Long noncoding RNAs in the p53 network

Abstract

The tumor-suppressor protein p53 is activated in response to numerous cellular stresses including DNA damage. p53 functions primarily as a sequence-specific transcription factor that controls the expression of hundreds of protein-coding genes and noncoding RNAs, including microRNAs (miRNAs) and long noncod-ing RNAs (lncRNAs). While the role of protein-coding genes and miRNAs in mediating the effects of p53 has been extensively studied, the physiological function and molecular mechanisms by which p53-regulated lncRNAs act is beginning to be understood. In this review, we discuss recent studies on lncRNAs that are directly or indirectly regulated by p53 and how they contribute to the biological outcomes of p53 activation. Published 2016. This article is a U.S. Government work and is in the public domain in the USA.

INTRODUCTION

The advent of RNA-sequencing (RNA-seq) has revealed the transcriptional complexity of the human genome. Thousands of long noncoding RNAs (lncRNAs), a heterogeneous class of noncoding RNAs >200 nucleotides (nt) long with low or no coding potential, have been identified by RNA-seq. According to the most recent version of Gencode (v24) 15,941 lncRNAs are transcribed from the human genome. Although the function of the vast majority of lncRNAs has not been investigated, increasing evidence suggests that some lncRNAs regulate critical cellular processes^1,2 including cellular proliferation, differentiation and development, dos-age compensation, chromosomal imprinting, and genomic stability.^3–12

LncRNAs can mediate their effects by diverse mechanisms; they can act as molecular scaffolds to recruit chromatin modifiers at specific genomic loci and thereby activate or repress target gene expression.^13,14 They can regulate gene expression in cis^15–17 or in trans¹⁸ such as splicing,⁶ mRNA translation,¹⁹ and mRNA degradation.²⁰ In addition, lncRNAs can act as decoys for RNA-binding proteins or as sponge for microRNAs (miRNAs)^21,22 or by directly interacting with RNA and DNA by base pairing.^23,24 LncRNAs can be localized in the nucleus and/or cytoplasm^25,26 and based on their genomic localization and proximity to protein-coding genes they have been classified as intergenic, intronic, bidirectional, sense, or antisense.²⁷

A major challenge in the lncRNA field is figuring out the molecular mechanisms by which they act. Unlike miRNAs that act via the RNA interference (RNAi) pathway,²⁸ each lncRNA has the potential to mediate its effect via a unique molecular mechanism. Identifying the molecular interactors of an lncRNA may also provide insights on the function of the lncRNA. Another approach to identify the function of an lncRNA is by identifying the transcription factors that drive its expression. For example, if an lncRNA is transcriptionally regulated by p53, it can mediate the effects of p53 by regulating the expression of select genes in the p53 network.

The master regulatory transcription factor p53 is the most frequently mutated protein in human cancer.²⁹ p53 is the most extensively studied transcription factor that is activated during the stress response including replicative stress, oxidative stress, hypoxia, DNA damage, nutrient deprivation, and telomere shortening.^30–32 The outcomes of p53 activation are diverse and include induction of cell cycle arrest, apoptosis, autophagy and senescence, and inhibition of invasion, migration, stemness, and metabolic reprogramming, all together leading to tumor suppression.^33,34 Transcriptional regulation of protein-coding genes by p53 is very well documented in literature.^31,33 In addition, others and we have shown that several miRNAs are direct targets of p53 and function as downstream effectors of p53.^35–40 More recently, specific lncRNAs have been identified as p53 targets and some of these have been shown to function in p53 signaling. In this review, we provide an update on the role of lncRNAs in p53 signaling. In particular, we summarize the known roles of lncRNAs in the p53 network.

p53-REGULATED LncRNAs LncRNAs at the p21 Locus

p21 (also known as CDKN1A, cyclin-dependent kinase inhibitor 1) mediates p53-dependent G1 growth arrest.^41,42 The lncRNAs, lincRNA-p21 and PANDA, have been identified to be located and transcribed from the p21 locus. LincRNA-p21 regulates gene expression both at the transcriptional level as well as post-transcriptionally by associating with the RNA-binding protein HuR. PANDA on the other hand regulates the expression of proapoptotic genes by interacting with the transcription factor NF-YA and regulates senescence by interacting with scaffold-attachment-factor A (SAFA).

LincRNA-p21

The LincRNA-p21 was identified by microarrays performed from p53^LSL/LSL MEFs (mouse embryonic fibroblasts) and lung tumor cells derived from mice expressing KRAS (p53^LSL/LSL). In this study, the endogenous p53 locus was inactive by insertion of a STOP cassette flanked by loxP sites (LSL) in the first intron and the p53 expression could be restored by Cre recombination. p53^LSL/LSL MEFs (Cre+/−) were treated with the DNA-damaging agent Doxorubicin and KRAS (p53^LSL/LSL) tumor cells were treated with hydroxytamoxifen for p53 restoration. LincRNA-p21 was identified as one of the 11 lncRNAs induced by p53 in both the systems.⁴³

LincRNA-p21 is located on chromosome 17, ~15 kb upstream of the Cdkn1a (p21) gene and hence named as lincRNA-p21. This lncRNA was the first identified p53-regulated intergenic lncRNA and was reported to have a proapoptotic function in the p53 network⁴³ (Figure 1(a)). Gene expression analysis following independent knockdown of lincRNA-p21 and p53 in MEFs identified a large subset of com-monly repressed genes, including regulators of apoptosis and cell cycle, suggesting that lincRNA-p21 is a downstream effector of p53-dependent transcriptional response.⁴³ Knocking down lincRNA-21 in MEFs and treatment with Doxorubicin increased cell survival without significantly altering the cell cycle.⁴³ In this study, lincRNA-p21 was shown to mediate its function in trans by associating with the RNA-binding protein hnRNPK and this interaction is needed for proper localization of hnRNP-K at repressed genes loci and regulation of p53-mediated apoptosis⁴³ (Figure 1(a)).

FIGURE 1.

