Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jul 15.
Published in final edited form as: Genomics. 2018 Jun 1;111(4):762–771. doi: 10.1016/j.ygeno.2018.05.006

The p53-signaling pathway and colorectal cancer: interactions between downstream p53 target genes and miRNAs

Martha L Slattery 1, Lila E Mullany 1, Roger K Wolff 1, Lori C Sakoda 2, Wade S Samowitz 3, Jennifer S Herrick 1
PMCID: PMC6274615  NIHMSID: NIHMS977307  PMID: 29860032

Abstract

Introduction

We examined expression of genes in the p53-signaling pathway. We determine if genes that have significantly different expression in carcinoma tissue compared to normal mucosa also have significantly differentially expressed miRNAs. We utilize a sample of 217 CRC cases.

Methods

We focused on fold change (FC) >1.50 or <0.67 for genes and miRNAs, that were statistically significant after adjustment for multiple comparisons. We evaluated the linear association between the differential expression of miRNA and mRNA. miRNA:mRNA seed-region matches also were determined.

Results

Eleven dysregulated genes were associated with 37 dysregulated miRNAs; all were down-stream from the TP53 gene. MiR-150–5p (HR=0.82) and miR-196b-5p (HR 0.73) significantly reduced the likelihood of dying from CRC when miRNA expression increased in rectal tumors.

Conclusions

Our data suggest that activation of p53 from cellular stress, could target downstream genes that in turn could influence cell cycle arrest, apoptosis, and angiogenesis through mRNA:miRNA interactions.

Keywords: Colorectal Cancer, TP53, miRNA, mRNA

Introduction

The tumor suppressor gene, TP53, is one of the most commonly mutated genes in many cancers including colorectal cancer (CRC) (1). P53 is normally expressed at low levels, in part through negative feedback loops that involve MDM2, a transcriptional target of p53 that when activated or overexpressed, mediates the degradation of p53 (2). Low levels of p53 expression maintains homoeostasis of the cell cycle and cell death. MDMX (also known as MDM4), a homolog of MDM2, is not regulated by p53, but by forming heterodimers with MDM2 can enhance MDM2 to induce p53 degradation (2). In response to stress factors, such as DNA damage, hypoxia, UV irradiation, oxidative free radicals, oncogenes, or deficiencies in growth factors or nutrients, the p53 protein is activated. Activation of p53 can either transactivate or repress downstream target genes that in turn regulate cell cycle arrest, apoptosis, DNA repair, and angiogenesis and metastasis (3).

As a transcription factor (TF), TP53, is involved in the regulation of many genes. Similarly, miRNAs which are short non-coding RNA molecules, are involved in the regulation of genes expression through translational repression or degradation of mRNAs (4). Investigations into the mechanisms whereby p53 is activated in tumors have examined miRNAs as important components of the p53 transcriptional network (5). MiR-25 and miR-30d, have been shown to negatively regulate the TP53 gene (5). Several other miRNAs, including miR-16–1, miR-143, miR-145, miR-34, miR-194, miR-192, miR-215, and miR-29, have been identified as miRNAs that are involved in the TP53 network, either by being directly altered by p53 or through their associations with downstream genes targeted by p53.

In this study, we examined how miRNAs are associated with dysregulated genes in the p53-signaling pathway during CRC tumorigenesis. Genes in the p53-signaling pathway were identified using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. We determined, which genes in this pathway were dysregulated based on their expression in paired carcinoma and normal mucosa, i.e. the expression levels in the carcinoma were significantly different than the expression in the normal mucosa. Differentially expressed miRNAs were examined with dysregulated mRNA and seed-region matches between the mRNA and the miRNA were used to help identify associations that had a greater likelihood of a direct effect in that the miRNA directly influences gene expression. MiRNA:mRNA associations without a seed-region match and without a negative beta coefficient were more likely to have an indirect biological function involving feedback loops, i.e. the miRNA and mRNA do not bind but through feedback loops the miRNA is still influencing the gene expression. We further examined the impact of the differentially expressed mRNA and miRNA on CRC-specific survival.

Methods

Study Participants

This study incorporates data from two population-based case-control studies of all incident colon and rectal cancer patients who were between 30 to 79 years of age living in Utah or who were members of Kaiser Permanente Northern California (KPNC). Participants were non-Hispanic white, Hispanic, black, or Asian (rectal cancer study only) (6, 7). SEER (Surveillance, Epidemiology, and End Results) registries were used to verify cases as a first primary adenocarcinoma of the colon or rectum within the study-specific dates of October 1991 to September 1994 (colon study) or between May 1997 and May 2001 (rectal study) (8). SEER registries also provided information on tumor site, date of diagnosis, date of death or last follow-up, months of survival, and cause of death. This study consist of individuals from the parent studies who had sufficient tumor and normal mucosa tissue for RNA-related analysis.

Ethics, consent, and permissions

Study participants signed informed consent. The Institutional Review Boards at the University of Utah and at KPNC approved this study. Individual-level participant data are not reported.

RNA processing

Formalin-fixed paraffin embedded tissue from the initial biopsy or surgery was used for RNA extraction. RNA was extracted, isolated and purified from carcinoma tissue and adjacent normal mucosa as previously described (9). The study pathologist (WS) reviewed all slides and identified normal and carcinoma tissue that was used to extract RNA. Variation in RNA quality by date of block was not observed.

mRNA: RNASeq Sequencing Library Preparation and Data Processing

Total RNA from 245 colorectal carcinoma and normal mucosa pairs was chosen for sequencing based on availability of RNA and high quality miRNA data in order to have both mRNA and miRNA from the same individuals; the 217 pairs that passed quality control (QC) were used in these analyses (10). RNA library construction was performed with the Illumina TruSeq Stranded Total RNA Sample Preparation Kit with Ribo-Zero. The samples were then fragmented and primed for cDNA synthesis, adapters were then ligated onto the cDNA, and the resulting samples were then amplified using PCR; the amplified library was then purified using Agencount AMPure XP beads. A more detailed description of the methods can be found in our previous work (11). Illumina TruSeq v3 single read flow cell and a 50 cycle single-read sequence run were performed on an Illumina HiSeq instrument. Reads were aligned to a sequence database containing the human genome (build GRCh37/hg19, February 2009 from genome.ucsc.edu) and alignment was performed using novoalign v2.08.01. Total gene counts were calculated for each exon and untranslated region (UTR) of the genes using gene coordinates obtained from http://genome.ucsc.edu. Genes that were not expressed in our RNASeq data or for which the expression was missing for the majority of samples were excluded from further analysis (11).

We have previously compared RNASeq data and qPCR data with dietary intake data with similar results observed for both methods (10). Analysis of expression of 20 genes analyzed with RNASeq were compared to expression data from qPCR. In these analyses we found no differences in associations between the two methods using the Wilcoxon Rank Sum test. These analyses support the validity of the RNASeq data.

miRNA

The Agilent Human miRNA Microarray V19.0 was used. Data were required to pass QC parameters established by Agilent that included tests for excessive background fluorescence, excessive variation among probe sequence replicates on the array, and measures of the total gene signal on the array to assess low signal. Samples failing to meet quality standards were re-labeled, hybridized to arrays, and re-scanned. If a sample failed QC assessment a second time, the sample was excluded from analysis. The repeatability associated with this microarray was extremely high (r=0.98) (8); comparison of miRNA expression levels obtained from the Agilent microarray to those obtained from qPCR had an agreement of 100% in terms of directionality of findings and the fold changes (FCs) were almost identical (12). To normalize differences in miRNA expression that could be attributed to the array, amount of RNA, location on array, or factors that could erroneously influence miRNA expression levels, total gene signal was normalized by multiplying each sample by a scaling factor which was the median of the 75th percentiles of all the samples divided by the individual 75th percentile of each sample (13).

