BioModels Database logo

BioModels Database


Benson et al., (2014). A Systems Pharmacology Perspective on the Clinical Development of Fatty Acid Amide Hydrolase Inhibitors for Pain

January 2016, model of the month by Vincent Knight-Schrijver
Original model: BIOMD0000000512


Validation of druggable targets is an essential role of Systems Pharmacology. Quantitative Systems Pharmacolgy (QSP) models are designed using our current understanding of biology in conjunction with the focus of drug development. The benefit of a proposed clinical strategy can be assessed and predicted by a number of facets highlighting the requirements for further knowledge and risks in drug development. The case presented here by Benson et al., (2014) [1] illustrates where QSP modelling seeks to evaluate the inhibition of fatty acid amide hydrolase (FAAH) for pain.

The FAAH inhibitor PF-04457845 was entered through clinical trials for its pain preventative effects in osteoarthritis. FAAH acts to rapidly degrade endocannabinoids (ECs) which effectively terminates their agonist binding to the cannabinoid receptors. Only one EC, N-arachidonoylethanolamine (AEA), is discussed in this model. AEA typically binds to the CB1 receptor eliciting the closure and inhibition of several Ca2+ channel classes and purportedly resulting in analgesia. Perhaps then, the inhibition of FAAH may potentiate the effects of AEA with clinically relevant antinociception.


The primary goals of the model were to describe the AEA EC system in order to understand the use of FAAH inhibitors in the treatment of pain and, additionally, to identify the key gaps in our biological understanding of the system. To describe the AEA EC system, the model was constructed using a physiologically based pharmacokinetics (PBPK) approach including the relevant target tissues constituting the brain, the rest of body (RoB) and the blood-brain barrier (BBB). The dynamic EC system itself was comprised of the AEA and FAAH synthesis and degradation pathways as well as the cognate EC receptor, CB1.

The maximal occupancy of the CB1 receptor plateaued just below 25 %. This was shown in figure 3 that as the dose increased, only the duration of and not the maximal receptor occupancy was increased. The authors declare that other experiments such as imaging techniques may be necessary to shed light upon this occupancy value, which points to the requirements for measurable parameters in FAAH inhibitor effectiveness. However it is duly noted that the relationship between CB1 receptor occupancy and analgesia is unsubstantiated.

Figure 2 Figure 2b

Figure 2. Left, simulated in Copasi, time course data for differing AEA clearance hypotheses after FAAH inhibition. Either the AEA is degraded solely by FAAH (dotted line) or AEA clearance contains additional elements (solid line) such as saturable fast cellular mechanisms or non-saturable systemic elimination routes. Right, assuming now that the unknown enzyme or elimination mechanism is apparent in biology, the relative levels of all XEA species can be simulated according to clinical data following a 10 mg dose of PF-04457845. Figure 2 (right) was taken from the original paper [1].

Figure 1

Figure 1. Four compartment model. The XEA represents the total pool of AEA and precursors. The active AEA is synthesised through the NAPE-phospholipase D biosynthetic pathway. AEA, in both RoB and brain, can reversibly bind to CB1. FAAH in both these compartments can also catalyse the degradation of AEA, a reaction which is competitively inhibited by PF-04457845. The plasma compartment allows the transfer of XEA between both target tissue compartments via the BBB. Figure 1 was taken from the original paper [1].


Initial simulations were carried out to examine the assumptions on degradation pathways for AEA. Typically it was assumed that the majority of AEA degradation is mediated by FAAH activity. To show that FAAH-independent degradation routes are essential for the endogenous regulation of AEA, simulations were performed which included additional clearance routes, resulting in a concentration-time profile that more successfully track clinical data. Candidate elimination routes suggested by the authors are cyclooxegenases, cytochromes P450 and hydrolase enzymes as well as potential renal clearance. Once the unknown elimination had been implemented, the model now successfully simulates all XEA species' time profiles after FAAH inhibition (Figure 2, right).

Figure 3

Figure 3. Simulated in Copasi, the CB1 receptor occupancy is increased following oral PF-04457845 administration. The range of doses here indicate that, whilst the fraction of agonist-receptor complexes remains just below 0.25, the duration of this occupation fraction increases with dose.


By construction of a QSP model, the effectiveness of FAAH inhibition for pain relief is brought into question. To understand the drug effect, initial speculations about the biological system had to be scrutinised and novel assumptions implemented. Designed to address the questions of risk in moving a compound forward the model acted to warn of considerable gaps in knowledge. This suggested that, based on this objective QSP analysis, there was minimal conclusive evidence to suggest that FAAH inhibition would cause clinically significant analgesia. This was also reported in a study involving osteoarthritis patients [2].


  1. Benson, N, Metelkin, E, Demin, O, Li, GL, Nichols, D, van der Graaf, PH (2014). A systems pharmacology perspective on the clinical development of Fatty Acid amide hydrolase inhibitors for pain. CPT Pharmacometrics Syst Pharmacol, 3:e91.
  2. Huggins, JP, Smart, TS, Langman, S, Taylor, L, Young, T (2012). An efficient randomised, placebo-controlled clinical trial with the irreversible fatty acid amide hydrolase-1 inhibitor PF-04457845, which modulates endocannabinoids but fails to induce effective analgesia in patients with pain due to osteoarthritis of the knee. Pain, 153, 9:1837-46.