Diverse functions of lincRNA-p21 in nucleus and cytoplasm. (a) In response to DNA damage, p53 is upregulated and activates the transcription of coding as well as noncoding transcripts including lincRNA-p21. LincRNA-p21 functions in cis- as well as in trans. In cis it regulates the expression of its neighboring gene p21. It has proapoptotic function and can regulate gene expression in trans by interacting with RNA-binding protein hnRNP-K. (b) LincRNA-p21 expression is induced during hypoxia. Under normal oxygen levels HIF-1α binds to VHL and is polyubiquitinated by VHL, leading to its degradation via proteasome. During hypoxic conditions, HIF-1α levels increase, which in turn increases the levels of lincRNA-p21. LincRNA-p21 can now bind to VHL and HIF-1α, therefore disrupting VHL-HIF-1α interaction. Once this interaction in disrupted the levels of HIF-1α accumulate, suggesting the existence of feedback loop between HIF-1α and lincRNA-p21. (c) LincRNA-p21 regulates translation in the cytoplasm. When cytoplasmic HuR levels are high, it associates with lincRNA-p21 and recruits let-7/Ago2 to lincRNA-p21 leading to its degradation and thus activates translation of select mRNAs. However, when the levels of HuR are low, lincRNA-p21 levels accumulate and it can now inhibit translation.

However, in another study, via conditional knockout of lincRNA-p21 in MEFs, the authors found that lincRNA-p21 functions predominantly in cis and regulates the expression of its neighboring gene p21¹⁵ (Figure 1(a)) unlike the previous study which reported that lincRNA-p21 regulates gene expression in trans by binding to hnRNP-K.⁴³ It is unclear if this discrepancy in the two studies could be because of different approaches used to deplete lincRNA-p21. Huarte et al. used RNAi to knock-down lincRNA-p21 whereas Dimitrova et al. used lincRNA-p21 conditional knockout MEFs.

The confusion on the mechanism of action of lincRNA-p21 was resolved in a recent study on the in vivo characterization of this lncRNA.⁴⁴ Using a knockout mouse model and a massively parallel enhancer assay, the authors delineated the functional elements at the lincRNA-p21 (designated as Linc-p21 in this study) locus. They showed that even in tissues with no detectable linc-p21 expression, deletion of this locus significantly affects local gene expression, including that of p21. The authors systematically interrogated the underlying DNA sequence for enhancer activity and identified multiple enhancer elements. In sum, the findings from this elegant in vivo study suggests that the cis-regulatory effects mediated by linc-p21 may also be due to DNA enhancer elements.

LincRNA-p21 also regulates genes at the post-transcriptional level.¹⁹ For instance, the RNA-binding protein HuR associates with lincRNA-p21 and recruits let-7/Ago2 to lincRNA-p21 leading to its degradation (Figure 1(b)). Knocking down HuR leads to an increase in the lincRNA-p21 levels both in nucleus as well as in cytoplasm. In the absence of HuR, lincRNA-p21 can associate with select mRNAs including JUNB and CTNNB1 and inhibit their translation. However, an increase in cytoplasmic HuR levels decreases lincRNA-p21 expression, resulting in upregulation of JUNB and CTNNB1 mRNA translation.¹⁹ Thus, HuR regulates translation of select mRNAs by influencing lincRNA-p21 levels (Figure 1(b)).

LincRNA-p21 is also reported as a key regulator of cell proliferation and apoptosis during atherosclerosis.⁴⁵ The expression of lincRNA-p21 was significantly lower in atherosclerotic plaques of ApoE (−/−) mice, an animal model for atherosclerosis, as compared to wild-type mice. LincRNA-p21 represses cell proliferation and induces apoptosis in human smooth muscle cells and mouse mononuclear macro-phage cells. In addition, the authors of this study showed that this lncRNA interacts with MDM2 in both human and mouse and knocking down lincRNA-p21 decreased p300/p53 interaction and increased MDM2/p53 interaction, thus affecting p53-transcriptional activity.⁴⁵ Furthermore, lincRNA-p21 is a mediator of both UVB-induced and p53-mediated apoptosis in mouse and human keratinocytes and also ING1b-induced apoptosis.^46,47 LincRNA-p21 is also induced during hypoxia and there is positive feedback loop between HIF-1α and lincRNA-p21 that promotes glycolysis under hypoxia⁴⁸ (Figure 1(c)). Finally, lincRNA-p21 also plays a role in regulating somatic reprogramming by forming a complex with hnRNPK and either H3K9 methyltransferase SETDB1 or DNA methyltransferase DNMT1, at the pluripotency gene promoters.⁴⁹

PANDA

The p53-regulated lncRNA PANDA refers to ‘P21 Associated NcRNA DNA damage Activated.’ PANDA was identified in a study where the authors created an ultra-high density tiling arrays comprising the promoters of 56 cell cycle genes including CDKN2A (p16), p14ARF, and CDKN2B (p15) to examine 54 pairs of polyade-nylated RNA samples. These RNA samples were obtained after diverse perturbations including cell cycle synchronization, DNA damage, differentiation stimuli, oncogenic stimuli, or carcinogenesis.²² They identified five lncRNAs including PANDA at the p21 promoter that are induced by DNA damage.

PANDA is a monoexonic lncRNA located ~5 kb upstream of the p21 transcriptional start site and ~2.5 kb upstream of the major p53 response element of p21. PANDA plays a role in maintaining cell cycle arrest and survival following DNA damage.²² Knocking down PANDA in human fibroblasts sensitized the cells to DNA damaged-induced apoptosis and increased the expression of several genes encoding canonical activators of apoptosis.²² PANDA interacts with the nuclear transcription factor NF-YA and thereby restricts binding of NF-YA to the chromatin, limiting the activation of apoptotic genes²² (Figure 2). PANDA also interacts with SAFA and this interaction has a crucial role in senescence.⁵⁰ The authors found that PANDA is present at ~130 copies per nucleus in proliferating fibro-blasts and ~1200 copies per nucleus in senescent fibroblasts. In proliferating cells, PANDA interacts with SAFA and this interaction represses senescence-promoting genes by recruiting polycomb repressive complex (PRC) proteins BMI1–PRC1 and EZH2– PRC2 to promoters of these genes e.g., p21 and PANDA itself.⁵⁰ Also, in proliferating cells, the binding affinity of NF-YA to PANDA is low and therefore it can activate E2Fs (Figure 2). However, in senescent cells, the SAFA-PANDA-PRC complex is disrupted and the expression of prosenescence genes including p21 is upregulated. PANDA can now interact with NF-YA and block proliferation by sequestering NF-YA away from occupying E2F/NF-YA target gene promoters.⁵⁰

FIGURE 2.

