A modular network model of aging
Huiling Xue1,2,a, Bo Xian1,2,a, Dong Dong1, Kai Xia1, Shanshan Zhu1, Zhongnan Zhang1, Lei Hou1, Qingpeng Zhang1, Yi Zhang1 & Jing-Dong J Han1
- Chinese Academy of Sciences Key Laboratory of Molecular and Developmental Biology, Center for Molecular Systems Biology, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing, China
- Graduate School, Chinese Academy of Sciences, Beijing, China
Correspondence to: Jing-Dong J Han1 Chinese Academy of Sciences Key Laboratory of Molecular and Developmental Biology, Center for Molecular Systems Biology, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Datun Road, Beijing 100101, China. Tel.: +8610 6484 5843; Fax: +8610 6484 5797; Email: jdhan@genetics.ac.cn
Received 13 June 2007; Accepted 17 October 2007; Published online 4 December 2007
aThese authors contributed equally to this work
Top of pageArticle highlights
- By examining the modular structure of the protein-protein interaction (PPI) networks during fruit fly and human brain aging, we identified a pair of transcriptionally anti-correlated modules associated with the cellular proliferation to differentiation temporal switch and another pair associated with reductive and oxidative metabolic switch.
- These network modules and their relationships demonstrated that aging is mainly associated with a small number of biological processes and some modular changes might be reversible.
- We found that genes connecting different modules through PPIs are more likely to affect aging and/or longevity and experimentally tested the RNAi effect of some of these genes on the C. elegans lifespan.
- Network simulations suggest that aging might preferentially attack key regulatory nodes that are important for the network stability, providing a potential molecular basis for the stochastic nature of aging.
Synopsis
Aging is a prominent factor associated with many complex human diseases. Genetic mutants have implicated a large number of different biological processes in the aging process of both human and model organisms. However, some fundamental questions on aging remain unanswered, because they require a systems level view of the process (Kirkwood, 2005; Sinclair, 2005; Hekimi, 2006). Through integrating protein–protein interaction (PPI) networks with gene expression profiles during human brain and fruitfly aging, we first extracted subnetworks that are 'active' during aging. Genes in these active subnetworks aggregate into a small number of network modules. In both human brain and fruitfly aging networks, there are two modules associated with the cellular proliferation to differentiation of temporal switch, which display opposite aging-related, possibly reversible changes in expression. In fruitfly aging network, another couple of modules are associated with the oxidative-reductive metabolic temporal switch and display opposite linear, possibly irreversible change with age. The transcriptional relationships among these modules can be modified by caloric restriction in fruitfly.
The topology of the aging network is related to the gene functions. Transcription regulatory genes and known 'aging genes' (genes that have been observed to affect cellular or organism aging), are overrepresented among PPIs connecting different modules, especially among those connecting transcriptionally anticorrelated modules. Such an enrichment of aging-regulatory functions at the module interface (PPIs connecting two modules) suggests that aging genes might be the key nodes to maintain the stability of the modular network. Consitent with this hypothesis, we found that the percentage of known aging genes and transcription regulators increase with increasing PPI interaction degrees or node betweeness (Yu et al, 2007) in both the full PPI network and the modular aging network (Figure 3A–D).
Figure 3
Correlations of the percentage of aging genes and regulatory genes to PPI degree, betweeness of the proteins and the AvgPCC of the hubs. (A, B) The percentage of aging genes (A) or transcription regulators (B) among the NP or HPRD proteins of distinct PPI degrees are plotted against the PPI degrees of the proteins. The linear regression significance P-values are indicated above the polynomial fitted trend lines. (C, D) The percentage of aging genes (C) or transcription regulators (D) among the NP or HPRD proteins that are within each betweeness value interval of 2000 is plotted against minimal betweeness value of the interval. The linear regression significance P-values are indicated above the polynomial fitted trend lines. (E, F) The percentage of aging genes (E) or transcription regulators (F) among the NP or HPRD hubs that are within each AvgPCC value interval of 0.2 is plotted against minimal AvgPCC value of the interval. The linear regression significance P-values are indicated above the connected lines in (E).
Full figure and legend (266K)Figures & Tables indexInteraction dynamics of the a hub (a protein that interacts with many other proteins) can be estimated by its level of coexpression with their interactors (Han et al, 2004). High coexpression level (as measured by average Pearson correlation coefficient of their expression profiles, AvgPCC) indicates more stable binding between the hub and its interactors, whereas relatively lower coexpression level indicates more dynamic binding (Han et al, 2004; Bertin et al, 2007). Both aging genes and transcription regulators are preferentially associated with the dynamic binding hubs, with the aging genes biased for hubs of extremely low AvgPCC (Figure 3E, F).
Characteristic path length (CPL) of a network is a measurement for network connectivity. Rapid increases in CPL indicate destabilization of the network. Consistent with a role in maintaining network stability, we found that the CPL increases much more rapidly after attacking the aging genes than removal of random genes in the same network.
The role of aging genes in coordinating the modules in the aging network and in maintaining network stability suggests a potential molecular basis for the stochastic nature of aging.
Acknowledgements
We thank Dr Thomas E. Johnson and Dr James R. Cypser at the University of Colorado at Boulder for numerous helpful discussions of the data, Dr Ralph Greenspan (The Neuroscience Institute, San Diego), Dr Nicholas Baker (Albert Einstein College of Medicine), Dr Li Cai (Rutgers University) and Dr Yong Liu (CAS Institute of Nutrition) for helpful discussions and three anonymous reviewers for their valuable comments. We are also grateful to Dr Chonglin Yang for sharing C. elegans resources. This work was supported by NSFC Grant 30588001, National Basic Research Program of China (2006CB910700) and CAS funds to J-DJH.
References
- Bertin N, Simonis N, Dupuy D, Cusick ME, Han JD, Fraser HB, Roth FP, Vidal M (2007) Confirmation of organized modularity in the yeast interactome. PLoS Biol 5: e153 | Article | PubMed | ChemPort |
- Han JD, Bertin N, Hao T, Goldberg DS, Berriz GF, Zhang LV, Dupuy D, Walhout AJ, Cusick ME, Roth FP, Vidal M (2004) Evidence for dynamically organized modularity in the yeast protein–protein interaction network. Nature 430: 88–93 | Article | PubMed | ISI | ChemPort |
- Hekimi S (2006) How genetic analysis tests theories of animal aging. Nat Genet 38: 985–991 | Article | PubMed | ChemPort |
- Kirkwood TB (2005) Understanding the odd science of aging. Cell 120: 437–447 | Article | PubMed | ISI | ChemPort |
- Sinclair DA (2005) Toward a unified theory of caloric restriction and longevity regulation. Mech Ageing Dev 126: 987–1002 | Article | PubMed | ChemPort |
- Yu H, Kim PM, Sprecher E, Trifonov V, Gerstein M (2007) The importance of bottlenecks in protein networks: correlation with gene essentiality and expression dynamics. PLoS Comput Biol 3: e59 | Article | PubMed | ChemPort |
