Hello,
I have 5 blood samples each from 6 patients treated with a drug: on day 1 (before treatment), 7, 14, 21 and 30. In these samples I have measurement of both the drug concentrations and quantification of many genes by RNAseq, which we believe are the targets of this drug.
What is the most appropriate way to analyze the correlations between the drug concentrations and the gene expression level across these 30 samples (6 patients x 5 time points)?
Would you recommend another PK/PD analysis other than the one I mentioned above? Please note that the genes are supposed to be "tagret engagement markers" due to the known mechanism of the drug so we hypothesize that they are correlated with the drug level.
Thanks
Iddo
Hi,
If you have 5 samples per (single dose and per patient within 30 days) looks like this could be a long half-life drug like mAb? If the mechanism of action (MOA) may have some relationship with gene expression then you can try to find in the first step the best PK parameter to try to find the best relationship - with a high dynamic range. It means the PK parameter which correlates with a very high level of change in specific genes (upregulation or downregulation related to expression-specific protein related to MOA).
Generally, the best parameters are parameters resistant to any source of BIAS like AUC. In your case AUC(0-30) or if you have no data immediately after drug administration as a first measurement then you can use fraction AUC(7-30).
The best way (depending on the drug) could be a comparison between different doses but I'm not sure it's your case. Then the model is more sensitive to explain differences.
If it's mAb with a complex mode of action (downstream effects), here I'm not talking about the mechanism (binding to the target). Then analysis of the relationship between Cmax and/or Clast per patient and gene expression relationship may give you mixed information. It may happen because pharmacodynamic effects may persist even if the concentration of the drug in the bloodstream is very very low (close to Clast or Ctroug in multiple dosing).
So I would try first AUC(7-30), Cmax, and Clast, but I would like to be sure there is no relationship with total body clearance (calculation is based on AUC) or even mean residence time. Clearance could be interesting (but its drug dependent) especially in biotechnological drugs if TMDD may explain better interaction of the drug with the immune system or immunocompetent cells etc (The assumption is that we talk about mAb for example).
So generally try to collect PK parameters (per patient) and then try to test all of them taking into account as first genes exactly related to the mechanism or mode of action and try to find the relationships with high dynamic range at the first step per patient. I'm sure you know that 6 patients could be only some kind of pilot study so probably in this group outliers will be present as well.
Hope maybe some ideas will be helpful,
Best regards
Tomasz
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning