 genes revealed multiple regulators for p-RPS6 (S235/236) in the pancreatic cancer cell line MIA PaCa-2, which contains a constitutively high p-RPS6 level, followed by confirmation in TSC1-null mouse embryonic fibroblasts (MEFs) [79]. Of the 1046 identified candidate regulators of p-RPS6 (S235/236), 632 were validated via individual RNAi. However, most of these p-RPS6 regulators should be further experimentally validated.
(1) Extracellular proteins that regulate RPS6 phosphorylation Various extracellular proteins such as cytokines and growth factors have been identified as regulators of RPS6 phosphorylation (Table 1). Notably, mitogens, including epidermal growth factor (EGF) and insulin, are well-established activators of RPS6 phos-phorylation. However, spatiotemporal fine regulation of RPS6 phosphorylation by multiple extracellular proteins still remains to be elucidated. PDGF-BB (platelet-derived growth factor-BB) induces p-RPS6 (S235/236) through the AMPK/mTORC1 pathway in vascular smooth muscle cells [92] Serum induces p-RPS6 (S235/236 and S240/244) through mTOR-and MEK-dependent mechanisms [48] increases p-RPS6 via S6Ks and translation of 5 -terminal oligopyrimidine track (5 -TOP) mRNAs in Swiss mouse 3T3 cells [93] TNFα (tumor necrosis factor α) induces p-RPS6 (S235/236) by activating S6K [94,95] WNT3A (WNT family member 3A) induces p-RPS6 (S240/244) and p-S6K (T389) in a manner independent on the β-catenin-dependent transcription [96] (2) Kinases and phosphatases that regulate RPS6 phosphorylation The human genome is known to encode 656 kinases and approximately 184 phosphatases [97]. Multiple kinases and phosphatases contribute to fine regulation of RPS6 phosphorylation ( Table 2). The upstream regulators of RPS6 are controlled in a contextand/or cell-type-dependent manner. In addition, the spatiotemporal regulation is more complex than the simple snapshot illustrated in Figure 2. For example, the activity of AKT is generally regulated by PI3K in response to growth factors, leading to activation of the mTORC1/S6K/RPS6 signaling by inhibiting TSC2 [98,99]. However, DNA damage induces AKT activation via the DNA-dependent protein kinase (DNA-PK) and activated AKT suppresses the RAF/MEK/ERK/S6K1 signaling, thereby reducing the p-RPS6 (S235/236) level, independently of TSC2, mTORC1, or p53 [100]. Under these conditions, the DNA damage-inducible transcript 4 protein (DDIT4)/regulated in development and DNA damage responses 1 (REDD1) inhibits the AKT signal to mTORC1 by sequestering the 14-3-3 protein to activate TSC. In addition, the interaction of Sestrin-2 with the 5 -AMPactivated protein kinase (AMPK) inhibits the AKT-mediated inhibition of AMPK upon DNA damage. Therefore, DNA damage downregulates p-RPS6 (S235/236) by inhibiting the RAF/MEK/ERK/S6K1 signaling through DNA-PK-activated AKT even in the presence of growth factors. These results suggest that changes in p-RPS6 level that are induced by drugs or other stimuli do not always reflect changes in the mTORC1 signaling [100]. The downstream effectors of mTORC1, S6K1, and eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) have been also reported to be differentially regulated by exogenous signals. For example, DNA damage suppresses S6K1-mediated RPS6 phosphorylation in a DNA-PK/AKT-dependent manner, whereas p-4E-BP1 does in a p53 and p63 activity-dependent manner [100].
Results derived from experiments involving small-molecule PKIs have suggested new protein kinases as regulators of RPS6 phosphorylation. For example, pan-PIM kinase inhibitors such as AZD1208 [215], LGB321 [132], and SGI-1776 [133], have been reported to downregulate p-RPS6. BI2536, an inhibitor of Polo-like kinases (PLKs), also inhibits RPS6 phosphorylation [136]. These results suggest that PIM and PLKs regulate p-RPS6 either directly or indirectly. Other potential upstream regulators of RPS6 phosphorylation, such as STK33 [143] and IKK [115], may also be identified by the applications of kinase-specific PKIs.
Various protein kinases crosstalk with each other, forming both positive and negative feedback and feedforward loops ( Figure 2). For example, pharmacological inhibition of mTORC1/S6K suppresses the negative feedback loops leading to the activation of compensatory upstream signaling molecules including AKT [216][217][218]. Notably, targeting mTORC1 with rapalogs as monotherapies to treat tumors has had limited success in hundreds of clinical trials [219]. Therefore, targeting RPS6 and its downstream rather than upstream effectors might be an alternative approach for cancer treatment. Table 8. Selected examples of the agents regulating the upstream of the RPS6 signaling.

Subcellular Localization of RPS6
As a component of the 40S ribosome, RPS6 is transported from the cytoplasm, where its translation occurs, to the nucleolus, where it is assembled into the 40S ribosome [321]. Then, the 40S ribosome is released to the nucleoplasm before its transport into the cytoplasm through the nuclear pores. RPS6 has three putative NLSs (Figure 1), and removal of all these NLSs results in the failure of its nuclear import [322]. The functions of RPS6 in all these cellular locations remain largely unknown.
In addition, p-RPS6 has been found in both the cytoplasm and nucleus [69], but differential nuclear/cytoplasmic distribution of p-RPS6, which depends on the phosphorylated sites, has also been demonstrated [323] with still unknown physiological significance. The nucleocytoplasmic localization of total (t)-RPS6 and p-RPS6 (S240/244) has been investigated. In TSC1 wildtype MEFs, the level of p-RPS6 (S240/244) was regulated in a cell cycle-dependent manner; the lowest levels were observed in the G 1 phase, a strong increase and a moderate decline in S phase, and the second peak in the G 2 /M phase [324]. The levels of p-RPS6 and the accompanying t-RPS6 were increased in the nucleus during the G 1 phase. The nuclear level of p-RPS6 reached the maximum in the early S phase, whereas the level of t-RPS6 rapidly declined. Concomitantly, the cytoplasmic RPS6 was strongly upregulated and peaked in the mid (t-RPS6)-to-late S (p-RPS6) phase.

Proteasome-Dependent Degradation of RPS6
RPs have been found excessively expressed beyond the amount required for efficient ribosome biogenesis, and their levels are controlled through continuous proteasomal degradation in the nucleoplasm [325]. However, the corresponding E3 ligase responsible for RPS6 stability has not been identified yet. Although the Drosophila Pallbearer (PALL) has been identified as an F-box protein that regulates proteasome-dependent RPS6 turnover [326], the role of the human homolog has not been investigated yet (see Section 3.2.11. Functions of RPS6 in Other Higher Eukaryotes).
Cyclic accumulation of RPS6 in the nucleolus, in accordance with the cell cycle, has been reported; it starts in the S phase, culminates in the G 2 phase, and diminishes in the M phase with the disintegration of the nucleoli [327]. In addition, mammalian RPS6 is dephosphorylated upon heat shock [328,329]. Heat shock protein 90 (HSP90) has been reported to interact with RPS3 and RPS6 and protect them from proteasomal degradation [330]. Consistently, the HSP90 inhibitor geldanamycin blocked the HSP90-RPS6 interaction and induced the degradation of RPS6. However, the mechanism for the regulation of RPS6 stability during cell cycle progress and by phosphorylation remains elusive.
Downregulation of t-RPS6 has been found in TNBC cells treated with the EGFR inhibitor (EGFRi) gefitinib and the MET inhibitor (METi) SU11274 [21] (see Section 5.2.1. RPS6-KD in Breast Cancer Cells). Although the reduction in t-RPS6 level was augmented until 16 h post-treatment, 4 h treatment of the cells with the proteasome inhibitor MG132 was not sufficient to inhibit the reduction in t-RPS6 level in the presence of gefitinib and SU11274 [21]. Further studies will be necessary to reveal the role of RPS6 stability in the combinatorial PKI therapies against cancer.