KEGG-identified TP53 Genes

The KEGG (www.genome.jp/kegg/pathway/hsa/hsa04115.html) pathway map for the p53 signaling pathway was used to identify genes included in this analysis. We identified 68 genes (Supplemental Table S1) in this pathway, 67 of which had sufficient expression in CRC tissue for inclusion.

Statistical Methods

We performed negative binomial mixed effects modeling in SAS (accounting for carcinoma/normal status as well as subject effect) to determine which genes in the KEGG-identified p53-signaling pathway had a statistically significant difference in expression between individually paired CRC and normal mucosa. We offset the overall exposure as the log of the expression of all identified protein-coding genes in the negative binomial model (n=17461). The Benjamini and Hochberg (14) method was used to control the false discovery rate (FDR) with a value of <0.05. A FC greater than one indicates a positive differential expression (i.e. upregulated in carcinoma tissue); a FC less than one indicates a negative differential expression (i.e. downregulated in carcinoma tissue). We calculated the level of expression of each gene by dividing the total expression for that gene in an individual by the total expression of all protein-coding genes per million transcripts (RPMPCG or reads per million protein-coding genes). We considered overall CRC differential expression as well as differential expression specific for microsatellite unstable (MSI) and microsatellite stable (MSS) tumors.

We focused our analysis with miRNAs on those genes with FCs of >1.50 or <0.67 (or 1/1.50) to have a greater degree of meaningful biological differences between carcinoma and normal samples. Analysis included the 814 miRNAs expressed in greater than 20% of the population of their normal colorectal mucosa samples. For both miRNA and mRNA data, differential expression was calculated using subject-level paired data as the expression in the carcinoma tissue minus the expression in the normal mucosa. In these analyses, we fit a least squares linear regression model to the RPMPCG differential expression levels and miRNA differential expression levels. P-values were generated using the bootstrap method by creating a distribution of 10,000 F statistics derived by resampling the residuals from the null hypothesis model of no association between gene expression and miRNA expression using the boot package in R. Linear models were adjusted for age and sex. The linear model allows us to more dynamically capture changes in miRNA expression that correspond to changes in mRNA expression. This furthers our understanding of potential target genes of miRNAs. A negative beta coefficient suggest an inverse association between the changes of miRNA and mRNA; i.e. the miRNA binds to the mRNA to alter gene expression. Multiplicity adjustments for gene/miRNA associations were made at the gene level using the FDR by Benjamini and Hochberg (14).

Survival analysis was conducted using survival months calculated from the difference between the diagnosis date and date of death or date of last follow-up (December 2001 for colon cancer and April 2007 for rectal cancer); at date of last follow-up all cases had over five years of follow-up since diagnosis. CRC-specific follow-up included deaths where the primary or secondary cause of death was listed as CRC. Individuals dying of other causes or who were lost to follow-up were censored at their time of death or date of last contact. The R package “survival” was used to calculate p-values based upon 10,000 permutations of the likelihood ratio test from the Cox proportional hazards model adjusted for age at diagnosis, gender, and AJCC tumor stage. Hazard Ratios (HR) are reported based on the difference between the 75th and 25th percentile of expression. While survival analysis of mRNA was limited to the 217 CRC cases, examination of key pathway miRNAs with survival was able to use a much larger sample. For differentially expressed miRNA survival analysis, data were available for 1134 colon cancer cases and 721 rectal cancer cases.

Bioinformatics Analysis

We further analyzed significantly associated genes (mRNA) with a more meaningful FC for seed-region matches with miRNAs by examining the mRNA 3’ UTR FASTA with the seed region sequence of the associated miRNA. MiRNA seed regions (seeds of 6–8 nucleotides in length) were calculated as described in our previous work (15). We believe that a seed match would increase the likelihood that an identified miRNA:mRNA interaction indicates a greater likelihood of a direct biological effect on expression given a higher propensity for binding, especially when a negative beta coefficient was observed. This implies that as the differential expression in the miRNA increased, the differential expressed in the mRNA decreased. Since miRTarBase (16) uses findings from many different investigations and alignments, we used FASTA sequences generated from both GRCh37 and GRCh38 Homo sapiens, using UCSC Table Browser (https://genome.ucsc.edu/cgi-bin/hgTables) (17). FASTA sequences that matched our Ensembl IDs and had a consensus coding sequences (CCDS) available were downloaded. Scripts in R 3.2.3 and in perl 5.018002 were used to conduct the analysis.

Results

The study population consisted of 169 colon cancer cases and 48 rectal cancer cases (Table 1). The majority of these cases were male (54.4%) and non-Hispanic white (74.2%). By tumor molecular phenotype, 13.4% were MSI and 86.6% were MSS. At the end of follow-up, 57.4% were alive.

Table 1.

Description of Study Population

N %
Site
Colon 169 77.9
Rectal 48 22.1
Sex
Male 118 54.4
Female 99 45.6
Age
Mean (SD) 64.8 10.1
Race
non-Hispanic White 161 74.2
Hispanic 14 6.5
non-Hispanic Black 8 3.7
Unknown 34 15.7
Tumor Phenotype
MSS 188 86.6
MSI 29 13.4
AJCC Stage
1 58 27.1
2 61 28.5
3 72 33.6
4 23 10.8
Vital Status
Dead 92 42.6
Alive 124 57.4
Open in a new tab

When examining all CRC cases (Table 2), five genes, IGF1, TP53I3, FAS, SESN2, and THBS1 were significantly down-regulated with a FC <0.67; 22 genes were up-regulated with a FC >1.5. Six genes, RRM2 (FC 3.15), PMAIP1 (FC 3.32), CDK1 (FC 3.45), CCNB1 (FC 3.50), PERP (FC 3.68), and SERPINB5 (FC 4.36) were up-regulated in tumors with a FC of >3.0. Examination of differential gene expression by tumor molecular phenotype showed almost identical results for MSS tumors and for MSI and MSS tumors combined (Supplemental Table 2). However for MSI tumors, two genes, RPRM (FCMSI 0.23; FCALL 2.03) and SESN1 (FCMSI 0.84 and FCALL 1.54) were down-regulated in MSI tumors (Supplemental Table 3) while being up-regulated overall and in MSS tumors, as in the case of RPRM (FC 2.03) or minimally down-regulated overall. Four other genes, MDM2, DDB2, and EI24, were more strongly up-regulated in MSI tumors than in all CRC tumors overall (FCMSI 1.56, 1.50, and 1.79 respectively and FCALL 1.19, 1.08, and 1.47 respectively). Figure 1 highlights the dysregulated mRNA in the KEGG p53-signaling pathway.

Table 2.

Differentially expressed genes in KEGG TP53 pathway

Gene Name Tumor Mean Normal Mean Fold Change P-Value Adjusted P-Value
IGF1 13.85 38.80 0.36 5.43E-28 3.31E-27
TP53I3 27.31 59.98 0.46 2.39E-39 3.21E-38
FAS 28.16 50.95 0.55 2.75E-24 1.23E-23
SESN2 19.04 30.80 0.62 3.25E-18 9.47E-18
THBS1 585.51 881.12 0.66 3.69E-18 1.03E-17
CCNG2 52.36 71.48 0.73 4.69E-13 1.01E-12
CASP9 13.05 17.74 0.74 1.52E-08 2.67E-08
TP53AIP1 0.61 0.81 0.75 1.59E-01 1.78E-01
ATM 239.32 312.70 0.77 4.82E-16 1.19E-15
CYCS 95.12 122.36 0.78 1.03E-10 2.08E-10
MDM4 282.34 351.94 0.80 1.27E-15 3.05E-15
RCHY1 28.51 34.63 0.82 2.42E-06 3.68E-06
CCND3 33.57 40.72 0.82 1.67E-06 2.66E-06
PTEN 125.77 151.99 0.83 2.25E-09 4.18E-09
CDKN1A 84.50 98.10 0.86 3.83E-03 4.94E-03
SFN 46.49 53.89 0.86 1.01E-02 1.26E-02
GORAB 16.88 19.51 0.87 4.27E-03 5.40E-03
GADD45G 2.79 3.18 0.88 2.07E-01 2.20E-01
SESN3 125.93 135.93 0.93 1.78E-01 1.92E-01
PIDD 28.99 31.19 0.93 1.08E-01 1.25E-01
APAF1 85.50 90.73 0.94 7.64E-02 8.98E-02
SIAH1 44.02 46.13 0.95 1.63E-01 1.79E-01
CASP3 42.94 44.90 0.96 3.39E-01 3.48E-01
TSC2 149.20 153.46 0.97 2.55E-01 2.67E-01
PPM1D 22.06 22.54 0.98 6.54E-01 6.54E-01
GADD45B 12.90 12.05 1.07 3.43E-01 3.48E-01
CCNG1 99.48 92.28 1.08 7.13E-02 8.53E-02
DDB2 19.14 17.69 1.08 1.53E-01 1.73E-01
GADD45A 13.07 11.37 1.15 3.07E-02 3.73E-02
RFWD2 99.18 84.17 1.18 1.63E-08 2.79E-08
MDM2 234.16 196.94 1.19 1.56E-06 2.55E-06
BAX 36.23 29.75 1.22 1.59E-04 2.22E-04
CASP8 70.64 57.92 1.22 3.90E-09 7.06E-09
CD82 33.29 27.02 1.23 1.55E-04 2.21E-04
CCNE2 25.72 20.34 1.26 2.47E-06 3.68E-06
BID 35.72 27.95 1.28 2.19E-09 4.18E-09
RRM2B 38.14 28.77 1.33 2.92E-07 4.89E-07
CDKN2A 6.06 4.39 1.38 3.14E-03 4.12E-03
CHEK2 17.67 12.81 1.38 2.44E-06 3.68E-06
ATR 133.15 96.36 1.38 1.07E-20 3.42E-20
SHISA5 118.78 82.99 1.43 8.67E-24 3.42E-23
TP73 6.29 4.37 1.44 5.69E-04 7.78E-04
ZMAT3 78.77 54.70 1.44 2.10E-12 4.40E-12
SERPINE1 40.60 27.80 1.46 8.68E-05 1.26E-04
EI24 92.68 62.99 1.47 5.23E-21 1.75E-20
TNFRSF10B 135.53 89.65 1.51 3.45E-21 1.22E-20
SESN1 62.44 40.42 1.54 2.18E-13 4.87E-13
STEAP3 54.77 34.22 1.60 5.82E-18 1.56E-17
CCND2 773.45 483.06 1.60 3.23E-16 8.33E-16
CCNB2 21.30 12.39 1.72 1.18E-13 2.73E-13
CDK6 289.15 166.12 1.74 1.71E-35 1.43E-34
TP53 105.07 59.63 1.76 3.25E-24 1.36E-23
BBC3 20.58 11.49 1.79 2.02E-18 6.14E-18
CCNE1 8.70 4.80 1.81 2.93E-10 5.77E-10
CDK2 45.70 24.96 1.83 8.37E-33 5.61E-32
IGFBP3 99.13 50.34 1.97 9.03E-25 4.65E-24
RPRM 1.53 0.75 2.03 8.38E-04 1.12E-03
GTSE1 24.59 11.82 2.08 4.46E-23 1.66E-22
CHEK1 37.56 15.63 2.40 4.25E-34 3.17E-33
CDK4 66.65 26.90 2.48 9.86E-58 2.20E-56
CCND1 317.79 122.64 2.59 1.53E-59 5.12E-58
RRM2 87.93 27.87 3.15 5.82E-41 9.75E-40
PMAIP1 11.00 3.31 3.32 6.06E-26 3.38E-25
CDK1 41.36 11.94 3.46 1.20E-36 1.34E-35
CCNB1 32.40 9.25 3.50 1.54E-36 1.47E-35
PERP 276.98 75.18 3.68 3.02E-65 2.02E-63
SERPINB5 42.33 9.70 4.36 2.57E-24 1.23E-23
Open in a new tab

Fig. 1.

Fig. 1.

Open in a new tab

KEGG-defined p53 signaling pathway with dysregulated mRNA and miRNA highlighted.

Eleven dysregulated genes were associated with 37 dysregulated miRNAs (Table 3); all of these genes were downstream of TP53 and none of the genes had unique associations with MSI or MSS tumors. Five genes (CCNB1, CCND1, CDK1, CDK4, and GTSE1) in the arm leading to cell cycle arrest; TFNFRS10B, PERP, and IGF1, in the arm leading to apoptosis; THBS1 leading to inhibition of angiogenesis; RRM2 leading to DNA repair, and STEAP3 in exosome mediated secretion. CCND1 was associated with 11 miRNAs of which nine had a seed-region match between the mRNA and miRNA; none of these matches had a negative beta coefficient suggesting that the effect is likely indirect. PERP was associated with24 miRNAs. Although 14 of these miRNAs had a seed-region match with PERP, only miR-650 had a negative beta coefficient suggesting a direct biological effect. RRM2 was associated with 14 miRNAs, and two of the eight miRNAs with a seed-region match (miR-150–5p and miR-650) also had a negative beta coefficient. TNFRSF10B, was associated with two miRNAs, and had a seed-region match with miR-196b-5p; in this interaction the differential expression of the mRNA was inversely associated with the differential expression of the miRNA as demonstrated by the negative beta coefficient. Figure 2 shows the scatter plots of the four miRNAs that have a negative beta coefficient with mRNA and also has a seed region match. This provides support that the miRNA is directly binding to the mRNA to alter gene expression. Four miRNAs, including miR-145–5p, miR-17–5p, miR-20b-5p, and miR-93–5p each were associated with five or more genes.

Table 3.

