NMF decomposes high-dimensional gene expression data into positive metagenes with a local biological representation based on non-negative constraints, which is more natural than representing gene expression profiles. From the PCA results, the identified significant AD genes showed that they are mainly related to immunoreactions, metal proteins, membrane proteins, lipoproteins, neuropeptides, cytoskeleton proteins, binding proteins, ribosomal proteins and phosphoric proteins [ 34 ].
A large number of these genes were related to metal metabolism and inflammation, cell growth, cell cycle, apoptosis, cellular fission and cell repair [ 35 ]. The shortcomings of the NMF method were that the number of significant genes was hard to reduce and that the local characteristics that appeared among the metagenes were not clear in our data.
Two Faces of Evil: Cancer and Neurodegeneration
From the biological analysis of the FastICA results on the same dataset, we found that the significant genes were involved in immunoreactions, metal proteins, membrane proteins, lipoproteins, neuropeptides, cytoskeleton proteins, binding proteins and ribosomal proteins and play prominent roles in AD phenotypes. FastICA also found many oncogenes and phosphoric proteins that were significantly lowly expressed in AD.
Based on the molecular biological analysis of these three types of results, significant gene extraction by ICA was better than those by PCA and NMF at identifying known and novel genes in meaningful biological processes for AD. The comparison indicated that ICA is more efficient at extracting potentially relevant genes from microarray data as well as mapping data that is closer to AD pathogenesis. Moreover, all three of these methods are based on purely statistical constraints and do not use any biological knowledge or transcriptional regulatory information.
Therefore, their results cannot contain biological transcriptional regulatory networks, which is also a primary reason that we used network component analysis NCA to detect the disease transcriptional regulatory mechanisms after feature gene extraction. In this study, besides the above main experiments of discovering the inverse TFs activities between AD and BC, two other experiments for exploring the regulatory mechanism for different subtypes or grades for the same disease AD or BC are studied as well.
To explore the differently regulatory mechanism between varying severities of AD, another AD dataset, series GSE [ 37 ], was studied as well. GSE dataset includes hippocampal gene expression of 9 control and 7 incipient, 8 moderate and 7 severe AD subjects. FastICA was performed to these three varying severities data and , and significant genes were extracted respectively from the incipient, moderate and severe AD samples [ 32 ]. There are shared genes between these two datasets and 50 out of them are also shared with the significant genes.
Then, NCA algorithm was performed to explore the similarities and differences of the activities of the 34 TFs between these two datasets. For the first 17 TFs, their regulatory activities are similar, while the regulatory activities of other 17 TFs are in opposite directions. Table 3 shows the upregulated or downregulated TF activities in these two datasets by the up or down arrows.
A gives the flowchart of the two-stage procedure in comparing AD-HIP dataset with AD in 3 severities dataset; B gives the flowchart of the two-stage procedure in comparing BC with no metastasis dataset with BC in 3 grades dataset. The molecular biological analysis showed that the changes of these TF activities and their target genes in the interactions of signaling proteins in cell cycle, chronic inflammation and immune response play important roles in the deterioration of AD.
The biological analyses of the regulatory activities show the transcriptional changes for the deterioration of AD. As we know that, in eukaryotic, the way to initiate transcription eukaryotic RNA polymerase requires the assistance of proteins. It suggests that with the aggravation of AD, the regulation of RNA level changed greatly and they will lead to the changes of the expression of many proteins. In this experiment, we compare the BC dataset with no metastasis with BC in 3 grades to explore the regulatory mechanism for the deterioration of BC.
The two-stage procedure was performed on BC series GSE, which is the same dataset in current study but in a different classification rule.
The breast cancer samples were divided into 3 categories according to their histologic grading: 11 tumours were grade 1; 40 were grade 2 and 53 were grade 3. To compare this experiment with the above results of BC data with no metastasis, the shared genes between these two BC datasets were further analysed. There are shared genes between these two datasets and 47 out of them are also shared with the significant genes of the first experiment.