Regulation of proliferation and survival by the p53-regulated lncRNA PANDA. PANDA is a direct target of p53 and its expression is induced following p53 activation after DNA damage. In response DNA damage PANDA has a prosurvival function as it promotes cell survival and inhibits apoptosis. PANDA interacts with the NF-YA and thus restricts the binding of NF-YA to chromatin, limiting the activation of apoptotic genes. In proliferating cells, PANDA interacts with SAFA and recruit PRC2-EZH2 and BMI1–PRC1 complexes that inhibit transcription of PANDA and prosenescence genes including p21. When the levels of PANDA decrease, its affinity to bind with NF-YA is low and NF-YA can now activate E2F and promote cell proliferation. PANDA acts in an autoregulatory loop with SAFA–BMI1–PRC1. In senescent cells, SAFA–PRC2-EZH2 and SAFA–BMI1–PRC1 complexes disintegrate and expression of PANDA and p21 is upregulated. PANDA now blocks proliferation by decoying NF-YA, which is also required for survival of senescent cells.

In summary, p21, which is a well-studied target of p53 and a major cell cycle regulator, houses two lncRNAs upstream of its promoter and these two lncRNAs regulate contrasting phenotypes. While lincRNA-p21 is proapoptotic and functions in cis, PANDA is antiapoptotic and functions in trans, in the context of DNA damage.

Pint

The lncRNA Pint (p53-induced noncoding transcript), also known as lincRNA-Mkln1, was identified in the same study that discovered lincRNA-p21.⁴³ Pint is predominantly nuclear and a direct p53 target induced after p53 activation. Knocking down Pint followed by Doxorubicin treatment decreases the number of cells in S phase, and simultaneously increases the population of cells in G1 and apoptotic cells.⁵¹ Overexpressing Pint increased cell proliferation and decreased cellular apoptosis, suggesting that Pint is a positive regulator of proliferation during DNA damage.⁵¹ Gene expression analysis in MEFs transfected with antisense oligos to deplete Pint or control antisense oligos, and subsequent treatment with Doxorubicin revealed that Pint regulates the expression of genes involved in cellular growth and proliferation and cell survival, and several genes in the mitogen-activated protein kinase (MAPK), transforming growth factor (TGF)-β, and p53 signaling pathways⁵¹ (Figure 3). By performing RNA immunoprecipitation (RIP) experiments from Doxorubicin-treated MEFs followed by IP with Suz12 subunit of PRC2, the authors found enrichment of Pint RNA in PRC2 immunoprecipitates. In addition, the authors confirmed the interaction between Pint and PRC2 by performing RNA pull-down experiments using in vitro synthesized biotinlabeled Pint RNA. They showed that the interaction between Pint and PRC2 is required for the targeting of PRC2 to repress the expression of subset of genes, thus regulating cell proliferation.⁵¹

FIGURE 3.

LncRNAs in the p53 network. DNA damage activates p53 and its downstream transcriptional targets including lncRNAs. The human and mouse orthologs of lncRNA PINT have contrasting functions. Human PINT has a tumor-suppressor function, whereas mouse ortholog increases cell survival and proliferation. PR-lncRNA1 and PR-LncRNA10 have proapoptotic function and they inhibit cell survival and proliferation. The LncRNA TUG1 has diverse role in different cancer types. TUG1 can promote or inhibit cell proliferation and cell cycle progression depending on the tumor type. The LncRNA NORAD is not a direct transcriptional target of p53 and it interacts with PUMILIO proteins to maintain genomic stability.

Furthermore, the authors explored the possibility of existence of Pint ortholog in humans. They found an lncRNA at a syntenic location in the human genome that shared high 5ʹ end sequence homology to the mouse Pint RNA. The human ortholog of Pint is also transcriptionally regulated by p53 and significantly downregulated in colorectal tumors compared with normal tissue suggesting a potential tumor-suppressor lncRNA (Figure 3).⁵¹

Interestingly, human Pint has the opposite phenotype of the mouse ortholog. Contrary to what was observed in MEFs, HCT116 (colorectal cancer) cells overexpressing Pint showed a significant decrease in proliferation, both in the presence or absence of Doxorubicin-induced DNA damage, as well as in the number of cells in S phase and increase in the percentage of cells in both G1 and G2/M phases.⁵¹ In conclusion, lncRNA Pint represents an interesting example of p53-regulated lncRNA with high sequence homology in both human and mouse species. Although both the orthologs are predominantly nuclear, interact with PRC2, and regulate similar pathways, knocking down Pint expression in MEFs has a phenotype similar to HCT116 cells overexpressing human Pint; this unexpected result may repre-sent species-specific molecular interactors of this lncRNA.

PR-lncRNA-1 and PR-lncRNA-10

Sanchez et al., 2014 performed RNA-seq and p53-ChIP-seq from HCT116 cells treated with the che-motherapeutic drug 5-FU (5-Flurouracil) and identified 18 lncRNAs directly regulated by p53, of which 3 showed conserved regulation in mouse cells. Among these 18 lncRNAs, RP11–467J12.4 (PR-lncRNA-1) was identified as a nuclear, multiexonic, intergenic lncRNA regulated in a p53-dependent manner in both human and mouse cells. Another lncRNA, PR-lncRNA-10 was identified as a nonan-notated intergenic nuclear lncRNA with no mouse orthologs.⁵²

Both PR-lncRNA-1 and PR-lncRNA-10 are negative regulators of cell survival and proliferation as knocking down these lncRNAs increased prolifer-ation and decreased apoptosis in the presence or absence of Doxorubicin or 5-FU.⁵² Gene expression profiling after depletion of PR-lncRNA-1 or PR-lncRNA-10, following DNA damage suggested that PR-lncRNA-1 regulates the expression of genes involved in apoptosis and proliferation such as BCL2L and BIRC3, the DNA polymerase subunit POLA1 and the cytokine TGFB2. PR-lncRNA-10 was found to regulate the expression of p21, the transcription factor JUNB and the apoptosis regulators BIRC6, TP53I3, and FAS.⁵² Furthermore, p53-ChIP experiments in the presence and absence of PR-lncRNA-1 or PR-lncRNA-10 followed by 5-FU treatment revealed that these lncRNAs are also required for the efficient binding of p53 to some of its target genes. However, the molecular mechanism by which these two lncRNAs regulate binding of p53 to its targets remains unclear.

