Introduction

Long non-coding RNAs (lncRNAs), classified as RNAs larger than 200 nucleotides that do not code for proteins, are involved in regulating various aspects of life, including development and disease1. Colon cancer, one of the most prevalent tumors worldwide, is featured with poor prognosis2. Aberrant Expression of a large number of lncRNAs have been identified in colon cancer and are applied as novel biomarkers for prognosis and diagnosis3. The roles and mechanisms of lncRNAs in colon tumorigenesis and progression center on regulating cell proliferation, apoptosis, cell cycle, cell invasion, epithelial-mesenchymal transition (EMT), metastasis, cancer stem cells and drug resistance4,5.

TP53, a transcription factor tetramer, functions as the “guardian of the genome” through managements of cell cycle, DNA repair, senescence and apoptosis6. Mutations in the p53 gene are widespread in approximately half of all human malignancies, particularly in highly prevalent cancers such as breast, colon, lung and liver cancers6. Mutations in tumor repressor TP53 contribute to a series of events of cancer progressions, including proliferation, escaping apoptosis, invasion and metastasis7. The top six most frequent TP53 mutations found in cancers are referred to as “hotspot mutations”. Approximately 60% of colon cancers harbor p53 mutations, with the majority being hotspots8. Among these, missense mutations at position R175 or R273 account for more than 20% of all TP53-mutated colon tumors9. TP53 R175H mutant, which lose the cell cycle control ability, has the highest occurrence and is very malignant in promoting tumor cell growth, migration, invasion, metastasis, cancer stemness, tumor drug resistance, inflammation, and angiogenesis, suggesting an oncogenic role through gain-of-function mechanisms10. The R175P mutant, a mutation at the same site in TP53, functions in a completely different way, retaining the capability to arrest the cell cycle, but losing the ability to induce apoptosis11. Different hotspot p53 mutants produce distinct oncogenic effects9,12. Therefore, the mechanisms underlying the differences in tumorigenesis between TP53 R175H, R175P, and TP53 wild-type (WT) are worth exploring13.

lncRNAs, directly or indirectly regulated by TP53, have been discovered as crucial regulators in shaping p53 activity and biological outcomes in cancers14,15,16. Actually, the functions of a serial of TP53-regulated lncRNAs in multiple cancers including lung cancer, squamous cell cancer, colorectal cancer cells and nasopharyngeal carcinoma cells have been characterized17,18,19,20,21. The landscape of TP53-regulated lncRNAs in several types of cancer are systematically defined by varied groups22,23,24,25. However, how lncRNAs mediate the oncogenic functions of TP53 mutants in colon cancer remains unknown.

This study aimed to determine the comprehensive changes in lncRNA targets of TP53 following two types of hotspot mutations through deep sequencing. Additionally, alterations in transcriptional expression in colon cancer cells of TP53 mutations were also investigated. By combining functional enrichment analyses of both lncRNA targets and mRNAs, critical pathways potentially modulated by the two different TP53 mutants in colon cancer development were revealed. Among the differential expression lncRNAs, those predicted to be TP53 targets were involved in the regulating signaling pathways related to DNA replication and cell cycle. These findings provide a global view of the regulatory pathways of lncRNA targets of TP53 hotspot mutants in colon cancer.

Results

Systematically determination of the expression changes of lncRNAs and mRNAs upon TP53 mutation in colon cancer cells through deep sequencing

To search for the potential lncRNA targets of TP53 mutants, four DLD1 colon cancer cell lines were utilized: TP53-null control cells lacking endogenous TP53 protein expression; TP53-WT (wild type) cells overexpressing wild type exogenous TP53 protein; TP53-R175H mutant cells overexpressing TP53 protein with R175H mutation; TP53-R175P mutant cells overexpressing TP53 protein with R175P mutation. All TP53 WT and mutant cell lines were constructed in the TP53-null background. Triple samples of each cell line were subjected to deep sequencing for both lncRNA and mRNA. Quality control (Table S1) measures were implemented. Principal Component Analysis (PCA) revealed that the three samples of each cell lines clustered uniquely together in both lncRNA (Fig. 1a) and mRNA (Fig. 1b) sequencing, suggesting reasonable variation within the same cell lines. The TP53 mRNA level was significantly higher in TP53-WT cells compared to TP53-null control cells (Table S3). However, no significant difference in TP53 mRNA expression were observed between TP53-WT DLD1 and TP53-R175H DLD1 cells (Table S5) or between TP53-WT DLD1 and TP53-R175P DLD1 cells (Table S7). The protein expression of TP53 in the four cell lines (Fig. 1c) were consistent with the mRNA expression. This excluded the dosage effect of TP53 expression when comparing wild-type and mutant cells. To systematically determine the functions of lncRNAs and mRNAs upon TP53 mutation in colon cancer cells, the function enrichment of differential expression mRNAs and the targets of differential expression lncRNAs as well as prediction TP53-regulated lncRNAs were performed and the overlapping pathways were compared (Fig. 1d).

