BioModels Database logo

BioModels Database


Benson et al., (2013). Systems pharmacology of the nerve growth factor pathway: use of a systems biology model for the identification of key drug targets using sensitivity analysis and the integration of physiology and pharmacology.

February 2017, model of the month by Varun Bhaskar Kothamachu
Original model: BIOMD0000000588


NGF (Nerve growth factor) belong to a class of proteins known as neurotrophins, and are a subset of neurotrophic growth factors [1]. Other members of this group include GDNF (Glial cell Derived Neurotrophic Factor), BDNF (Brain Derived Neurotrophic Factor), NT-3 (Neurotrophin-3) and NT-4 (Neurotrophin 4). These proteins play an important role in neuronal growth, functioning and survival. They bind to a variety of receptors viz. TrkA, TrkB and TrkC on the neuronal cell surface and regulate cell death and developmental processes in the peripheral and central nervous system [1]. Figure 1 is a schematic representation of the binding between different neurotrophins and neuronal cell surface receptors viz. TrkA, TrkB,TrkC and P75 [2].

As shown in figure 1, the binding of NGF to its cognate receptor TrkA (Tropomyosin receptor Kinase A) results in an auto-phosphorylation of the TrkA domain on the receptor. This results in the activation of intracellular signalling cascades involving Ras, Rap-1, phosphatidylinositol 3-kinase , phospholipase C-γ1 and MAP (Mitogen Activated Protein) kinase pathways [3]. The autophosphorylated NGF-TrkA complex is then internalized and transported to the neuronal cell body. This triggers the accumulation of dppERK (Di-phosphorylated Extracellular signal Regulated Kinase) in the neuronal nucleus which in turn activates additional downstream pathways that regulate neuronal cell survival and pain [4]. NGF pathway is very well studied as it plays a key role in neuronal development . Its identified involvement in triggering accumulation of dppERK, and a resultant regulation of pain mechanisms has made it the subject of several investigations to identify new drug targets for managing pain [5].

Motivation for this study

The authors in this study [6, BIOMD0000000588], based on existing knowledge about the role of NGF pathways in regulating pain have integrated two different models built by Sasagawa et al. [7, BIOMD0000000049] and Fujioka et al. [8] to develop a combined model that can be used to identify additional drug targets in the pathway. The concentration of dppERK in the neuronal nucleus is used as a reference and all reactions and species in the model are examined by carrying out a time dependent sensitivity analysis on the accumulation of dppERK dimers in the nucleus. This allows the identification of key reactions and species in the model that impact downstream pain response and neuronal survival mechanisms which are triggered by dppERK. A ranking of different species in the pathway with respect to their impact on regulating the concentration of dppERK can be used to identify the most druggable target in the pathway. By carrying out a study like this, the authors attempt to ensure a more informed selection of drug targets. Having a ranked list of targets is important as it can be used to screen the most druggable new drug molecules, remove ineffective or low impact drug targets early in the drug development process and has the potential to aid in the development of polypharmacological therapeutic strategies when developing drugs.

Model Description

The authors have examined the NGF-TrkA pathway using ODE (ordinary differential equation) models developed based on the current understanding of the system. The first model (model 1) contained reactions in the NGF pathway occurring inside the neuronal cell and it did not include any physiological information. This model was then expanded to create a second one (model 2; Figure 2 ) with compartments for extracellular body water and the neuron. Reactions in model 1 occur inside the neuron, while reactions involving the binding of antibodies with NGF with its antibody and inhibitors targeting TrkA are part of the extracellular compartment. With model 2, it is possible to study the inter-compartmental reactions occurring in the pathway with a physico-chemical separation of species and reactions into two compartments which reflects the environment of these reactions in the nervous system. Using published pharmacokinetic data for different parameters in the model, the authors study the impact of the binding of antibodies on NGF and the action of inhibitors targeting TrkA on the accumulation of dppERK inside the neuronal nucleus.

Figure 2

Figure 2 Diagram of Model 2. (1) refers the compartment 1, the extra-cellular body water compartment. (2) is the neuronal intracellular compartment. Model 2 is built by extending the NGF pathway model (model 1) occurring within compartment 2 (neuronal cell). Figure taken from [6].

In addition to these, the impact of all the remaining reaction parameters and concentrations of molecular species on temporal changes in the concentration of dppERK within the nucleus is examined to arrive at a ranked list of reactions and species based on the sensitivity of the concentration dppERK in the nucleus (of a neuron) to changes in them. This is important as it allows the identification of potential drug targets in the network with the most impact in regulating pain response mechanisms occurring downstream from the accumulation of dppERK. Models like this represent different molecular interactions, employ measured pharmacokinetic data and study the impact of different reactions on drug response.