NORAD

NORAD (noncoding RNA activated by DNA damage) has recently been identified as a lncRNA with high degree of evolutionary conservation in mammals. The mouse ortholog of NORAD was identified among several lncRNAs induced after Doxorubicin treatment in p53^+/+ and p53^−/− MEF.⁵³ The function and mechanism of this lncRNA has recently been identified.³

Interestingly, although NORAD is induced after Doxorubicin treatment in a p53-dependent manner, it is not a direct p53 target. NORAD is a very abundant, cytoplasmic lncRNA expressed at ~300–1400 copies per cell. Genetic deletion of NORAD in HCT116 cells suggested that this lncRNA regulates chromosomal stability.³ NORAD^−/− cells can have tetraploid DNA content, variable chromosome numbers with 4N DNA content and both diploid and tetraploid NORAD^−/− clones undergo chromosomal rearrangements.³ To search for functional domains in NORAD, the authors used an elegant approach by aligning the sequence of NORAD to itself using the BLAST algorithm. This approach uncovered a ~400 nt domain that is repeated five times in the NORAD transcript. Using in vitro RNA pulldowns followed by mass spectrometry, the authors found that NORAD binds to PUMILIO protein PUM2 and this interaction occurs in each of the five domains. Moreover, the authors confirmed the NORAD-PUM2 interaction by PARCLIP on endogenous PUM2 in HCT116 cells and this analysis identified 15 binding sites for PUM1/PUM2 protein per molecule of NORAD. To determine the functional significance of this interaction, the authors performed RNA-seq in NORAD^+/+ and NORAD^−/− HCT116 cells and found that PUM2 CLIP targets were significantly downregulated in NORAD^−/− cells, representing genes that are reguators of the cell cycle, mitosis, DNA repair, and DNA replication. These results were further supported by the RNA-seq with PUM1/2 overexpressing HCT116 cells that depicted expression profiles with similar patterns as NORAD^−/− cells and PUM1/2 overexpression also increased level of aneuploidy.³ More importantly, depleting PUM1/2 in NORAD^−/− cells using CRISPR/Cas9 or TALENs, partially suppressed chromosomal instability and surprisingly, the double knockout led to considerable aneuploidy, highlighting that tight control of PUM1/2 levels is necessary to maintain genomic stability. Altogether, this study demonstrates that PUM1/2 proteins are downstream of NORAD and play an important role in the maintenance of genomic stability (Figure 3). In sum, NORAD is a very abundant lncRNA indirectly regulated by p53. NORAD has multiple binding sites for cellular protein PUMILIO. NORAD sequesters a significant fraction of PUMILIO proteins and making them unavailable to repress the target genes. These targets represent genes that are critical for mitosis, DNA repair, and DNA replication. In the absence of NORAD, there is increased genomic instability leading to aneuploidy.

TUG1

The lncRNA TUG1 (Taurine Upregulated Gene 1) was identified in a screen for genes upregulated by taurine (a cysteine derivative important for neural development), and siRNA-based depletion of TUG1 in the developing mouse eye blocked retinal development.⁵⁴ Later, TUG1 was identified among the 39 lncRNAs induced after DNA damage in both human and mouse fibroblasts in a p53 dependent manner and its promoter contains several conserved binding sites for p53.^13,53 By performing a RIP-ChIP assay in the HeLa cells as well as human fibroblast cells, this lncRNA was found to be associated with PRC2.¹³ SiRNA-mediated knock down of TUG1 in human fibroblasts leads to upregulation of genes involved in cell cycle control. Knockdown of TUG1 significantly increased cell proliferation, in vitro clo-nogenicity, and cell cycle progression in the nonsmall cell lung cancer cell lines SPC-A1 and NCI-H1650.⁵⁵ TUG1 regulates tumorigenicity in vivo, as nude mice injected with scramble/shTUG1-transfected SPC-A1 cells develop bigger tumors and displayed higher Ki-67 staining.⁵⁵ Knockdown of TUG1 upregulates homeobox B7 (HOXB7) expression and ChIP assays showed that knockdown of TUG1 resulted in the loss of H3K27 trimethylation and PRC2 binding to the genomic loci of HOXB7, confirming that HOXB7 was a bona target of TUG1/PRC2-regulated genes.⁵⁵ The exact mechanism how TUG1 limits cell proliferation is not known, however, one possible mechanism is via epigenetically regulating HOXB7 expression.

Interestingly, some other studies in different cancer types reported a contrasting phenotype. They showed that TUG1 promotes cell proliferation of esophageal squamous cell carcinoma,⁵⁶ urothelial carcinoma of bladder,⁵⁷ osteosarcoma,⁵⁸ and hepatocellular carcinoma.⁵⁹ In hepatocellular carcinoma cells, TUG1 silencing resulted in decreased proliferation, colony formation, tumorigenicity, and induced apoptosis. TUG1 epigenetically repressed Krüppel-like factor 2 (KLF2) transcription by binding to PRC2 and recruiting it to KLF2 promoter region.⁵⁹ In conclusion, as mentioned previously that lncRNAs can function in a cell type specific manner, TUG1 is an example of p53-regulated lncRNA that has opposite effects in different cancer types.