Fig. 1
figure 1

Deep sequencing systematically determined the differential expression of lncRNAs and mRNAs in colon cancer cells upon TP53 overexpression, TP53 R175H mutation or TP53 R175H mutation. (a, b) Principal Component Analysis (PCA) of lncRNA and mRNA expression of the four groups of DLD1 colon cancer cell lines. Sample variability is the driving force for PCA to extract principal components, directly determining the effectiveness and reliability of downstream analyses. (c) Western blotting showed the protein expression of TP53 and GAPDH in the DLD1 cells of TP53-null control, TP53-WT, TP53-R175H and TP53-R175P. (d) The flowchart of the study design and data analysis pipeline. Three cell group pairs, including TP53-null vs TP53-WT, TP53-R175H vs TP53-WT, and TP53-R175P vs TP53-WT, were analyzed.

Full size image

597 significantly differentially expressed lncRNAs (adjusted p value or false discovery rate (FDR) < 0.05; Fig. 2a, Table S2) and 1969 significantly differentially expressed mRNAs (FDR < 0.05; Fig. 2d, Table S3) between TP53-null control and TP53-WT DLD1 cells were identified. Among these, known p53 targets NEAT1, TUG1, and LINC-PINT15 were found in the significantly differentially expressed lncRNAs upon TP53 overexpression (Table S2), supporting the reliability of our results. Between TP53-WT DLD1 and TP53-R175H DLD1 cells, 300 significantly differentially expressed lncRNAs (FDR < 0.05; Fig. 2b, Table S4) and 899 significantly differentially expressed mRNAs (FDR < 0.05; Fig. 2e, Table S5) were identified. Between TP53-WT DLD1 and TP53-R175P DLD1 cells, 294 significantly differential expression lncRNAs (FDR < 0.05; Fig. 2c, Table S6) and 1016 significantly differential expression mRNAs (FDR < 0.05; Fig. 2f, Table S7) were determined.

Fig. 2
figure 2

The differential expression of lncRNAs and mRNAs in colon cancer cells upon TP53 overexpression, TP53 R175H mutation or TP53 R175H mutation. (a, b, c) Scatter Plots of the differentially expressing lncRNAs between TP53-null control and TP53-WT cells, TP53-WT and TP53-R175H cells, TP53-WT and TP53-R175P, respectively. The number of gene up-regulated and down-regulated is labeled. (d, e, f) Scatter Plots of the differentially expressed mRNA between TP53-null control cells and TP53-WT cells, TP53-R175H cells and TP53-WT cells, TP53-R175P cells and TP53-WT cells, respectively. The number of gene up-regulated and down-regulated is labeled.

Full size image

Identification of common pathways enriched by differentially expressed lncRNA targets and mRNAs regulated by the overexpressed wild type TP53

To analyze downstream signaling pathways, functional enrichments of KEGG and Reactome were conducted. The top 20 pathways were displayed. Enriched KEGG and Reactome pathways (Fig. 3a,c) of differential expression mRNAs between TP53-null control DLD1 and TP53-WT DLD1 cells were shown. The pathway enrichments of trans-regulatory targets (Fig. 3b,d) of differentially expressed lncRNAs between TP53-null control DLD1 and TP53-WT DLD1 cells were indicated as scatterplots. Common pathways enriched by the differentially expressed lncRNA targets and mRNAs were identified. Notably, DNA replication (mRNA FDR = 6.24E-07, lncRNA trans-regulatory target FDR = 9.88E-07) and cell cycle (mRNA FDR = 0.006416604, lncRNA trans-regulatory target FDR = 1.04E-06) (Fig. 3a,b) were among the common pathways found in KEGG enrichments, suggesting these pathways mediate TP53 regulation of cell proliferation. Reactome enrichments revealed 59 common pathways (Table S8 sheet 1, Table 1) including HCMV early events, HCMV late events, meiotic recombination, and deposition of new CENPA-containing nucleosomes at the centromere (Fig. 3c,d).

