Biochemical Pathway Network Anticancer Drug Efficacy
Disclaimer: This work has been submitted by a student. This is not an example of the work written by our professional academic writers. You can view samples of our professional work here.
Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.
Published: Tue, 08 May 2018
Abstract and Keywords
As cancer is a complex disease, recent cancer research is increasingly carried out as systems-level. And this conducts novel polypharmacological strategies to evaluate anticancer drugs which will alter the entire pathway network rather than inhibit or activate single target. In current work, a novel approach was developed by integrating multi-target binding free energy calculation and biological network efficiency analysis to estimate the biological potency. The obvious advantage of the network based prediction method is that it can assess the role of different proteins by biological pathway network analysis. This approach can evaluate the drugs’ efficacy more comprehensively than traditional single target evaluation methods and would be very useful for discovering novel drugs against complex disease like cancer.
Keywords: anticancer drug, network efficiency, cancer pathway, system biology
Translational research promotes basic biological discoveries from the fundamental research into clinic applications, and it uses the clinic observations to indicate future directions for fundamental research (1). A great portion of biological molecules with specific physiological changes such as proteins, nucleic acids and small ligands have important impact on the prevention and treatment of diseases. As most complex diseases are related to biological pathways, the studies on related pathways could discover the relation between the biological molecules, pathways, cellular entities and clinic targets (2).
Abiological pathwayis a series of actions among biological molecules in a biological process that are organized in a specified manner and can perform certain biological functions (3). As cancer is a complex disease, recent cancer research is increasingly carried out as systems-level (4). And this conducts novel polypharmacological strategies to evaluate anticancer drugs which will alter the entire pathway network rather than inhibit or activate single target (5). Since many proteins or ligands whose biological functions are not explored completely in the cancer pathways, it is time-consuming and costly to determine biological functions through biological experiments for each (6). More importantly, it can hardly to evaluate a drug’s effect on tens of targets and pathways by experimental techniques. Therefore, it is urgent need of developing a computational approach to solve this problem.
With the progress of system biology and bionetwork, we know that the biological potency of an ideal drug may not merely determined by the inhibition of a single target, but rather by the rebalancing of several proteins or events, which contribute to the etiology, pathogeneses, and progression of a complex disease (7). The available methodologies of in silico screening based on a single target seem not effective in studying ligands’ effects on biological process comprehensively for complex disease like cancer (8-9). In current work, a novel approach was developed by integrating multi-target binding free energy calculation and biological network efficiency analysis to estimate the biological potency. The obvious advantage of the network based prediction method is that it can assess the role of different proteins by biological pathway network analysis. This approach can evaluate the drugs’ efficacy more comprehensively than traditional single target evaluation methods and would be very useful for discovering novel drugs against complex disease like cancer.
The network was constructed by manually curating scientific literatures (10-41), and information from Reactome (42) and KEGG (43) knowledgebase. In order to construct a proper network for anticancer drug evaluation, only proteins and their interactions that have been reported directly relate to cancer were included. The proteins which participate in the cancer pathway were proposed as nodes. The relations of these proteins were proposed as connections between nodes. The relations means that protein in upstream enhances the function of the downstream protein.
The drug library for performing biochemical pathway network based anticancer drug evaluation comprises 683 drugs from The Chinese Pharmacopoeia 2010. The structures of these drugs were built by Marvin Sketch (ChemAxon). And then, all of these structures were minimized with the MMFF force field by openbabel (44). In the minimization procedure, the threshold of root mean square deviation (RMSD) of potential energy was set to 0.01 kcal/Å·mol. The optimized structures of all drugs were saved as a Multi-MoleculeSDF file and ready for further binding energy calculation.
Binding free energy calculations
The binding free energy calculations were performed by Autodock 4.2 software package (45). For each protein structures, all water molecules were removed. After that, polar hydrogen atoms were added and non-polar hydrogen atoms were merged using the Hydrogen module in Autodock Tools (ADT). All of the atoms were assigned Kollman partial charges. Grid maps of the pre-defined active site for each atom type were calculated by AutoGrid. The grid map of the molecular docking was established by a 45×45×45 box centered on the pre-defined active site, with a spacing of 0.375 Å between each grid points. Molecular docking and calculation of drug/protein dissociation constant based on Lamarckian genetic algorithm (LGA) were performed by Autodock. The docking parameters were set as follows: the ligand translation step was set to 1.0 Å, the ligand quaternion and torsion step were both set to 60 degrees, the maximum number of energy evaluations was set to 1.0 × 107, the number of individuals in population of genetic algorithm was set to 150, the maximum number of genetic algorithm operations was set to 2.5 × 105, the rate of mutation and crossover were set to 0.02 and 0.8, respectively. All other parameters were set to defaultvalues. When searching the conformational and orientational spaces of the ligand with rotatable bonds full flexibility, the structure of the proteins was kept rigid. For each docking works, 10 independent runs were performed to evaluated different ligand poses and only the most favourable pose was dumped to the result file.
Network efficiency calculation
The damage induced by the attacks on the network is characterized by the network efficiency (NE), which is defined as the sum of the reciprocals of the shortest path lengths between all pairs of nodes. Due to a global topological property of a network which could be applied to measure the integrity of the network, the network efficiency was assumed to be used as a measure for drug efficiency. The NE of a graph G is measured by the shortest paths between pairs of nodes with the following expression:
where dij is the length of the shortest path between node i and j and the sum is over all N(N −1)/2 pairs of nodes with total number N of nodes in the graph G. If the network is weighted, dij is the path with the minimum weight. The initial line values of every edge were arbitrarily set to 10. To give relative network efficiency, this quantity NE is divided by the initial network efficiency. Thus we considered the network efficiency of the initial network as 100% and measured the relative network efficiency after each attack. We have chosen the clotting cascade network as the network models.
