Abstract

Background/objective

To explore the anti-tumour activity of combining AKT inhibition and docetaxel in PTEN protein null and WT prostate tumours.

Methods

Mechanisms associated with docetaxel capivasertib treatment activity in prostate cancer were examined using a panel of in vivo tumour models and cell lines.

Results

Combining docetaxel and capivasertib had increased activity in PTEN null and WT prostate tumour models in vivo. In vitro short-term docetaxel treatment caused cell cycle arrest in the majority of cells. However, a sub-population of docetaxel-persister cells did not undergo G2/M arrest but upregulated phosphorylation of PI3K/AKT pathway effectors GSK3β, p70S6K, 4E-BP1, but to a lesser extent AKT. In vivo acute docetaxel treatment induced p70S6K and 4E-BP1 phosphorylation. Treating PTEN null and WT docetaxel-persister cells with capivasertib reduced PI3K/AKT pathway activation and cell cycle progression. In vitro and in vivo it reduced proliferation and increased apoptosis or DNA damage though effects were more marked in PTEN null cells. Docetaxel-persister cells were partly reliant on GSK3β as a GSK3β inhibitor AZD2858 reversed capivasertib-induced apoptosis and DNA damage.

Conclusion

Capivasertib can enhance anti-tumour effects of docetaxel by targeting residual docetaxel-persister cells, independent of PTEN status, to induce apoptosis and DNA damage in part through GSK3β.

Discussion

Here, we show that combining the AKT inhibitor capivasertib with docetaxel increases anti-tumour effects in PTEN null and PTEN WT prostate tumour models and cell lines. In vivo the combination increased tumour growth inhibition in all models assessed, with significant tumour regressions in 3/6 models. In vitro sequential addition of capivasertib following short-term (24 h) docetaxel incubation reduced cell growth in PTEN null and WT cells. In vitro, acute docetaxel treatment killed a substantial number of cells and induced G2/M arrest. However, in a residual docetaxel-persister cell fraction that remained adherent to the plastic, a dose-dependent increase in phosphorylation of p70S6K, 4E-BP1 and in some cell lines, GSK3β occurred, which was reversed by capivasertib. These cells had progressed through G2/M arrest and remained in cycle without cell division. Generally, capivasertib monotherapy induces a G1/S arrest, and in combination capivasertib reduced cell cycle in docetaxel-persister cells. Moreover in 5/7 cell lines combination treatment increased apoptosis and induction of DNA damage. The fact that the increased apoptosis induced by capivasertib was reversed or reduced by inhibiting GSK3β suggests a direct role for AKT signalling. Although the PI3K/AKT inhibitor taxane combination is effective in preclinical models [1, 46, 47], PI3K-AKT pathway inhibition prior to taxane reduces efficacy as the PI3K-AKT mediated G1/S arrest blocks progression through S phase and G2 where taxanes induce cell death [26]. Inhibiting AKT post-taxane treatment increased tumour growth inhibition in breast, gastric and prostate cancer models [1, 25, 27, 29].

It has been suggested that long term resistance to docetaxel is associated with an increase in AKT phosphorylation [25, 30, 35]. Here in PTEN WT and PTEN null prostate cancer cells, following acute treatment with docetaxel the increased AKT phosphorylation in the docetaxel-persister cells was less apparent in the cell lines tested. Commonly phosphorylation of p70S6K(T421/S424) and 4E-BP1(T37/46), downstream of AKT, was observed. Induction of phospho-GSK3β (S9) was also observed with docetaxel treatment in some cell lines, while in cell lines where increased phosphorylation of GSK3β was less apparent baseline phospho-GSK3β levels were high prior to treatment. Interestingly increased phosphorylation of S6 was not seen in all cell lines, which may indicate that increased PI3K-AKT pathway activation in docetaxel-persister cells influences cell cycle or survival rather than general PI3K-mTOR activation including effects on protein synthesis [36, 48]. Modulation of 4E-BP1 also suggests a cell cycle or cell stress response following docetaxel treatment [37, 38]. One other study has examined the acute response to docetaxel in ER+ breast cancer MCF7 cells where a transient increase in pAKT was observed [49]. However, the induction of signalling downstream of AKT in the absence of increased pAKT signalling has been observed in other settings. In ER+ breast cancer cell lines that have become oestrogen independent following long term oestrogen deprivation, or resistant to the CDK4/6 inhibitor palbociclib increases in pS6 and other markers downstream of AKT are observed but little pAKT is detected or minimal to no change in pAKT [50, 51]. The significance of this warrants further investigation. While the data show persister cells are impacted by capivasertib treatment, we have not performed unbiased phospho-site profiling or reverse phase protein array phospho analysis to look at all changes following docetaxel and combination treatment, therefore changes in other proteins or pathways may be associated with survival of the docetaxel persister cells.