Fig. 3
figure 3

Identification of the common pathways via comparison of pathway enrichments of the trans-regulatory targets of differential expression lncRNAs between TP53-null control cells and TP53-WT cells versus pathway enrichments of differential expression mRNAs between TP53-null control cells and TP53-WT cells. (a, b) Scatterplot of enriched KEGG pathways of the differentially expressed mRNA and the trans-regulatory targets of differentially expressed lncRNAs between TP53-null control cells and TP53-WT cells. The vertical axis represents the enriched KEGG pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of differential expression genes (DEGs) enriched in certain KEGG pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. The red underlines indicate the common pathways. (c, d) Scatterplot of enriched Reactome pathways of the differentially expressed mRNA, and the trans-regulatory targets of differentially expressed lncRNAs between TP53-null control cells and TP53-WT cells. The vertical axis represents the enriched Reactome pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of DEGs enriched in certain Reactome pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. The red underlines indicate the common pathways.

Full size image
Table 1 Summary of the differential expression gene numbers and the overlapping pathways.
Full size table

Common pathways enriched by differentially expressed lncRNAs and mRNAs regulated by overexpressed TP53-R175H mutant in colon cancer cells

TP53-R175H is the most frequent mutation type and is referred to as a gain-of-function mutation in multiple cancer types 10. To search for the potential lncRNAs that mediate the function of TP53-R175H in cancer development, the differentially expressed lncRNAs between TP53-WT DLD1 and TP53-R175H DLD1 cells were detected and pathway enrichments of the differentially expressed lncRNAs were computed. The top 20 pathways enriched by KEGG and Reactome for the differentially expressed mRNAs between TP53-WT DLD1 and TP53-R175H DLD1 cells were determined (Fig. 4a,d). Enriched KEGG (Fig. 4b,c) and Reactome (Fig. 4e,f) pathways of the trans-regulatory and cis-regulatory targets of the differentially expressed lncRNAs were also shown. Comparison of enriched pathways of differentially expressed lncRNA and mRNAs discovered common KEGG signals related to DNA replication (mRNA FDR = 3.57E-06, lncRNA trans-regulatory target FDR = 5.94E-11), and the cell cycle (mRNA FDR = 0.000584871, lncRNA trans-regulatory target FDR = 2.14E-08) (Fig. 4a,b). Reactome pathway enrichments showed 53 same pathways (Table S8 sheet 2, Table 1) including HDACs deactylate histones, HCMV early events, HCMV Late Events, RNA polymerase I promoter opening, RNA polymerase I promoter escape, RMTs methylate histone arginines, meiotic recombination and Transcriptional regulation of granulopoiesis (Fig. 4d,e). The result suggests that lncRNAs and mRNAs regulated by TP53-R175H may work synergistically through these pathways to promote the progression of tumor progression driven by TP53-R175H mutations.

Fig. 4
figure 4

Identification of the common pathways via comparison of pathway enrichments of the trans-regulatory, cis-regulatory targets of differential expression lncRNAs between TP53-R175H cells and TP53-WT cells versus pathway enrichments of differential expression mRNAs between the same two cells. (a, b, c) Scatterplot of enriched KEGG pathways of the differentially expressed mRNA, the trans-regulatory targets of differentially expressed lncRNAs and the cis-regulatory targets of differentially expressed lncRNAs between TP53-R175H cells and TP53-WT cells. The vertical axis represents the enriched KEGG pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of DEGs enriched in certain KEGG pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. The red underlines indicate the common pathways. (d, e, f) Scatterplot of enriched Reactome pathways of the differentially expressed mRNA, the trans-regulatory targets of differentially expressed lncRNAs and the cis-regulatory targets of differentially expressed lncRNAs between TP53-R175H cells and TP53-WT cells. The vertical axis represents the enriched Reactome pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of DEGs enriched in certain Reactome pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. The red underlines indicate the common pathways.

Full size image

The same pathways enriched by the differentially expressed lncRNAs and mRNAs regulated by the overexpressed TP53-R175P mutant in colon cancer cells

Another hotspot mutation TP53-R175P leads to defect of cell apoptosis26. Similarly, the differentially expressed lncRNAs and mRNAs between TP53-WT DLD1 and TP53-R175P DLD1 cells were determined and the corresponding pathway enrichments were analyzed. Enriched KEGG pathways (Fig. 5a) of the differentially expressed mRNAs, and the pathway enrichments of Reactome of the differentially expressed mRNAs (Fig. 5b) were shown. However, no significant pathway (adjusted P value < 0.05) is found in KEGG and Reatome enrichments of either cis-regulatory targets or trans-regulaotry targets of differentially expressed lncRNAs. Therefore, no overlapping pathways were found between lncRNA targets and mRNAs (Table 1).