Ribosome Biogenesis and Protein Synthesis
RPS6 plays a role in the maturation of pre-rRNA. The siRNA-based RPS6-KD in HeLa cells resulted in the accumulation of 30S pre-rRNAs and the decrease in mature 18S rRNAs without perturbing the formation of mature 28S and 5.8S rRNAs [331]. Similar to this observation, in a mouse model with conditional deletion of both Rps6 alleles in liver, the decease in 18S rRNAs was observed, and 34S pre-rRNAs (equivalents of human 30S pre-rRNAs) were accumulated in liver cells [332]. However, the hepatic hypertrophy but not hyperplasia in fasted animals in response to nutrients was not blocked by the absence of RPS6. A recent study has demonstrated that p-RPS6 is involved in the endonuclease processing of 30S pre-rRNAs in the nucleolus [327]. The splicing of 30S pre-rRNAs was only mediated by fully phosphorylated RPS6 in the C-terminal five serine residues. Additional genetic manipulations of RPS6, including tissue-specific knockouts of both alleles or conditional deletion of one allele, further support its indispensable role in the ribosome in the thymus, spleen, and lymph nodes [333]. More importantly, Rps6 gene haploinsufficiency leads to embryonic lethality during gastrulation, preceded by failure in entering mitosis and induced apoptosis [334]. As in heterozygous Rps6 T cells [333], the embryonic lethality of heterozygous Rps6 mice might be due to a p53-mediated checkpoint during gastrulation [334]. P-RPS6 controls translation at the level of mRNA-binding in dividing cells [335][336][337]. P-RPS6 (S235/236) is also induced by RAS/RAF/ERK/RSK signaling, and p-RPS6 is recruited to the 7-methylguanosine cap structure [48].
RPS6 has been reported to play a role in the translation of mRNAs with a polypyrimidine tract at their 5 -terminal oligopyrimidine track (5 -TOP mRNAs) [338]. High levels of RPS6 have been found in primary diffuse large B-cell lymphoma (DLBCL) cell lines and patient samples. RPS6-KD reduced the proliferation of DLBCL cell lines [338] and increased the 5 -TOP mRNA translation [338,339]. A study reports that the translation of 5 -TOP mRNAs is downregulated by rapalog treatment, suggesting a role of p-RPS6 in the regulation of 5 -TOP mRNA translation [93]. Another study has reported that MEFs with the non-phosphorylated mutant RPS6 in all five serine sites (rpS6 P−/− ) and 70 kDa S6K1-null (p70S6K1 −/− ) display efficient translation of 5 -TOP mRNAs in response to mitogens [90]. RPS6 has been found to associate with 5 -TOP mRNAs, such as RPS8, RPL11, RPL16, and RPS24, through the 5 -TOP sequences to inhibit their translation [338]. RPS6-KD in the human breast cancer cell MCF7 and the human cervical carcinoma cell HeLa increases the number of these 5 -TOP mRNAs in actively translating polysomes.

Extra-Ribosomal Functions of RPS6
In addition to the roles in translational control, RPS6 has extra-ribosomal functions ( Figure 3). A study has reported that approximately 5% of endogenous RPS6 is detected in ribosome-free subcellular fractions [340]. Here, we briefly describe the functions of RPS6 in various cellular processes.

Regulation of Cell Cycle, Proliferation, and Growth
The relationship between p-RPS6 and cell cycle regulation was first identified in Xenopus eggs [341]. Hyperphosphorylation of RPS6 was correlated with the oncogenic RAS-induced cell-cycle arrest.
RPS6 deficiency has been demonstrated to induce the p53-dependent cell-cycle arrest that is antagonized by the depletion of p53 [339]. This induction of p53 is dependent on the upregulation of ribosomal protein L11 (RPL11). Loss of RPS6 disrupts the 40S ribosome biogenesis, leading to selective upregulation of the translation of 5 -TOP mRNAs [339]. RPL11, which is translated from one of these 5 -TOP mRNAs, binds to MDM2, the E3 ligase of p53. Under normal conditions, the activity of the tumor suppressor p53 is tightly regulated by MDM2, through a ubiquitin-dependent proteasome pathway [342,343]. RPL11-MDM2 association stabilizes and activates p53 by preventing MDM2-p53 binding [6]. Induction of p53 led to the expression of its target genes including cyclin-dependent kinase inhibitor 1a (Cdkn1a; the gene encoding p21), BCL2-associated X (Bax), and Mdm2, in the liver of Rps6-deleted mice [339]. Previous results also support the p53-dependent cell-cycle arrest and apoptosis in other tissues with the deletion of one Rps6 allele [333,334]. Notably, RPS6-KD is also known to upregulate RPL11 in the human breast cancer cell line, MCF7, and the human cervical carcinoma cell line, HeLa [338].
Failure of hepatic cell proliferation or cyclin E induction after partial hepatectomy has been reported in RPS6-deficient mice despite the presence of active cyclin D-CDK4 complexes [332]. These results suggest the induction of a checkpoint control by abrogation of 40S ribosome biogenesis, leading to the prevention of cell cycle progression. The depletion of RPS6 induced the p53-dependent cell-cycle arrest [339]. Newborn rpS6 P−/− mice have been demonstrated to contain high DNA content [77]. In addition, rpS6 P−/− MEFs displayed a shorter population-doubling time with a short G 1 phase compared with those of the wild-type MEFs. However, the mechanism of its regulation of cell proliferation remains to be determined.
The suppressive effect of short hairpin RNA (shRNA)-based RPS6-KD on cell proliferation has been observed to increase with time in two lung cancer cell lines, A549 and H520 [37]. Inversely, the expression of senescence-associated β-galactosidase (SA-β-gal) was increased by RPS6-KD in both cell lines. Additionally, RPS6-KD increased the number of cells in the G 0 /G 1 phase and decreased the number of cells in the G 2 /M phase, concurrently with reduced levels of p-RB and cyclin D1, whereas no significant changes were observed in the levels of cyclin A, cyclin E, and total RB. More interestingly, the levels of CDK inhibitors (CKIs), including p16, p21, p27, and p57, were increased. However, RPS6-KD did not induce apoptosis and altered the expression of B-cell lymphoma-extra large (BCL-xL), BAX, and caspase-3. The anticancer effect of RPS6-KD was reproduced in a xenograft model: A549 lung cancer cells with RPS6-KD resulted in reduced tumorigenicity with an increase in the number of SA-β-gal(+) cells in xenograft tissues. Similar changes were observed in the expression of cell cycle regulators and CKIs [37].
Functional implications of RPS6 in cell cycle regulation were also demonstrated by RPS6-KD in a variety of cancer cells. RPS6-KD induced G 0 /G 1 cell cycle-arrest in human NSCLC cells [37,38], ovarian cancer cells [40], and drug-resistant melanoma cells [39] (see Section 5.2. RPS6-KD in Cancer Cells).
The levels of p-RB, cyclin D1, cyclin E, CDK2, CDK4, or CDK6 by RPS6-KD were reduced in NSCLC cells [38], ovarian cancer cells [40], and drug-resistant melanoma cells [39]. Hypophosphorylated RB inhibits cell cycle progression by binding to the transcription factor E2F and repressing the transcription of E2F target genes that are required for G 1 -to-S transition [36]. The phosphorylation of RB is mediated by the cyclin D/CDK4/6 complex and derepresses the transcription of E2F target genes [36]. It remains to be studied how RPS6 regulates the G 0 /G 1 checkpoint regulators in NSCLC, ovarian cancer cells, and melanoma cells. In contrast, the levels of the CKIs p21 and p27 were increased by RPS6-KD in NSCLC cells [38]. These CKIs inhibit CDK activity either by disrupting the CDK4/6-cyclin complexes (p16) or by binding to both the cyclin and CDK in the complexes (p21, p27, and p57) [344]. Inversely, overexpression of RPS6 in the normal human bronchial epithelial (HBE) cell line promoted the cell proliferation with concurrent increases in the levels of p-RPS6, p-RB, cyclin D1, cyclin E, CDK2, and CDK4 and decreases in CKIs, p21 and p27, and the number of cells at the G 0 /G 1 phase [38].
It has been demonstrated that cell growth (increase in size/volume) is regulated mainly by the mTORC1 pathway. RPS6 phosphorylation by S6K1 has been demonstrated to be directly involved in the positive control of cell size. Various cells derived from rpS6 P−/− mice, including MEFs, fetal liver cells, pancreatic β-cells [345], and muscle myotubes [78], have been reported to be significantly smaller than the wild-type controls. However, pancreatic acinar cells displayed a similar size regardless of S6K1 deficiency [346] or RPS6 phosphorylation defect [77].

Regulation of Cell Migration
The migration of NSCLC cells is also reduced by RPS6-KD, which has been shown to downregulate the level of proteins involved in cell migration, including N-cadherin, vimentin, matric metallopeptidase-9 (MMP-9), MMP-2, and p-paxillin. Conversely, Ecadherin is increased by RPS6-KD [38]. In contrast, the migration of HBE cells is enhanced by RPS6 overexpression. Additionally, RPS6 overexpression upregulates N-cadherin, vimentin, MMP-2, and p-paxillin in HBE cells [38]. Induction of the epithelial-mesenchymal transition (EMT) markers suggests that overexpression of RPS6 contributes to metastasis of cancers [347].
Cell migration is also linked to angiogenesis and the tumor microenvironement. In fact, activation of the mTOR/RPS6 pathway has been associated with cancer cell survival, inflammation, and neoangiogenesis through various upstream regulators [84,95,310,348]. Since tumor growth is a result of the constant crosstalk between a tumor and its surrounding microenvironment, leading to neoangiogenesis and immune escape, further studies on the role of RPS6 in this crosstalk contribute to developing additional strategies combining anti-angiogenic therapy and immunotherapy against cancer [349,350].