Associations between differentially expressed mRNA and differentially expressed miRNA

Gene Name Tumor Mean Normal Mean Fold Change miRNA Tumor Mean Normal Mean Fold Change Beta Raw p-value FDR p-value
CCNB1 32.40 9.25 3.50 hsa-miR-145–5p 132.97 223.14 0.60 -0.30 <.0001 0.0203
hsa-miR-195–5p 3.59 12.18 0.29 -0.28 0.0003 0.0305
hsa-miR-25–3p 30.05 12.78 2.35 0.27 0.0002 0.0233
hsa-miR-93–5p 41.72 15.20 2.74 0.30 <.0001 0.0203
CCND1 317.79 122.64 2.59 hsa-miR-106b-5p 15.90 5.19 3.06 0.25 0.0006 0.0349
hsa-miR-17–5p 61.04 16.38 3.73 0.30 <.0001 0.0244
hsa-miR-19b-3p 29.80 10.42 2.86 0.28 0.0003 0.0244
hsa-miR-203a 12.52 3.70 3.38 0.27 0.0002 0.0244
hsa-miR-20a-5p 70.78 17.61 4.02 0.28 <.0001 0.0244
hsa-miR-20b-5p 17.65 3.30 5.35 0.29 <.0001 0.0244
hsa-miR-21–5p 463.11 167.37 2.77 0.25 0.0006 0.0349
hsa-miR-221–3p 13.53 4.12 3.28 0.26 0.0003 0.0244
hsa-miR-27a-3p 56.26 23.29 2.42 0.27 0.0002 0.0244
hsa-miR-29b-3p 24.31 9.83 2.47 0.27 0.0003 0.0244
hsa-miR-93–5p 41.72 15.20 2.74 0.26 0.0002 0.0244
CDK1 41.36 11.94 3.46 hsa-miR-106b-5p 15.90 5.19 3.06 0.24 0.0006 0.0488
hsa-miR-17–5p 61.04 16.38 3.73 0.28 <.0001 0.0203
hsa-miR-19b-3p 29.80 10.42 2.86 0.29 <.0001 0.0203
hsa-miR-20b-5p 17.65 3.30 5.35 0.27 0.0002 0.0271
hsa-miR-221–3p 13.53 4.12 3.28 0.24 0.0006 0.0488
hsa-miR-25–3p 30.05 12.78 2.35 0.35 <.0001 0.0203
hsa-miR-93–5p 41.72 15.20 2.74 0.33 <.0001 0.0203
CDK4 66.65 26.90 2.48 hsa-miR-17–5p 61.04 16.38 3.73 0.33 <.0001 0.0271
hsa-miR-20a-5p 70.78 17.61 4.02 0.28 <.0001 0.0271
hsa-miR-20b-5p 17.65 3.30 5.35 0.26 0.0002 0.0407
hsa-miR-3651 58.66 25.92 2.26 0.26 0.0003 0.0488
hsa-miR-93–5p 41.72 15.20 2.74 0.33 <.0001 0.0271
GTSE1 24.59 11.82 2.08 hsa-miR-145–5p 132.97 223.14 0.60 -0.33 <.0001 0.0116
hsa-miR-199a-5p 20.18 9.28 2.17 -0.23 0.0007 0.0438
hsa-miR-365a-3p 8.43 4.33 1.94 -0.26 0.0003 0.0222
hsa-miR-99a-5p 6.30 3.70 1.71 -0.26 0.0002 0.0181
IGF1 13.85 38.80 0.36 hsa-miR-145–5p 132.97 223.14 0.60 0.28 <.0001 0.0407
PERP 276.98 75.18 3.68 hsa-miR-1291 5.52 3.67 1.51 0.22 0.0017 0.0407
hsa-miR-151a-3p 5.15 1.56 3.31 0.25 0.0002 0.009
hsa-miR-17–5p 61.04 16.38 3.73 0.28 0.0002 0.009
hsa-miR-199b-5p 4.69 1.53 3.07 0.26 0.0003 0.0129
hsa-miR-19b-3p 29.80 10.42 2.86 0.24 0.001 0.0291
hsa-miR-203a 12.52 3.70 3.38 0.29 0.0002 0.009
hsa-miR-20a-5p 70.78 17.61 4.02 0.28 0.0002 0.009
hsa-miR-20b-5p 17.65 3.30 5.35 0.31 <.0001 0.009
hsa-miR-21–3p 22.68 9.89 2.29 0.22 0.0021 0.0471
hsa-miR-21–5p 463.11 167.37 2.77 0.29 0.0002 0.009
hsa-miR-221–3p 13.53 4.12 3.28 0.30 <.0001 0.009
hsa-miR-222–3p 19.45 11.08 1.76 0.28 <.0001 0.009
hsa-miR-23a-3p 174.68 87.53 2.00 0.25 0.0005 0.0163
hsa-miR-24–3p 106.75 62.39 1.71 0.22 0.0016 0.0407
hsa-miR-27a-3p 56.26 23.29 2.42 0.23 0.0011 0.0309
hsa-miR-29a-3p 110.29 51.04 2.16 0.29 <.0001 0.009
hsa-miR-29b-3p 24.31 9.83 2.47 0.28 <.0001 0.009
hsa-miR-324–5p 5.20 2.27 2.29 0.21 0.0022 0.0471
hsa-miR-34a-5p 25.15 12.32 2.04 0.23 0.0008 0.025
hsa-miR-3651 58.66 25.92 2.26 0.21 0.0022 0.0471
hsa-miR-501–3p 7.07 2.95 2.39 0.27 0.0002 0.009
hsa-miR-650 4.51 16.60 0.27 -0.26 0.0004 0.0142
hsa-miR-663b 65.50 32.21 2.03 0.28 <.0001 0.009
hsa-miR-92a-3p 121.60 41.18 2.95 0.36 <.0001 0.009
RRM2 87.93 27.87 3.15 hsa-miR-106b-5p 15.90 5.19 3.06 0.29 0.0002 0.0163
hsa-miR-145–5p 132.97 223.14 0.60 -0.33 <.0001 0.0102
hsa-miR-150–5p 14.90 39.17 0.38 -0.23 0.0009 0.0318
hsa-miR-17–5p 61.04 16.38 3.73 0.28 0.0003 0.0163
hsa-miR-195–5p 3.59 12.18 0.29 -0.25 0.0004 0.0203
hsa-miR-19b-3p 29.80 10.42 2.86 0.25 0.0005 0.0226
hsa-miR-20a-5p 70.78 17.61 4.02 0.23 0.0012 0.0362
hsa-miR-20b-5p 17.65 3.30 5.35 0.28 <.0001 0.0102
hsa-miR-21–3p 22.68 9.89 2.29 0.24 0.0008 0.0296
hsa-miR-221–3p 13.53 4.12 3.28 0.24 0.0011 0.0344
hsa-miR-25–3p 30.05 12.78 2.35 0.30 <.0001 0.0102
hsa-miR-650 4.51 16.60 0.27 -0.23 0.001 0.0339
hsa-miR-93–5p 41.72 15.20 2.74 0.35 <.0001 0.0102
hsa-miR-99a-5p 6.30 3.70 1.71 -0.23 0.0011 0.0344
STEAP3 54.77 34.22 1.60 hsa-miR-196a-5p 6.70 4.21 1.59 0.25 0.0002 0.0407
hsa-miR-20b-5p 17.65 3.30 5.35 0.27 0.0002 0.0407
hsa-miR-583 6.61 3.22 2.05 0.25 <.0001 0.0407
THBS1 585.51 881.12 0.66 hsa-miR-145–5p 132.97 223.14 0.60 0.29 <.0001 0.0407
TNFRSF10B 135.53 89.65 1.51 hsa-miR-196b-5p 17.89 5.53 3.24 -0.25 0.0003 0.0488
hsa-miR-3124–5p 1.37 2.27 0.60 -0.26 0.0003 0.0488
Open in a new tab

Figure 2A.

Figure 2A.

Open in a new tab

Scatter plot of RRM2 and has_miR-650

Figure 2B.

Figure 2B.

Open in a new tab

Scatter plot of RRM2 and has_miR_150_5p

Figure 2C.

Figure 2C.

Open in a new tab

Scatter plot of PERP and has_miR_650

Figure 2D.