Fig. 5
figure 5

Identification of the common pathways in prediction TP53-regulated lncRNAs with differential expression. (a) Scatterplot of enriched KEGG pathways of the differentially expressed mRNA between TP53-R175P cells and TP53-WT cells. The vertical axis represents the enriched KEGG pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of DEGs enriched in certain KEGG pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. (b) Scatterplot of enriched Reactome pathways of the differentially expressed mRNA between TP53-R175P cells and TP53-WT cells. The vertical axis represents the enriched Reactome pathways. The horizontal axis represents the rich factor (enrichment score)—the ratio of the number of DEGs enriched in certain Reactome pathway to the total number of annotated genes. The higher degree of DEGs enrichment gives the greater value. The size of the dots indicates the number of DEGs enriched in certain pathway, and the color of the dots indicates the range of the adjusted p value which is significant when below 0.05. (c) Scatterplot of enriched KEGG pathways of the trans-regulatory targets of differentially expressed prediction TP53-regulated lncRNAs between TP53-null control cells and TP53-WT cells. The red underlines indicate the same pathways detected in previous pathway enrichments. (d) Scatterplot of enriched Reactome pathways of the trans-regulatory targets of differentially expressed prediction TP53-regulated lncRNAs between TP53-null control cells and TP53-WT cells. The red underlines indicate the same pathways detected in previous pathway enrichments. (e) Scatterplot of enriched KEGG pathways of the trans-regulatory targets of differentially expressed prediction TP53-regulated lncRNAs between TP53-R175H cells and TP53-WT cells. The red underlines indicate the same pathways detected in previous pathway enrichments. (f) Scatterplot of enriched Reactome pathways of the trans-regulatory targets of differentially expressed prediction TP53-regulated lncRNAs between TP53-R175H cells and TP53-WT cells. The red underlines indicate the same pathways detected in previous pathway enrichments.

Full size image

KEGG analysis of the trans-regulatory targets of differentially expressing TP53 prediction target lncRNAs highlighted DNA replication and cell cycle

To predict direct lncRNA targets of TP53, the online website tool hTFtarget-HUST (hust.edu.cn)27 was applied. A total of 5340 lncRNAs were predicted as TP53’s downstream targets (Table S9). Among these, 74 lncRNAs were differentially expressed between TP53-null control DLD1 and TP53-WT DLD1 cells, 42 lncRNAs were differentially expressed between TP53-WT DLD1 and TP53-R175H DLD1 cells, and only 22 lncRNAs were differentially expressed between TP53-WT DLD1 and TP53-R175P DLD1 cells (Table 1). The top 20 KEGG enriched pathways of the trans-regulatory targets of the differentially expressed lncRNAs specifically regulated by TP53-WT when compared to TP53-null control were displayed (Fig. 5c). Alignment with previous KEGG pathway enrichments of the differential expression mRNA between the same two groups of cells revealed two identical pathways: DNA replication and cell cycle. Similarly, KEGG analysis of the trans-regulatory targets of the differentially expressing TP53 prediction target lncRNAs between TP53-WT DLD1 and TP53-R175H DLD1 cells was performed (Fig. 5e). Alignment with previous KEGG pathway enrichments of the differential expression mRNA between TP53-WT cells and TP53-R175H mutation cells revealed two shared pathways, including DNA replication and cell cycle. Reactome enrichments of the trans-regulatory targets of the differentially expressed prediction TP53-regulated lncRNAs were aligned to Reactome pathway enrichments of the differential expression mRNA between TP53-WT and TP53-null control cells, as well as between TP53-WT DLD1 and TP53-R175H DLD1 cells (Table 1). Between TP53-WT and TP53-null control cells, 62 common pathways (Table 1, Table S10) like HCMV Early Events, HCMV Late Events, HDACs deacetylate histones, RNA Polymerase I Promoter Escape, RNA Polymerase I Promoter Opening, Senescence-Associated Secretory Phenotype (SASP) and SIRT1 negatively regulates rRNA expression et al. (Fig. 5d), were found. Between TP53-WT DLD1 and TP53-R175H DLD1 cells, 52 common pathways (Table 1, Table S11) such as Meiotic recombination, HCMV Early Events, HCMV Late Events and NoRC negatively regulates rRNA expression et al. were identified (Fig. 5f). These data indicate that DNA replication and cell cycle are prominent in both TP53’s management of cell fates and TP53-R175H’s regulation of cancer development.

To confirm TP53’s regulation of cell proliferation, a CCK8 assay was conducted. Restoration of TP53 significantly reduce the cell growth rate compared to TP53-null control cells (Fig. 6a). The R175H mutation of TP53 significantly increased DLD1 cell proliferation (Fig. 6b). Although the R175P mutation also increase the cell proliferation, its effect was less pronounced compared to R175H mutation (Fig. 6c). These results validate the findings from transcriptomic sequencing that TP53 and its R175H variant regulate the proliferation pathways, including cell cycle and DNA replication.

Fig. 6
figure 6

TP53 and its mutants regulate cell proliferation and lncRNA expression in DLD1 cells. (a) CCK8 assay the growth of TP53-null control cells and TP53-WT cells in 7 days. Data are presented as mean ± STDEV; biological replicates n = 3, error bars: standard deviation. 2-way ANOVA test is applied, *** indicates P < 0.001. (b) CCK8 assay the growth of TP53-R175H cells and TP53-WT cells in 7 days, ** indicates P < 0.01, *** indicates P < 0.001. (c) CCK8 assay the growth of TP53-R175P cells and TP53-WT cells in 7 days, *** indicates P < 0.001. (d) ChIP-qPCR analyses of TP53 binding to the promoter of lncRNA H19 and lncRNA LINC00969, respectively. TP53-WT DLD1 cells were used. Enrich folds indicate the ratio of immunoprecipitation (IP) versus input in qPCR assay. Data are presented as mean ± STDEV. Student t-test is applied, *** indicates P< 0.005, **** indicates P < 0.001.

Full size image

Common lncRNAs found in the different cell group pairs

When screening for common lncRNAs across different cell group parirs (Table S12), we found 46 lncRNAs both differentially expressed both in DLD1 cells of TP53-null control vs TP53-WT and TP53-R175H vs TP53-WT, 30 in DLD1 cells of TP53-null control vs TP53-WT and TP53-R175P vs TP53-WT, and 22 in DLD1 cells of TP53-R175P vs TP53-WT and TP53-R175H vs TP53-WT. Among these, 11 lncRNAs, including TNFRSF14-AS1, TARID, PP14571, LINC00969, LINC00471, H19, EPHA1-AS1, CCDC18-AS1, CASC8, AC125611.4, and AC091544.5 (Table S12) showed consistent differential expression across all three cell group pairs. The web tool of Tfsitescan (http://www.ifti.org/Tfsitescan/) is applied to predict TP53 transcription factor binding sites in the promoter regions of these lncRNAs. Binding sites for TP53 were found ~ 1.1 kb upstream of transcription start site (TSS) of lncRNA H19 and ~ 4.4 kb upstream of TSS of lncRNA LINC00969. Chromatin immunoprecipitation (ChIP) assay using IgG and TP53 antibodies was performed to validate the binding. Subsequent qPCR analyses revealed significant enrichment of the binding sites on H19 promoter by ~ 33-fold with the TP53 antibody and ~ 14-fold on LINC00969 promoter (Fig. 6d). The results suggest that TP53 directly modulate the expression of lncRNA H19 and LINC00969.

Discussion

Given the prevalence of TP53 mutations in over 50% of cancers, extensive research has been conducted to examine TP53-modulated lncRNAs across multiple cancer types. Employing genome-wide approaches to TP53 binding profiles in response to DNA damage has unveiled numerous novel TP53-regulated lncRNAs28. Furthermore, transcriptome profiling of TP53-overexpressed nasopharyngeal carcinoma cell line HNE2 identified 1190 differentially expressed lncRNAs along with numerous TP53 downstream signaling cascades29. Ribosome profiling and RNA-seq techniques have also revealed more than 300 novel TP53-modulated lncRNAs in liver cancer HepG2 cells30. Similarly, a genome-wide screen in a colon cancer cell line identified a substantial set (over 1000) of lncRNAs as potential TP53 targets31, with 1596 lncRNAs being dysregulated in colorectal cancer tissues5.

Herein, we initiated a comprehensive genome-wide exploration of TP53-regulated lncRNAs and their associated signaling pathways in a colon cancer cell line, DLD1, both with and without TP53 mutations. This investigation unveils the lncRNA landscape and mRNA profile of colon cancer cells harboring TP53 hotspot mutations, specifically R175H and R175P. We characterized the transcriptomes of four cell lines TP53-null control, stably overexpressing TP53-WT, TP53-R175H mutant, and TP53-R175P mutant. For each cell line, we delineated the functional enrichments of differentially expressed mRNAs and the targets of differentially expressed lncRNAs, revealing the underlying molecular regulatory mechanisms of these three cellular statuses in colon cancer. Their precise roles in colon cancer deserve further scientific investigation.

