Autoinhibition in PK-Sim

[Jesminna]

It was indicated in the manual that "If an inhibition is set up for an enzyme or transporter, all processes with the same name will be linked and affected by the inhibition. An autoinhibition cannot be set-up because measured Ki values will already be altered due to autoinhibition."

Could you please explain this a little bit more regarding 1) why an autoinhibition (competitive inhibition) cannot be set up in PK-Sim? 2) why the measured Ki values will be altered due to autoinhibition?

If a drug shows autoinhibition (competitive inhibition) effect on an enzyme or transporter, how can I incorporate the autoinhibition effect in the model?
Thanks!

[sfrechen]

Hi @Jesminna
It appears that the manual contains confusing information here with regard to "[...] autoinhibition cannot be set-up because measured Ki values will already be altered due to autoinhibition." I guess this sentence does not make sense.

Please note that there is no auto-inhibition in the narrow sense for a competitive inhibitor. A competitive inhibitor binds tightly to the enzyme of interest. If there is metabolism of the inhibitor via this enzyme, the kinetics of this are fully considered via Michaelis-Menten kinetics. In the case of a very simple (but common) model assuming simple binding to the enzyme to the one and only active site, then Km is actually equal to Ki. Km characterizes the metabolism of the inhibitor itself, Ki charcaterizes its effect on other substrates.

Auto-inhibition in the narrow sense may happen in case of mechanism-based inactivation. Here, the inactivator alters the amount of the enzyme which obviously has an additional effect on the rate of metabolism. This is automatically accounted for in PK-Sim if your inactivator features metabolism via the enzyme of interest.

Kind of auto-inhibition may also happen in case of competitive inhibitors if inhibiting metabolites are formed (typical example is itraconazole, please have a look here https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/v9.1/Itraconazole). Of course, this can also easily be implemented in PK-Sim if your inhibitor features metabolism via the enzyme and of the metabolite features (competitive) inhibition of the enzyme.

Hope that helps
Sebastian

[Jesminna]

Thanks a lot for the detailed explanation! This is really helpful.

[msevestre]

@sfrechen @Yuri05
Let's move this as an issue in the docs repo. It would be good to remove any confusion if we can

[AyaSaleh90]

Dear @sfrechen,

I have a similar question here Open-Systems-Pharmacology/Forum#1185:
Based on your explanation, I thoroughly looked at the itraconazole model. I understood that itraconazole's slow elimination (non-linear PK) was mainly due to the formation of potent inhibitor metabolites with their accumulation plus its own inhibition parameters.

This is almost similar to my assumption regarding our parent compound. Still, the metabolites are not metabolised via the same enzyme as the parent (they only inhibit the parent's CL). The in vitro analyses showed that the parent is metabolised via three CYP enzymes. Still, it inhibits 2 of those CYP enzymes with a competitive mechanism and the other one with a non-competitive mechanism. And when I tried to apply the Ki of the parent only as 1st step to its own enzymes, it did not change the simulated concentration-time profile. I'm still in process of the adding the metabolites since we don't have much information about them except their inhibition effects on the parent.

However, back to the itraconazole model, I still have a question to fully understand how to implement such an interaction in my model. Could you please explain to me why the Km for the 4 molecules (parent & 3 metabolites) were optimised?? like if I have a similar situation, what is the assumption behind this step and what type of clinical data shall be used; to infer such information?

Thanks a lot in advance for your help!
Aya

[StephanSchaller]

@AyaSaleh90 , it helps opening a new thread and then referencing an old one you want to follow-up on to get answers.

Two points:

1. as Sebastian pointed out, the Km already considered (competitive/reversible) autoinhibition (only for mechanism-based inhibition one might add a process)
2. Km could be optimized if there is data for multiple dose levels available. Of course, ITZ is a complex example, so in this case also metabolite PK data would be required to really gain certainty on the identified values (i.e. identifiability of the fitted parameters given the model structure and available data might be an issue).