Figure 2D.

Open in a new tab

Scatter plot of TNFRSF10B and has_miR-196b-5p

TP53 was not associated with any miRNAs after adjustment for multiple comparisons (See Supplemental Table 4). However, prior to adjustment for multiple comparisons 63 miRNAs were associated with TP53 differential gene expression.

Only two mRNAs, (CHEK1 HR 0.56 95% CI 0.38, 0.83, Praw 0.004, Padj 0.12 and SESN3 HR 0.70 95% CI 0.50, 0.99 Praw 0.04 and Padj 0.64) were associated with CRC-specific survival prior to adjustment for multiple comparison (Supplemental Table 5). Nine miRNAs were associated with CRC-specific survival in all CRC cases, although the adjusted p values were slightly over 0.05 for all but one miRNA (miR-20b-5p) (Table 4). Likewise, evaluation of CRC-specific survival by tumor location showed that no statistically significant associations after adjustment for multiple comparisons for colon cancer, while 19 miRNAs were associated with CRC-specific survival for those diagnosed with rectal cancer.

Table 4.

Associations between differentially expressed miRNAs associated with mRNA and CRC survival

miRNA Q1 Q3 HR1 (95% CI) P-value Q-value FDR- p value
All Colorectal Cancer
hsa-miR-145–5p -1.94 -0.15 1.13 (1.01, 1.26) 0.03 0.07 0.14
hsa-miR-17–5p 0.95 2.40 0.91 (0.84, 0.98) 0.02 0.07 0.14
hsa-miR-19b-3p 0.61 2.35 0.91 (0.84, 0.99) 0.02 0.07 0.14
hsa-miR-20a-5p 1.02 2.61 0.91 (0.84, 0.98) 0.02 0.07 0.14
hsa-miR-20b-5p 0.96 3.11 0.83 (0.75, 0.91) 0.0002 0.07 0.01
hsa-miR-29a-3p 0.39 1.74 0.91 (0.84, 0.98) 0.01 0.07 0.14
hsa-miR-34a-5p 0.27 1.57 0.90 (0.84, 0.97) 0.01 0.07 0.09
hsa-miR-92a-3p 0.63 1.87 0.91 (0.83, 0.99) 0.03 0.07 0.14
hsa-miR-93–5p 0.71 1.99 0.93 (0.86, 1.00) 0.04 0.07 0.17
Colon
hsa-miR-145–5p -1.88 -0.03 1.18 (1.03, 1.36) 0.02 0.20 0.37
hsa-miR-99a-5p -0.94 1.53 1.17 (1.02, 1.34) 0.03 0.20 0.37
Rectal
hsa-miR-106b-5p 0.71 2.64 0.80 (0.68, 0.94) 0.01 0.01 0.03
hsa-miR-1291 0.00 1.40 0.85 (0.75, 0.97) 0.02 0.02 0.04
hsa-miR-150–5p -2.49 -0.73 0.82 (0.70, 0.96) 0.01 0.01 0.04
hsa-miR-151a-3p 0.00 2.01 0.77 (0.63, 0.93) 0.01 0.01 0.03
hsa-miR-17–5p 1.13 2.43 0.80 (0.70, 0.91) 0.002 0.003 0.01
hsa-miR-196a-5p -0.41 1.91 0.81 (0.68, 0.96) 0.02 0.02 0.04
hsa-miR-196b-5p 0.00 3.21 0.73 (0.59, 0.89) 0.002 0.003 0.01
hsa-miR-19b-3p 0.82 2.39 0.80 (0.70, 0.91) 0.001 0.003 0.01
hsa-miR-203a 0.00 2.76 0.82 (0.68, 0.99) 0.04 0.04 0.07
hsa-miR-20a-5p 1.20 2.70 0.79 (0.69, 0.91) 0.001 0.003 0.011
hsa-miR-20b-5p 1.24 3.30 0.67 (0.57, 0.79) 0.0001 0.003 0.004
hsa-miR-21–3p 0.50 1.62 0.85 (0.75, 0.97) 0.02 0.02 0.04
hsa-miR-221–3p 0.63 2.57 0.81 (0.69, 0.94) 0.01 0.01 0.03
hsa-miR-25–3p 0.63 1.84 0.86 (0.76, 0.98) 0.02 0.02 0.04
hsa-miR-29a-3p 0.50 1.74 0.77 (0.68, 0.88) 0.0003 0.003 0.01
hsa-miR-29b-3p 0.33 1.90 0.85 (0.73, 0.99) 0.046 0.046 0.08
hsa-miR-34a-5p 0.23 1.45 0.86 (0.76, 0.97) 0.02 0.02 0.04
hsa-miR-365a-3p 0.00 2.31 0.82 (0.68, 0.97) 0.02 0.02 0.048
hsa-miR-501–3p 0.57 1.73 0.81 (0.69, 0.94) 0.01 0.01 0.03
hsa-miR-92a-3p 0.78 1.86 0.78 (0.67, 0.91) 0.003 0.003 0.01
hsa-miR-93–5p 0.88 1.98 0.85 (0.76, 0.96) 0.01 0.01 0.03
Open in a new tab
1

Hazard Ratios (HR) and 95% Confidence Intervals (CI) adjusted for age, sex, and AJCC. HR calculated as the difference between the lower and u`pper quartile

Discussion

We found that miRNAs mainly influenced the p53-signaling pathway through associations with genes downstream of TP53. Downstream genes from TP53 that lead to cell cycle arrest (CCND1, CDK4, CCNB1, CKD1, and GTSE1), apoptosis (TNFRSF10B, PERP, and IGF1), inhibition of angiogenesis (THBS1), DNA repair and angiogenesis (RRM2), and exosome mediated secretion (STEAP3) were associated with miRNAs. Several miRNAs associated with these genes also were associated with CRC-specific survival, especially after being diagnosed with rectal cancer. Taken together, these findings provide further insight into how miRNAs work in the p53-signaling pathway to alter CRC risk.

A validation of these findings of the identified interactions between the mRNAs and the miRNAs is beyond the scope of this study; however, many of the association detected have face validity in that miRNA expression targets specific genes. In examining dysregulated genes in the pathway, expression of TP53 is up-regulated. MDM2, one of the main regulators of TP53, is up-regulated in tumors relative to normal mucosa, possibly because of the feedback loop causes MDM2 levels to increase when TP53 is up-regulated. MDMX (also known as MDM4), on the other hand, binds to MDM2 to inhibit p53 activity and is not regulated by p53 directly. Since p53 is activated in response to cellular stress, the MDM2 and MDMX complex must be inhibited (18). In our data, MDMX was down-regulated which could result in a less efficient MDM2/MDMX complex to inactivate p53. As we stated, having TP53 up-regulated could result in increased expression of down-stream target genes (TG) of TP53. We observed that several of these TGs were up-regulated, possibly in response to TP53 up-regulation. Interestingly, IGFBP3 which is affected by TP53 was up-regulated, which in turn could block IGF1. In our data, IGFBP3 was up-regulated while IGF1 was down-regulated in CRC tumors.