Using the initial model (model 1) built on known reactions occurring in the NGF signalling pathway, the authors examined the significance of different species and reaction parameters in model 1. The sensitivity of dppERK (SdppERK,an) with respect to different parameters or species (an) was calculated using eq.1

Eqn 1

Figure 1

Figure 1. Shows the binding of different neurotrophins with receptors P75, TrkA, TrkB and TrkC. All the different pre-processed (nascent) neurotrophins have strong binding affinity towards P75 (solid arrows) which drops by nearly 1/100th when they mature (indicated by dashed arrows). Processed (mature) form of these neurotrophins have a more specific high affinity binding with NGF binds to TrkA, BDNF & NT-f bind to TrkB and NT-3 binds to TrkC. This figure was adopted from figure 2 of [2].

A plot of SdppERK,an over a time period of 100mins was generated for all reaction parameters and species in the model. The authors calculate the Area Under the Curve (AUC) to derive a ranking of species and parameters where a large AUC value indicates greater sensitivity. We find that after NGF, and complexes (viz., Grb2_SOS_pShc_pTrkA, pShc_pTrkA and Shc_pTrkA), concentration of phosphorylated TrkA (pTrkA) has the highest impact on the concentration of dppERK. With respect to the importance of different reactions parameters in the model, the authors find that the rate of synthesis (NGFext) viz. Kf147 has the most significant impact on the concentration of dppERK (See Table 1).

Table 1

Table 1. Shows the ranking of different species concentrations and model parameters based on the sensitivity of concentration of dppERK with respect to these. This table was derived from the tables in [6].

Using model 2, and including physiological information regarding the compartments in which different reactions occur and employing antibodies and inhibitors at NGF and TrkA respectively, we find that targeting these two species is effective in blocking the accumulation of dppERKnuc (See figure 3).

Figure 3

Figure 3 Model 2 simulated time course of dppERKnuc response to NGF (solid line). The dashed line shows the response in the presence of a TrkA binding inhibitor Ki = 0.1 nM (at 100 × Ki of the inhibitor) and the dashed-dotted line at 1000 × Ki, both given at t = 0. Figure taken from [6].


This study employs known understanding about the NGF pathway and undertakes a systems pharmacology approach towards identifying effective drug targets in the NGF signalling pathway. By carrying out a sensitivity analysis on species (concentrations) and reaction parameters (reaction rates) occurring in a representative model without pharmacokinetic data, the authors identified a ranked list of targets. This model was then enriched with both physiological and pharmacokinetic data (model 2). Using the ranked list of most impactful target species, they confirmed that introducing a virtual antibody (Tanezumab which acts on NGF) and an inhibitor for TrkA, produces the desired reduction in dppERK accumulation. In a model with 76 species, identifying specific targets like this is extremely useful as it aids in filtering non-impactful targets from those which produce the desired effect and a successful marketable drug. The benefits of screening potential targets using an approach like this increases the efficiency of downstream stages in drug development, especially in systems where there are hundreds of species in a pathway.


  1. Conner and Tuszynski. Neurotrophins: Physiology and Pharmacology Encyclopedia of Neuroscience , L. R. Squire, Ed. Oxford: Academic Press, 2009, pp. 1101–1106.
  2. Segal. Selectivity in neurotrophin signaling: theme and variations Annu. Rev. Neurosci. 2003; vol. 26, pp. 299–330.
  3. Chao. Neurotrophins and their receptors: A convergence point for many signalling pathways Nat. Rev. Neurosci. 2003 Apr;4(4):299-309.
  4. Mizumura and Murase. Role of nerve growth factor in pain Handb Exp Pharmacol. 2015;227:57–77.
  5. Hefti et al. Novel class of pain drugs based on antagonism of NGF. Trends Pharmacol Sci. 2006 Feb;27(2):85-91.
  6. Benson et al. Systems pharmacology of the nerve growth factor pathway: use of a systems biology model for the identification of key drug targets using sensitivity analysis and the integration of physiology and pharmacology. Interface Focus. 2013 Apr 6;3(2):20120071.
  7. Sasagawa et al. Prediction and validation of the distinct dynamics of transient and sustained ERK activation. Nat Cell Biol. 2005 Apr;7(4):365-73.
  8. Fujioka et al. Dynamics of the Ras/ERK MAPK cascade as monitored by fluorescent probes. J Biol Chem 2006 Mar 31;281(13):8917-26.