PVT1

Plasmacytoma Variant Translocation 1 (PVT1) locus encodes several noncoding RNAs including PVT1 and a cluster of six miRNAs, miR-1204, miR-1205, miR-1206, miR-1207–5p, miR-1207–3p, and miR-1208.⁶⁰ The PVT1 locus was first discovered in mouse, frequently associated with translocations in plasmacytomas.⁶¹ In humans, the PVT1 locus maps to chromosome 8 (8q24), which is a site of genetic alterations, including translocations, amplifications, and viral integration in a wide variety of malignancies and lies ~60 kb downstream of the MYC oncogene.⁶² p53 induces transcription on the PVT1 locus and there is a conserved p53-binding site ~1200 bp downstream of the PVT1 transcriptional start site and 172 bp upstream of the miR-1204 stem-loop sequence.⁶³ PVT1 functions as a competitive endog-enous RNA in miRNA-mediated sponge interactions in normal breast tissues, but not in the cancerous cells.⁶⁴ It has a strong binding preference to the miR-200 family.⁶⁴ PVT1 also interacts with MYC protein and prevents its degradation by reducing phosphorylation of its T-58 residue.⁶⁵

LINP1

Similar to NORAD, the lncRNA LINP1 (lncRNA in nonhomologous end joining pathway 1) is not directly regulated by p53. LINP1 is overexpressed in human triple-negative breast cancer and associates with two proteins of the NHEJ DNA repair pathway, Ku80 and DNA-PKcs.⁶⁶ Knocking down LINP1 delays ionizing radiation (IR)-induced DNA damage and increases γ-H2AX levels. Wild-type p53 negatively regulates LINP1 expression as its expression is higher in the breast cancer tumors with mutant p53. Activation of p53 by Nutlin-3a treatment in MCF10A (breast cancer) and HCT116 cells that express wild-type p53 significantly decreased the expression of LINP1. However, Nutlin-3a treatment of MDA-MB-231 cells that express mutant p53 or HCT116 cells with homozygous TP53 deletion, does not alter LINP1 levels.⁶⁶ No p53-binding site was found on the LINP1 promoter suggesting that it is not a direct transcriptional target of p53. Interestingly, exon 2 of LINP1 has binding sites for miR-29, which is positively and directly regulated by p53 and therefore, miR-29 may be the mediator of p53-regulated LINP1 expression.⁶⁶

DDSR1

LncRNA-DDSR1 (DNA damage sensitive RNA1) is a ~1.6-kb long transcript that is induced in response to different DNA-damaging agents including camptothecin, etoposide, and Neocarzinostatin (NCS).⁶⁷ DDSR1 has NF-kB-binding site ~1.1 kb upstream of the transcription start site and its induction after DNA damage is regulated by the ATM-NF-kB signaling pathway. Although p53 can promote DDSR1 induction upon DNA damage, it is not necessary for DDSR1 induction. Additionally, depletion of DDSR1 followed by camptothecin treatment leads to upregulation of several p53 target genes including p21, PUMA, and GADD45 suggesting that DDSR1 negatively regulates p53-mediated gene expression. Knocking down DDSR1 reduced cell proliferation in several cell lines including U2OS, PC3, and A549. Furthermore, depletion of DDSR1 followed by camptothecin treatment significantly reduces the levels of γ-H2AX, phosphor-RPA (Ser4/8), phospho-CHK1 (Ser345), and phosphor-p53 (Ser15). However, there was no increase in phosphor-ATM suggesting that DDSR1 acts downstream of ATM and regulates DNA damage response signaling. By using MS2-TRAP-tagged RNA affinity purification followed by mass spectrometry, the authors identified several DNA-damage related proteins to be associated with DDSR1 including heterogeneous nuclear ribonucleoprotein U-like 1 (hnRNPUL1) and BRCA1 which are involved in homologous recombination (HR). The authors further showed that DDSR1 regulates the recruitment of HR repair factors BRCA1 and RAP80 to the double-strand breaks and thus modulating DNA repair by HR. In summary, DDSR1 is a p53-regulated lncRNA that is not directly regulated by p53 and plays an important role in maintaining genome stability.

LncRNAs THAT REGULATE p53

MALAT1

MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1) also known as nuclear-enriched abundant transcript 2 (NEAT2) is an abundant nuclear lncRNA overexpressed in many cancers, and its high expression is associated with hyperproliferation and metastasis.⁶⁸ MALAT1 regulates gene expression and pre-mRNA splicing by interacting with pre-mRNA splicing factors including serine arginine dipeptide-containing SR family splicing factors and recruitment of SR splicing factors from nuclear speckles to sites of transcription.^6,69 Overex-pression of MALAT1 promotes cell proliferation and migration in vitro, and promotes tumor growth and metastasis in vivo⁷⁰ whereas silencing MALAT1 expression results in reduced migration, metastasis, and tumorigenicity.⁷¹

In human fibroblasts, MALAT1 levels are differentially regulated during cell cycle with low levels during G1 and G2 and high levels during G1/S and mitosis.⁵ Depletion of MALAT1 using antisense oli-gos severely affects G1 to S transition and mitotic progression, and results in increased accumulation γH2AX.⁵ Gene expression profiling from control or MALAT1-depleted cells results in downregulation of several genes involved in G1/S transition and S-phase progression. Furthermore, MALAT1-depleted cells showed enhanced β-galactosidase (β-gal) staining indicative of cellular senescence.⁵

Depletion of MALAT1 also results in the activation of p53 and its targets such as p21.⁵ These results indicated that p53 could be a downstream mediator of MALAT1. To test this, the authors depleted MALAT1 in the tumorigenic cell lines including HeLa, U2OS, and WI-38-VA13, having weak p53, p16, and Rb activity. These cells did not show G1 or G1/S arrest and showed a normal S-phase progression. However, these cells displayed nuclear breakdown phenotype due to defective chro-mosomal segregation and mitotic abnormalities. Their data suggest that in normal diploid human lung fibroblasts, the cell cycle arrest observed upon MALAT1 depletion is primarily due to p53 activation, because it can be rescued by knocking down p53 expression in these cells. However, the molecular mechanism by which MALAT1 influences p53 signaling is not clear yet.

