Network Medicine and Multiomic Approaches to Health and Disease¶
Chapter 499 | Harrison's 22e · Parts 19-20 – Consultative & Emerging Topics · Chapter 499
Key Clinical Points¶
- Systems biology provides a holistic approach to understanding complex biological systems by analyzing their topology (structure) and dynamics (time-dependent response).
- Reductionism, while useful for simple pathways, fails to explain 'off-target effects' of drugs because it ignores the integrated context of the system.
- Scale-free networks are common in biology; they contain 'hubs' (highly connected nodes) that are critical for system stability and serve as primary targets for drug intervention.
- Multiomic integration combines genomic, transcriptomic, proteomic, and metabolomic data to map the human interactome and resolve phenotypic differences.
- Endophenotypes (I) represent intermediate physiological/biochemical states between genetic/environmental inputs and clinical pathophenotypes (P).
- Disease modules in the interactome allow for the identification of shared pathways between different diseases, facilitating drug repurposing.
- Precision medicine aims to move from 'one-size-fits-all' treatments to tailored therapies based on computationally assisted molecular and phenotypic subtyping.
- The Interactome is governed by four principles: Local, Disease Module, Functional Coherence, and Shared Components.
- Network analysis has successfully identified novel targets in conditions such as asthma, COPD, and SARS-CoV-1.
DEFINITION & CLASSIFICATION¶
• Systems Biology:
Definition: The holistic study of living organisms or their cellular or molecular network components to predict their response to perturbations. • System Pathobiology: Definition: The study of how genetic or environmental perturbations produce disease and drug perturbations restore normal system behavior. • System Complexity Classification: ◦ Simple Systems: Nodes are linearly linked; the behavior is predictable and equal to the sum of its parts (e.g., a single metabolic pathway or gene network). ◦ Complex Systems: Nodes are linked in non-linear networks; behavior depends on a complex set of initial conditions and interactions; behavior eq sum of parts.
ETIOLOGY & PATHOPHYSIOLOGY¶
• Limitations of Reductionism: ◦ Historically, reductionism has been the dominant approach due to its simplicity. ◦ Failure: Cannot explain 'off-target effects' because drugs are not studied in an integrated context (failure to explore all possible actions aside from the primary target). • Network Topology & Dynamics: ◦ Topology: The structure or form of the network (e.g., scale-free vs. random). ◦ Dynamics: The time-dependent response to perturbations. • Scale-Free Networks: ◦ Biological systems often follow a power law distribution (P(k) = k^{-gamma}). ◦ These networks contain 'hubs' (highly connected nodes) that are critical for system stability and serve as primary targets for drug intervention. • Interactome Organizing Principles (Table 499-2): ◦ Local Principle: Proteins involved in the same disease tend to interact. ◦ Disease Module Principle: Proteins involved in the same disease tend to cluster in connected subnetworks. ◦ Functional Coherence Principle: Proteins in a disease module are often involved in the same biological process. ◦ Shared Components Principle: Related diseases are located in the same interactome neighborhood; unrelated diseases are separated.
DIAGNOSTIC APPROACH¶
• Methodology for Identifying Disease Modules (Figure 499-6): 1. Interactome Reconstruction: Reconstruct the interactome to provide a foundation for analysis. 2. Seed Gene Identification: Ascertain potential seed (disease) genes from: ◦ Curated literature ◦ Online Mendelian Inheritance in Man (OMIM) database ◦ Genomic analyses (e.g., GWAS, transcriptomics) ◦ Proteomic profiling 3. Module Identification: Identify disease modules using various modeling or statistical approaches to identify overlapping regions between different diseases. 4. Pathway Analysis: Identify pathways and the specific roles of disease genes/modules within those pathways. 5. Validation & Prediction: Validate identified modules (e.g., via "dynamic" analysis or cross-referencing with known protein interactions) and predict novel proteins associated with the disease state (distinguishing 'known' from 'predicted').