Regulation of Apoptosis
Unphosphorylated RPS6 has been demonstrated to induce apoptosis via a mechanism involving the tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) by inducing the expression of death receptor 4 (DR4) [340] (see Section 5.2.2. RPS6-KD in Cervical Carcinoma Cells and Section 5.2.7. RPS6-KD in Hematopoietic Cancer Cells). MEFs that express unphosphorylated RPS6 have been demonstrated to be more sensitive to TRAILinduced apoptosis than the control MEFs. Similarly, the human HCC cell line, SK-HEP-1, expresses a low level of p-RPS6 and is more vulnerable to TRAIL-induced apoptosis than other tumor cells that express a high level of p-RPS6. Consistently, ectopic expression of the phospho-defective RPS6 mutant (RPS6 SS235/236AA ) in HeLa cells resulted in an increase in TRAIL-induced apoptosis compared to that of the phospho-mimic RPS6 mutant (RPS6 SS235/236DD ) [340]. Conversely, the overexpression of RPS6 SS235/236DD had no effect on the TRAIL-induced apoptosis. Since S6K1, the upstream kinase of RPS6, has anti-apoptotic activity, other apoptotic regulators, rather than RPS6, may contribute to the S6K1-dependent apoptosis induced by TRAIL [340]. More importantly, the N-terminus of RPS6 (aa 1-70) without the phosphorylation residues has been reported to carry pro-apoptotic activity to sensitize HeLa cells to TRAIL. These results imply a phosphorylation-dependent negative regulatory effect of the C-terminus of RPS6 on the pro-apoptotic activity of RPS6 [340].
DNA damage-regulated autophagy modulator protein 1 (DRAM1) is a direct target of p53 during DNA damage. It induces autophagy and is essential for the p53-dependent apoptosis [351]. Recently, negative regulation of p-RPS6 by DRAM1 has been reported [166]. Overexpression of DRAM1 downregulated the levels of p-RPS6 (S235/236) and p-RPS6 (S240/244) in an mTORC1-dependent manner in HEK293T cells. DRAM1 was also found to be localized at the plasma membrane to regulate IGF1R phosphorylation. Additionally, the overexpression of DRAM1 reduced the viability and inhibited the colony formation of the human colon cancer cell line SW480 [166]. The role of the DRAM1/RPS6 axis in apoptotic regulation remains to be determined.

Roles in the DNA Damage Response
The expression of rpS6 P−/− in mice with the oncogenic Kristen rat sarcoma (KRAS) gene background has shown an increase in p53 expression along with the increased staining of phosphorylated H2A histone family member X (γ-H2AX) and p53-binding protein 1 (53BP1) in areas of acinar ductal metaplasia [17]. First, the administration of rapamycin, the selective inhibitor of mTORC1, once every other day for 1 month to mice implanted with a 7,12-dimethylbenz[a]anthracene (DMBA)-soaked cotton pledget in the pancreas significantly decreased the score of pancreatic intraepithelial neoplasia (PanIN) lesions compared with the score in the untreated mice. Histological examination has revealed that no RPS6 phosphorylation is observed in the pancreas of DMBA-treated mice when treated with rapamycin, whereas strong p-RPS6 is observed without rapamycin treatment. Transgenic mice expressing the oncogenic KRAS G12D in the pancreatic epithelium develop PanIN lesions that infrequently progress to pancreatic ductal adenocarcinoma (PDAC) [371]. These mice have a strong and uniform expression of p-RPS6 in the acinar cells [17]. Additionally, rpS6 P−/− mice treated with DMBA showed attenuated development of PanIN lesions. Furthermore, mice with both a KRAS G12D mutation and rpS6 P−/− had a significantly lower score of PanIN lesions compared with the score in the mice with a KRAS G12D mutation and one or two wild-type RPS6 alleles. Further analysis revealed that frequent p53-positive cells were detected in mice with both KRAS G12D mutation and rpS6 P−/− . Finally, the DNA damage markers, γ-H2AX and 53BP1, were highly expressed in the mice with both KRAS G12D mutation and rpS6 P−/− . γ-H2AX is recruited to the chromatin domains near DNA double-strand breaks (DSBs) [372] and 53BP1 is colocalized with γ-H2AX and required for p53 accumulation in response to DNA damage [373]. Taken together, these results suggest the potential role of p-RPS6 in the attenuation of DNA damage in a mutant KRAS background, leading to the reduction of p53-dependent tumor suppression [17]. In relation to this observation, BEZ235, a dual inhibitor of PI3K and mTOR, has previously been reported to downregulate p-53BP1 (S25) and 53BP1 foci induced by the poly (ADPribose) polymerase inhibitor (PARPi) olaparib in breast cancer type 1 susceptibility protein (BRCA1)-mutant breast cancer cells [374].
ATM, the master regulator of the DNA damage response, is a serine/threonine protein kinase [375]. Upon DSB formation, ATM is phosphorylated and activated to induce cellcycle arrest, DNA repair, or apoptosis. The S247 of RPS6 is a substrate of ATM in response to UV irradiation or the treatment with the genotoxic drug doxorubicin and is inhibited by the ATM inhibitor KU-55933 [57]. ATM-mediated RPS6 phosphorylation has been reported to be S6K-independent.
Phosphorylation of RPS6 has been reported to attenuate DSBs in BRCA1-deficient breast cancer cells both in vitro and in vivo [376]. In olaparib-resistant BRCA1-deficient breast cancer cells, ionizing radiation (IR) decreased the number of γ-H2AX foci and increased the number of DNA repair protein RAD51 homolog 1 (RAD51) foci compared with the numbers of them in the parental cells. A remarkable increase in p-RPS6 is the distinct feature of olaparib-resistant BRCA1-deficient cells, and abrogation of RPS6 phosphorylation completely reverses the formation of foci for γ-H2AX or RAD51. These results suggest that p-RPS6 is crucial to load the RAD51 recombinase onto DNA damage sites in BRCA1-deficient cells. In fact, BRCA1 is indispensable for RAD51 loading onto DNA damage sites, leading to homologous recombination repair (HRR) [377]. More importantly, data support that p-RPS6 confers olaparib-resistance to BRCA1-deficient breast cancer cells [see Section 4.3.4. RPS6 in Resistance to PARP Inhibitors (PARPis)].

Response to Oxidative Stress
Insulin has been demonstrated to induce the interaction of mTORC2 with the ribosome, leading to enhancement of mTORC2 activity in a translation-independent manner [378]. It has been reported that RPS6 plays a cardioprotective role in response to oxidative stress [52]. IPC, and insulin or opioid treatment, induced phosphorylation of RPS6 at S235/236 through the AKT/mTORC1/S6K pathway in perfused mouse hearts or neonatal rat ventricular myocytes. The p-RPS6 interacted with the rapamycin-insensitive companion of mTOR (RICTOR) to enhance mTORC2 kinase activity. RPS6-KD reduced the insulin-induced phosphorylation of mTORC2 and AKT (S473). Inversely, RPS6 overexpression upregulated p-AKT (S473). Additionally, RPS6-KD abrogated insulin-induced cardioprotection against the H 2 O 2 -induced oxidative stress [52]. Interestingly, S6K1 negatively regulates mTORC2 via the phosphorylation of RICTOR in the mouse adipocyte cell line 3T3-L1 [379]. Taken together, mTORC2 activity may be tightly controlled by RPS6-mediated positive feedback and S6K-induced negative feedback.

RPS6 in Cellular Senescence
Cellular senescence is one of the nine hallmarks for aging, which is characterized by constant G 1 cell-cycle arrest and an inflammatory response, the senescence-associated secretory phenotype (SASP) [380][381][382]. The mTORC1/S6K pathway has been recognized as the key signaling pathway responsible for aging and cellular senescence [383,384]. Since RPS6 is the downstream effector of this mTORC1/S6K pathway, p-RPS6 has been used as a marker for aging and premature senescence [385].
Contrary to this notion, a study found that there are no significant differences in the levels of mTOR or p-mTOR (S2448) between young and old human dermal fibroblasts (HDFs), whereas downregulation of S6K1, p-S6K1 (T389), and p-RPS6 was observed in replicative senescent HDFs [386]. In addition, p-RPS6, S6K1, and p-S6K1 were also downregulated in HDFs with autophagy impairment-induced premature senescence (AIPS) that was induced by siRNA-based knockdown of the autophagy related 7 (ATG7) or lysosomeassociated membrane glycoprotein 2 (LAMP2). Interestingly, reactive oxygen species (ROS) scavenging (through treatment of the antioxidant N-acetyl cysteine (NAC)) or p53 inhibition (either through treatment of a p53 inhibitor, pifithrin-α, or knockdown) restored the levels of p-RPS6, S6K1, and p-S6K1 in AIPS HDFs. However, the exact roles of RPS6 and its phosphorylation in cellular senescence have not been elucidated yet.