Interpretation of miRNA:mRNA interactions is complicated and can be facilitated by examining seed-region matches between the miRNA and the 3’UTR of the mRNA; this greater propensity for binding may increase the likelihood that miRNAs directly alter gene expression. However, indirect associations between miRNAs and mRNAs, indicated by a positive beta coefficient with a seed-region match and/or no identified seed matches, exists. Indirect effects most likely operate through a feed-forward loop (19–21). In feed-forward loops, regulators such as miRNAs can have either the same effect (repression of expression) or opposite effects on each other (20). In feed-forward loops, a transcription factor (TF) such as TP53 can regulate the miRNA and the TG, which in turn is regulated by the miRNA. In this instance, the miRNA may regulate the TG directly, through seed region binding leading to mRNA degradation or translational repression, or indirectly, through repression of the TF that is influencing transcription of the same TG. Studies suggest that regulatory pathways involving miRNAs are prevalent mechanisms of altering gene expression (20).

Seed-region matches are an important component of identifying target sites for binding between miRNA and mRNA and are the basis of several computational programs used to identify target sites such as TarPmiR, miRanda, and the program we used (15, 22, 23). However, other factors are important in identifying target sites such as binding at energy-based sites. Ding and colleagues used energy based sites as the second set of criteria for identifying candidate target sites after seed matching which they considered their first line of candidate site identification (23). In our analysis, we used seed matching to provide additional information on our miRNA and mRNA associations and taken together with a negative beta coefficient between the two, provide additional support for their direct binding. It has been shown that as much as 16% of miRNA:mRNA interactions may not involve contacts within seed regions; these interactions are considered seedless interactions (24). Because we only looked at seed matching, we did not consider seedless interactions. We did however observe that 10 miRNAs had a negative beta coefficient with a mRNA that did not have a seed match. The expression data suggest that increase of miRNA expression in tumors resulted in a decreased expression of mRNA in tumors. These could have been false negatives in our seed region analysis and possibly the result of seedless interactions. Based on what we know about seedless interactions, most likely one or two of the 10 was missed by only using seed region matches (24). Functionality test of the interaction for the 14 miRNA:mRNA associations with a negative beta coefficient is a logical step in the progression of understanding these interactions.

In our data, cell cycle arrest was one of the major cellular responses that contained dysregulated genes associated with dysregulated miRNAs downstream from TP53. Up-regulated genes, GTSE1, CCND1, CKD4, CCNB1, and CKD1, were associated with miRNAs that were up-regulated. Only miR-145–5p (associated with CCNB1 and GTSE1) and miR-195–5p (associated with CCNB1) were down-regulated. This suggest that most of these associations are through feedback or feedforward loops. Some of the miRNAs associated with these genes have been previously associated with downstream target genes of the p53-signaling pathway. Both miR-34a and miR-145 have been associated with cell-cycle related factors such as CDK4 and Cyclin E2 and miR-15a and miR-16–1have been showed to be up-regulated after p53 activation (25, 26). We did not observe miR-34a to be associated with any of these genes, while miR-145 was down-regulated and associated with two genes with seed-region matches, both of which were up-regulated. Neither miR-15a nor miR-16–1 were associated with any of the genes analyzed that were reported to be involved in cell cycle arrest. Differences in findings between this study and the literature could be from many sources, including number miRNAs assessed. Given our large number of targeted mRNA and miRNAs our level of adjustment for multiple comparisons could influence associations detected. We also used a cutoff for meaningful level of differential expression that could influence findings. Additionally, few other studies have conducted similar types in research, but have often relied on cell lines and targeted miRNAs that could easily influence findings.

Many of the downstream dysregulated genes in the p53-signaling pathway, were intermediaries to the apoptosis process. Among these, PERP (TP53 apoptosis effector or p53 apoptosis effector related to PMP-22), TNFRSF10B (Tumor necrosis factor receptor superfamily, member 10b), and IGF1 (Insulin-like growth factor 1) also were associated with miRNAs. PERP, a documented target gene of p53 (27, 28) and thus thought to be an important regulator of apoptosis in response to p53 activation (27) was up-regulated in our data. PERP was associated with 24 miRNAs, however, only miR-650 had both a seed-region match and a negative beta coefficient suggesting a greater likelihood of direct biological effect. It is thus likely that the other miRNAs associated with PERP, are working indirectly with PERP through feedback loops that could ultimately influence apoptosis. TNFRSF10B (also called DR5 or death receptor 5) was associated with two miRNA, one of which miR-196b-5p, had both a seed-region match and a negative beta coefficient between the differential expression of the mRNA and the miRNA. Part of the TNF-receptor superfamily, TNFRSF10B comprises part of the death receptor domain that leads to apoptosis. IGF1 was down-regulated in our data; however IGFBP3 was up-regulated. IGFBP3 blocks IGF1, which could contribute to its down-regulation. IGF1 has been shown to inhibit apoptosis (29); down-regulation of IGF1 would thus promote apoptosis, possibly as a result of p53 activation which could be responsible for the increased expression of IGFBP3 as well. IGF1 was associated with miR-145–5p with a seed-region match, although both IGF1 and miR-145–5p were down-regulated suggesting an indirect effect of the miRNA on the gene, possibly through a feedback loop. Several miRNAs have previously been associated with p53 through its role in apoptosis. MiR-34 has been cited widely as associated with down-stream gene expression that promotes apoptosis (30–33). Other miRNAs, such as miR-145–5p, miR-192, miR-194, miR-215, miR-125b, miR-504, miR-25, and miR-30d (5, 34, 35), also have been associated with the p53-signaling pathway through their association with downstream targets of p53. In our data, miR-34a-5p (without a seed-region match) was only associated with PERP, miR-145–5p was associated with CCNB1, GTSE1 (without an identified seed-region match), and IGF1 (with an identified seed-region match), and miR-25 (without an identified seed-region match) was associated with CCNB1 and CDK1 in regards to cell cycle arrest and apoptosis.

Other genes that were downstream of p53 and associated with the p53-signaling pathway through their role in DNA repair and angiogenesis, i.e RRM2 (ribonucleotide reductase M2) was up-regulated and THBS1 (thrombospondin-1 also called TSP-1), was down-regulated and associated with miRNAs. Overexpression of RRM2 has been shown to enhance the metastatic potential of tumors, thus being involved in tumor growth and angiogenesis (36). THBS1 is an endogenous inhibitor of angiogenesis and low THBS1 expression may be an indicator of metastases and poor prognosis in colorectal cancer (37); high levels of RRM2 have been shown to inhibit THBS1 (36). We found that differential expression of RRM2 was up-regulated while THBS1 was down-regulated. RRM2 was associated with more miRNAs (N=14) than any gene other than PERP. Of the miRNAs associated with RRM2, both miR-150–5p and miR-650 had a seed-region match and the differential expression of RRM2 was inversely related to the differential expression of the miRNAs (i.e. negative beta coefficient), suggesting that these miRNAs influence RRM2 expression. THBS1 was down-regulated and had a seed-region match with miR-145–5p, which as noted earlier is one of the miRNAs associated with p53 target genes. These conditions would lead to a promotion of angiogenesis and metastasis. MiR-150–5p also was associated with slightly better survival when the rectal tumor expression was increased. In our data, increased expression of RRM2 was associated with less expression of miR-150–5p in tumors.

We have previously reported that TP53 expression levels did not differ significantly by TP53 mutational status of tumors (38, 39). However, we have previously shown that 26 miRNAs were significantly differentially expressed between TP53-mutated vs TP53-wildtype tumors (40). Others have suggested that dysregulated or mutated TP53, influences downstream miRNAs and mRNAs (30). Our data would support these previous findings, in that no significant associations were observed between differential TP53 expression and differential miRNA expression after adjustment for multiple comparisons. However, genes downstream of TP53 that demonstrated dysregulation simultaneously with TP53 up-regulation, were associated with miRNAs. These results taken together suggest that TP53 is up-regulated in CRC tissue relative to normal mucosa, most likely as a result of various stress factors contributing to tumor initiation. Up-regulated TP53 in turn targets down-stream genes that in conjunction with miRNAs, favor apoptosis or cell cycle arrest. However, the activated system could also be considered as ineffective in its goal of cell cycle arrest and apoptosis, given that approximately half of these tumors have a mutated TP53 gene.