By integrating the enriched pathways of TP53-regulated mRNAs, we observed that the KEGG pathway enrichments of gene targets derived from altered TP53-modulated lncRNAs exhibited two consistent signaling pathways documented in prior research—DNA replication and cell cycle—all pivotal in TP53’s tumorigenic modulation in colon cancer. TP53 regulates cellular senescence –the permanent arrest of cell cycle32, which represents a novel treatment of colon cancer33,34. Reactome enrichments in TP53-WT cells pointed to formation of beta-catenin and TCF transactivating complex35, conferring a unique phenotype to colorectal cancer cells36. Other pathways including HCMV early events, HCMV late events, DNA methylation, RNA polymerase I promoter opening, RNA polymerase I promoter escape, condensation of prophase chromosomes, meiotic recombination, deposition of new CENPA-containing nucleosomes at the centromere, B-WICH complex positively regulates rRNA expression, ERCC6 and EHMT2 positively regulate rRNA express are all indicated in Reactome enrichments of TP53-overexpressing cells. Human cytomegalovirus (HCMV) infection is common across different types of cancer including breast cancer, colorectal cancer, brain cancer37,38. Aberrant DNA methylation is accumulated during colorectal cancer development and is applied as biomarker for diagnosis39,40. Abnormal deposition of new CENPA-containing nucleosomes at the centromere were found in colon cancer41,42. B-WICH is involved in rDNA transcription and is associated with the nucleolar stress response43. Overexpression of ERCC6 and EHMT2 has been observed in several cancer types, and anti-EHMT2 is developed as an anticancer strategy44,45,46.

Moreover, the same two common KEGG signals - DNA replication, cell cycle were identified in KEGG analyses of TP53-R175H cells. This revealed that TP53-R175H promotes cancer cell proliferation via DNA replication and cell cycle pathways. The p53 signaling pathway manages cellular senescence, apoptosis and cell cycle, dysregulation of these pathways were always detected in colon cancers8,47,48. Reactome pathway enrichments in TP53-R175H cells determined 53 shared pathways, including those mentioned in TP53-WT cells, as well as HDACs deactylate histones, activated PKN1 stimulates transcription of AR (androgen receptor) regulated genes KLK2 and KLK3, SIRT1 negatively regulates rRNA expression, RMTs methylate histone arginines, meiotic recombination and RUNX1 regulates genes involved in megakaryocyte differentiation and platelet function. Reactome enrichments proposed that TP53-R175H might regulate cancer progression through HDACs-mediated histone deacetylation, which is vital in diverse cancer developments and represents a potential therapeutic target49,50. The aberrant expression of KLK2 and KLK3 found in many human cancers highlight their significance as cancer biomarkers for early diagnosis and prognosis51. High SIRT1 expression promoted tumorigenesis in colorectal cancers52, and SIRT1 inhibitors were applied in the treatments of colorectal cancers53. Arginine methylation on histones had major implications for cancer therapy54. RUNX1 functioned directly and indirectly in cancer metastasis, angiogenesis, proliferation and cancer stemness55.

By integrating the prediction of lncRNA targets of TP53, we identified 74 differential expressed lncRNAs upon restoration of TP53 and 42 differential expressed lncRNAs upon TP53-R175H mutation. Functional enrichment analyses display the same two pathways in the 2 pairs of cell groups: DNA replication and cell cycle. Both signaling are relative to cell proliferation, highlighting its importance in TP53 modulation of cancer development. Growth curve assay of DLD1 cells validated this speculation. Interestingly, we detected consistent differential expression of 11 lncRNAs across all three cell group pairs. Among them, H19 and LINC00969 are directly regulated by TP53 in DLD1 cells. The H19 lncRNA expression are significantly upregulated in a diverse range of human malignancies, including breast, colorectal, pancreatic, glioma, and gastric cancer56. Known as an oncogene57, H19 has been confirmed to be involved in tumorigenesis and malignant progression by promoting cell growth, migration, invasion, epithelial-mesenchymal transition, metastasis, and resistance of therapy58. In colon cancer, H19 promotes the proliferation, migration and invasion of cancer cells and drug resistance59,60,61, suggesting it may mediate the gain-of-function roles of TP53 mutants in colon cancers. Similar to H19, LINC00969 confers cancer cell drug resistance62,63. On the other hand, LINC00969 is beneficial to patient via inhibition of cancer progress and proliferation64,65. Through LINC00969, TP53 may repress cell proliferation and delay cancer progression in colon tumors. The functional diversity of TP53-regulated lncRNAs indicated the intricacies of TP53’s roles in colon tumorigenesis. The TP53/miRNA/lncRNA axes could also contribute to the regulatory network of TP53 in tumorigenesis.

