Journal Name
Dynamic Article Links ►
Cite this: DOI: 10.1039/c0xx00000x
ARTICLE TYPE
www.rsc.org/xxxxxx
Induced genome maintenance pathways in pre-cancer tissues describe an anti-cancer barrier in tumor development Éder Maiquel Simão,a Marialva Sinigaglia,b Cristhian Augusto Bugs,ac Mauro Antonio Alves Castro,d Giovani Rubert Librelotto,e Ronnie Alves, f and José Carlos Merino Mombach*a 5
10
15
Received (in XXX, XXX) Xth XXXXXXXXX 20XX, Accepted Xth XXXXXXXXX 20XX DOI: 10.1039/b000000x A recent model proposing that a barrier is raised against tumor evolution in pre-cancer tissues is investigated. For that we quantify expression alterations in genome maintenance pathways: DNA damage response, death pathways and cell cycle and also differentially expressed genes in transcriptomes of precancerous and cancerous lesions deposited in the GEO database. We find that the main alterations in precancer samples comprising the barrier are: (1) DNA double strand-breaks signaling and repair pathways induction, (2) upregulation of cyclin-dependent kinases, (3) p53 dependent (and independent) repair and apoptosis pathways induction and (4) replicative senescence induction early in tissue transformation. In the cancer samples we find that the induced pathways in pre-cancer are systematically inhibited and the only remaining induced pathway is p53, whereas the retinoblastoma pathway arises induced in most samples. The results give support to the model, furthermore they reveal the involvement of additional mechanisms in pre-cancer, including the early induction of replicative senescence and of p53 independent apoptosis.
Introduction 20
25
30
35
40
45
A major goal in biology is the description of tissue phenotype in terms of molecular pathways behavior, especially in cancer research where the molecular alterations are complex. Cancer is a genetic disorder that involves multiple steps characterized by the accumulation of slow and gradual changes in genes involved in key processes for the maintenance of a normal cell, e.g., the cell cycle, apoptosis and repair. Within this context, we might characterize tumor progression as follows: starting with a normal epithelium, through pre-malignant lesions (adenomas) and culminating in malignant lesions (carcinomas) that invade surrounding tissues and may also acquire the ability to spread to other parts of the body (metastasis). Thus, the study of pre-cancer tissues represent a useful model to investigate genetic changes involved in tumor progression, mostly because they originate and proceed through a series of well-defined standard histological changes. In the early stages, cancer is associated with stress in DNA replication which leads to DNA double-strand breaks (DSBs) and genomic instability1. The genome maintenance mechanisms (GMM) ensure the integrity of DNA, they involve pathways that present specific functions in DNA damage response and repair, programmed cell death2 and senescence3,4,5. Halazonetis and coworkers3 proposed a qualitative model based on studies of precancerous lesions on how these tissues resist to tumor evolution. According to their model, precancerous lesions have increased DNA damage response (DDR), apoptosis and senescence activity which play the role of a barrier to tumor progression 2,4. They This journal is © The Royal Society of Chemistry [year]
50
55
60
65
suggest that oncogenes induce a stress in DNA replication leading to the formation of double-strand breaks (DSBs) and consequent activation of TP53 that induces apoptosis and senescence. They speculate that the origin of replicative stress is due to the deregulation of cell cycle by oncogenes increasing the activity of cyclin-dependent kinases (CDKs) that affect the checkpoints G1 and S6. The constant formation of DSBs associated with genomic instability leads to the accumulation of mutations that eventually breach the barrier and the tissue evolves towards cancer. Our main aim in this work is to describe the common mechanisms and the genes involved in the barrier in pre-cancer tissues using publicly available transcriptome data where we focused on cell cycle, DDR and cell death pathways that are important in genome maintenance and cancer development. We determined common pathways alterations in these tissues using two statistical methods: (1) directly, using a pathway analysis introduced by our group3,7 and (2) through consensuses of upregulated genes6,9 among adenomas and among cancers.
Methods Selection of cellular pathways 70
Here we will use the word subpathway to refer to a given pathway with a specific function in one of the three main cellular pathways implicated in the development of cancer: DDR, [journal], [year], [vol], 00–00 | 1
5
10
apoptosis and cell cycle (see Table 1). We used as the main reference databases the National Cancer Institute (NCI) - Nature Interaction Pathway Database, Reactome, BioCarta, and Ontocancro. The Replicative senescence subpathway was developed by us based on the scientific literature. The NF-kB activation and RB/E2F pathways were excerpted from the works of Sabatel10 and Calzone11. In Table 1 we list the 3 main sets of pathways with their respective subpathways. This information is also available in the supplementary files† with the genes within each subpathway. Information about subpathways and genes can also be obtained from the Ontocancro 2.0 database (http://www.ontocancro.inf.ufsm.br).