Roles of RPS6 in Erythropoiesis
Congenital (Diamond-Blackfan anemia; DBA) and acquired (5q-syndrome) hypoproliferative macrocytic anemia share a common erythroid phenotype of RP haploinsufficiency [387]. Although mutations of the RPS6 gene have not been reported in this anemia, mice lacking one Rps6 allele postnatally display features of the 5q-syndrome, such as macrocytic anemia, erythroid hypoplasia, and megakaryocytic dysplasia with thrombocytosis [388]. Additionally, mice with heterozygously deleted Rps6 have also been demonstrated to phenocopy the 5q-syndrome [389]. However, the clinical significance of these findings remains elusive.

Roles of RPS6 in the Central Nervous System
Increasing evidence indicates p-RPS6 as a marker of neuronal activation during synaptic plasticity [390,391]. Various pharmacological stimuli also induced the p-RPS6 in neurons [392]. For example, massive phosphorylation of RPS6 at S235/236 and S240/244 by a proconvulsant drug, such as kainite, pilocarpine, pentylenetetrazol, or dopamine D1 receptor (DRD1) agonist SKF81297, has been demonstrated. In addition, drugs of abuse and antipsychotics also regulate RPS6 phosphorylation [392]. Interestingly, the p-RPS6 level in schizophrenia is reduced [393]. Elevated expression or phosphorylation of RPS6 has also been found in multiple sclerosis [394]. The roles of RPS6 in diseases of the central nervous system remain to be determined.

Roles of RPS6 in Response to Infection
Mounting evidence suggests the roles of RPS6 during pathogen infection [395]. The induction of p-RPS6 in HeLa cells by vaccinia virus infection was reported as early as 1976 [396].
RPS6 has been demonstrated to be associated with latency-associated nuclear antigen (LANA). LANA is a multifunctional protein of Kaposi's sarcoma-associated herpesvirus (KSHV) that is tightly associated with primary effusion lymphoma (PEL), Kaposi's sarcoma I, and multicentric Castleman's disease (MCD) [397]. RPS6 binds to the N-terminal domain of LANA in a nucleic acid-independent manner in a PEL cell line, BC-3 [398]. Interestingly, RPS6 increased the transcriptional activity of the LANA protein on the LANA promoter. Moreover, shRNA-based RPS6-KD reduced the stability of the LANA protein, while increasing the stability of p53 in BC-3 cells. The half-life of the LANA protein was markedly reduced from several days [399] to 0.6 h by RPS6-KD [398]. Consistent with the increased p53 stability, the upregulation of p21 protein levels was also induced by RPS6-KD. These effects led to anti-proliferative effects in BC-3 cells [399]. The underlying mechanism of the RPS6-dependent regulation of LANA stability is subject to further study.
The open reading frame 45 (ORF45) of KSHV has been reported to induce p-RPS6 (S235/236) by directly activating RSKs through the mTORC1/S6k-dependent signaling [400]. Phosphorylation of RPS6 and eukaryotic translation initiation factor 4B (eIF4B) through the ORF45/RSK axis has been suggested to contribute to the translational control during KSHV lytic replication.
An indispensable role of RPS6 has been found in hepatitis C virus (HCV) propagation. HCV is a single-stranded RNA virus whose translation is initiated in a cap-independent manner by its internal ribosome entry site (IRES) [401]. Reduction of the 40S ribosome abundance by RPS6-KD was reported to selectively suppress HCV IRES-mediated translation without affecting the translation of the host cells [402]. Only knockdown of the 40S RPS genes, but not the 60S RPL genes, of the ribosomal subunit inhibited HCV translation.
Protein phosphatase 2A (PP2A) is a heterotrimeric serine/threonine phosphatase, which is composed of a structural A subunit, a catalytic C subunit, and >20 regulatory B-type subunits resulting in >80 different PP2A holoenzymes [403]. The composition and function of the PP2A holoenzyme are regulated through methylation and demethylation by leucine carboxyl methyltransferase 1 (LCMT1) and protein phosphatase methylesterase 1 (PME1), respectively [404]. Cellular transformation either by polyomavirus or by simian virus 40 (SV40) occurs through activation of p-RPS6 (S235/236) via inhibiting PP2A by replacing the regulatory B-type subunits from the PP2A heterotrimer by viral oncoproteins such as the polyomavirus middle (PyMT) and small (PyST) tumor antigen and the SV40 small tumor antigen (SVST) [137]. PyST and SVST preferentially bind to PP2A instead of the methylated PP2A catalytic subunit, which is mediated by LCMT1, whereas the binding of PyMT is not. The methylated PP2A downregulates p-RPS6 by inactivating AKT and S6K. The methylated PP2A also reduced p-RPS6 in an S6K-independent manner. Similarly, PyMT and PyST enhance RPS6 phosphorylation [405,406].
RPS6 phosphorylation is also increased in cells infected by pathogens such as Rift Valley Fever (RVF) Virus, Herpesvirus, Plasmodium, and Toxoplasma [407][408][409][410]. Pharmacological inhibitions of the mTORC1/S6K/RPS6 pathway inhibits the infection or pathogenesis of these pathogens [407,408]. Additionally, p-RPS6 is activated by an unknown regulator of the non-canonical pathway in Plasmodium-infected hepatocytes [407]. These data suggest that parasites override the RPS6 pathway to support their growth in infected cells.

Functions of RPS6 in Other Higher Eukaryotes
In the plant Arabidopsis, RPS6 negatively regulates rDNA transcription by associating with the histone deacetylase, AtHD2B [411], which might be antagonized by the interaction of RPS6 with the histone chaperone, AtNAP1 [412].
RPS6 has been identified as a negative regulator of efferocytosis in Drosophila. Efferocytosis is a process of clearing apoptotic cells via phagocytosis by professional or amateur phagocytes [413]. Since approximately 3 × 10 11 -5 × 10 11 cells in our body die daily via apoptosis, clearing dead cells is crucial to maintain tissue homeostasis [414,415]. PALL is an F-box protein in Drosophila, and a component of the Skp-Cullin-F-box (SCF) E3 ubiquitin ligase complex [416]. PALL has been found to bind to p-RPS6 to promote RPS6 ubiquitylation and proteasomal degradation [326]. Degradation of RPS6 promoted efferocytosis through Rac family small GTPase 2 (RAC2) activation and F-actin remodeling in Drosophila cells. Consistently, RPS6-KD enhanced the efferocytosis and RPS6 overexpression reduced the clearance of apoptotic cells. Although F-box only protein 28 (FBXO28) has been identified as a human homolog of PALL [417], whether FBXO28 binds to and degrades p-RPS6 in mammalian cells remains to be determined.

RPS6 in Cancer
Although it still needs further investigation whether p-RPS6 is just a byproduct of tumorigenic pathway activation or a prerequisite for tumorigenesis in various types of cancers, we provide evidence here supporting the potential roles of RPS6 in tumorigenesis in humans and suggesting this protein as a potential therapeutic target against cancer.

Roles of RPS6 in Tumorigenesis
A knock-in mouse model of a non-phosphoylatable RPS6 mutant (rpS6 P−/− ) has demonstrated that p-RPS6 is indispensable to develop PDAC in KRAS G12D mutation background [17]. The importance of p-RPS6 in tumorigenesis was further demonstrated in mice expressing constitutively active AKT (MyrAKT1) in pancreatic β-cells in the background of rpS6 P−/− . Defects in the phosphorylation of RPS6 rescued the MyrAKT1-induced reduction in the nuclear level of p27 [418], a CKI [419,420]. In addition, p-RPS6 deficiency reduced the development of the MyrAKT1-induced hyperplasia and tumor formation in β cells (insulinoma). RPS6 phosphorylation deficiency also increased the overall protein synthesis with concomitant reduction in translation fidelity, leading to unexpected resistance to proteotoxic (MG132) or genotoxic (etoposide) stress-induced apoptosis in MyrAKT1expressing cells. However, p-RPS6 deficiency failed to inhibit, and instead enhanced, the MyrAKT1-triggered aneuploidy, and increased the number of β cells and amount of insulin secretion [418]. Although several nuclear proteins, including PC4 and SFRS1-interacting protein 1 (PSIP1), serine/arginine-rich splicing factor 1 (SRSF1), and DNA topoisomerase IIβ (TOP2B), have been identified as binding partners of unphosphorylated RPS6, the functional significance of these interactions has not been elucidated yet [418]. Since rpS6 P−/− could not prevent thymic lymphomatogenesis in mice with a constitutively active AKT2 in immature T cells [421], the anti-tumorigenic effect of rpS6 P−/− may be tissue-specific.
P-RPS6 has been reported to promote the translation, not transcription, of hypoxiainducible factor 1-alpha (HIF1α) upon the activation of an environmental carcinogen, arsenite, leading to carcinogen-induced transformation of the normal mouse epidermal Cl41 cells [422]. Arsenite treatment induced the increase in p-RPS6 (S235/236). P-RPS6-dependent HIF1α expression was negatively regulated by the CKI, p27, via the RAS/RAF/MEK/ERK/RSK pathway, but not via the AKT/S6K pathway. P27 is a cell cycle regulator that binds and inhibits cyclin D/CDK4 [423,424], cyclin D/CDK6 [425], and cyclin E/CDK2 complexes [426]. CDKN1B (the gene encoding p27)-KD induced p-RPS6 even in the absence of arsenite. Arseniteinduced p-RPS6 was augmented by CDKN1B-KD and in an arsenite dose-dependent manner. Mutational analyses further demonstrated that the induction of HIF1α translation in p27deficient cells was p-RSP6 (S235/236)-dependent. Interestingly, CDKN1B depletion abrogated the expression of PHPLL, a negative regulator of RAS [427], RAF1 [428], and AKT [429,430], resulting in the activation of the RAS/RAF/MEK/ERK/RSK pathway [422]. Notably, SRC, the upstream regulator of the RAS/RAF/MEK/ERK pathway, phosphorylates p27 and induced its proteolysis [431]. In addition, hypoxia is an important inducer of neoangiogenesis, and activated RPS6 may play an important role in the crosstalk between endothelial cells and tumor cells in the tumor microenvironment [84,95,310,348]. For example, it has been reported that RPS6 is activated by tumor microenvironment conditions and hypoxia in endothelial cells [432].