Through these comprehensive genome-wide analyses of TP53-modulated lncRNAs and mRNAs, we have identified crucial pathways in colon tumorigenesis. Notably, H19 and LINC00969 emerge as promising targets for further study in colon cancer.

Methods

Cell Lines. Human TP53-/- colon cancer DLD1 cells obtained from Sigma-Aldrich (St. Louis, MO; catalog No. # CLL1002) were cultured with RPMI-1640 containing 10% Fetal Bovine Serum (FBS), at 37ºC, 5% CO2. The GFP fragment in the vector of pHAGE2-EF1a-GFP1-10-IRES-Puro (Addgene Plasmid #211467) was replaced by the fragments of TP53-Flag or TP53-R175H-Flag mutant or TP53-R175P-Flag mutant via NotI and XbaI endonuclease. The cells were infected with lentivirus carrying the phage vector control, TP53-Flag or TP53-R175H-Flag mutant or TP53-R175P-Flag mutant at MOI (multiplicity of infection) of 10 for 8h, and treated with 1ug/mL puromycin for 7 days to select the positively infected clones.

RNA Sequencing. Total RNA of each cell sample was isolated via the RNA mini kit (Qiagen, Germany). RNA quality was determined by agar gel electrophoresis and with Qubit (Thermo, Waltham, MA, USA). Sequencing libraries were constructed from 1μg of total RNA using VAHTSTM Total RNA-seq (H/M/R) Library Prep Kit for Illumina®. The libraries were sequenced as 150-bp paired-end reads using Illumina Novaseq6000 according to the manufacturer’s instructions by the commercial service of Genergy Biotechnology Co. Ltd. (Shanghai, China).

Reads Mapping, Transcript Assembly and Long Non-Coding RNA Prediction. Adapters and poor quality bases in raw pair-end reads were trimmed off using Skewer version 0.2.2, https://sourceforge.net/projects/skewer/files/?source=navbar (-parameter: q 30 -l 50 -Q 30 -k 5). Software FastQC (v0.11.5, http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was applied to control the quality of pre-treated data and count the base percentage of Q20 and Q30. The quality control is provided in Table S1 and FastQC reports were included in the original report of this sequencing project. The values of Q30 of 12 samples in the quality control were all higher than 95.3% (Table S1). Clean bases numbers of 12 samples were all larger than 1.5G (Table S1). The clean reads were aligned to the human reference genome (human genome version: GRCh38) from the Ensembl (release-98) website (http://ensemblgenomes.org/) via the stochastic targeted (STAR) version 2.5.3a (parameter: --twopassMode Basic --chimSegmentMin 12 --alignMatesGapMax 200000 --alignIntronMax 200000) and annotated in the gene transfer format (GTF) file from the same website. The mapped reads were assembled into transcripts according to the reference guided assembly strategy using StringTie version1.3.1 (parameter: --rf)66. The assembled transcripts were then merged to obtain a non-redundant unified set of transcripts. The set of transcripts was compared to Ensembl gene annotations with gffcompare version 0.9.9c67. Unknown intergenic transcripts, intron transcripts, antisense exon transcripts overlapping within reference exons and antisense intron transcripts overlapping within reference introns were refereed as preliminary candidate lncRNA transcripts. Transcripts that were single-exon or short (<200bp), were filtered out due to short-read sequencing, alignment or assembly errors. The coding potential of the remaining transcripts was evaluated by PLEK version 1.2, CPAT version 1.2.2, CNCI version 1.2.2 and CPC2 version 0.168,69,70. To reduce the false positives, the intersection of the prediction results of all four software were referred as final candidate lncRNAs.

Differential Expression Genes and Functional Pathway Enrichment Analyses. The expression levels of all annotated genes and newly predicted lncRNAs were quantified with StringTie version 1.3.1. Differential expression analysis between matched cell lines were performed by the R package DESeq2 (v1.16.1) https://bioconductor.org/packages/release/bioc/html/DESeq2.html (parameter: fitType=“local”, cooksCutoff=FALSE) with the likelihood ratio test option using raw count matrix as input71.If the normalized read count value of a specific gene was close to 0, log2 transformation was performed after adding 1. Genes exhibiting higher than two-fold changes, and false discovery rate (FDR) (adjusted P-value) ≤ 0.05 were marked as differentially expressed genes. The expression values were visualized by the R package pheatmap (1.0.12) https://cran.r-project.org/web/packages/pheatmap/index.html. Gene ontology enrichment analysis for these significantly differentially expressed genes was performed by TopGO R package (2.59.0) http://www.bioconductor.org/packages/release/bioc/html/topGO.html. Reactome and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using the Hypergeometric test in R. Significantly enriched GO terms and pathways were selected by a threshold false discovery rate (FDR) (adjusted P-value) ≤ 0.05.