55
2- The relative diversity (hα) measures alterations in a subpathway without necessarily a change in its relative activity in relation to a control 3. For the calculation of diversity we used the equation: −1 Hα = ln Mα
Mα
p i, α ln p(i, α) I
60
where Mα is the number of genes in the pathway , p(i,α) is the frequency of the activity of gene i in pathway α. p(i,α) is defined as p i, α =
15
Microarray Data Selection
20
For this study we selected transcriptome data of adenomas and carcinomas having control samples (normal tissues) using the Affymetrix Human Genome U133 Plus platforms GPL57012 and GPL96. The four studies were selected from Gene Expression Omnibus database (GEO) and normalized by the Robust Multichip Average (RMA)13 available in the affy R package. The samples used correspond to the following GEO series:
s i, α , Nα
65
25
30
1. GSE4183 of colorectal: 8 samples of colorectal normal, 15 samples of inflammatory bowel diseases, 15 samples of precancerous (adenoma) and 15 samples of colorectal carcinomas; 2. GSE10927 of adrenocortical (adrenal): 10 samples of adrenocortical cortex normal, 22 samples of adrenocortical adenomas and 33 samples of adrenocortical carcinomas;
where s(i,α) is the sum of the activities of gene i and N α is the total activity of the genes in pathway 𝛼. Analogously, the relative diversity is defined as hα = 70
75
3. GSE19650 of pancreas: 7 samples of main pancreatic duct normal, 6 samples of intraductal papillary-mucinous adenoma e 6 samples of intraductal papillary-mucinous carcinoma of the pancreas; 80 35
40
4. GSE27155 of follicular thyroid: 4 samples of thyroid normal, 17 samples of follicular thyroid adenoma, 13 samples of follicular thyroid carcinoma. Additional information about these studies is available in the supplementary files†. The data samples are also available in the Ontocancro 2.0 database.
85
ViaComplex 1.0 and Ontocancro 2.0 Tools 90 45
We used two measures of the alteration in a subpathway: 1- The relative activity (nα) measures an increase or decrease in the cumulative activity of the genes in a subpathway a in relation to a control (normal tissue):
50
nα = Nαe
Nαe
Hαe
Hαe γ, + Hα
γ
where Hαe and Hα are the diversities of sample and its control, respectively. Similarly to the relative activity, the value of h α γ varies between 0 < hα < 1. If hα < 0.5, then Hαe < Hα , i.e., the relative diversity of the pathway in the sample is less than its control, while hα > 0.5 is the opposite case. To determine whether an alteration in relative activity or diversity in a subpathway is statistically significant we use a random resampling (bootstrap) with 10,000 repetitions involving all genes for each transcriptome for each subpathway studied. The confidence level used in the identification of alterations is 5%. For the analysis of alterations in subpathways we used the ViaComplex software7 used to build functional landscapes of gene expression. It can be freely downloaded from: http://www.lief.if.ufrgs.br/pub/biosoftwares/viacomplex. The Ontocancro 2.0 is a database that compiles gene pathways involved in genome maintenance mechanisms comprising several pathways and subpathways of cell cycle, DNA repair mechanisms, apoptosis and chromosome stability from NCINature, BioCarta, KEGG and Reactome. In the database the data used in this paper is normalized with MAS 5.0 Call P/M/A12. The calculations of relative activity and diversity (in construction) can be obtained at the database and complete information about pathways and genes used in this paper can be found and downloaded from: http://www.ontocancro.inf.ufsm.br.
95
Genes fold change analysis
γ
Nαe + Nα
γ Nα
where is a sample and its control sample. The value of nα γ varies between 0 < nα < 1. If nα < 0.5, then Nαe < Nα , i.e., the relative activity of the pathway in the sample is less than in its control, while nα > 0.5 is the opposite case. 2 | Journal Name, [year], [vol], 00–00
100
All microarray data used was normalized with Robust Multichip Average (RMA)13. To complement the analysis of the pathways alterations, we calculated the fold change of each gene in the subpathways with altered relative activity and diversity. In a matrix, we grouped by arithmetic mean the sets of samples (in each of the tissues) of normal tissue (control), inflammation This journal is © The Royal Society of Chemistry [year]
5
10
(colorectal), cancer and adenoma (see Microarray data selection section). From the average of the samples we calculated the fold change of each gene by the ratio between the experimental and control tissue8: inflammation vs normal, adenoma vs normal and cancer vs normal. Finally, we selected the top-5 upregulated genes on each comparison for further biological analysis. Typically, for these samples, the fold changes ranged from 1.3 to 1.5 (see Supplementary Material 1). All analyses were performed using R programming language through Bioconductor packages such as limma, biobase, annaffy and hguplus2.db available in http://bioconductor.org/ (See Script in the Supplementary Material 1).