MEG3

A lncRNA that functions in both p53-dependent and independent manner is human maternally expressed gene 3 (MEG3).⁷² Overexpression of MEG3 in HCT116 cells induces p53 protein levels, stimulates transcription from p53-dependent promoter and selectively regulates p53 target gene expression.^73,74 MEG3 overexpression stimulates the expression of the p53 target GDF15 by enhancing p53 binding to the p53 response element in its promoter region; however, MEG3 does not influence p21 levels in HCT116 cells. The mechanism of p53 regulation by MEG3 is not clear, but MEG3 overexpression also downregulates MDM2 expression and this could be one of the mechanisms whereby MEG3 activates p53.⁷³ Whether MEG3 directly interacts with MDM2 and p53 and regulate their levels or there are specific RNA-binding proteins that mediate these interactions need to be tested.

MEG3 also inhibits cell proliferation in a p53-independent manner.⁷³ In breast cancer cells, ectopic expression of MEG3 induces cell cycle arrest, decrease cell proliferation and colony formation potential of breast cancer cells, and promotes cell migration and invasion.⁷⁵ Furthermore, ectopic expression of MEG3 in MCF7 and MB231 cells decreases the RNA as well as protein levels of MDM2 and increases those of p53, p21, Maspin, and KAI1, and also increases p53 transcriptional activity on the promoters of p21, Maspin, and KAI1.⁷⁵ The discrepancy in the p21 upregulation after MEG3 overexpression in the above two reports could be due to cell type-specific function of MEG3; however, this needs further investigation.

H19

H19 is a maternally expressed lncRNA ~2.7 kb long that is reciprocally imprinted and regulated with its neighboring gene IGF2.⁷⁶ H19 has oncogenic properties and during hypoxic stress, and it regulates the expression of genes involved in angiogenesis, survival, and tumorigenesis.⁷⁷ H19 and p53 mutually counter regulate each other. p53 negatively regulates the expression of H19 in vivo by inducing DNA demethylation of the imprinting control region upstream to the H19 gene,⁷⁸ and also suppress its promoter activity.⁷⁹ Presence of wild-type p53 inhibits the elevation of H19 RNA under hypoxic stress in the nucleus⁸⁰ (Figure 4).

FIGURE 4.

LncRNA-mediated regulation of p53. MEG3 induces p53 protein levels resulting in cell cycle arrest and inhibition of cell proliferation and thus inhibit tumorigenesis. MEG3 downregulates MDM2 expression that can possibly activate p53. LncRNAs H19, RoR, and 7SL act in an autoregulatory loop with p53. MT1JP regulates p53 protein by associating with RNA-binding protein TIAR.

RNA-IP using an antibody against p53 from nuclear extracts of gastric cancer cell line AGS showed significant enrichment of H19, suggesting an association between H19 and p53.⁸¹ This association causes partial inactivation of p53 and suppressed the protein level of the p53 target Bax.⁸¹ A highly conserved miRNA, miR-675, resides within exon-1 of the H19 gene.⁸² In bladder cancer cells, H19 derived miR-675 has a major role in inhibiting p53 activation and also decreases the ratio of Bax/Bcl-2 and cyclin D1 expression.⁸³ These findings indicate the counter-regulatory relationship between p53 and H19.

LincRNA-RoR

LincRNA-ST8SIA3 or LincRNA-RoR (Regulator of Reprogramming) is another example of lncRNA that influences p53 proteins levels and is itself under p53 transcriptional control (Figure 4). RoR is upregulated during somatic cell reprogramming to induced pluripo-tent stem cells (iPSCs). Knockdown of LincRNA-RoR leads to upregulation of genes involved in the p53 response, oxidative stress response, and DNA damage-inducing agents.⁸⁴ In addition, knockdown of LincRNA-RoR leads to apoptosis and simultaneous knockdown of p53 partially rescued the apoptotic phe-notype indicting that LincRNA-RoR plays a role in promoting survival in iPSCs and ESCs, likely by preventing the activation of stress pathways including the p53.⁸⁴

RoR does not regulate p53 protein levels in the unstressed cells, however, following DNA damage, it represses the induction of p53 protein and similarly siRNA-mediated knockdown of lincRNA-RoR increases p53 protein levels.⁸⁵ However, RoR does not influence p53 mRNA levels in the presence or absence of Doxorubicin, suggesting post-transcriptional regulation. RoR also represses the expression of p53 targets p21 and miR-145. Furthermore, RoR reduced p53-mediated G2/M arrest and apoptosis, following Doxorubicin treatment. RNA pulldown using biotin-labeled lincRNA-RoR identified its interaction with hnRNP I that occurred mainly in the cytoplasm.⁸⁵ Knocking down hnRNP I expression caused a decrease in p53 protein levels in the Doxorubicin-treated cells. A 28-base RoR sequence carries hnRNP I-binding motifs to directly interact with hnRNP I and repress p53.⁸⁵ More interesting finding in this study is the autoregulatory feedback loop between p53 and RoR. There is a p53-binding site within the 1-kb region upstream of RoR promoter and thus p53 transcriptionally controls RoR expression.⁸⁵ In summary, p53 activation can induces RoR expression and the increased RoR expression suppress p53 by a novel negative feedback loop and thus maintaining the cellular homeostasis as unwanted induction of p53 could be deleterious to the cells.

7SL

7SL is a 300-bp lncRNA that forms a ribonucleoprotein complex with six signal-recognition proteins. 7SL is highly expressed in several cancer cell types including liver, lung, breast, and stomach, and siRNA-mediated knockdown of 7SL reduces cell growth.⁸⁶ Silencing 7SL increases the population of cells in the G1 and G2 phases of cell cycle, decreases the proportion of cells in S phase, and promotes cellular senescence and autophagy. Overexpressing 7SL increased proliferation as measured by [³H]-thymidine incorporation, indicating increased DNA replication.⁸⁶ RNA pulldown experiments that utilized biotinylated antisense oligomers complementary to p53 mRNA revealed that 7SL interacts with p53 mRNA and lowers p53 protein levels.⁸⁶ 7SL competes with HuR for binding with p53 and silencing 7SL leads to increased HuR binding to p53 mRNA, thus promoting p53 mRNA translation and accumulation.⁸⁶ These data suggest that the mechanism by which 7SL promotes cell cycle progression and sup-presses cellular senescence and autophagy in the cancer cells is by lowering p53 expression. However, in order to conclude that this competitive interaction could actually exist in a cell, it will be necessary to evaluate the stoichiometry ratios of 7SL and HuR in the cancer cells.