The network efficiency was calculated for each drug. The drugs’ effects to the network rely on their binding ability to the target protein. We supposed that one drug could inhibit the target protein well while the binding ability was relatively high. For a drug, we transformed the binding energy to network flux that directly downstream the target protein of the network and then calculated the network efficiency. In other word, the network flux of all edges, which point to the other targets protein from this target protein, was re-assigned based on the binding energy between the drug and the target protein. For each target protein, the drug with the highest binding energy was chosen as the reference standard. Network flux of each edge in the network was calculated with the expression:
ΔGs represents the binding free energy of the most potent drug, ΔG represents the binding free energy of other drugs, and Fedge is the flux of the edges come out of the protein in the network. Default flux value of the edges between nodes was set to 10.
Results & Discussion
The cancer related biochemical pathway network was constructed by manually curating scientific literatures. By analyzing these literatures, 50 important cancer-associated proteins or ligands have been figured out. At the same time, a series of classical cancer-associated signaling pathways and their interactions were also included. All of the proteins and their interactions have been reported directly relate to cancer by scientific literatures. The schematic diagram of the network we have constructed is shown in Figure 1. In this regulation network, all of the 50 important cancer-associated proteins or ligands act as positive factors for cancer progression. Although these proteins or ligands directly relate to cancer, most of them can not promote their function independently. There are 54 pairs of relationships between these proteins make them link to others and form an integrated regulation network.
On the other hand, in the regulation network of cancer we have built, there are some alternative processes can enhance cancer progression independently. For example, a series of receptor tyrosine kinases such as EGFR, FGFR, HGFR, c-kit, IGFR and PDGFR, can be stimulated by their extracellular protein ligands (growth factors) and then elicits downstream activation and signaling by phosphorylated tyrosines through the Grb2 phosphotyrosine-binding SH2 domains. Among these receptor tyrosine kinases, complete single-target inhibition can not shut down the downstream activation and signaling by phosphorylated tyrosine because other receptor tyrosine kinases can trigger downstream activation simultaneously.
Degree distribution and characteristics of nodes with different degree values
As the nodes in the pathway network covered most of the important proteins or ligands relate to cancer, we performed degree analysis for each node in the cancer network (Figure 2). The results of degree distribution analysis showed that a great portion of the nodes (74.5%) in our cancer network have lower degrees (equal or lower than 2). On the other hand, there are also several nodes with higher degrees which link the nodes with lower degrees and make the whole cancer network integrated together. These results showed that the cancer network composed of multi-level protein or ligands with different features. Firstly, none of the nodes in our cancer network with degree value equal to 0. This result implied that proteins or ligands in the network can not exert their function without others. They showed a multi-pathway network manner. Secondly, the nodes which degree value equal to 1 were always growth factors or ligands that can trigger different subpathway activation. It is worth noting that these nodes exert their function by activating their downstream subpathway and finally enhance cancer progression respectively. Elimination of one node can not inhibit cancer progression because other node will attenuate the inhibition effect by their functions. Thirdly, the nodes which degree value equal to 2 are always signal transduction related proteins or ligands. They always act as linking nodes and can transfer flux from an upstream node to downstream node. Lastly, the nodes with higher degrees in the network mainly act as linking centers which combine different subpathway together.
In order to proceed with the network efficiency based prediction, a three step protocol has been setup. Firstly, every drug that needs to be evaluated was docked into the proteins in our cancer pathway network and got its binding free energies to all the cancer related proteins in the pathway network. Secondly, the binding free energies were converted to flux of edges in the network. Finally, the network efficiency was calculated from each flux of all the edges in the pathway network. Because all of the nodes in our cancer pathway network were proteins or ligands which act as positive enhancer for cancer progression, drugs which can reduce the network efficiency more significantly were evaluated as drugs with better anticancer efficacy. We have evaluated 683 drugs molecules from the Chinese Pharmacopoeia 2010 because all of these molecules were clinically used drugs. The drugs were sorted by their abilities of reducing the network efficiency. The top and last 20 drugs were taken to further analysis (Figure 3). Among top 20 drugs which reduced the network efficiency most significantly, 8 out of 20 drugs (40%) were clinical used anticancer drugs or drugs that reported have anticancer activities. In contrast, among the last 20 drugs, none of them was clinical used anticancer drugs or has anticancer activities. The ratio of anticancer drugs was significantly different between the top and last 20 drugs. By using our cancer pathway related network efficiency prediction protocol, compounds with anticancer activities could be clearly distinguished from those without.
Like all molecular evaluation method based on computational strategy, our approach has many advantages as well as some limitations. One of the obvious advantages of the network based prediction method is that it considers the role of different proteins by biological pathway network analysis. On the other hand, the disadvantage of this method is that the accuracy of the evaluation technique relies on the reliability of network construction and the accuracy of binding free energy assessment.
We developed a novel approach to evaluating anticancer drug efficacy based on biochemical pathway network. This method combined multi-target binding free energy calculation, network flux and network efficiency for the prediction of anticancer drugs efficiently. The obvious advantage of the network based prediction method is that it can assess the role of different proteins by biological pathway network analysis. This approach can evaluate the drugs’ efficacy more comprehensively than traditional single target evaluation methods and would be very useful for discovering novel drugs against complex disease like cancer. It remains to be determined that what different composition and complexity of the cancer pathway network takes effect to the activity, and the relevant work is underway.
Cite This Work
To export a reference to this article please select a referencing stye below: