Combinatorial PI3K/AKT Pathway Inhibition As a Therapeutic Approach in Diffuse Large B-Cell Lymphoma

Background. Diffuse large B-cell lymphoma (DLBCL) is the most frequent type of non-Hodgkin lymphoma (NHL, ~ 25-30% of NHL cases). Currently, DLBCL treatment is based on non-specific chemotherapy combination (cyclophosphamide, doxorubicin, vincristine, and prednisone) with anti-CD20 monoclonal antibody. Although the DLBCL survival dramatically improved during last decades, approximately 40 % of DLBCL patients could not be cured with the currently used therapeutic approaches. Novel approaches are, therefore, needed to improve DLBCL survival and to decrease treatment related toxicity. Since many DLBCL tumors depend on phosphatidylinositol 3-kinase /protein kinase B (PI3K/AKT) signaling pathway, it represents one of the possible novel therapeutic targets. Despite promising results in various cancers, PI3K/AKT pathway inhibitors show only partial efficacy, low specificity, and resistance development in lymphomas. Since the PI3K/AKT activity is highly regulated, we hypothesize that the PI3K/AKT pathway needs to be inhibited at multiple levels simultaneously to achieve its full and lasting inhibition.

Methods. To provide a proof of principle of multilevel vertical PI3K/AKT inhibition, we tested the effect of four inhibitors targeting the PI3K/AKT pathway at four different levels: PI3Kδ inhibitor idelalisib, PDPK1 inhibitor GSK2334470, AKT inhibitor ipatasertib, and mTOR inhibitor rapamycin on a spectrum of commonly used DLBCL model cell lines. Precise measurement of AKT activity in living cells was done using our recently published genetically encoded Förster Resonance Energy Transferred (FRET) based AKT activity reporter including novel methodology for flow cytometry-based FRET measurement (PMID: 34128311) allowing medium throughput and measurement in kinetics setting.

Results. First, we measured how single inhibitors affect the AKT activity. The AKT activity IC50 concentrations were in 100 nM range except for rapamycin that had no AKT inhibitory effect. These concentrations were approximately 10 times lower in comparison to IC50 values calculated from net cell growth inhibition assessing the overall inhibitor toxicity. This suggests compensatory mechanisms and/or off target effects of PI3K/AKT inhibitors. Precise measurement of AKT activity allowed us to identify the minimal concentration of each inhibitor at which AKT activity was fully inhibited. Kinetics and short time course experiments showed that maximal AKT inhibition is achieved within hours of inhibitors treatment initiation. Importantly, pharmacological inhibition of one or two members of PI3K/AKT signaling cascade did not lead to lasting AKT activity inhibition. Cellular contra-regulatory mechanisms compensated for the decrease of AKT activity within 24 hours even in the presence of the inhibitors. Moreover, we detected a massive increase in AKT activity following release of the cells from media with PI3K/AKT pathway inhibitors. This further illustrates the powerful cellular contra-regulatory mechanisms and highlights the need for combinatorial approach. Combination of at least three inhibitors was necessary to achieve lasting AKT activity decrease (long time course experiment, Fig. 1A). Cross titration experiments further confirmed synergistic effect between tested PI3K/AKT pathway inhibitors on cell growth inhibition and allowed to determine the concentration ratios of their maximal synergism. Importantly, testing of single inhibitors and two, three, or all four inhibitors in fixed concentration combinations identified a substantial decrease in cell growth/toxicity IC50 concentrations of individual inhibitors (Fig. 1B).

Conclusions. Concept of inhibiting more members of PI3K/AKT pathway was suggested before using combinations and/or dual inhibitors of AKT and mTOR, however, our data suggest that more than two inhibitors are needed for lasting PI3K/AKT inhibition. Moreover, we provided a proof of principle for further multilevel PI3K/AKT signaling inhibition testing, including in vivo cell line and/or patient derived xenograft models. Our data suggest that proposed combinations should allow significant decrease of inhibitors concentrations limiting off-target effect. This approach might overcome above mentioned issues of PI3K/AKT pathway inhibition and allow full therapeutical utilization of this pathway inhibition possibly applicable to other lymphoma types.

Havranek:Bristol Myers Squibb: Research Funding.