Dual Inhibition of PI3K-Delta and Gamma By Duvelisib (IPI-145) Impairs CLL B- and T-Cell Migration, Survival and Proliferation in a Murine Xenograft Model Using Primary Chronic Lymphocytic Leukemia Cells

Duvelisib (IPI-145), a dual inhibitor of phosphoinositide 3-kinase (PI3K)-δ and -γ, has shown clinical activity in treatment-naïve and relapsed/refractory chronic lymphocytic leukemia (CLL) patients. Clinically, duvelisib results in a redistribution of malignant B cells and concomitant reduction in nodal enlargement. These effects are believed to be due to important roles of PI3K-δ and -γ in CXCL12-mediated CLL cell migration (Peluso 2014), cytokine-induced CLL B-cell proliferation, and BCR-stimulated B-cell survival (Balakrishnan 2015). Additional data suggest an effect of duvelisib on the tumor supporting cells of the CLL microenvironment. This includes preclinical studies demonstrating that PI3K-γ inhibition blocks normal T cell migration toward tumor chemokines and prevents murine bone marrow-derived M2 macrophage polarization (Peluso 2014), as well as clinical data in CLL patients receiving duvelisib showing reduced serum levels of myeloid and T cell-secreted cytokines and chemokines (Douglas 2015).

To further characterize duvelisib’s effect on CLL cells and the tumor microenvironment (TME), a murine xenograft model using primary human CLL cells was employed. We first studied duvelisib’s effect on CLL B- and T-cell migration in vivo. CLL PBMCs (n=2; 1 IGHV unmutated (U)-CLL, 1 IGHV mutated (M)-CLL) pre-treated with duvelisib for 48 hours were injected retro-orbitally into NOD-scid IL2Rgamma^null (NSG) mice. B- and T-cell localization in tissues and circulation was studied 1 and 24 hours post-injection.  Duvelisib treatment (1000 nM) prevented the egress of CLL B and T cells from the circulation into the spleen, indicating impaired homing of CLL B and T cells. To better define the effect of duvelisib on T-cell migration, T cells from CLL patients (n=3; 2 U-CLL, 1 M-CLL) treated ex vivo with duvelisib at 1, 10, 100 and 1000 nM were injected into mice and analyzed for their trafficking 24 hours later. Inhibition of T-cell homing to spleen was dose dependent, with only 100 and 1000 nM having significant effects. Given duvelisib’s cellular IC₅₀s for PI3K isoforms, these results suggest that impaired T-cell migration is due to PI3K-γ inhibition, and studies with isoform-selective PI3K-δ and PI3K-γ inhibitors are currently underway to examine this possibility.

The effect of duvelisib on CLL T-cell proliferation was evaluated after in vitro activation with anti-CD3/28 Dynabeads plus IL2 (n=6; 3 U-CLL, 3M-CLL). In duvelisib treated cells, CD4+, but not CD8+, T-cell proliferation was inhibited at doses of 100 and 1000 nM, suggesting a role for PI3K-γ. The effects of duvelisib on CLL B- and T-cell growth in vivo (n=4; 2 U-CLL, 2 M-CLL) were then studied. Autologous CLL T cells were stimulated as above and injected with CLL PBMCs into NSG mice. Animals treated orally with duvelisib for 3 weeks at 100 mg/kg/day had preferentially reduced CD4+ T-cell recovery from spleens, thereby decreasing the CD4 to CD8 ratio. In each case, duvelisib treatment reduced the number of splenic CLL B cells. This reduction reflected inhibition of both CLL cell proliferation and survival, since duvelisib treatment decreased the percentage of cycling CLL cells and increased the percentage of apoptotic B cells. Thus, duvelisib may target CLL B-cell growth directly, or indirectly by inhibiting the support of CD4+ T cells in the TME.

The potential effect of duvelisib on the tumor-supporting myeloid compartment was also tested.  Because of limited human myeloid-cell engraftment in our NSG model, we studied the effect of duvelisib on murine macrophages. Mice receiving duvelisib had reduced numbers of splenic CD11b⁺GR-1^lowLY-6C^lowLY-6G^negmacrophages compared to controls, suggesting duvelisib altered macrophage development.

Prior in vitro studies demonstrated inhibition of CLL B-cell survival and proliferation by duvelisib, as well as blockade of T-cell migration and M2 macrophage polarization (Balakrishnan 2015; Peluso 2014). Our current in vivo studies further support duvelisib’s effect on CLL B-cell growth and survival through inhibition of cellular homing to supportive tissue niches and alterations in the TME.  The latter, in part, is through suppression of T-cell support and alterations in the macrophage compartment. Overall, these preclinical results suggest that inhibition of PI3K-δ and PI3K-γ by duvelisib affects CLL cell survival through direct and indirect mechanisms.