There was limited statistical power to evaluate CRC-specific survival with dysregulated genes. However, evaluation of miRNAs with survival enabled us to use a much larger sample of CRC cases and therefore evaluate associations with all CRC cases combined as well as for colon and rectal cancer cases specifically. Since these miRNAs were associated with several genes within the pathway, these analyses provided some insight into possible impact on prognosis. Of the miRNAs associated with survival, miR-150–5p appeared to regulate expression of RRM2, and miR-196b-5p appeared to regulate TNFRSF10B expression. MiR-150–5p and miR-196b-5p were both associated significantly with survival after diagnosis with rectal cancer, whereas increased differential expression in the tumor improved survival. Other miRNAs that were associated with a given mRNA is plausibly in a feedback loop, also influenced survival, with those being diagnosed with rectal cancer being most affected by these associations. As we have previously reported, differential expression of miRNAs appears to have the greatest impact on those diagnosed with rectal cancer (41). It is unclear why this occurs, but it could possibly be that rectal tumors are often radiated prior to surgery and that this radiation may somehow influence the response to miRNAs that have altered expression.

Our study had both strengths and weaknesses. While our sample of paired carcinoma and normal mucosa data is one of largest available, it is still small and limited some of our analysis, such as examining survival. The normal colonic mucosa is the closest normal tissue available for a matched-paired analysis, although non-carcinoma mucosa may have undergone changes from healthy colonic mucosa. Thus, while there is caution when using the normal samples, it is the best option available to determine mRNA and miRNA expression in non-cancer CRC tissue. The normal colonic mucosa was taken from the same colonic site as the tumor to prevent differences in expression being the result of tumor location. A general limitation in studies such as this, is the lack of a set cut point or FC at which differential mRNA or miRNA expression is deemed meaningful. In this study, we focused on those genes that were statistically significant after adjustment for multiple comparisons that also had a FC of greater than 1.50 or less than 0.67. Thus by restricting the genes we evaluated with miRNAs, we could have missed mRNA:miRNA associations. We exclusively used the KEGG pathway database to identify p53-signaling pathway genes; other genes related to the pathway but not included in KEGG may also be important. Since miRNAs have their impact post-transcriptionally, a study limitation is availability of gene expression data but not protein expression data. However, much of the current information on miRNA target genes comes from gene expression data, and while associations could have been missed, those identified may have important biological meaning (16, 42). While it is beyond the scope of this study, it is important that other studies conduct more detailed functionality assessment between mRNA and miRNA associations observed.

Conclusions

In summary, these data support previous findings that downstream target genes are influenced by activated p53. These downstream target genes interact with miRNA, both direct and through feedback and feed-forward loops, to possibly influence cell cycle arrest, apoptosis, and angiogenesis. The face validity of our findings support the role of miRNAs in these cell processes within the p53-signaling pathway; further laboratory-based validation of these results may lead to an improved understanding of the p53-signaling pathway as well as identify potential therapeutic targets for CRC.

Supplementary Material

Supplemental TAbles 1 to 5

Acknowledgements

The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the National Cancer Institute. We acknowledge Sandra Edwards for data oversight and study management, and Michael Hoffman and Erica Wolff for miRNA analysis. We acknowledge Dr. Bette Caan and the staff at Kaiser Permanente Northern California for sample and data collection.

Funding:

This study was supported by NCI grant CA163683.

List of abbreviations

CRC

colorectal cancer

FDR

false discovery rate

FC

fold change

KPNC

Kaiser Permanente of Northern California

KEGG

Kyoto Encyclopedia of Genes and Genomics

HR

Hazard Ratio

miRNA

microRNA

MSI

microsatellite instability

MSS

microsatellite stable

QC

quality control

SEER

Surveillance Epidemiology and End Results

TG

target gene

TF

transcription factor

UTR

untranslated region

Footnotes

Declarations

Ethics approval and consent to participate:

Study participants signed informed consent prior to study participation. The Institutional Review Boards at the University of Utah and at KPNC approved this study.

Consent for publication:

This manuscript does not contain individual data presented in a manner that could identify them. All data are presented in aggregate.

Availability of data and materials:

Data can be released in conjunction with study participant signed consent form. The datasets analyzed during the current study are not publicly available given the restrictions arising from the signed consent. To discuss feasibility of obtaining data contact Dr. Slattery.

Competing Interests:

The authors declare that they have no competing interests.

Author’ Contributions:

MLS oversaw all aspects of study, obtained funding, wrote paper; LEM conducted bioinformatics analysis that included seed-region matches between mRNA and miRNA and helped write the manuscript; RKW oversaw laboratory genomic analysis; LCS contributed data and assisted in editing and writing the manuscript; WS provided pathology input; JSH managed data and conducted statistical analysis. All authors read and approved final manuscript.