Consensuses of differentially expressed genes9
60
65
70 15
Results Pathways alterations
20
25
30
35
40
45
50
55
Fig. 1 shows the alterations in pathways of DNA damage response (DDR), apoptosis and cell cycle obtained with the calculation of relative activity (n α) and relative diversity (hα) (see ‘Methods’ section) for the following transcriptomes: colorectal inflammation; colorectal, adrenocortical, pancreatic and thyroid adenomas; colorectal, adrenocortical, pancreatic and thyroid cancers. In each of the altered subpathways marked in red to increase and blue to decrease, we found the top-5 upregulated genes using the fold change analysis (see ‘Genes fold change analysis’ section). In a first global analysis of Fig. 1 we observed that the DDR and apoptosis pathways have more altered subpathways in adenomas than in cancer, while the opposite seems to occur with the cell cycle pathway. The colorectal adenoma sample has a high number of altered subpathways, while no significant alterations were found in adrenocortical adenoma. In cancer we found fewer altered subpathways. Mainly, we found alterations of colorectal inflammation only in the apoptosis pathway and a consistent alteration in the Replicative senescence subpathway for almost all samples, what is examined more detailed at the end of the Discussion section. In the cell cycle pathway there is no clear alteration pattern and it seems to depend on tissue specific features. The method did not identify significant alterations in pathways for some tissues as seen in Fig. 1, they are the DDR pathway for adrenocortical adenoma and adrenocortical and thyroid cancers, and in the Apoptosis pathway for pancreatic cancer and in the Cell cycle pathway for pancreatic adenoma. For these tissues we used the differential expression analysis in association with the consensus analysis to determine the altered pathways. For these tissues only we determine the top-5 upregulated genes among all genes in the pathway and used them to infer which of its subpathways could be possibly altered corresponding to the subpathways most often altered in the other tissues. As a first overall observation, the increased activity in DDR and apoptosis pathways in adenomas agree with the barrier described in reference 4. This journal is © The Royal Society of Chemistry [year]
75
80
85
90
95
100
105
In Fig. 2 we present the Venn diagrams of each consensus of upregulated genes between at least two adenomas or cancers for each altered pathway. Here we will focus on the most relevant genes, additional details of this analysis can be found in the supplementary materials 1 and 2 † accompanying this article. In Fig. 2a we present the differential genes for the DDR pathway in adenomas. Noticeably, most of the genes of the consensus (MRE11A, TP53BP1, BRCA2, LIG4, ATM and XRCC4) belong to the Double-strand break repair and to the Non-homologous end joining repair subpathways. The other genes of the consensus (MSH2, MSH3 and MSH6) belong to the Base excision repair (BER) subpathway. For the DDR pathway in cancer samples (Fig. 2b) the consensus genes are also linked to the Double-strand break repair subpathway: H2AFX, XRCC4, XRCC5 and RAD52. In the consensus of Fig. 2c for the Apoptosis pathway we found that many of the upregulated genes are associated with regulation and activation of apoptosis through the tumor necrosis factor and TNFR2 signaling subpathways we found the upregulated genes CASP8, CASP2, RIPK1 and TANK, RIPK1, TNFRSF1B, respectively, and in the Regulation of apoptosis subpathway the genes PAK2 and DAPK2. Three of the consensus genes belong to the Replicative senescence subpathway: CDK6, CDKN1A and MDM2. Another altered subpathway in adenomas is the Activation of NF-kB subpathway with the upregulated genes NEMO and CHUK. Other upregulated genes are: FAS in Death receptor signaling, Apoptosis - Homo sapiens and Extrinsic Pathway for apoptosis and CFLAR in all adenomas. In cancer (Fig. 2d) the upregulated genes DAPK2 and PAK2 that belong to the Regulation of apoptosis, CDK6, TP53, ATR and H2AFX in the Replicative senescence pathway, and NEMO in the Activation of NF-kB subpathway. Fig. 2e presents the consensus of the Cell cycle pathway in adenomas where the altered subpathways are: the Mitotic phases M-M/G1 with three upregulated genes SMC3, MAD2L1 and CDC7, the G2G2M with WEE1, the RB/E2F subpathway with genes WEE1, EP300 and EDD. TP53 is upregulated in Cell Cycle: G1/S checkpoints and G1/S DNA damage checkpoints subpathways. As the most common feature in cancer is aberrant proliferation, we find the activation of some proteins important for maintaining the cell cycle in cancer, in particular, we find changes in the Rb/E2F pathways containing the genes EP300, CREBBP, APC and TP53 highlighted in the consensus of Figure 2f. In the next section we will propose how these pathways interact and how all that can be linked to the model proposed by Halazonetis4.
