Synergism Between PDE4 and PI3Kδ Inhibitors in DLBCL: Improved Clinical Activity with the Potential for Lower Toxicity

Aberrant activation of the B cell receptor (BCR) is a hallmark of mature B-cell tumors. A better understanding of this process will spearhead effective clinical translation. The initiation and amplification of BCR signaling are well-defined events, and the successful deployment of BTK and PI3Kδ inhibitors in the clinic capitalizes on this knowledge. Conversely, the intricacies of the termination of BCR signals are less well-understood, and to date no rational therapeutic approach has been developed that exploit this aspect of the oncogenic BCR. Cyclic-AMP (cAMP) is a second messenger with marked growth suppression properties towards immune cells, including neoplastic mature B lymphocytes. In earlier work, we showed that inhibition of phosphodiesterase 4 (PDE4), the enzyme that hydrolyzes cAMP, downmodulates the activity of classical effectors of BCR signals, including SYK and PI3K. Herein, we attempted to gain further mechanistic understanding on how cAMP suppresses the proximal BCR activity, and built on this information to pre-clinically test therapeutic strategies that simultaneously attack the BCR at its amplification and termination points.

Using diffuse large B cell lymphoma (DLBCL) as a model, we focused our attention on the interplay between cAMP and LCK, as we unexpectedly found that this cAMP-regulated canonical T-cell kinase is also widely expressed in DLBCL. Working with LCK-positive PDE4-low/null DLBCL cell lines, we found a marked increase in the phosphorylation of the inhibitory Y505 of LCK following elevation of intra-cellular cAMP. Next, we showed that ectopic expression of wild-type (WT) PDE4B, but not of a phosphodiesterase-inactive (PI) mutant, abrogated the cAMP-mediated, CSK-dependent, phosphorylation of LCK. Active LCK can phosphorylate PI3K's p85 regulatory subunit, thus freeing the catalytic domain from its structural constraints to promote lipid kinase activity. Thus, we tested whether the cAMP-mediated inhibition of LCK, by suppressing p85 phosphorylation, down-modulated PI3K activity. In LCK-positive PDE4B-null DLBCL, we showed that cAMP readily decreased the phosphorylation of p85 that followed BCR engagement; using the WT and PI PDE4B genetic models, we demonstrated that PDE4B expression abrogated cAMP effects and led to sustained PI3K activity following BCR engagement.

These data suggested that inhibition of PDE4, by unleashing the negative effects of cAMP on LCK/p85, could accelerate the termination of PI3K activation that follows BCR engagement. If this hypothesis was correct, then the combination of PI3K and PDE4 inhibitors by attacking the BCR at its amplification and termination points, respectively, may synergistically suppress the growth of DLBCL. In in vitro studies with multiple DLBCL cell lines (WSU-NHL, OCI-Ly7, OCI-Ly18, OCI-Ly3, HBL-1, and OCI-Ly10) we showed that the combination of the FDA-approved PDE4 inhibitor roflumilast with idelalisib synergistically suppresses DLBCL growth (combination index < 1). This synergism was associated with a significant suppression of PI3K and AKT activities (p<0.05, cells treated with the drug combination vs. single agents). We expanded this observation to an in vivo xenograft model of human DLBCL, and showed that mice treated with roflumilast and idelalisib had a significantly smaller tumor burden than those receiving single agents (p< 0.01, two cohorts, n=47 mice). We also found a greater suppression of PI3K activity in the xenografts from mice treated with the combination of PDE4 and PI3Kδ inhibitors (p< 0.0001), as well as increased apoptosis.

Together, these data further delineated how cAMP suppresses the BCR and showed that the rational combination PDE4 and PI3Kδ inhibitors synergistically suppresses DLBCL growth. These results are particularly important given the recent evidence of inflammatory/immune toxicity associated with the use of idelalisib, which we propose could be countered by the well-established anti-inflammatory properties of PDE4 inhibitors. Thus, we hypothesize that combining PDE4 and PI3Kδ inhibitors will enhance anti-lymphoma activity while decreasing clinical toxicity. This concept is ripe for clinical testing as we have recently completed a phase Ib trial showing that roflumilast is safe and active in patients with advanced B cell malignancies.