References

  • 1.Slattery ML, Curtin K, Wolff RK, Boucher KM, Sweeney C, Edwards S, et al. A comparison of colon and rectal somatic DNA alterations. Dis Colon Rectum. 2009;52(7):1304–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Mandinova A, Lee SW. The p53 pathway as a target in cancer therapeutics: obstacles and promise. Sci Transl Med. 2011;3(64):64rv1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Harris SL, Levine AJ. The p53 pathway: positive and negative feedback loops. Oncogene. 2005;24(17):2899–908. [DOI] [PubMed] [Google Scholar]
  • 4.Paroo Z, Ye X, Chen S, Liu Q. Phosphorylation of the human microRNA-generating complex mediates MAPK/Erk signaling. Cell. 2009;139(1):112–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kumar M, Lu Z, Takwi AA, Chen W, Callander NS, Ramos KS, et al. Negative regulation of the tumor suppressor p53 gene by microRNAs. Oncogene. 2011;30(7):843–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Slattery ML, Potter J, Caan B, Edwards S, Coates A, Ma KN, et al. Energy balance and colon cancer--beyond physical activity. Cancer research. 1997;57(1):75–80. [PubMed] [Google Scholar]
  • 7.Slattery ML, Caan BJ, Benson J, Murtaugh M. Energy balance and rectal cancer: an evaluation of energy intake, energy expenditure, and body mass index. Nutr Cancer. 2003;46(2):166–71. [DOI] [PubMed] [Google Scholar]
  • 8.Slattery ML, Herrick JS, Pellatt DF, Stevens JR, Mullany LE, Wolff E, et al. MicroRNA profiles in colorectal carcinomas, adenomas and normal colonic mucosa: variations in miRNA expression and disease progression. Carcinogenesis. 2016;37(3):245–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Slattery ML, Herrick JS, Mullany LE, Valeri N, Stevens J, Caan BJ, et al. An evaluation and replication of miRNAs with disease stage and colorectal cancer-specific mortality. Int J Cancer. 2015;137(2):428–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Pellatt AJ, Slattery ML, Mullany LE, Wolff RK, Pellatt DF. Dietary intake alters gene expression in colon tissue: possible underlying mechanism for the influence of diet on disease. Pharmacogenet Genomics. 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Slattery ML, Pellatt DF, Mullany LE, Wolff RK, Herrick JS. Gene expression in colon cancer: A focus on tumor site and molecular phenotype. Genes, chromosomes & cancer. 2015;54(9):527–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Pellatt DF, Stevens JR, Wolff RK, Mullany LE, Herrick JS, Samowitz W, et al. Expression Profiles of miRNA Subsets Distinguish Human Colorectal Carcinoma and Normal Colonic Mucosa. Clin Transl Gastroenterol. 2016;7:e152. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Agilent Technologies I. Agilent GeneSpring User Manual Santa Clara, CA: Aglient Technologies Inc; 2013. [ [Google Scholar]
  • 14.Benjamini Y H Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. 1995;57:289–300. [Google Scholar]
  • 15.Mullany LE, Herrick JS, Wolff RK, Slattery ML. MicroRNA Seed Region Length Impact on Target Messenger RNA Expression and Survival in Colorectal Cancer. PLoS One. 2016;11(4):e0154177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chou CH, Chang NW, Shrestha S, Hsu SD, Lin YL, Lee WH, et al. miRTarBase 2016: updates to the experimentally validated miRNA-target interactions database. Nucleic Acids Res. 2016;44(D1):D239-47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, et al. The UCSC Table Browser data retrieval tool. Nucleic Acids Res. 2004;32(Database issue):D493-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Shadfan M, Lopez-Pajares V, Yuan ZM. MDM2 and MDMX: Alone and together in regulation of p53. Transl Cancer Res. 2012;1(2):88–9. [PMC free article] [PubMed] [Google Scholar]
  • 19.Lin Y, Zhang Q, Zhang HM, Liu W, Liu CJ, Li Q, et al. Transcription factor and miRNA co-regulatory network reveals shared and specific regulators in the development of B cell and T cell. Sci Rep. 2015;5:15215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Martinez NJ, Walhout AJ. The interplay between transcription factors and microRNAs in genome-scale regulatory networks. Bioessays. 2009;31(4):435–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Mangan S, Alon U. Structure and function of the feed-forward loop network motif. Proc Natl Acad Sci U S A. 2003;100(21):11980–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS. Human MicroRNA targets. PLoS Biol. 2004;2(11):e363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ding J, Li X, Hu H. TarPmiR: a new approach for microRNA target site prediction. Bioinformatics. 2016;32(18):2768–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Helwak A, Kudla G, Dudnakova T, Tollervey D. Mapping the human miRNA interactome by CLASH reveals frequent noncanonical binding. Cell. 2013;153(3):654–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Tarasov V, Jung P, Verdoodt B, Lodygin D, Epanchintsev A, Menssen A, et al. Differential regulation of microRNAs by p53 revealed by massively parallel sequencing: miR-34a is a p53 target that induces apoptosis and G1-arrest. Cell Cycle. 2007;6(13):1586–93. [DOI] [PubMed] [Google Scholar]
  • 26.Hermeking H MicroRNAs in the p53 network: micromanagement of tumour suppression. Nat Rev Cancer. 2012;12(9):613–26. [DOI] [PubMed] [Google Scholar]
  • 27.Attardi LD, Reczek EE, Cosmas C, Demicco EG, McCurrach ME, Lowe SW, et al. PERP, an apoptosis-associated target of p53, is a novel member of the PMP-22/gas3 family. Genes Dev. 2000;14(6):704–18. [PMC free article] [PubMed] [Google Scholar]
  • 28.Ihrie RA, Reczek E, Horner JS, Khachatrian L, Sage J, Jacks T, et al. Perp is a mediator of p53-dependent apoptosis in diverse cell types. Curr Biol. 2003;13(22):1985–90. [DOI] [PubMed] [Google Scholar]
  • 29.Kooijman R Regulation of apoptosis by insulin-like growth factor (IGF)-I. Cytokine Growth Factor Rev. 2006;17(4):305–23. [DOI] [PubMed] [Google Scholar]
  • 30.Chang TC, Wentzel EA, Kent OA, Ramachandran K, Mullendore M, Lee KH, et al. Transactivation of miR-34a by p53 broadly influences gene expression and promotes apoptosis. Mol Cell. 2007;26(5):745–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.He X, He L, Hannon GJ. The guardian’s little helper: microRNAs in the p53 tumor suppressor network. Cancer Res. 2007;67(23):11099–101. [DOI] [PubMed] [Google Scholar]
  • 32.Hermeking H p53 enters the microRNA world. Cancer Cell. 2007;12(5):414–8. [DOI] [PubMed] [Google Scholar]
  • 33.Corney DC, Flesken-Nikitin A, Godwin AK, Wang W, Nikitin AY. MicroRNA-34b and MicroRNA-34c are targets of p53 and cooperate in control of cell proliferation and adhesion-independent growth. Cancer Res. 2007;67(18):8433–8. [DOI] [PubMed] [Google Scholar]
  • 34.Liu J, Zhang C, Zhao Y, Feng Z. MicroRNA Control of p53. J Cell Biochem. 2017;118(1):7–14. [DOI] [PubMed] [Google Scholar]
  • 35.Pichiorri F, Suh SS, Rocci A, De Luca L, Taccioli C, Santhanam R, et al. Downregulation of p53-inducible microRNAs 192, 194, and 215 impairs the p53/MDM2 autoregulatory loop in multiple myeloma development. Cancer Cell. 2010;18(4):367–81. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  • 36.Zhang K, Hu S, Wu J, Chen L, Lu J, Wang X, et al. Overexpression of RRM2 decreases thrombspondin-1 and increases VEGF production in human cancer cells in vitro and in vivo: implication of RRM2 in angiogenesis. Mol Cancer. 2009;8:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Teraoku H, Morine Y, Ikemoto T, Saito Y, Yamada S, Yoshikawa M, et al. Role of thrombospondin-1 expression in colorectal liver metastasis and its molecular mechanism. J Hepatobiliary Pancreat Sci. 2016;23(9):565–73. [DOI] [PubMed] [Google Scholar]
  • 38.Slattery ML, Herrick JS, Mullany LE, Samowitz WS, Sevens JR, Sakoda L, et al. The co-regulatory networks of tumor suppressor genes, oncogenes, and miRNAs in colorectal cancer. Genes Chromosomes Cancer. 2017;56(11):769–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Slattery ML, Mullany LE, Sakoda LC, Samowitz WS, Wolff RK, Stevens JR, et al. Expression of Wnt-signaling pathway genes and their associations with miRNAs in colorectal cancer. Oncotarget. 2018;9(5):6075–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Slattery ML, Herrick JS, Mullany LE, Wolff E, Hoffman MD, Pellatt DF, et al. Colorectal tumor molecular phenotype and miRNA: expression profiles and prognosis. Mod Pathol. 2016;29(8):915–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Slattery ML, Herrick JS, Pellatt DF, Mullany LE, Stevens JR, Wolff E, et al. Site-specific associations between miRNA expression and survival in colorectal cancer cases. Oncotarget. 2016;7(37):60193–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Slattery ML, Herrick JS, Stevens JR, Wolff RK, Mullany LE. An Assessment of Database-Validated microRNA Target Genes in Normal Colonic Mucosa: Implications for Pathway Analysis. Cancer Inform. 2017;16:1176935117716405. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental TAbles 1 to 5
NIHMS977307-supplement-Supplemental_TAbles_1_to_5.docx (39.7KB, docx)

RESOURCES