RPS6 in Acute Myeloid Leukemia (AML)
The shRNA-based knockdown of CSNK1 (the gene encoding CK1) downregulated p-RPS6 (S244/247) concurrently with increased p53 activity in primary mouse MLL-AF9 leukemia cells [448]. CSNK1-KD had anti-leukemia efficacy both in vitro and in vivo. The anti-leukemic effects of CSNK1-KD were further supported by the rescue of CK1 functions by an shRNA-resistant CSNK1 cDNA, which inhibits the anti-leukemic effects of CSNK1-KD, whereas kinase-dead CK1 D136N [449] cDNA did not [448]. The regulation of RPS6 activity by CK1-mediated phosphorylation has been reported previously [56]. Interestingly, CSNK1-KD also induced the levels of p53 and p21 with concurrent induction of apoptosis and G 1 arrest in MLL-AF9 leukemia cells [448]. Previously, it has been demonstrated that dephosphorylation of RPS6 causes p53 activation in pancreatic cancer cells in response to DNA damage [17]. In addition, RPS6 depletion has been reported to induce p53-dependent cell-cycle arrest [339]. The p53 induction was mediated by the upregulation of RPL11, which binds to and inhibits MDM2 in RPS6 depleted cells [6,339] (see Section 3.2.1. Regulation of Cell Cycle, Proliferation, and Growth). Overexpression of a phosphomimetic mutant RPS6 S5D [56] partially rescued the CSNK1-KD-induced proliferation defect in leukemia cells [448]. Taken together, the relationship between p-RPS6 and p53 in AML or other cancer cells will be an interesting topic in the future.

RPS6 in Breast Cancer
High levels of p-RPS6 (S235/236), but not p-ERK (T202/Y204) or p53, have been found to strongly correlate with high Ki-67 expression in ER(+)/HER2(−) breast cancer clinical samples [434]. However, there was no significant difference in relapse-free survival (RFS) between cancer patients with high p-RPS6 levels and those with low p-RPS6 levels. In addition, high-level expression of p-RPS6 (S240) correlated with high Ki67 expression and short overall survival (OS) of breast cancer patients [450].
The t-RPS6 protein has been reported to be downregulated in TNBC cells treated by the EGFRi gefitinib and the AKT inhibitor MK2206 [451]. TNBC cells, especially mesenchymal stem-like (MSL) subtype cells, have intrinsic resistance to EGFRi [21,[451][452][453][454][455][456]. Treatment of gefitinib alone did not reduce the viability or colony-forming ability of the MSL subtype cells [21,451,455]. The combination of gefitinib and MK2206 induced synergistic anti-proliferative and anti-colony forming effects in HS578T and MDA-MB-231 cells. No significant suppression of the PI3K/AKT and MAPK pathways was observed in TNBC cells treated with gefitinib alone. The treatment of MK2206 alone reduced the levels of p-AKT (S473), p-mTOR (S2448), p-PRAS40 (T246), p-4E-BP1 (T37/46), and p-RPS6 (S235/236). Interestingly, MK2206 alone induced downregulation of t-RPS6 [451]. The reduction of t-RPS6 was induced by the gefitinib+MK2206 treatment in a dose-dependent manner. Since no significant change in the level of p-ERK1/2 (T202/Y204) was observed, RSK may not contribute to the t-RPS6 downregulation by the gefitinib+MK2206 treatment. Moreover, siRNA-based RPTOR-KD, but not RICTOR (rapamycin-insensitive companion of mTOR), downregulated the t-RPS6 protein and this reduction was potentiated by gefitinib addition, supporting the idea that mTORC1 controls the level of t-RPS6 in TNBC cells. RPTOR-dependent downregulation of t-RPS6 inhibited the proliferation of TNBC cells in the presence of gefitinib [451]. Although these results suggest RPS6 as an attractive therapeutic target to treat TNBC, the underlying mechanism maintaining the RPS6 level and its contribution to EGFRi resistance remain to be determined.
RPS6 has been reported to contribute to the regulation of intraductal colonization of basal-like breast cancer cells [457]. In an immortalized basal-like breast epithelial cell line, RPS6 dephosphorylation initiated the keratinization process. In addition, dephosphorylated RPS6 triggered detachment-induced keratin 5 (KRT5) upregulation posttranscriptionally. P-RPS6 was not detected in suspension cells in 3D cultures, whereas reconstitution of p-RPS6 by a constitutively active S6K mutant attenuated keratinization. Consistently, pharmacological inhibition of RPS6 phosphorylation caused sporadic keratinization in attached cells. Reduced RPS6 phosphorylation was also identified in a prolonged detached TNBC cell line, MDA-MB-468 cells, and pharmacological inhibition of p-RPS6 also caused keratinization in attached MDA-MB-468 cells [457].

RPS6 in Gastric Cancer
AMPK negatively regulates mTORC1 activity through direct phosphorylation of TSC2 and RPTOR ( Figure 2): (1) It phosphorylates and activates TSC2, which inhibits the mTORC1 kinase activity [96,458]; and (2) it phosphorylates RPTOR, leading to its binding to 14-3-3, which suppresses the mTORC1 activity [459]. High levels of p-RPS6 and low levels of p-AMPKα have been reported to be associated with gastric tumor progression and to be independent predictors of patient survival after resection of primary cancer [437]. The median OS values of 28 vs. 73 months and 28 vs. 78 months in patients with positive vs. negative expression of p-RPS6 and with negative vs. positive expression of p-AMPKα, respectively.

RPS6 in Glioblastoma
High-level expression of RPS6 is strongly correlated with GBM stem cell (GSC) markers, Nestin and SRY-box transcription factor 2 (SOX2), and an oligodendrocyte progenitor cell marker, oligodendrocyte transcription factor 2 (OLIGO2), in GBM tissues [438]. High levels of RPS6 and SOX2 are detected in high-grade GMB samples compared with the levels in the low-grade samples. RPS6 may contribute to the stemness of GSC through activation of the JAK/STAT3 pathway to induce these stemness-related proteins (see Section 5.2.5. RPS6-KD in Glioblastoma Cells). RPS6-KD reduced p-JAK2 and p-STAT3 and suppressed the tumor sphere formation, a characteristic of GSCs, of GBM cells in vitro. Inversely, RPS6 overexpression facilitated the tumor sphere formation, and the STAT3 inhibitor, AG490, antagonized the RPS6-enhanced sphere formation.

RPS6 in Head and Neck Cancer
In patients with HNSCC, double-positive p21 and p-RPS6 present better diseasespecific survival [439]. The p21-p-RPS6 double-positiveness was determined to be mTORC1dependent but not p53-dependent. The inhibitory phosphorylation of 4E-BP1 by mTORC1 stabilizes p21 by inhibiting its degradation that is induced by its interaction with 4E-BP1 [460]. Consistently, the PI3K/AKT/mTORC1 pathway is frequently activated in HNSCC [439].
High levels of p-RPS6 (240/244) are associated with shorter disease-free survival (DFS) than low levels in patients with ESCC [436]. There was no difference in the OS rate. The association was more significant in the early-stage ESCC patients than in the latestage patients. In addition, the high ratio of p-RPS6/t-RPS6 resulted in a more significant correlation with adverse prognosis than the p-RPS6 level alone [436]. Interestingly, RPS6-KD resulted in anticancer effects in ESCC cell lines (see Section 5.2.3. RPS6-KD in Head and Neck Cancer Cells).

