Hase T, Tanaka H, Suzuki Y, Nakagawa S, Kitano H. “Structure of Protein Interaction Networks and Their Implications on Drug Design.”について知りたい

The study by Hase T et al. (2009) demonstrates that protein interaction networks (PINs) are structured as Highly Optimized Tolerance (HOT) networks rather than simple scale-free networks, with drug targets primarily localized to a central "backbone" of middle-degree proteins (Direct, High; PMID: 19876376) «✓ PMID:19876376».

Global Architecture of Protein Interaction Networks

• HOTnet Topology: Human and yeast PINs conform to HOT theory, which describes systems optimized for specific perturbations but fragile against unexpected ones (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Structural Components:
  • Stratus: A core backbone formed by tightly interconnected middle-degree nodes (degrees 6–38 in yeast; 6–30 in human) (Direct, High; PMID: 19876376) «✓ PMID:19876376».
  • Altocumulus: High-degree nodes (hubs) that rarely connect to each other but instead connect to a large number of low-degree nodes (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Graph-Theoretic Validation: The normalized likelihood of high-degree node connectivity, $S(g)$, was found to be low for both yeast (0.25) and human (0.38) PINs, matching the profile of HOTnets (e.g., Abilene Internet2 router topology, $S(g)$ = 0.34) rather than preferential-attachment scale-free networks ($S(g)$ = 0.61) (Direct, High; PMID: 19876376).
• Scale-Richness: PINs are "scale-rich," meaning different functional categories (based on GO slim ontology) exhibit distinct degree-distribution patterns rather than a uniform self-similar structure across the entire network (Direct, High; PMID: 19876376) «✓ PMID:19876376».

Distribution of Essential and Lethal Genes

• Node Importance: Essential and synthetic lethal genes are concentrated in middle-degree and high-degree nodes (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Severity Gradient: The average degree of essential proteins (4.95) and synthetic lethal proteins (4.40) is significantly higher than that of synthetic sick genes (4.07), suggesting that genes with less severe impacts are distributed toward lower-degree regions (Direct, High; PMID: 19876376) «✓ PMID:19876376».

Implications for Drug Design and Target Prioritization

• Optimal Targets: Most currently marketed drug targets are concentrated on middle-degree nodes within the network backbone and certain low-degree nodes that perform specific functions (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Avoidance of High-Degree Hubs: High-degree nodes are generally avoided as drug targets because their elimination affects a vast number of low-degree nodes, likely leading to unacceptable side effects or lethality (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Cancer vs. Non-Cancer Targets:
  • Cancer Targets: Tend to have higher connectivity, with an average degree of 7.82 (Direct, High; PMID: 19876376).
  • Non-Cancer Targets: Have a lower average degree of 4.24 (Direct, High; PMID: 19876376).
  • The highest degree target identified was Src (degree 41), targeted by drugs such as Dasatinib (Direct, High; PMID: 19876376) «✓ PMID:19876376».
• Combinatorial Strategy: Low-degree nodes, which show a distribution similar to random expectation, are proposed as advantageous targets for combinatorial or "long-tail" drugs because they have less severe impacts on the overall system (Direct, High; PMID: 19876376) «✓ PMID:19876376».

What specific middle-degree proteins were identified as potential novel drug targets in the human PIN?

How does the HOTnet architecture of protein interaction networks contribute to biological robustness and fragility?

What are the quantitative differences in average node degrees between essential, synthetic lethal, and synthetic sick genes?