110
Discussion In Fig. 3 we propose a diagram of interactions among pathways summarizing and explaining our results for preJournal Name, [year], [vol], 00–00 | 3
5
10
15
20
25
30
35
40
45
50
55
cancer (Fig. 3a) and for cancer (Fig. 3b). It also includes the consensus of the upregulated genes (in red) associated with the barrier. The figure illustrates the chain of mechanisms connecting the detection of DNA damage activation, cell cycle checkpoints, pathways related to apoptosis, replicative senescence and the DSB repair that uses the Homologous recombination (HR) and Non-homologous end joining (NHEJ) subpathways. Fig. 3b describes the scenario in cancer including the genes and pathways that are downregulated (in gray). In what follows, we discuss the diagrams for precancer and cancer and at the end, the results for colorectal inflammation.
60
65
70
Pre-cancer The DSBs trigger a cascade of events responsible for the cell cycle arrest and activation of DDR, apoptosis and senescence. The activation of the double-strand break repair subpathway in DDR (Fig. 1) and the consensus gene MRE11A - belonging to the MRN complex and a sensor of DSBs - triggers the activation of ATM and ATR and their substrates CHEK1 and CHEK2 to coordinate the response to damage14,15. ATM and ATR are also sensors of damage and activators of signaling to checkpoints in response to DNA replication stress4. Fig. 3a shows ATM participating in the signal transduction, in the control of the cell cycle arrest through CHEK2 and TP53, and in the activation of apoptosis by the NF-kB subpathway activation, whereas ATR regulates CHEK116,17,18. The signaling of ATM and ATR is dependent on TP53BP1, the low level of TP53BP1 in human cells reduces the phosphorylation of p53 and of ATM/CHEK2. TP53BP1 is also necessary for the recruitment of ATR and for the activation of NHEJ15. In the diagram is indicated the regulation of ATM and ATR and the activation of NHEJ by TP53BP1. In the pathway of activation of ATM and ATR, TP53 is activated by CHEK1 and CHEK2 forcing the cell cycle arrest. In normal tissues p53 is maintained at low expression levels due to its inhibition by MDM2, however, in the presence of replication stresses, MDM2 is downregulated and TP53 is activated by CHEK2 and CHEK1, forcing the cell cycle arrest and activation of DDR, Apoptosis and the Replicative senescence subpathways19,20. CHEK2 and CHEK1 induce the cell cycle arrest to allow the DNA damage repair, as well as their participation in the regulation of apoptosis or senescence in the case of irreparable damage. Cell cycle arrest can result from the downregulation of CDC25A or from the upregulation of the TP53 by CHK1/CHK2. Both CDC25A and TP53 are responsible for the inactivation of CDK217, required for phosphorylation of Rb and transcription initiation of important genes for DNA replication in the S phase. The cell cycle arrest in G1/S and G2/M checkpoints is controlled mainly by p53 that activates CDKN1A to prevent cell cycle progression. CDK6 is inhibited by CDKN1A to induce cell cycle arrest in the G1/S checkpoint17, 21, 22. The Inhibition of cell cycle in G2/M is regulated by the interaction between CDKN1A and the checkpoint kinase WEE1 4 | Journal Name, [year], [vol], 00–00
75
80
85
90
95
responsible for controlling the transition from the G2/M checkpoint. WEE1 inhibits the activity of CDK's responsible for the cell cycle progress and activates the Replicative senescence subpathway23. As we can see in the figure, the mechanisms of DSB repair by Homologous recombination (HR) and Non-homologous end joining repair (NHEJ) subpathways can also be activated independently by TP53. The other consensus genes involved in the repair are: BRCA2, LIG4 and XRCC4. BRCA2 participates in the HR pathway and is a tumor suppressor involved in the repair of damaged chromosomes, with an important role in repairing errors caused by DSB's24. XRCC4 and LIG4 genes form a complex responsible for the binding of the NHEJ to the DNA25. The activation of apoptosis can proceed from two independent pathways: the NF-kB pathway and TP53. The NF-kB pathway is activated by ATM in response to a high level of DSB's. NEMO and CHUK that belong to this pathway interact with genes of the TNFR1 and TNFR2 signaling pathways to activate CASP826. CASP8 (regulated by FAS) participates in the death-induced signaling complex (DISC) which activates the cascade of proteases of apoptosis27. Remarkably, the only gene found upregulated in all adenomas was CFLAR. CFLAR inhibits the activation of CASP8 and consequently it inhibits apoptosis. Despite this anti-apoptotic action of CFLAR, recent evidences suggest that CFLAR may facilitate apoptosis forming a heterodimer with CASP8 that participates in the cleavage of procaspases 826,28, possibly acting as the main activator of caspases through an alternative pathway independent of FAS. We performed a separated investigation of this gene in two other independent adenoma studies (pituitary and colon) deposited in GEO and it was also differentially expressed in both, suggesting that maybe it could be used as a marker of precancer. Finally, the activation of apoptosis by CASP8 is regulated by PAK229. PAK2 and DAPK2 belong to the Regulation of apoptosis subpathway, these genes are positive regulators of apoptosis and are activated by extracellular stimuli generated by stresses30.
100
Cancer
105
110
115
The diagram of Fig. 3b describes the scenario for cancer samples. In light gray the downregulated genes that were upregulated in adenomas. An overall observation is that nearly all of the mechanisms in cancer were shutdown and that most of the upregulated genes belong to the RB/E2F subpathway that regulates the cell cycle in the transition from the G1 to the S phase11. TP53 arises upregulated in three of the four cancers, colorectal, adrenocortical and thyroid. H2AFX, a cancer marker31, arises in the consensus for apoptosis, it is regulated by ATR and is a sensor of DSBs 32. For the DDR pathway we found XRCC4, LIG4, MRE11A and TP53BP1 upregulated. XRCC4 remains active in cancer and also NEMO which is involved in tumor This journal is © The Royal Society of Chemistry [year]
5
necrosis factor signalling through TNFRSF1B. TNFR signaling remains active and below we will discuss this feature in terms of tissue inflammation. The results show that the cell cycle remains induced but independent on the DDR machinery.
Colorectal inflammation 10
15
20
25
30
35
40
In the analysis of colorectal inflammation, we observed alterations in three apoptosis subpathways: Replicative senescence, Apoptosis - Homo sapiens and Granzyme A mediated apoptosis, see Fig. 1. We can explain these alterations as a consequence of the dynamics of colorectal cell renewal for being involved in inflammatory bowel diseases, since activated apoptosis and epithelial cell proliferation is found in colitis 33 and are involved in the regulation and activation of inflammatory processes, a characteristic of ulcerative colitis (UC) and Crohn's disease (CD)33,34,35,36. The analysis of upregulated genes in the Replicative senescence pathway shows upregulation of CDK6, CDC25A and CDKN1A, indicating the regulation of the senescent state in colorectal inflammation by CDKN1A (see Supplementary Material 2). In the Apoptosis - Homo sapiens subpathway we found TNFRSF10D/B, BIRC3, NTRK1 and BAX upregulated. TNFRSF10D/B encode the receptors TRAILR2 and TRAILR4 whose mutual interaction is capable of inhibiting apoptosis by the activation of NF-kB28 (see Supplementary Material 2). Recent works have used these as targets for anti-cancer drugs37. BIRC3 encodes CIPA2 and is also responsible for the inhibition of apoptosis through the activation of NF-kB, contributing to prevent the activation of CASP8 through RIPK138. The analysis of the Granzyme A mediated apoptosis subpathway yielded four upregulated genes: SET, HMGB2, APEX1 and ANP32A indicating that in the colorectal inflammation the activation of apoptosis is also coordinated by T cells by the pathway Granzyme A, reinforcing the apoptotic activity of cells present in colorectal inflammation39,40. For more details of the fold change analysis results in the colorectal inflammation, see the supplementary material 2 †.
