No replies to this topic

[Christopher Mehl]

Q: ­how to know whether Kin is inhibited or Kout is getting inhibited?­

A: Ideally the selection of the model is done by knowing the mechanism of action of the agonist or antagonist to determine whether it acts on Kin or Kout. If this mechanism is not known, then the model can be selected based on pattern recognition in the exploratory plot, i.e. the shape of the response curve. There is a good discussion of this in the Indirect models section of Gabrielsson and Weiner.


Q: Is it % inhibition or concentration of endpoint? at time zero, it is ~60%?

A: Per the authors, the Y-axis label of %GSECR inhibition is a bit confusing. The concentration of soluble beta-amyloid [Ab1-40] (which is called ‘response’ from now on) in brain serves as a biomarker of enzyme activity. The soluble beta-amyloid concentration was normalized to a pre-dose concentration measure over time to get the response values.


Q: ­How do we build an overlay plot?­

A: This was shown during the demo. If all the input data is in one worksheet, use the column you want to overlay (e.g. Subject ID) as a Group variable in the XY plot. If you want to read in a second worksheet to make the overlay, go to the bottom Options panel in the XY plot and select Plot, Graphs tab, then click the Add button to get a mapping panel for a second dataset.


Q: ­Can we use the same model for stimulation of Kin?

A: Stimulation of Kin is in the PK/Indirect grouping, but it is a different model. In the menu options, select stimulation limited of buildup. This will modify Kin using a stimulatory function:

E(C ) = (1 + Emax*C^γ/ (EC50^γ+ C^γ))



Q: ­Can we combine both inhibition of kin and stimulation of Kin­?

A: Not simultaneously, but you could enter both inhibitory and stimulatory functions for Kin in separate equations, and then implement a sequence statement that switches between the stimulatory and inhibitory functions at a certain time or response level.


Q: ­how can you get Kout initial estimate­?

A: The initial estimate for Kout is determined from the exploratory plot, using the slope of the downswing (onset) of the response. The model then fits all the parameters simultaneously to get the final estimate for Kout.


Q: In past, i had an issue with gamma estimation. Have you had any trouble in past?­

A. Try fitting the reduced model (no gamma, analogous to gamma=1) to get good estimates for the other parameters, Kin, Kout, IC50. Then add gamma to the model. If confidence is very high for one of the other parameters, set the value to Fixed for that parameter and then try to estimate gamma.


Q: ­Can we apply this model for enzyme inhibition and enzyme induction?­

A: Yes, the PK/Indirect model group can handle either enzyme inhibition or enzyme induction.