Table 4 displays the activities of the shared TFs in these two datasets by up or down arrows. It is clear to find that the continuous activation or inhibition of these TFs are closely associated with the deterioration of breast cancer in the signal transduction pathways like cell proliferation, mitosis, apoptosis, Ras signal transduction, DNA replication, cholesterol homeostasis and growth regulation.
The biological analyses of the inverse regulatory activities help us to exploring the transcriptional mechanism from BC with no metastasis to the deterioration of BC. From the results of the above experiments, we can conclude that the quantification of the changes of significant TFs provide an increased understanding to the regulatory laws of disease in varying severity, different grades or even different diseases.
The extracted significant genes and TFs on different datasets of one disease were not the same, since FastICA extracts the statistically independent biological process and feature genes based on the gene expression profiles of each dataset. However, it is interesting that the molecular biological analysis showed that the TFs related pathways were closely similar like mitosis, apoptosis, cell cycle, chronic inflammation and immune disorder and so on.
That means the method can find out the key regulatory changes of one disease. To exploring the regulatory mechanism for different severities, grades, subtypes or different diseases, selecting the common significant TFs and reconstructing their quantitatively regulatory networks are effective to gain insights into the pathogenesis of diseases.
There are many kinds of neurodegenerative diseases, among them AD is the largest proportion of the diseases, and younger trend.
- Pillole di Cina - Zhōngguó (Italian Edition);
- No-brainer Sales Tips and Techniques?
- Two Faces of Evil: Cancer and Neurodegeneration / Edition 1!
- Through the History of the Cold War: The Correspondence of George F. Kennan and John Lukacs;
Parkinson disease PD is the second most common neurodegenerative disease. The causes of neurodegenerative diseases are not completely clear, but a lot of evidence have proved with the related inflammation and immune changed. In cancer studies, many results show the Immune response plays an important role in tumour development, treatment and prognosis. Tumours in different stages have different characteristics.
However, inflammation and immune changes in tumour growth and also plays an important role in the process of treatment [ 43 , 44 ]. In our research, the transcriptional regulatory pathway of inflammation and immune disorders of AD and BC were first studied, and other kinds of neurodegenerative diseases and cancers will be studied in the next step. Transcription factors are a diverse family of proteins that generally function in multi-subunit protein complexes that are vital to the normal development of an organism and are involved in routine cellular functions and responses to disease.
The function of TFs allows for the unique expression of each gene in different cell types and during development. Therefore, it is very important to study TFs while analyzing pathways to understand disease pathogenesis. Specifically, 10 out of 17 TFs showed inverse regulatory activities between AD and BC top 10 TFs in Table 2 , which showed that these two diseases shared many genes and biological pathways, but that the genes and pathways are regulated in different directions of the same spectrum. Combined with the current understanding of the functions of the innate and adaptive immune systems, transcriptional biological analyses related to the immune response were reviewed below to determine the opposite pathogenic regulatory mechanisms of AD and BC.
Ascl1 ASH1L , the activity of which increased in AD and declined in BC compared to control samples in the NCA results, is central to the differentiation of neuroblasts and the lateral inhibition mechanism, which inherently creates a safety net in the event of damage or death in these incredibly important cells, as well as neuronal commitment [ 45 ]. Ascl1 regulates astrocytes and oligodendrocytes by density and distribution in neurodegenerative diseases [ 46 , 47 ]. Under certain conditions, ASH1L can regulate interleukin-6 production.
As a TF, if ASH1L displays abnormal regulation, it will lead to many diseases, such as cancer and neurological diseases [ 51 — 52 ]. ASH1L has been implicated in facioscapulohumeral muscular dystrophy. In this disease, human muscles will experience progressive wasting [ 53 ], which is a common feature of AD, i. The p53 gene is a tumor suppressor gene that is involved in anti-tumor formation and inducing cell apoptosis.
Data from the literature show that CFLAR is also regulated by the IL-2 and MAPK pathway [ 55 , 56 ] and that abnormal expression is related to some diseases, such as cancer and autoimmune diseases [ 57 ]. Cold-inducible RNA binding CIRB protein CIRBP is a TF that plays a critical role in controlling cellular response upon confronting a variety of cellular stresses, including short wavelength ultraviolet light, neuroinflammation, hypothermia and hypoxia [ 58 — 61 ].
Some studies have indicated that CIRP regulates multiple pathways, such as MAPK, Wnt, apoptosis and many cancer-related signaling pathways in cerebral ischemia [ 62 , 63 ]. It has also been reported to mediate neuroinflammation [ 64 ]. Hmgb3 is a member of high mobility group DNA-binding motifs, which have been found to increase the transcriptional regulatory process in AD and decrease this process in BC.
It is known that Hmgb3 over-expression can inhibit B-cell and myeloid differentiation. Therefore, reducing the regulation of the HMGB3 protein levels is an important step for myeloid and B-cell differentiation [ 65 ]. HMGBs bind to all of the immunogenic nucleic acids examined with a correlation between affinity and immunogenic potential. HMGB expression disorders can lead to immunological disorders and some diseases, such as cancer, neurological diseases, and microphthalmia syndromic [ 67 ].
In the transcriptional regulatory network figures from our research, lipoma preferred partner LPP was shown to increase in AD and decline in BC. LPP plays a structural role at sites of cell adhesion in maintaining cell shape and motility. In addition to these structural functions, it is also implicated in signaling events and gene transcription activation. The LPP protein is localized at sites of cell adhesion, such as focal adhesions and cell-cell contacts, and shuttles to the nucleus where it has transcriptional activation capacities [ 67 — 69 ].
LPP and the expression of fusion proteins probably mediate tumor growth [ 70 ]. Some studies have reported that LPP correlated innate and T cell-mediated immune responses [ 71 ]. This may be additional evidence that AD and BC are both closely associated with the immune response, but at opposite ends. Their members have helicase and ATPase activities and are thought to regulate the transcription of certain genes by altering the chromatin structure around those genes. During transcription elongation, the FACT complex acts as a histone chaperone that both destabilizes and restores the nucleosomal structure.
WD-repeat protein 1 WDR1 or actin-interacting protein 1 AIP1 is a highly conserved WD-repeat protein in eukaryotes that promotes cofilin-mediated actin filament disassembly [ 76 ]. It is an emerging regulator of the actin cytoskeleton that is vital to filament disassembly. When WDR1 loses its function, it leads to embryonic lethality, macrothrombocytopenia and autoinflammatory disease [ 77 ]. This TF represses the tlr4 gene, a protein encoded by the Toll-like receptor TLR family, which plays a fundamental role in pathogen recognition and activation of innate immunity by cytokines [ 78 ].
It may play a role in transcription or interact with other nuclear matrix proteins to form the internal fibrogranular network. A proteomic screen revealed that MATR3 is bound to calmodulin and suggested that it is cleaved by both caspase-3 and caspase-8 [ 81 , 82 ]. Fig 5 displays the related target genes, pathways and common pathophysiological mechanisms with the opposing ends of the 10 inversely regulated TFs in the immune response between AD and BC. From the biological analysis we can know that they are closely related to innate and adaptive immune response.
In Fig 5 , the common TFs between AD and BC with inverse regulatory activities are displayed on both sides of the horizontal axis towards these two diseases. The results are showed in Tables 5 and 6. Table 5 shows the changes of activities of the 17 TFs during varying severities of AD. Table 6 provides the activities of the 17 TFs in 3 grades of BC dataset. The changes of the activities of these 17 TFs are more complicated than those in AD. Among them, CIRBP is a TF which have the ability to control cellular response upon confronting a variety of cellular stresses, mediate neuroinflammation and regulate MAPK, Wnt, apoptosis and many cancer-related signaling pathways.
ZNF is another important TF which is closely related to pathogen recognition and activation of innate immunity by cytokines. From Tables 7 to 9 we can see that ZNF is inversely associated during different courses of these two diseases with the increasing in AD and declining in BC.
Although the regulatory directions of these TFs are different in different states, even in different regions for the same disease, they play important roles in many immunological disorders including regulatory processes in inflammation, apoptosis, cellular response, innate and adaptive immune response. Considering the diverse changes in the different datasets of diseases, deep analysis is needed based on more numbers of microarray datasets and more detailed molecular biology research.