replicative senescence induction in almost all sample lesions and related to tissue inflammation. In (1), our results show that apoptosis is also invoked in pre-cancer in a p53 independent way through the NF-kB pathway and furthermore seems to remain active even in the cancer stage. The activity of NF-kB may have beneficial or detrimental effects in relation to tissue homeostasis, in the tumor suppression it may regulate cell cycle arrest, the senescent state and the signaling of the immune system to remove damaged cells. On the other hand, it can be responsible for the resistance to apoptosis by regulating the anti-apoptotic genes, as well as, participating in the activation of angiogenesis mechanisms and invasion. In (2), it is suggested strong and very early involvement of telomere erosion in the tissue transformation process. Remarkably, the colorectal inflammation and the colorectal adenoma activate different mechanisms, while in the inflamed sample apoptosis and senescence are activated, in pre-cancer samples are the DNA repair and apoptosis that ensure the integrity of the tissue and tumor suppression.
60
65
70
75
a 80
85
90
Departamento de Física, Universidade Federal de Santa Maria, Brazil. Tel: +55-55-3220-9521; E-mail:
[email protected] b Departamento de Genética , Universidade Federal do Rio Grande do Sul, Brazil. c Universidade Federal do Pampa, Campus São Gabriel, Brazil. d Departamento de Bioquímica, Universidade Federal do Rio Grande do Sul, Brazil. e Departamento de Eletrônica e Computação, Universidade Federal de Santa Maria, Santa Maria, Brazil. f Instituto de informática, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil. † Electronic Supplementary Information (ESI) available: This article presents two additional files. See DOI: 10.1039/b000000x/ 1
95
2 3 100
4 5
Conclusions
105
45
50
55
In this work we investigated and compared differentially expressed pathways and genes in pre-cancer and cancerous lesions, our predictions are based on transcription control that is important in cancer, however they should receive additional support from experimental verification. The results, in general, suggest support to the model proposed by Halazonetis4 for corroborating the mechanisms that activate the anti-tumor barrier. Moreover, our study identified two important additional features of the transformation that seem to complement the model: (1) the NF-kB (p53 independent) apoptosis activation, primarily related to the high level of DSBs in DNA and (2) the This journal is © The Royal Society of Chemistry [year]
6
7
M.A.A. Castro, J. Filho, R.J. Dalmolin, M. Sinigaglia, J.C.F. Moreira, J.C.M. Mombach and R.M.C. de Almeida, Bioinformatics, 2009, 11,1468-1479.
8
C. Steinhoff and M. Vingron, Briefings in Bioinformatics, 2006, 7, 166-177. H. Chen, P.C. Boutros, VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R, BMC Bioinformatics, 2011, 12, 35. H. Sabatel, C. Pirlot and J.H. Piette, Biochemical Pharmacology, 2011, 82, 1371-1383. L. Calzone, A. Gelay, A. Zinovyev, F. Radvanyi and E. Barillot, Molecular Systems Byology, 2008, 173, 1-12. Affymetrix, Statistical Algorithms Description Document, 2002, 127. R. A. Irizarry, Biostatistics, 2003, 4, 249-264.
110
9 115
10 11 120
V. G. Gorgoulis, I.V.F. Vassiliou, P. Karakaidos, P. Zacharatos A. Otsinas T. Liloglou, M. Venere, R.A. Ditullio, N.G. Kastrinakis, B. Llevy, D. Kletsas, A. Yoneta, M. Herlyn, C. Kittas and T.D. Halazonetis, Nature, 2005, 434, 907-913. I.N. Lavrik, R. Eils, N. Fricker, C. Pforr, P.H. Kramer, Molecular BioSystems, 2009, 5, 1105-1111. M.A.A. Castro, J.C.M. Mombach, R.M.C. de Almeida and J.C.F. Moreira, Nucleic Acids Res, 2007, 35, 1859-1867. T.D. Halazonetis, V.G. Gorgoulis and J. Bartek, Science, 2008, 319, 1352-55. E.M. Simão, H.C. Cabral, M.A.A. Castro, M. Sinigaglia, J.C.M. Mombach and G.R. Librelotto, Physica A, 2010, 389, 4188-4194. F. al-Mulla, M.S. Bitar, Z. Taqi, O. Rath, W. Kolch, RAF kinase inhibitory protein (RKIP) modulates cell cycle kinetics and motility, Molecular BioSystems, 2010, 7, 928-941.