MANAGEMENT & TREATMENT¶
• Precision Medicine Strategy: 1. Transition from "one-size-fits-all" medicine to individually tailored medicine. 2. Utilize computationally assisted molecular and phenotypic subtyping. 3. Match specific treatments to identified patient subgroups based on unique profiles. • Clinical Applications of Network Analysis (Table 499-1): ◦ Hereditary ataxias: → Identify shared partners in neurodegeneration; → Link three unique metabolite abnormalities in prediabetics to seven type 2 diabetes genes via four enzymes; → Identify NFATC4 as a regulator of diabetes-associated genes. ◦ Epstein-Barr virus (EBV): → Map viral proteome effects through links to the host interactome; → Identify microRNA 21's role in suppressing the rho kinase pathway. ◦ Asthma: → Use disease modules to explain phenotype heterogeneity and drug response; → Identify vitamin D signaling as a central role in offsetting disease risk. ◦ Calcific aortic valve disease: → Identify first molecular regulatory networks in vascular calcification; → Identify NEDD9 as a key regulator in pulmonary vascular fibrosis. ◦ Chronic obstructive pulmonary disease (COPD): → Identify microRNA dysregulation as a determinant of vascular remodeling; → Utilize network-based approaches for in silico drug repurposing. ◦ SARS-CoV-2: → Use network-based approaches to identify novel treatments.
KEY PEARLS & HIGH-YIELD POINTS¶
• Scale-Free Networks: Biological systems often follow a power law distribution; these networks contain "hubs" (highly connected nodes) that are critical for system stability and serve as primary targets for drug intervention. • Endophenotypes (I): These represent intermediate physiological/biochemical states between genetic/environmental inputs (D, E) and the final clinical pathophenotype (P). • Multi-scale Integration: Network medicine integrates data from cell-specific, tissue-specific, and gene-specific levels to understand how mutations propagate through networks to cause systemic disease. • Drug Repurposing: Identifying shared modules between different diseases allows for the identification of common therapeutic targets.
Reference Tables¶
TABLE 499-1 Examples of Systems Biology Application to Disease and Therapy DISEASE Hereditary ataxias Diabetes mellitus¶
Harrison's 22e, p.3961
| DISEASE | ANALYSIS | REFERENCE |
|---|---|---|
| Hereditary ataxias | Many ataxia-causing proteins share interacting partners that affect neurodegeneration |
Lim et al: Cell 125:801-814, 2006 |
| Metabolite-protein network analysis links three unique metabolite abnormalities in prediabetics to seven type 2 diabetes genes through four enzymes The transcription factor NFATC4 regulates diabetes associated genes |
||
| Epstein-Barr virus infection | Viral proteome exerts its effects through linking to host interactome | Gulbahce et al: PLoS One 8:e1002531, 2012 |
| Network analysis indicates adaptive role for microRNA 21 in suppressing rho kinase pathway |
||
| Asthma | Disease module in the interactome explains phenotype heterogeneity and drug response |
Sharma et al: Hum Mol Genet 24:3005-3020, 2015 |
| Network analysis demonstrates the central role of vitamin D signaling in offsetting disease risk |
||
| Calcific aortic valve disease | Network and systems biology approaches identify the first molecular regulatory networks in vascular calcification |
Schlotter et al: Circulation 138:377-393, 2018 |
| Network analysis identifies novel scaffold protein NEDD9 as a key regulator in pulmonary vascular fibrosis |
||
| Chronic obstructive pulmonary disease |
MicroRNA dysregulation identified from network analysis as a determinant of vascular remodeling |
Musri et al: Am J Respir Cell Mol Biol 59:490-499, 2018 |
| Network-based approach to in silico drug repurposing | ||
| SARS-CoV-2 | Network-based approach to identifying novel treatments for SARS-CoV-2 infection | Morselli Gysi et al: Proc Natl Acad Sci USA 118:e2025581118, 2021; Patten et al: iScience 25:104925, 2022 |
TABLE 499-2 Organizing Principles That Tie the Interactome to Human Disease Local Principle: proteins involved in the…¶
Harrison's 22e, p.3961
- Local Principle: proteins involved in the same disease tend to interact.
- Disease Module Principle: proteins involved in the same disease tend to cluster in
connected subnetworks. - Functional Coherence Principle: proteins in a disease module are often involved in the
same biological process. - Shared Components Principle: related diseases are located in the same interactome
neighborhood from which unrelated diseases are separated.