MP84-06 CXCL8 COULD INDUCE AND MAINTAIN PROSTATE CANCER STEM CELLS VIA PI3K/AKT PATHWAY

Yun Cui,Mengkui Sun,Jixuan Liu,Shuai Hu,Yu Fan,Tianjing Lv,Wei Yu,Jie Jin
DOI: https://doi.org/10.1016/j.juro.2016.02.2230
2016-01-01
Abstract:approaches as well as to identify resistance mechanisms in PC to CR.We hypothesize that aberrant activity of a small subset of kinases promotes PC growth in obese hosts and resistance to caloric restriction.To identify targetable kinases for PC therapy in a host specific manner, we conducted a functional genomic kinome screen.METHODS: LAPC-4 cells were inoculated (MOI of 0.3) with a shRNA library of ~5000 lentivirus targeting 513 kinases.5x106 cells (~1,000 cells per shRNA) were grafted to chronically obese mice.Tumors were established to ~200 mm3 and a portion collected for baseline reference.Remaining mice were randomized to continue on ad lib WD or 25% CR diet.Mice were sacrificed 25 days after randomization.Genome-integrated shRNA inserts from tumoral DNA were amplified using nested barcoded (6 bp) PCR primers.Amplified shRNAs were sequenced using Illumina Hi-Seq 2000 and quantified.We focused on depleted probes with 2-fold less reads in diet arm tumors vs. reference tumors with a p-value <¼ 0.05.RESULTS: Tumor volumes and weights at sacrifice were larger in WD mice compared to 25% CR mice.Of the >5000 shRNAs screened, 99 were significantly (p¼0.05)depleted in obese, 103 in CR and 37 in both arms.As expected, Akt, mTOR, Insulin Receptor and IGF-1R were found significantly depleted.Among the depleted hits, we focused on RIOK2 which was significantly depleted (-5.46 log2fold, p¼ 0.0062) in tumors from obese mice, but not 25% CR mice.RIOK2 expression varies between AR null (DU145) and AR wildtype (LnCaP) PC cells and was induced with synthetic androgen R1881.RIOK2 knockdown with shRNAs shows dramatic cell death in multiples PC cells and RIOK2 positively correlates with Gleason score in a radical prostatectomy TMA.CONCLUSIONS: RIOK2 is an AR target gene that may promote PC growth in obese hosts.Our next aim is to validate RIOK2 in obesity driven PC and asses other hits.Altogether, our in vivo screen has yielded host specific actionable targets.
What problem does this paper attempt to address?