12 13
Journal Name, [year], [vol], 00–00 | 5
5
10
15
20
25
30
35
40
45
50
14 T.Bohgaki T, M. Boagaki and R. Hakem, Genome Integrity, 2010, 1, 1-15. 15 J. H. Lee, A.A. Goodarzi, P.A. Jeggo and T.T. Paull, The EMBO Journal, 2010, 29, 574-585. 16 M. Boxus and L. Willems, Cells Retrovirology, 2012, 9, 2-4. 17 G. P. Studzinski, X. Wang and M. Danilenko, Cancer Biol Ther, 2010, 9, 253-255 18 A. Salminen and K. Kaarniranta, Cellular Signaling, 2011, 23, 747752. 19 A. Dey, V. Tergaonkar V and D.P. Lane, Nature, 2008, 7, 10311040. 20 L.A. Donehower, Cold Spring Harb Perspect Biol, 2009, 4, 1-17. 21 J..H. Houtgraaf, J. Versmissena, W.J. van der Giessen, Cardiovascular Revascularization Medicine, A concise review of DNA damage checkpoints and repair in mammalian cells, 2006, 7, 165–172. 22 J. Bartek, J. Lukas, DNA damage checkpoints: from initiation to recovery or adaptation, Current Opinion in Cell Biology, 2007, 19, 238– 245. 23 C.S. Sorensen, R.G. Syljuasen, Nucleic Acids Research, 2012, 40, 477-86. 24 J.A. Duncan, J.R. Reeves and T.G. Cooke, Clin Pathol: Mol Pathol, 1998, 51, 237–247. 25 S. Sengupta S and C.C. Harris, Nature, 2005, 6, 44-55. 26 S. Biton and A. Ashkenazi, Cell, 2011, 145, 92-103. 27 C. Falschlehmer, C.H. Emmerich, B. Gerlach and H. Walczak, The International Journal of Biochemistry & Cell Biology, 2007, 39, 1462–1475. 28 G. Jing, K. Yuan, Q. Liang, Y. Sun, X. Mao, J.M. McDonald and Y. Chean, Laboratory Investigation, 2012, 92, 82–90. 29 Y. Hsuuw, T. Kuo, K. Lee, Y. Liu, Y. Huang and C. Lai, Annals of the New York Academy of Sciences, 2009, 1171, 501–508. 30 T.Li, J. Zhang, F.Zhu, W. Wen, T. Zykova, K. Liu, C. Peng, W. Ma, G. Shi, Z. Dong and A.M. Bode, Carcinogenesis, 2009, 32, 659-666. 31 J.H. Kim, S.S. Kim, S.W. Byun, Y.J. Chang, J.S.Kim, J.K. Kim, H.J. Cho, K.W. Lim and E.S. Jung, Journal of Gastroenterology, 2010, 55, 21-25. 32 W. Bonner, C. Redon, J. Dickey, A. Nakamura, O. Sedelnikova, S. Solier and Y. Pommier Y, Nat Rev Cancer, 2008, 8, 957-967. 33 L. Chen, S.M. Park, J.R. Turner and M.E. Peter, Inflamm. Bowel Dis. 16, 1071 - 1076. 34 S. Danese and C. Fiocchi, N. Engl. J. Med., 2011, 365, 1713-25. 35 A. Di Sabatino, P. Biancheri, L. Rovedatti, T.T. MacDonald and G.R. Corazza, Intern. Emerg. Med, 2011, 7, 103-111. 36 S. Kraus and N. Arber, Current Opinion in Pharmacology, 2009, 9, 405–410. 37 A.L. Lee, S.H. Dillon, Y. Wang, S. Pervais, W. Fan, Y.Y. Yang, Synergistic anti-cancer effects via co-delivery of TNF-related apoptosis-inducing ligand (TRAIL/Apo2L) and doxorubicin using micellar nanoparticles, Molecular BioSystems, 2011, 7, 1512-1522. 38 M. Gyrd-Hansen and P. Meier, Nature Reviews Cancer, 2010, 10, 561 – 574. 39 S. Elmore, Toxicol. Pathol, 35, 495 – 516. 40 M. Bots and J. Medema, Journal of Cell Science, 2006, 119, 50115014.
75
80
85
90
95
100
105
cancers, respectively. (c) and (d) are the consensuses of the genes of the apoptosis pathways in adenomas and cancers, respectively. (e) and (f) are the consensuses of the genes of the cell cycle pathway in adenomas and cancers, respectively. In red the genes associated with the barrier.