LncRNA Target Gene Prediction. To predict the trans-acting results, the Pearson’s correlation coefficient was calculated to determine the correlation between lncRNAs and mRNAs. Pearson correlations ≥0.95 and FDR <0.05 were considered to indicate a significant correlation between lncRNAs and mRNAs. To predict the cis-acting results, the sliding window strategy was used to search cis-acting target genes 100 kb upstream and 100 kb downstream of mRNAs. The different lncRNA-mRNA modules including lncRNAs located within 100 kb upstream or 100 kb downstream of mRNA, were counted and the overlap between lncRNA and mRNA was determined.

Identification of the lncRNA targets of TP53. The method is cited from the website of hTFtarget-HUST27 (http://bioinfo.life.hust.edu.cn/hTFtarget). The beta-model score indicated by Tang et. al72 (https://doi.org/10.1158/0008-5472.CAN-11-2091) was applied to identify the TF-target regulation. The beta-model score is calculated as the following:

$$S_{g} = \sum\limits_{i = 1}^{k} {e^{ - (0.5 + 4\Delta i)} }$$
(1)

Sg is the beta-model score represented as the sum of the weighted scores of peaks nearby the transcription start site (TSS) of gene g;

k the number of transcription factors binding sites (TFBSs) within the 50 kb distance from the TSS;

Δi is the distance between the summit of peak i and the TSS (which was normalized to 50 kb; e.g., the value 1 represents a 50 kb distance, while 0.04 indicates 2 kb).

If only one peak was detected within 50 kb distance, and the summit of the peak was exactly at 2 kb before the TSS, the beta-model score was 0.517. Thus, the value 0.517 was applied as the cutoff for regulatory capacity. If the putative functional peak with a beta-model score (Sg) ≥ 0.517 locates upstream of the TSS, we consider the TF-target regulation is reliable.

Cell proliferation assay. CCK8 (Cell Count Kit-8, Takara Bio, Inc.) assay was applied to determine the growth of DLD1 cells. Briefly, 500 cells per well were seeded into 96-well plates and cultured in RPMI-1640 containing 10% FBS, at 37ºC, 5% CO2. After 1day, 2days, 3days, 4days, 5days, 6days and 7days, 10  μL CCK-8 solution was added into each well of the 96-well plates and incubated for 1.5 h at 37 °C. The absorbance at 450 nm was measured with spectrophotometer (BioTek Instruments) and calculated. 2-way ANOVA test is applied for statistical analysis. Data are presented as mean ± STDEV; n = 3.

Western blotting. Total protein of cells was collected with RIPA buffer supplemented with protease inhibitor cocktail (Cat#04693159001, Roche). The protein concentrations were determined with the Pierce BCA assay kit (Thermo Scientific, Cat# 23225). Then, the proteins of same amount in various cells were separated on 10% SDS-PAGE and transferred to PVDF membranes. The membranes were immunoblotted with TP53 antibody (Cell Signaling Technology, Cat#2527) and GAPDH antibody (Cell Signaling Technology, Cat#5174) and visualized in the imaging system of Tanon 4600 (Tanon Science and Technology, China).

Chromatin Immunoprecipitation (ChIP). ChIP assay was performed with ChIP-IT® Express Chromatin Immunoprecipitation Kits (Active Motif Inc.) according to the manufacturer’s manual. Briefly, confluence cells in 10-cm dishes were cross-linked in 1% formaldehyde in PBS for 10 minutes at room temperature. The fixation was stopped by adding glycine to a final concentration of 125 mM. The cells were washed twice with ice-cold PBS and resuspended in lysis buffer supplemented with proteinase inhibitor cocktails (PIC) and PMSF on ice for 30 minutes to release nucleus. The nuclei pellets were spun down and resuspended in supplied digestion buffer with PIC and PMSF. The Chromatin was sheared at 37°C for 20 minutes by adding enzymatic shearing cocktails. The lysate was cleared by centrifugation (10 minutes, 18,000RCF, 4°C). Supernatant was subjected to immunoprecipitation with specific antibodies TP53 (Cell Signaling Technology #2527) and control IgG via incubation with protein G magnetic beads on a rotator for 4hours at 4°C. Supernatant (10%) was used as ChIP input. After 3 times wash, immunoprecipitated chromatin was eluted with elution buffer AM2 and crosslinks were reversed in reverse cross-linking buffer at 95°C for 10minutes. The samples were sequentially used in quantity PCR assay with specific primer sets for the predicted transcription factor binding sites on human genome.