RPS6 in Lung Cancer
The positivity of RPS6 and p-RPS6 (S235/236) is significantly higher in NSCLC clinical samples than in normal tissues (82.4% vs. 55.8% and 62.6% vs. 53.3%, respectively) [37]. It has been correlated with shorter median OS and DFS in NSCLC patients with high p-RPS6 levels than in patients with low levels (10 months vs. 60 months and 21 months vs. 48 months, respectively). However, although the positiveness of t-RPS6 showed differences, they did not reach statistical significance (30 months vs. 43 months and 26 months vs. 40 months, respectively; p > 0.05) [37]. In a separate study, hyperphosphorylation of RPS6 (S235/236) or the ratio of p-RPS6 to t-RPS6 was reported to be a predictive marker for survival of patients with NSCLC [38]. From the analysis of patient samples, the 5-year survival rate and median OS of patients with high levels of p-RPS6 were found to be significantly lower than those with p-RPS6(-) (3.0% vs. 26.5%; 20 months vs. 42 months, respectively). The significance was greater in patients with a high p-RPS6/t-RPS6 vs. patients with a low p-RPS6/t-RPS6 (2.1% vs. 32.0%; 12 months vs. 48 months, respectively) [38]. The predictability of the p-RPS6/t-RPS6 ratio was more powerful in patients of stage I NSCLC (8 months vs. 61 months). In addition, hyperphosphorylation of RPS6 is correlated with unfavorable clinical survival outcomes in NSCLC, lung adenocarcinoma, and GC [38,437,461].

RPS6 in Melanoma
The suppression of p-RPS6 has been identified as a predictive marker for improved progression-free survival (PFS) in BRAF-mutant melanoma treated with RAF or MEK inhibitors. A study conducted sensitivity assays of growth inhibition and apoptosis induction for the RAF inhibitor (RAFi), vemurafenib, and the MEK inhibitor (MEKi), selumetinib, against a panel of 16 BRAF-mutant melanoma cell lines. The results suggested that the inhibition of p-ERK is necessary but not sufficient to predict the sensitivity of melanoma cell lines to the RAFi or MEKi [440]. Further analysis revealed that the sensitivity of melanoma cell lines to these drugs correlates well with the decrease in the levels of p-RPS6 (S235/236). Resistant cell lines maintain the p-RPS6 levels after RAFi or MEKi treatment. The degree of suppression of RPS6 phosphorylation was a predictable indication of the sensitivity to these drugs not only in xenograft models but also in melanoma patients in a prospective evaluation. Although the number of patients was small (reduction vs. no reduction, n = 6 vs. n = 3, respectively), the reduction of p-RPS6 levels after treatment of the RAFi vemurafenib was correlated with an approximately 5-fold improvement in PFS [440].

RPS6 in Ovarian Cancer
High levels of RPS6 are associated with poor OS and PFS in OEC [40]. The RPS6 protein level was evaluated in clinical tissue samples, and poor clinical outcomes were found with median OS values of 30 vs. 42 months and median PFS values of 15 vs. 21 months in patients with high vs. low RPS6 levels, respectively [40]. High levels of p-RPS6 and NOTCH3 are associated with poor prognosis and a shorter OS (12.3 months) than the OS (81.9 months) of patients with low p-RPS6 and NOTCH3 levels in OEC [443]. Although NOTCH is known to activate the PI3K/mTORC1 pathway, the details are still largely unknown [462]. In addition, the roles of RPS6 in conjunction with NOTCH-related cancers remains to be determined.

RPS6 in Renal Cell Carcinoma (RCC)
In RCCs, RPS6 has been identified as the key mediator of the antitumor effects of everolimus (RAD001), a rapalog [445]. Everolimus binds to the 12 kDa FK506-binding protein (FKBP12) to form a complex, which binds to mTOR to destabilize and inactivate the mTORC1 complex [348,463]. Everolimus reduces the clonogenicity and proliferation of RCC cells by blocking protein biosynthesis but does not induce apoptosis. Knockdown of RPS6, but not EIF4EBP1 (the gene encoding 4E-BP1) or CDKN1B, abolishes the everolimusinduced anti-proliferative effect and blocks translation. More importantly, high levels of RPS6 and p-RPS6 have been found inversely correlated with the survival of patients with RCC [445].

RPS6 in Vulvar Squamous Cell Carcinoma (VSCC)
The non-viral-associated VSCC results from genetic and epigenetic changes in the absence of papillomavirus infection. Genetic alterations in the differentiated vulvar intraepithelial neoplasia (VIN) include allelic imbalance, microsatellite instability, and TP53 (the gene encoding p53) mutations [464,465]. P-RPS6 (S235/236)-positiveness has been identified in VSCC and differentiated VINs but not in basal and parabasal cells of the healthy vulvar epithelium. Additionally, laminin γ 2 expression has solely been identified within p-RPS6 positive regions [447]. The expression of laminin γ 2 is not detectable via immunohistochemical analysis in the normal epithelial basal cells, whereas its expression is a well-established marker of carcinoma [466][467][468].

RPS6 in Anticancer Drug Resistance
As mentioned earlier, the overexpression of RPS6 has been implicated in intrinsic or acquired drug resistance of cancer cells (see Section 3.2.4. Regulation of Drug Resistance) [39,[352][353][354][355][356][357][358][359][360]. More importantly, constitutive phosphorylation of RPS6 has been found to confer drug resistance in breast cancer cells, GC cells, and melanoma cells [39,359,360] The level of p-RPS6 in trastuzumab-resistant HER2-positive (+) breast cancer cells inversely correlated with their susceptibility to HER2-targeting drugs [360]. Trastuzumab (Herceptin ® ), a US FDA-approved drug, is a humanized monoclonal antibody against HER2 to treat HER2+ breast cancer [469]. However, the benefit of trastuzumab is limited by intrinsic or acquired resistance [470,471]. The phospho-proteins in the PI3K/mTORC1 pathway, including p-mTOR (S2448), p-AKT (S473), and p-RPS6 (S235/236), have been tested in response to trastuzumab in both sensitive and resistant HER2+ breast cancer cells [360]. Among these proteins, the p-RPS6 level and the degree of its response to the treated drugs could predict the susceptibility of breast cancer cells to not only trastuzumab but also other drugs, including rapamycin, AZD2014, BEZ235, erlotinib, lapatinib, MK226, and OSI-906. Drugs that could not downregulate p-RPS6, such as trastuzumab, erlotinib, MK2206, and OSI-906, failed to inhibit the growth of trastuzumab-resistant breast cancer cells [360]. Moreover, the combination of trastuzumab and lapatinib further reduced the p-RPS6 level, alongside concurrent synergistic inhibition of cell growth. These results suggested the p-RPS6 level as a predictive marker for the susceptibility of HER2+ breast cancers to anti-HER2 drugs and a feasibility marker of drugs for a combination strategy for the trastuzumab-resistant HER2+ breast cancers [360].

RPS6 in Resistance to MEK Inhibitors (MEKis)
Selumetinib (Koselugo TM ) is a very recently FDA-approved MEK1/2 inhibitor for patients with neurofibromatosis [472]. Anticancer activity of selumetinib has been tested in a panel of 12 CRC cell lines [353]. The study found that the levels of p-S6K1 (T3889) and p-RPS6 (S235/236) remained unaffected or increased by selumetinib treatment in cell lines with intrinsic resistance. Resistant primary tumors from patients also demonstrated a similar response. Pharmacological inhibition of the PI3K/AKT/mTORC1/S6K1 pathway or S6K1-KD overcame the selumetinib resistance in these CRC cells [353].
The phosphorylation level of mTORC1 downstream effectors, such as S6K1, 4E-BP1, and RPS6, has been found to be associated with resistance to the MEKi PD0325901 in a panel of 48 GC cell lines [359]. Interestingly, MEKi sensitivity was independent of the level of p-ERK suppression but dependent on the degree of suppression in the levels of p-S6K1 (T389), p-4E-BP1 (S65), and p-RPS6 (S235). On the contrary, no significant suppression of phosphorylation in these proteins was observed in resistant cells after MEKi treatment. Among them, the change in p-RPS6 level by MEKi showed the strongest correlation with MEKi sensitivity in vitro. This resistance of GC cells to MEKi was similarly recapitulated in xenograft models [359]. In addition, since the p-AKT level in sensitive GC cells was not affected by MEKi, the MEKi-dependent suppression of p-RPS6 was suggested to be mediated by mTORC1 in these cells. Interestingly, ERK regulates the level of p-RPS6 not only through RSK but also the TSC1/2 complex, the upstream negative regulator of mTORC1: (1) ERK1 induces inhibitory phosphorylation of TSC2 [74]; (2) ERK also phosphorylates and activates S6K [73]; and (3) ERK activates mTORC1 through RSK-mediated RPTOR phosphorylation ( Figure 2) [75]. Previously, p-RPS6 has been demonstrated to be a potential predictive marker for the responsiveness of RCC cells to everolimus [445]. Taken together, further evaluation is warranted for RPS6 and/or p-RPS6 as a companion diagnostic for the therapeutic use of various inhibitors for mTOR and MEK.