It was clear that across a panel of tumour cells the response to docetaxel and hence the combination was heterogeneous. In vitro combination activity was enhanced in cells with monotherapy sensitivity to capivasertib. However, in vivo additive anti-tumour combination effects were seen in 5 out of 6 models and appeared related to the intrinsic response of tumour models to docetaxel. In models that were more sensitive to docetaxel, capivasertib addition drove regressions, whereas in less sensitive models the combination resulted in cytostatic effects.

Inhibition of AKT signalling can contribute to combination benefit through different mechanisms, and it is possible that more than one mechanism is important in a specific cell line or tumour model. AKT-mediated phosphorylation of GSK3β, p70S6K and 4E-BP1, can enable evasion of apoptosis, cell cycle progression in the face of a G2/M blocker and absence of cell division. These mechanisms may not always be induced by docetaxel, cells may have higher baseline signalling, contributing to both intrinsic and induced resistance that is reduced by capivasertib treatment. For example, high intrinsic GSK3β activity could render docetaxel less effective independent of other PI3K pathway functions. While GSK3β may be an important mediator of persistence following docetaxel treatment, other mechanisms may also contribute, e.g. changes in translation downstream of mTORC1, coupling through 4E-BP1 or alternate non-canonical p70S6K signalling. For example, it is possible that intrinsic activity of AKT in the context of the cell cycle status of the persister cell fraction drives the survival effect. Alternatively pAKT increases may be more transient than the increase seen in GSK3β and p70S6K, but phosphorylation is still regulated by AKT or finally that docetaxel induced cell stress dysregulates phosphatases such as PTEN and PHLPP that control AKT [52, 53]. In addition TSC1, TSC2 [54], and PP2A [55] that regulate P70S6K could also be disrupted which may reduce the activation threshold for signalling molecules down stream of AKT. Indeed, in breast cancer cells reduction of mTORC1-4EBP1 signalling results in a reduction in PTEN protein [56]. It will be important to further evaluate whether other mechanisms that are regulated directly or indirectly by AKT modulation by capivasertib also make a contribution to the combination benefit. Variable sensitivity of prostate cancer cells to taxanes can also be influenced by differences in drug uptake and consequent intracellular levels of drug [57]. Therefore, different features may be important in different cells, and it is likely there is not one single unifying mechanism driving combination benefit. For example, it is possible that on long term treatment there may be engagement of the immune system, changes in the tumour microenvironment (TME) or adaptive responses in the TME, or the tumour cells.

Here, we have examined the acute interaction between docetaxel and capivasertib and sought to mimic the Phase II ProCAID study schedule [23, 24] in vivo and in vitro, but have not assessed prevention of longer-term resistance. The data presented here support the ProCAID study observation that the combination could be broadly effective in prostate cancer. How the pathway is activated remains an interesting question. It may be as a result of cells being in a specific phase of the cell cycle when treated with docetaxel, through inactivation of phosphatase regulation following redox stress, or signalling through other pathways such as activation of DNA damage repair proteins in response to aberrant mitosis [58].

In summary, combining the AKT inhibitor capivasertib with docetaxel in prostate cancer improves anti-tumour effects by targeting the residual surviving cells following docetaxel treatment. The benefit can be driven through different mechanisms downstream of AKT, by reducing AKT mediated cell cycle progression and enhancing induction of apoptosis or DNA damage in cells that persist after docetaxel treatment.