Figure 3 – Diagram illustrating the mechanisms involved in the barrier against tumor development based on the results in Figures 1 and 2. In red the genes of the consensus analysis. (a) Pre-cancer: The barrier is activated by DNA replication stress pathways and the MRN protein complex that plays a central role in the repair of DSBs. It is a sensor of damage responsible for facilitating the activation of other proteins including ATM and ATR. This response involves the activation of checkpoints, DDR and apoptosis by NF-kB Activation Pathway. In the diagram TP53BP1 regulates ATM and ATR, the NHEJ subpathway and the checkpoint signaling through the regulation of its substrates CHEK1 and CHEK2. The substrate CHEK2 inhibits CDC25A that activates, with TP53, the CDK6 to promote cell cycle arrest. ATR also controls G2/M through of the activation of CHEK1 that controls the cell cycle in response to damage by two pathways: one involving the activation of MDM2 which inhibits TP53, and by direct activation of TP53. TP53 promotes the activation of cell-cycle arrest, DNA repair, replicative senescence and apoptosis. Cell cycle arrest is also activated by CDKN1A (p21) that is induced by TP53 in DDR. CDKN1A binds to CDK6 to block the cell cycle progression, and this process is regulated by WEE1 which is a negative regulator at the entry of mitosis. With the cell cycle arrest, DDR, apoptosis and senescence are activated. The activation of apoptosis can also occur through the NF-kB activation pathway and TNRF1 and TNFR2 signaling that activates CASP8 to initiate the apoptotic events. (b) Cancer: several upregulated genes in pre-cancer (Fig. 3a) are downregulated in cancer and incapacitated to activate the anti-cancer barrier. Most of the upregulated genes belong to RB/E2F pathway that regulates the cell cycle in the transition from the G1 to the S phase. Downregulated genes are shown in light gray.
110
115
120
55
Table 1 - List of subpathways used and the database (or article) from where they were obtained and the total number of genes in each subpathway. 60
Figure 1 – Subpathways expression profiles of microarrays studies of colorectal inflammation, colorectal, adrenocortical, pancreatic and thyroid adenomas and cancers. The relative activity (nα) and diversity (hα) were calculated at 5% significance level. Significant increases and decreases are in red and blue, respectively.
125
130
65
70
Figure 2 – Venn diagrams used to estimate the consensus between microarray probes (genes) in adenomas and cancers with the consensuses of the top 5 fold changes in each subpathway. Differential gene expression was evaluated by fold change analysis (see text) and the results are shown in Venn diagrams. (a) and (b) are the consensuses of the genes of the DDR pathway in adenomas and
6 | Journal Name, [year], [vol], 00–00
135
This journal is © The Royal Society of Chemistry [year]
TABLE 1
15
FIGURE 1
DDR Pathway SUBPATHWAYS
NUMBER OF GENES
DATABASE
Base excision repair
44
Ontocancro
Fanconi anemia pathway
24
Reactome
Homologous recombination
34
Ontocancro
Mismatch repair
28
Ontocancro
Nucleotide excision repair
51
Ontocancro
Non-homologous end joining
14
Ontocancro
Double-strand break repair
22
Reactome
Apoptosis / Senescence Pathway SUBPATHWAYS
NUMBER OF GENES
DATABASE
TNFR1 signaling pathway
13
BioCarta
TNFR2 signaling pathway
10
BioCarta
TNF receptor signaling pathway
40
NCI-Nature
Death receptor signaling
13
Reactome
12
BioCarta
20
BioCarta
Apoptosis - Homo sapiens
83
KEGG
Caspase cascade in apoptosis
52
Ontocancro
Regulation of apoptosis
60
Reactome
Apoptotic signaling in response to DNA damage Induction of apoptosis through dr3 and dr4/5 death receptors
Apoptotic execution phase
51
Reactome
Extrinsic pathway for apoptosis
13
Reactome
Granzyme A mediated apoptosis pathway
13
BioCarta
Replicative senescence
56
Ontocancro
20
(Sabatel et al. 2011)
NF-kB activation pathway
FIGURE 2
Cell Cycle Pathway SUBPATHWAYS
NUMBER OF GENES
DATABASE
Cell cycle: G1/S checkpoints
21
BioCarta
S phase
112
Reactome
Cyclins and cell cycle regulation
15
BioCarta
G1/S DNA damage checkpoints
60
Reactome
G2/M checkpoints
43
Reactome
Mitotic G2-G2/M phases
86
Reactome
Mitotic M-M/G1 phases
178
Reactome
Mitotic spindle checkpoint
19
Reactome
RB/E2F pathway
78
(Calzone et al. 2008)
5
20
10
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 7
FIGURE 3
5
8 | Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]