RPS6 in Resistance to RAF Inhibitors (RAFis)
Currently, BRAF V600E -selective inhibitors (BRAFis), such as dabrafenib, encorafenib, and vemurafenib, and MEK inhibitors (MEKis) including binimetinib, cobimetinib, and trametinib, have been approved by the US FDA for the treatment of melanoma [213]. However, similar to other clinical PKIs, acquired resistance to these inhibitors is rapidly developed after a transient clinical benefit [473]. The constitutive phosphorylation of RPS6 has been found in BRAF V600E -mutant melanoma cell clones with acquired resistance to PKIs [39]. In two human melanoma clones, A375-DR and A375-TR, with resistance to the BRAFi, dabrafenib, and the MEKi trametinib, respectively, p-RPS6 could not be downregulated by other MAPK pathway inhibitors. In contrast, an S6K inhibitor LY2584702 downregulated p-RPS6 in these cells. Additionally, the PI3K/mTOR inhibitor BEZ235 and the mTOR inhibitor AZD2014 could also inhibit the RPS6 phosphorylation in the resistant cells. Therefore, the switched regulation of p-RPS6, from the RAS/RAF/MEK/ERK pathway to the PI3K/AKT/mTORC1/S6K pathway, led to the resistance of these cells to BRAFis and MEKis by inducing the expression of G 0 /G 1 cell cycle checkpoint proteins, including RB, cyclin D1, and CDK6 [39]. The mechanism underlying this mechanistic switched regulation of p-RPS6 remains to be evaluated.

RPS6-KD in Cancer Cells
RPS6 or its phosphorylation has been proposed as an alternative target to treat cancer types with high levels of RPS6 [21,482] (Figure 5). The siRNA-or shRNA-based RPS6-KD has demonstrated anticancer effects in various cancer cells (Table 9).

RPS6-KD in Breast Cancer Cells
Studies on the TNBC cells HS578T and MDA-MB-231 have suggested RPS6 as a therapeutic target against TNBC [21]. Synergistic downregulation of both p-RPS6 (S235236) and t-RPS6 has been identified as a unique feature in TNBC cells co-treated with the EGFRi gefitinib and the METi SU11274, whereas no significant or consistent changes have been observed in the levels of p-EGFR (Y1068), t-EGFR, t-MET, p-AKT (S473), t-AKT, p-ERK1/2 (T202/Y204), and t-ERK1/2. Additionally, 4 h treatment with the proteasome inhibitor MG132 was not sufficient to inhibit the reduction of t-RPS6 in the TNBC cells in the presence of gefitinib and SU11274. The siRNA-based RPS6-KD reproduced the reduction of cell proliferation observed in TNBC cells co-treated with the gefitinib alongside SU11274. Unfortunately, the functional consequence of RPS6-KD in TNBC cells has not been further studied. However, it is noteworthy that the downregulation of t-RPS6 was augmented over time, whereas the initial reduction in the levels of p-EGFR and p-AKT was recovered with time in the TNBC cells co-treated with gefitinib and SU11274 [21]. Further research is needed to address the mechanism of regulating RPS6 stability in TNBC cells and its implication for TNBC treatment. RPS6-KD or rapamycin treatment has been reported to upregulate p-4E-BP1 (T37/46) and acetylated histone H3 (Ac-H3) in TNBC cells [477]. Cotreatment of the cells with valproic acid (VA) and tamoxifen suppresses RPS6-KD-mediated induction of p-4E-BP1 and Ac-H3 (K56) levels, and the triple combination of rapamycin, VA, and tamoxifen suppresses p-RPS6 (S235/236), p-4E-BP1 (T37/46), and Ac-H3 (K56) levels concurrently with a decrease in the viability of TNBC cells. Interestingly, this triple combination restored the expression of estrogen receptor alpha (ERα). In addition, both the triple combination and RPS6-KD with VA and tamoxifen reduced the cancer stem cell (CSC) population in TNBC cells. Furthermore, the triple combination reduced tumor growth, CSC subpopulation, and tumorigenesis in vivo [477].

RPS6-KD in Cervical Carcinoma Cells
Downregulation of RPS6 in the cervical carcinoma cell line HeLa resulted in the inhibition of TRAIL-dependent apoptosis in a DR4-dependent manner [340]. RPS6-KD downregulated DR4 and desensitized these cancer cells to TRAIL-dependent apoptosis (see Section 5.2.7. RPS6-KD in Hematopoietic Cancer Cells). Conversely, the overexpression of DR4 sensitized TRAIL-induced apoptosis in RPS6-KD cells. The detailed mechanism underlying the regulation of DR4 expression by RPS6 remains to be described. Similar to Jurkat cells expressing antisense RPS6, HeLa cells stably expressing antisense RPS6 or RPS6-shRNA were not resistant to apoptosis induced by doxorubicin or tunicamycin. In contrast, RPS6 overexpression in HeLa cells resulted in sensitization of the cells to TRAIL-dependent apoptosis, but not to apoptosis induced by doxorubicin, etoposide, staurosporine, or tunicamycin. Overexpression of other RPSs, such as RPS2, RPS3, and RPS20, did not affect TRAIL-induced apoptosis in HeLa cells. As mentioned earlier (see Section 3.2.3. Regulation of Apoptosis), the sensitization of TRAIL-induced apoptosis is mediated by unphosphorylated RPS6 not by p-RPS6 in HeLa cells.

RPS6-KD in Head and Neck Cancer Cells
The level of p-RPS6 (S240/244) and the ratio of p-RPS6/t-RPS6 are correlated with unfavorable prognosis in ESCC patients [436]. The important role of RPS6 in ESCC cells was further demonstrated by a knockdown experiment. RPS6-KD in the ESCC cell lines TE8 and TE10 reduced the numbers of these cells over time compared with the numbers of the control KD cells. Western blot analyses showed downregulation of cyclin D and CDK2 proteins in the RPS6-KD ESCC cells. In contrast, the levels of p21 and p27 were induced by RPS6-KD. Furthermore, RPS6-KD reduced the migration, invasion, and focal adhesion formation of ESCC cells, concurrently with a reduction of p-FAK (Y397), p-paxillin (Y118), p-ERK, and p-JNK (T183/Y185). S6K1-KD, an upstream kinase of RPS6, also resulted in similar effects in ESCC cells [436].

RPS6-KD in Gastric Cancer Cells
As mentioned earlier, RPS6-KD, in HER2i-resistant GC cell lines, resulted in a reduction in viability and induced sensitivity to HER2is through the downregulation of NRF2 and its target genes such as AKR1C1, AKR1C2, and AKR1B10 [354] (see Section 3.2.4. Regulation of Drug Resistance). The mechanisms of NRF2 downregulation by RPS6-KD remain largely unknown. Notably, RPS6 has been reported to bind to the La autoantigen (also known as Sjögren syndrome type B; SSB) upon oxidative stress, leading to increased translation of NRF2 [370]. La/SSB binds to the 5 -UTR of NRF2 mRNA in response to oxidative stress, and RPS6 consequently associates with the ribosomes.

RPS6-KD in Glioblastoma Cells
RPS6-KD suppressed the tumor-sphere formation of GBM cells through the inhibition of p-JAK2 and p-STAT3 and concomitantly downregulated the stemness-related proteins Nestin and SOX2 [438]. The size of the tumor spheres was also reduced by RPS6-KD. Inversely, transient overexpression of RPS6 enhanced the tumor-sphere formation of these cells [438]. Furthermore, the interleukin 6 (IL6)-induced tumor sphere formation was reduced by RPS6-KD. Taken together, RPS6 may regulate the IL6/IL6 receptor (IL6R) pathway. Since the tumor sphere formation and the expression of Nestin and SOX2 are the characteristics of GSCs, these results suggest that RPS6 plays critical roles in the development and maintenance of GSCs [438]. Interestingly, the stemness-related proteins Nestin and SOX2 have been established as transcriptional targets of STAT3 [483,484].

RPS6-KD in HCC Cells
Similar to HeLa cells (see Section 5.2.2. RPS6-KD in Cervical Carcinoma Cells), RPS6-KD in the human HCC cell line SK-HEP-1 suppressed TRAIL-induced apoptosis and downregulated DR4 [340]. The reconstitution of DR4 expression rescued the TRAILresistance in this cell line.

