In this work, we used human kinome siRNA library to screen for kinases that positively regulate Akt phosphorylation at the ser473 residue in the breast cancer cell line, MDA-MB 468. MDA-MB 468 cells have an intrinsic PTEN mutation resulting in high endogenous Akt activity in the absence of growth factors. The systematic silencing of individual kinases in these cells with the RNA interference library allows us to identify kinases that alter Akt(ser473) phosphorylation. In combination with the high content screening microscope, we found a total of 92 kinases that upon knock down, resulted in 20 to 60% decrease in Akt(ser473) phosphorylation. In the screen setup, due to the "edge" effect of the 96 well plates, we noted that the standard deviation of these wells were high. Hence, these samples were not considered further. Regardless, the screen enables us look at 500 kinases and their effect on Akt phosphorylation.
Further validation had shown that ChoKα, plays an important role in regulating Akt(ser473) phosphorylation. Our data showed that ChoK is unlikely to act on the components upstream of Akt such as the PI3K signaling axis. This is showed by the ability of PH-GFP fusion protein to be recruited to the peripheral membrane in the presence of IGF stimulation in ChoK-silenced cells. These results demonstrated that PI3K is functional and able to generate PIP3 for the recruitment of both Akt and PDK1 as shown with the intact Akt(thr308) phosphorylation in these cells. Interestingly, other than the reported effects on Akt(ser473) phosphorylation, we also observed a decrease in Erk phosphorylation in ChoK-silenced cells. Since silencing of ChoK does not affect PI3K activity, it is unlikely that the reduced Erk phosphorylation is due to an inactivation of the upstream Ras. It is however possible that the reduction of Erk phosphorylation is due to yet unknown effects of this lipid kinase upon the Raf/MEK pathway, which will requires further investigation. Alternatively, the downstream effect on Erk signaling could arise from the cross talk between PI3K/Akt pathway and the Raf/MEK pathway, as shown with PI3K inhibitor, LY294002 treatment.
Although our data from both the RNAi silencing and small molecule inhibitor studies clearly demonstrate an interesting role of ChoK on Akt(ser473) phosphorylation, it is unlikely that the lipid kinase phosphorylate Akt(ser473) directly since our data with the ChoK inhibitors demonstrated a distinct lag time between ChoK activity inhibition and Akt phosphorylation. Only ~50% reduction in Akt(ser473) phosphorylation was observed when 70% of ChoKα activity was inhibited after 2 h of Mn58b treatment. A similar observation was made for TCD828 treated cells with a 56% reduction in Akt(ser473) phosphorylation after 0.5 h incubation with TCD828 which inhibited ~85% of ChoKα activity. In addition, we did not observe a physical interaction between Akt and ChoKα via co-immunoprecipitation (data not shown). Nonetheless, the evidence presented by xenograft regression with reduced Akt(ser473) phosphorylation and strong inhibition in Akt(ser473) phosphorylation after prolong treatment with ChoK inhibitors strongly support our data, suggesting a probable role of ChoKα as a regulator of PDK2, controlling the phosphorylation of Akt at ser473. Alternatively, the effect of ChoKα on Akt(ser473) phosphorylation could also arise through the inactivation of the Akt phosphatase. Previously, PH domain leucine-rich repeat protein phosphatase, PHLPP was identified by Gao et al as the phosphatase that dephosphorylate Akt1(ser 473) . Further experiments will be required to definitively demonstrate these unexpected properties of ChoKα.
These findings are particularly exciting for two reasons. Firstly, there are several potential kinases that phosphorylate Akt(ser473). Of which, mTORC2 effect on Akt is significantly reproducible in many different cell types. In our work, we had shown that silencing of the lipid kinase, ChoK, resulted in reduced Akt(ser473) phosphorylation to a similar degree as observed following the silencing of Rictor, a member of the mTORC2 complex. Secondly, reminiscent of the regulators of the Akt pathway, there is evidence that ChoK can serve as marker for tumor progression. It has been shown that ChoK activity and its product, PCho, are increased in tumor cells relative to the normal cells [12, 23–27]. This has been established in tumors of different tissue origins and in particular those derived from the breast [28, 29]. It has also been demonstrated in vivo by NMR, where increase levels of PCho are frequently associated with cell malignancy [30–32]. All these results have established PCho as a malignancy marker with potential use in cancer diagnosis [28, 33]. Our data demonstrate the presence of a novel cross talk between the lipid kinase and Akt pathway
Although the precise role of ChoK in these cancer cells is still not fully understood, it has been postulated that this lipid kinase is likely to be upregulated in order to provide lipid components for the actively dividing cancer cells. In addition, the PCho appears to induce mitogenic signaling, promoting cellular proliferation. Currently, there is an active effort in the development of ChoK inhibitors. Results from Mn58b, a well characterized ChoK inhibitor with in vitro and in vivo antiproliferative and antitumoral effect in mice xenografts provides strong support to this concept.