Accumulating epidemiological evidence and meta-analysis data suggest that there is a strong inverse correlation between AD and cancer. This suggests that there are shared genes or biological pathways regulated by both AD and cancer with dramatically different directions. Convincing evidence suggests that both AD and cancer are age-related immune dysregulation diseases and that age plays a crucial role in their pathogenesis.
We know that high-throughput DNA microarray datasets can successfully investigate hundreds of thousands of gene expression profiles simultaneously; the high-dimensional data are typically the regulatory results of a small set of TFs through an interacting network. However, high-throughput technologies that measure TF activities are not yet available on a genome-wide scale. Therefore, some statistical computational methods were applied in this study to deduce the biologically significant information and underlying transcriptional regulatory structure for the inverse regulatory mechanisms between AD and BC.
Based on our current understanding of the innate and adaptive immune system in neurodegenerative diseases and cancer, we focused on the contribution of inverse TFs in the present study to determine the immune balance and pathogenesis of AD and BC. Our previous studies showed that as an unsupervised matrix decomposition technique, FastICA preceded PCA and NMF in capturing the potential biological processes via a statistically independent assumption.
Second, NCA was performed to determine the activities of the shared TFs and regulatory influences on TGs because understanding dynamic TF regulation is a key component of understanding disease pathogenesis. There were 17 TF activities with regulatory control strength acting on TGs. From the molecular biological analysis, we found that all 10 inversely associated TFs were closely related to cytokines and played important roles in the innate immunity, especially the adaptive immune response. Furthermore, some typical biological pathways related to the adaptive immune response were revealed by the reconstructed transcriptional regulatory networks based on the NCA results.
The experiments on two additional AD and BC datasets with different grades also show that the inverse associations of these TFs exists in the whole process of the diseases and play important roles in the deterioration of diseases. We believe that uncovering the inverse associations of cytokines and adaptive immune response in our work will add significant contributions to diagnosis, immunotherapy and pathogenic discovery in both AD and BC.
The transcriptional regulatory mechanisms for the inverse associations between AD and BC, as well as many other prevalent cancers based on immune dysregulation, will be investigated further in our future studies. Project administration: WK. Supervision: WK. Browse Subject Areas? Click through the PLOS taxonomy to find articles in your field. Abstract Although chronic inflammation and immune disorders are of great importance to the pathogenesis of both dementia and cancer, the pathophysiological mechanisms are not clearly understood.
Introduction Alzheimer's disease AD , which is the most common form of dementia, is a progressively fatal neurodegenerative disorder characterized by irreversible cognitive and memory deterioration that inevitably leads to death. Methods Independent Component Analysis ICA in gene expression data ICA is a high-order statistical and unsupervised algorithm that has been widely used in voice signal blind separation, array processing, image processing, and medical ICA is a high-order statistical and unsupervised algorithm that has been widely used in voice signal blind separation, array processing, image processing, and medical and biological signal analysis and has been recently successfully used in gene clustering, classification and pathway and biomarker identification.
The procedure of exploring the differences of regulatory activities The structure flowchart of the proposed two-stage procedure for comparing the significantly cellular behaviors of two diseases is showed in Fig 1. Download: PPT. Fig 1. Structure flowchart of the proposed two-stage procedure. NCA results and transcriptional regulatory process identification The second stage in this study was to find the activities of the shared TFs and their regulatory influence on TGs by NCA.
Fig 2. Dynamic transcriptional regulatory networks for the AD dataset. Table 1. Fig 3. Dynamic transcriptional regulatory networks for the BC dataset. Discussion Comparison of several feature gene extraction methods In recent years, several matrix decomposition methods have been widely used for feature gene extraction, gene clustering and disease classification, including principle component analysis PCA or singular value decomposition SVD , ICA and nonnegative matrix factorization NMF.
Comparison of different datasets with the same methods In this study, besides the above main experiments of discovering the inverse TFs activities between AD and BC, two other experiments for exploring the regulatory mechanism for different subtypes or grades for the same disease AD or BC are studied as well. AD data in varying severities To explore the differently regulatory mechanism between varying severities of AD, another AD dataset, series GSE [ 37 ], was studied as well.
Fig 4. Table 3. Regulatory activities comparison of the shared TFs in two different AD datasets. Table 4. The regulatory activities of the shared TFs in two different BC datasets. Transcriptional regulatory processes related to immune response between AD and BC There are many kinds of neurodegenerative diseases, among them AD is the largest proportion of the diseases, and younger trend. Fig 5. The inversely associated TFs in Table 2 and their related genes, pathways and biological processes.
International Conference on Alzheimer's & Parkinson's Diseases 2013
Table 5. The regulatory activities of the 17 TFs in the additional AD datasets. Table 6. The regulatory activities of the 17 TFs in the additional BC datasets. Table 7. Table 8. Table 9. Conclusions Accumulating epidemiological evidence and meta-analysis data suggest that there is a strong inverse correlation between AD and cancer.
- Top Authors.
- 1. INTRODUCTION?
- Wettbewerb der Regionen in Zeiten der Globalisierung (German Edition).
- Cancer and Dementia: It’s Complicated - Europe PMC Article - Europe PMC;
- Make a Comment.
- At Sundown.
References 1. Drug Discovery Today. Immune responses in rapidly progressive dementia: a comparative study of neuroinflammatory markers in Creutzfeldt-Jakob disease. J Neuroinflammation. Breast Cancer and the Immune System. J Soc Integr Oncol. Cancer immunotherapy: nanodelivery approaches for immune cell targeting and tracking.
Front Chem. Association between cancer and Alzheimer's disease: systematic review and meta-analysis. J Alzheimers Dis.
Driver JA. Inverse association between cancer and neurodegenerative disease: review of the epidemiologic and biological evidence. Inverse and direct cancer comorbidity in people with central nervous system disorders: a meta-analysis of cancer incidence in , participants of 50 observational studies.
Psychother Psychosom. Alzheimer disease and cancer risk: a meta-analysis. J Cancer Res Clin Oncol. View Article Google Scholar Cancer linked to Alzheimer disease but not vascular dementia. Tumor diagnosis preceding Alzheimer's disease onset: is there a link between cancer and Alzheimer's disease?
Two Faces of Evil: Cancer and Neurodegeneration : Thomas Curran :
The inverse association of cancer and Alzheimer's: a bioenergetic mechanism. J R Soc Interface. Inflammatory mediators may have divergent roles in cancer and in Alzheimer's disease. Clin Pharmacol Ther. Age-dependent metabolic dysregulation in cancer and Alzheimer's disease. Transcriptional evidence for the "Reverse Warburg Effect" in human breast cancer tumor stroma and metastasis: similarities with oxidative stress, inflammation, Alzheimer's disease, and "Neuron-Glia Metabolic Coupling".
Aging Albany NY. Antioxid Redox Signal. A common biological mechanism in cancer and Alzheimer's disease? Curr Alzheimer Res. Inverse association between cancer and Alzheimer's disease: results from the Framingham Heart Study. Pin1 dysregulation helps to explain the inverse association between cancer and Alzheimer's disease. Biochim Biophys Acta. Papazoglu C, Mills AA. J Pathol. RNA Biol. BMC Genomics. Mol Biosyst. Nat Struct Mol Biol. Trans Neurosci.
Molecular mechanisms of initiation and maintenance of long-term memory storage. Curran and Y. Christen eds. DOI Pgs Nucleic Acids Res. Eur J Biochem. J Biol Chem. Role of the neural visinin-like domain in regulating autophosphorylation and calmodulin binding affinity. Mohan, M. S Molecular mapping of a resistance-specific PCR-based marker linked to a gall midge resistance gene Gm4t in rice. Theoretical and Applied Genetics. Plant Mol Biol. Publications The following articles were published by Sathya Puthanveettil, his lab members and his collaborators.
PubMed Link Mohan, M. Search site.