It has been reported previously that p53 induces the expression of the transcription factor FOXP3 following DNA damage.⁸⁷ The FOXP3 is also expressed in tumor cells and is identified as a suppressor in several tumors.⁸⁸ Interestingly, it also binds to the promoter region of 7SL and represses its transcription suggesting that 7SL is a downstream target of FOXP3.⁸⁹ Overexpression of FOXP3 in MCF7 cells significantly increased p53 protein levels and this increase was mitigated by overexpression of 7SL RNA suggesting that there exist a feedback loop between p53 and FOXP3, and 7SL is involved in this regulatory mechanism of p53 by FOXP3.⁸⁹ There-fore, targeting 7SL could reconstruct the p53/FOXP3 feedback loop and may inhibit tumor growth in cer-tain cancer types such as breast cancer.

LOC572558

The LncRNA LOC572558 was initially identified as one of the highly downregulated lncRNAs in an expression analysis of aberrantly expressed lncRNAs in bladder cancer and its levels are also low in bladder cancer cell lines.^90,91 It is a negative regulator of bladder cancer cell growth. Overexpression of LOC572558 inhibits bladder cancer cell proliferation, induces apoptosis, and decreases cell migration and invasion.⁹¹ To determine the mode of action for LOC572558, the authors performed a phosphoprotein antibody array in bladder cancer cells after overexpressing LOC572558. They found several proteins, which are important for cell proliferation, migration, invasion, and apoptosis as differentially phosphorylated after LOC572558 overex-pression. Among these, overexpression of LOC572558 resulted in reduced phosphorylation of Akt, MDM2, and increased phosphorylation of p53.⁹¹ This increase in p53 activity could possibly account for decreased proliferation after LOC572558 overexpression. How-ever, the detailed mechanism of how LOC572558 influences p53 phosphorylation levels is not known.

MT1JP

MT1JP is another lncRNA, which is downregulated in several types of cancer compared to adjacent normal tissues, including gastric, liver, colon, and lung.⁹² RNAi-mediated knockdown of MT1JP in normal human hepatic cell line, L02, led to an increase in the proportion of cells in S phase and a decrease in G1 phase, whereas, overexpression of MT1JP in the liver hepatocellular carcinoma cell line, HepG2, gave opposite phenotype indicating that MT1JP has a sig-nificant effect on the G1/S check point.⁹² Further-more, knockdown of MT1JP promotes cell proliferation, decreases the percentage of apoptotic cells, and increases L02 cell migration and invasion. Gene expression analysis in L02 cells after knocking down MT1JP revealed that it regulates p53 target genes, for example, GADD45a and p21. Interest-ingly, MT1JP knockdown resulted in a decrease in the p53 protein level by affecting the polysome distribution of p53 mRNA, thereby influencing p53 trans-lational efficiency.⁹² Thus, suggesting that MT1JP is an upstream regulator of p53. The MT1JP is a cyto-plasmic lncRNA and RNA-pulldown experiments using biotin-labeled MT1JP followed by mass spec-trometry identified interaction between MT1JP and RNA-binding protein TIAR. RNA-IP experiments using IgG antibody and TIAR antibody revealed that TIAR interacts with p53 mRNA and knockdown of TIAR decreased p53 protein levels without affecting the RNA levels.⁹² Simultaneous knocked down of MT1JP and TIAR reduced p53 protein levels suggest-ing that MT1JP may regulate p53 protein by associating with TIAR.⁹² Additional experiments showed that altering the MT1JP levels affect the polysome distribution of p53 mRNA. Interestingly, knockdown of MT1JP does not influence TIAR mRNA or protein levels, but knockdown of TIAR increased the expression of MT1JP.⁹² Altogether, this data suggests that the translational regulation of p53 occur by a complex of MT1JP and TIAR, however, whether the regulatory effect of MT1JP on p53 translation depends on the presence of TIAR is yet to be determined.

ZFAS1

The LncRNA ZFAS1 is upregulated in colorectal cancer tissues. Silencing ZFAS1 expression in HCT116 and DLD-1 inhibits proliferation, cell cycle, and colony formation on soft agar, and increases PARP cleavage.⁹³ ZFAS1 interacts with CDK1 and is predicted to sponge miR-590–3p, which targets the 3ʹ UTR of CDK1. Furthermore, ZFAS1 silencing leads to decrease in the protein levels of p53 and cyclin B1. However, the molecular mechanism behind these interactions is elusive.⁹³

p53 and ENHANCER RNAs (eRNAs)

eRNAs are noncoding, nonpolyadenylated RNAs with sizes of up to 2 kb. They are produced from enhancer domains that are characterized by high levels of H3K4me1 and H3K27Ac, low levels of H3K4me3 and RNAPII and p300/CBP binding.⁹⁴ p53 can regulate transcription of multiple distantly located genes through binding to enhancers, although there is no p53 response element in the promoter of these genes. eRNAs are produced from these p53-bound enhancers regions in a p53-dependent manner and are involved in transcription enhancement of the neighboring genes.⁹⁵ Some eRNAs are also required for p53-mediated cell cycle arrest.⁹⁵ Global runon and sequencing (GRO-seq) of MCF-7 cells after treatment with Nutlin-3a identified genome-wide p53-regulated eRNAs.⁹⁶ In addition to eRNAs, the authors also identified 194 Nutlin-3a-responsive lncRNA genes. One of these lncRNAs LED sup-pressed cell proliferation and significantly influenced the G1 arrest, following Nutlin-3a treatment and knocking down LED increases the percentage of mitotic cells after Nutlin-3a treatment.⁹⁶ Furthermore, LED knockdown significantly reduced the levels of p21. By using oligos antisense to LED and performing chromatin isolation by RNA purification technique (ChIRP), the authors found that LED associates with enhancer domains in chromatin and regulates the production of eRNAs from these enhancer regions possibly by influencing the accumulation of H3K9 acetylation at specific enhancer loci.⁹⁶ Knockdown of LED in MCF7 cells also decreased the p53-binding affinity at distal and proximal enhancers located upstream of p21 tran-scription start site. However, if LED influences the epi-genetic features of regulatory elements before dysregulating the p53 occupancy remains to be eluci-dated. In cancer cell lines and in tumor samples, especially in leukemia, LED promoter was methylated, particularly in cells lines with wild-type p53, suggesting a potential tumor suppression function of LED.⁹⁶

