PF-3758309

This probe meets minimal criteria for use in cellular and in vivo assays. It has additional biochemical activity in the PAK subfamily: PAK2: IC50 = 190 nM, PAK3: IC50 = 99 nM (PMID 20439741). No antiproliferative activity was noted in primary human hepatocytes: IC50 >1,000 nM (PMID 20439741). PF-3758309 was also used to inhibit PAK1 in tumor cells (PMID 23535073). It is classified as a poorly cell permeable compound using the MDCKII-LE assay and is a substrate of the efflux transporters P-gp and BCRP. Of note, a strong correlation was observed between P-gp (but not BCRP) expression in tumor cell lines and resistance to PF-3758309 in cellular assays. A similar, but weaker, correlation was observed in vivo (PMID 23524533).

PF3758309 is reportedly potent against PAK4 (3-18 nM, depending on the assay format). Using a peptide phosphorylation assay, PF3758309 does not appear to be selective for PAK4 over PAK1, 5, and 6; it is 5-10x selective for PAK4 over PAK2-3.

Selectivity over the rest of the protein kinome remains a concern.  PF3758309 was active (1 nM <IC50 <60 nM) against 13 of 146 kinases screened in a broad panel, including 3 SRC family kinases at 45-60 nM.  The authors indirectly showed that SRC-family activity is not relevant for this compound when dosed in cells at 1 uM; however we do not have data about the remaining 10 liability kinases with IC50 <35 nM at 1 uM in cells. Furthermore, with only ~25% of the protein kinome screened, it is possible that ~40 kinases are inhibited with potencies IC50 <40 nM.

PF3758309 inhibits phosphorylation of a natural PAK4 substrate in cells with an IC50 ~1 nM and inhibits anchorage-independent growth (expected biology for PAK4 inhibition) at 240 pM.  It is possible that PAK4 adopts a conformation in cells that binds PF3758309 more potently than PAK4 that is overexpressed and purified; however, it is very rare for an ATP competitive kinase inhibitor to be more potent in cells (competing with 2 mM ATP) than it is in in vitro biochemical (0.04 mM ATP) or a biophysical (0 mM ATP) assays. This is a signficant disconnect that is not resolved in the publication and may very well result from polypharmacology.

Use this compound with caution.

There is a reported allosteric PAK4 inhibitor that, while less potent, could be considered as a potentially more selective alternative (KPT-9274).

PF-3758309 is reported as a potent, ATP-competitive inhibitor of PAK4 with additional activity against PAK1, PAK5 and PAK6, and, to a slightly lesser degree PAK3 and PAK2 (range of affinities PAK4 IC50=3 nM to PAK2 IC50=190 nM). PF-3758309 also shows binding to a number of representative kinases (20 with K_i <200nM) where there may be an overlapping relevant pharmacology. PF-3758309 clearly shows activity in both cellular and in vivo models at reasonable doses and using convenient routes of administration (oral & subcutaneous). PF-3758309 is commercially available, and there is a growing body of evidence with regard to its pharmacology and potential utility. If PF-3758309 is used, there needs to be a good understanding of both the on-target (PAKs) and off-target (other kinase activities) protein expression and function to interpret the results and draw a credible association between the compound, the target(s) and the outcomes.

I would not recommend PF3758309 as a specific probe.  While earlier data already have uncovered significant off-target activities, including potent CDK7 inhibition (J Med Chem 58, 111–129 (2015)), a newer study demonstrated that this compound inhibited the growth of Rosa26 and PAK4-KO melanoma cells equipotently (GI50 value of ~9 nM). Given that this drug is fully capable of killing cells in which its putative target has been deleted, the ability of PF-3758309 to block cancer cell growth must be through an off-target effect. (Lin, A. et al. Off-target toxicity is a common mechanism of action of cancer drugs undergoing clinical trials. Sci Transl Med 11, eaaw8412 (2019)).