CDK4/6 inhibition synergizes with inhibition of P21-Activated Kinases (PAKs) in lung cancer cell lines

Abstract

Theoretically, small molecule CDK4/6 inhibitors (CDK4/6is) represent a logical therapeutic option in non-small cell lung cancers since most of these malignancies have wildtype RB, the key target of CDKs and master regulator of the cell cycle. Unfortunately, CDK4/6is are found to have limited clinical activity as single agents in non-small cell lung cancer. To address this problem and to identify effective CDK4/6i combinations, we screened a library of targeted agents for efficacy in four non-small cell lung cancer lines treated with CDK4/6 inhibitors Palbociclib or Abemaciclib. The pan-PAK (p21-activated kinase) inhibitor PF03758309 emerged as a promising candidate with viability ratios indicating synergy in all 4 cell lines and for both CDK4/6is. It is noteworthy that the PAKs are downstream effectors of small GTPases Rac1 and Cdc42 and are overexpressed in a wide variety of cancers. Individually the compounds primarily induced cell cycle arrest; however, the synergistic combination induced apoptosis, accounting for the synergy. Surprisingly, while the pan-PAK inhibitor PF03758309 synergizes with CDK4/6is, no synergy occurs with group I PAK inhibitors FRAX486 or FRAX597. Cell lines treated only with Ribociclib, FRAX486 or FRAX597 underwent G1/G0 arrest, whereas combination treatment with these compounds predominantly resulted in autophagy. Combining high concentrations of FRAX486, which weakly inhibits PAK4, and Ribociclib, mimics the autophagy and apoptotic effect of PF03758309 combined with Ribociclib. FRAX597, a PAKi that does not inhibit PAK4 did not reduce autophagy in combination with Ribociclib. Our results suggest that a unique combination of PAKs plays a crucial role in the synergy of PAK inhibitors with CDK4/6i. Targeting this unique PAK combination, could greatly improve the efficacy of CDK4/6i and broaden the spectrum of cancer treatment.

Fig 1. Combination drug effects of clinical CDK4/6is in lung cancer cells.

A) Drug combination screening with a customized library of 240 targeted agents (0.5 μM and 2.5 μM, increasing wedge size) and either palbociclib (PD-0332991–2.5 μM) or abemaciclib (LY2835219–0.75 μM), depicted as a heat map, suggests synergy particularly with the PAK inhibitor PF03758309. B) Validation of drug synergy. Left panel: full dose-response matrix depicting inhibition of viability as determined using a CTG viability assay with varying drug concentrations (Ribociclib: 0–20 μM; PF03758309 or FRAX486: 0–10 μM). Right panel: deviation from calculated Bliss value for additivity (in % viability difference); needles pointing up indicate synergy, needles pointing down antagonism.

Fig 2. Cell cycle arrest and apoptosis.

A. Percentage of H322 cells in various stages of the cell cycle after 24 hours for single drug (PF03758309, FRAX486 and Ribociclib) or drug combination, as well as with consecutive drug addition at 24 hours and 18 hours before harvest. B. Percentage of H157 or H1299 cells in various stages of the cell cycle after 24 hours for single drug or combination drug with PAKi (including FRAX597), added for 24 hours, and Ribociclib for 18 hours.

Fig 3. Cleavage of Caspase 3 and PARP following treatment with PF03758309 and its combinations.

Apoptosis was determined by immunoblot detection of cleaved Caspase 3 and PARP measured in H322 after 48 hours treatment with the indicated drugs. Drug concentrations were the same as in Fig 2 with the addition of 10μM FRAX486.

Fig 4. PF03758309 and its combinations induces apoptosis based on Celigo imaging.

Colorimetric assay on a Celigo S (Nexcelom) showing apoptosis in H322 cells treated with vehicle control (DMSO), 2.5 μM PF03758309, 2.5 μM FRAX486 or 5 μM Ribociclib as well, as the combinations. The cells were stained with Cyto-ID green, to detect Caspase 3/7 (green), and Hoechst, to detect the nuclei of the cells (blue), and images were captured with the Celigo S.

Fig 5. Autophagy and cell survival.

Cells were treated with increasing concentrations of PF03758309, FRAX486 or FRAX597 in combination with 5 μM Ribociclib. After 6, 12, 24 and 48 hours, the cells were stained with Hoechst and Cyto-ID green to detect the cellular nucleus and autophagosomes, respectively. Cell survival was reported as a percentage of viable cells compared to the vehicle control, and autophagy was reported as number positive vesicles per viable cell normalized to vehicle control. Red lines % cell survival, blue lines % autophagy. Results shown in A and B are representative of H322 (n = 3) and H157 (n = 1) cell lines, respectively. The very high calculated percentage of cells in autophagy observed in H157 cells at high drug concentrations and later time points results from the fact that there are very few surviving cells under the conditions, and they all have autophagy.