Another example of p53-bound enhancers (p53BER)-regulated lncRNA is linc-475, which has a p53-binding peak in its first intron and there is production of stress-dependent antisense eRNAs from its intronic p53BER.⁹⁷ As linc-475 is produced from a p53BER, there is a possibility that it might regulate the transcription of neighboring genes. To test, this the authors treated MCF7 cells with or without Nutlin-3a and performed circularized chromosome conformation capture (4C) assay followed by deep sequencing. Using the p53BER as a viewpoint, they found high frequency of interactions within the pro-moter region of the linc-475 and the upstream (SPTLC and LOC100128076) and downstream (IARS) neighboring genes. Furthermore, linc-475 is required for p53-mediated G1 arrest following nutlin-3a treatment, and knocking down linc-475 increases the number of cells entering mitosis after nutlin-3a treatment.⁹⁷ Silencing as well as knocking out linc-475 leads to a significant reduction in the mRNA and protein levels of p21, without affecting localization and protein levels of p53 and linc-475 is required for proper binding of p53 and RNAPII to the p21 promoter.⁹⁷ Altogether, these results high-light the significance of lncRNAs originating from p53BER and their role in regulating p53 transcriptional activity (Table 1).

TABLE 1.

LncRNAs That Regulate p53 or Are Regulated by p53

	Annotation	Species	Accession Number	Reference
p53-regulated lncRNAs
LincRNA-p21	Trp53cor1	Mouse	NR_036469, HM210889.1	43
PANDA	PANDAR	Human	NR_109836.1, JF803844	22
PINT A, B, C	Lncpint	Mouse	KC860257, KC860259, KC860258, NR_110470.1	51
LncRNA	LINC-PINT	Human	NR_109854.1, FLJ43663	51
PR-lncRNA-1	RP11–467J12.4	Human	NR_136518.1	52
NORAD	LINC00657	Human	NR_027451.1	3
TUG1	TUG1	Human	NR_110492.1	13,53,55
PVT1	PVT1	Human	NR_003367.3	60,63–65
LINP1	RP11–554I8.2	Human	HG502647.1	66
DDSR1	DDSR1	Human	KT318134	67
LncRNAs that regulate p53
MALAT1 (NEAT2)	MALAT1	Human	NR_002819.3	5,6
MEG3	MEG3	Human	NR_003531.3	72–75
H19	H19	Human	NR_002196.2	78–81,83
LincRNA-ST8SIA3 or LincRNA-RoR	LINC-ROR	Human	NR_048536.1	84,85
7SL	RN7SL1	Human	NR_002715.1	86,89
LOC572558	PGM5-AS1	Human	NR_121192.1	90,91
MT1JP	MT1JP	Human	NR_036677.1	93
ZFAS1	ZFAS2	Human	NR_003606.2	93
p53 and enhancer RNAs
LED	Linc00086, SMIM10L2A	Human	NR_024359.2	96
Linc-475	Linc00475	Human	NR_027341.1	97

CONCLUSIONS AND PERSPECTIVES

In normal proliferating cells, p53 levels are low. However, when cells encounter stress, p53 signal-ing is activated to regulate DNA repair, metabolic reprogramming, cell cycle arrest, apoptosis, or senescence.³⁰ p53 regulates the expression of a repertoire of lncRNAs, and some of these lncRNAs are critical regulators of cell cycle arrest, DNA damage repair, chromosomal stability and apoptosis, altogether contributing to tumor-suppressor function of p53. LncRNAs are dysregulated in a variety of cancers and therefore can be utilized as biomarkers for disease progression and as thera-peutic targets.⁹⁸ Most lncRNAs are expressed at low levels, even after induction during the stress response. This low expression indicates that most lncRNAs may not function as global regulators of gene expression but they may regulate the expres-sion of a subset of genes expressed in a cell. If these genes happen to be important regulators of specific pathway, even low abundant lncRNAs can have critical functions.

In more than 50% of the cancers, p53 is mutated and a significant proportion of the remaining 50% inactivate p53 by a variety if mechanisms including overexpression of MDM2 ²⁹. However, more than 10 million patients across the world have intact p53 protein. In the tumors with intact p53 signaling, it would be important to understand the biology and molecular function of p53-regulated lncRNAs. Similar to the p53-regulated protein-coding genes, lncRNAs may have prosurvival or proapoptotic functions in tumor cells that can be utilized for future translational research. However, figuring out the function and mode of action of an lncRNA is not straightforward due to low abundance of most lncRNAs and the diverse mechanisms by which they function. Conservation and regulation by p53 in mul-tiple cell lines can be used to select lncRNAs that may have important functions in the p53 pathway. If a given p53-regulated lncRNA shows partial conserva-tion between human and mice, the partially conserved regions may represent functional domains that are common to the human and mouse homologs. Such functional domains may act as binding sites for RNA-binding proteins that mediate the effect of the lncRNA. The identification of such domains may act as a critical step in selecting candidate biologically rel-evant p53-regulated lncRNAs. Analyses of candidate lncRNA expression in patient samples and determin-ing the association with prognosis and tumor progression may indicate the role of the lncRNA in a given cancer type. Last but not the least, experiments in mouse models and patient-derived xenografts would establish the function of the lncRNA in vivo. Future genetic and molecular analyses of p53-regulated lncRNAs will be critical in elucidating the role of lncRNAs in the p53 network.