RPS6-KD in Hematopoietic Cancer Cells
A functional genetic screening for the death-rescue modifier has identified TRAILresistant T cell leukemia Jurkat clones that contain antisense RPS6 cDNA in their genome [340]. The reduced RPS6 expression in these cells resulted in the attenuation of TRAIL-induced apoptosis. This resistance was specific against TRAIL. No attenuation was observed in cell death caused by other signals, such as tumor necrosis factor-α (TNFα), FAS ligand (FASL), etoposide, and staurosporine.
As previously mentioned, RPS6-KD reduces the proliferation of the PEL cell line BC-3 (see Section 3.2.10. Roles of RPS6 in Response to Infection) [399]. Since BC-3 cells were infected by KSHV, the cells contain the viral genome and express the viral proteins as well [485]. RPS6 may contribute to the extraordinary stability (several days) of LANA. RPS6-KD remarkably reduced the LANA stability to 0.6 hr, leading to p53 stabilization and a p53-target gene, CDKN1A (the gene encoding p21) expression [399]. However, the molecular mechanism of RPS6 in the LANA stabilization needs further study.
RPS6-KD has decreased the levels of the proliferating cell nuclear antigen (PCNA) in p53 wild-type DLBCL cell lines [338]. However, RPS6-KD did not induce apoptosis in these cells.

RPS6-KD in Lung Cancer Cells
The suppression of proliferation by shRNA-based RPS6-KD was exaggerated by time in two lung cancer cell lines, A549 and H520 [37]. Inversely, the expression of senescence-associated β-galactosidase (SA-β-gal) was increased by RPS6-KD in both cell lines. Additionally, RPS6-KD increased the G 0 /G 1 phase and decreased the G 2 /M phase with reduced protein levels of p-RB and cyclin D1, whereas no significant changes were observed in the levels of cyclin A, cyclin E, and total RB. More interestingly, the levels of CKIs, including p16, p21, p27, and p57, were increased. However, RPS6-KD did not induce apoptosis or altered the expression of BCL-xL, BAX, or caspase-3. The anticancer effect of RPS6-KD was reproduced in a xenograft model in which A549 lung cancer cells with RPS6-KD resulted in reduced tumorigenicity with an increase in the number of SA-β-gal(+) cells in the xenograft tissues. Similar changes were observed in the expression of cell cycle regulators and CKIs [37].
RPS6-KD induced the G 0 /G 1 cell-cycle arrest, decreasing cell viability in the NSCLC cell lines SK-MEK-1 and H1650 [38]. The levels of p-RB, cyclin D1, cyclin E, CDK2, and CDK4 were also reduced, whereas the levels of CKIs, p21, and p27 were increased by RPS6-KD in these cells. The migration of NSCLC cells was also reduced by RPS6-KD, which also downregulated the proteins involved in cell migration, including N-cadherin, vimentin, MMP-9, MMP-2, and p-paxillin. In contrast, E-cadherin was upregulated by RPS6-KD. Inversely, the overexpression of RPS6 in a normal HBE cell line promoted cell proliferation concurrently with the upregulation of p-RPS6, p-RB, cyclin D1, cyclin E, CDK2, and CDK4, and decreases in CKIs and the number of cells in the G 0 /G 1 phase. The migration of HBE cells was also enhanced by RPS6 overexpression. Additionally, RPS6 overexpression upregulated N-cadherin, vimentin, MMP-2, and p-paxillin in these cells [38].
As mentioned earlier (see Section 3.2.1. Regulation of Cell Cycle, Proliferation, and Growth), RPS6-KD by siRNA in A549 cells induced the p53-mediated checkpoint in an RPL11-depenent manner [339]. Although RPS6-siRNA transfection was not enough to completely deplete the RPS6 protein due to the long half-life of ribosomes, it induced the expression of p53 and p21, leading to a partial decrease in DNA synthesis and an increase in the number of cells arrested at the G 1 phase. The concurrent depletion of TP53 by siRNA abrogated the effects of RPS6-siRNA transfection. RPS6 depletion increased the translation of RPL11 mRNA and the amount of ribosome-free RPL11, which can bind to MDM2, leading to the inhibition of MDM2 function to degrade p53. Consequently, the mRNAs and protein levels of p53 targets, such as CDKN1A (the gene encoding p21) and MDM2, were increased.

RPS6-KD in Melanoma Cells
RPS6-KD alone reduced the proliferation of two human melanoma cells A375-DR and A375-TR, which are resistant to the BRAF V600E -selective inhibitor dabrafenib and the MEK1/2 inhibitor trametinib [39]. In addition, RPS6-KD induced the G 0 /G 1 cellcycle arrest in these cells. More interestingly, RPS6-KD was sufficient to reduce the levels of G 0 /G 1 checkpoint regulators such as RB, cyclin D1, and CDK6. Additionally, the enhanced downregulation of p-RPS6 by drug combinations exerted a synergistic antitumor activity [39].
In a separate study, RPS6-KD significantly reduced the colony formation and induced the apoptosis of cutaneous malignant melanoma (CMM) cells in vitro [478]. Additionally, the same authors found that S6K1 was upregulated in clinical samples with advanced stages of CMM. Furthermore, the mTORC1 signaling pathway was identified to be downregulated by the combination of the EGFRi afatinib and the ALK/ROS1 dual inhibitor crizotinib in malignant melanoma. More interestingly, the level of t-RPS6 itself, in addition to p-RPS6 level, was also downregulated by the afatinib+crizotinib treatment in melanoma cell lines in vitro [478].

RPS6-KD in Ovarian Cancer Cells
It has been reported that the transfection of ovarian cancer cells with RPS6-siRNA reduces the activities of glutaminase (GLS) and glutamate dehydrogenase (GDH), leading to a reduction in glutamine-induced proliferation of the cells [479]. Since glutamine is the main nutrient used by cancer cells, targeting the glutamine metabolism has been recognized as a promising therapeutic alternative [486]. Glutamine depletion inhibits cell proliferation concurrently with the induction of G 1 cell-cycle arrest and apoptosis in ovarian cancer cells [479]. Glutamate increases cell proliferation in the presence of glucose. Interestingly, rapamycin reduced the glutamine-induced p-RPS6 (S235/236) and GLS expression in a dose-dependent manner in ovarian cancer cells. In the ovarian cancer cell line HEY, transfection of RPS6-siRNA was sufficient to deplete p-RSP6 (S235/236) although not t-RPS6 [479]. Under these conditions, RPS6-siRNA transfection reduced the glutamineinduced GDH activity and cell proliferation in HEY cells, whereas no significant effect of RPS6-siRNA transfection was observed in the absence of glutamine. Notably, rapamycin treatment only partially mimicked the effects of RPS6-siRNA transfection. The significance of this discrepancy remains to be elucidated.
RPS6-KD also inhibits the proliferation and long-term colony formation of the ovarian cancer cell lines SKOV-3 and HO-8910 [40]. The migration and invasion of these ovarian cancer cells are also reduced by RPS6-KD. Additionally, RPS6-KD induces the G 0 /G 1 cell-cycle arrest. Consistently, the G 0 /G 1 checkpoint regulators, such as cyclin D1, cyclin E, CDK2, CDK4, CDK6, and p-RB, are diminished by RPS6-KD in ovarian cancer cells [40].

RPS6-KD in Sarcoma Cells
An siRNA-based screening identified RPS6 as an effector of IGF1R inhibition in sarcoma cells [480]. RPS6-KD by siRNA reduced the survival of Ewing sarcoma cell lines over time. Interestingly, MTOR-KD recapitulated RPS6-KD, whereas TSC1-KD had little or no effect. In addition, RPS6-KD also induced cell death in rhabdomyosarcoma (RMS) cell lines [480]. The significance of p-RPS6 in these sarcoma cell lines is underlined by their sensitivity to the IGF1R inhibitor BMS-536924. RMS cell lines were resistant to BMS-536924 compared to Ewing sarcoma cell lines. BMS-536924 failed to downregulate p-RPS6 (S235/236) and p-4EBP1 in RMS cell line, but dose-dependently downregulated p-RPS6 (S235/236) and p-4EBP1 in sensitive Ewing sarcoma cell lines. Additional siRNA screening has found that the knockdown of macrophage stimulating 1 receptor (MST1R) sensitizes RMS cell lines to BMS-536924 with concordant reduction of p-RPS6 (S235/236). These data suggest that MST1R may be an alternative activator of RPS6 in IGF1R inhibitor-resistance sarcoma cells.

Conclusions
In addition to its role in protein synthesis, RPS6 may activate various pathways to induce the oxidative stress response, drug resistance, cancer stem cell stemness, and cell-cycle arrest ( Figure 6). Evidence from knockdown experiments supports the therapeutic potential of RPS6 targeting in various cancer cells including breast cancer, cervical carcinoma, head and neck cancer, gastric cancer, glioblastoma, HCC, hematopoietic cancer, lung cancer, melanoma, ovarian cancer, and sarcoma ( Figure 5). However, the signaling pathways downstream of RPS6 are currently largely unknown. In addition, the roles of activated RPS6 in tumorigenesis still remain elusive. The positive and negative feedback and/or feedforward mechanisms and novel regulatory pathways controlled by RPS6 may provide alternative therapeutic interventions to overcome the current limitations in cancer treatment.