Soluble Factors Produced by Stromal Cells Greatly Influence Adenoviral Oncolytic Efficacy

LOPEZ, MV; VIALE, DL; CAFFERATA, EGA; PODHAJCER, OL

San Diego, California, USA

Congreso; 12th Annual Meeting of the American Society of Gene Therapy; 2009

American Society of Gene Therapy

Since tumor growth is essentially the result of a continuouscross-talk between malignant and tumor-associated stromal cells,targeting both cell compartments may profoundly infl uence viraleffi cacy. Therefore, we developed SPARC promoter (F512Pr )-basedCRAd (Ad(I)-F512-TK) since the SPARC gene is expressed bothin malignant cells and in tumor-associated stromal cells. Next, wedemonstrated that Ad(I)-F512-TK combined with GCV exhibitedmoderate effi cacy on mixed melanoma tumors (Mel/endotheliumor Mel/fibroblast). Remarkably, treatment of SPARC-negativepancreatic tumors (Pan/endothelium) inhibited tumor growth in 5out of 6 mice, demonstrating that in contrast to what occurred withmelanoma tumors. The striking differences in CRAd activity observedbetween melanoma and pancreatic cancer when microendothelialcells were co-administered with malignant cells and the evidencethat microendothelial cells and fi broblasts hampered CRAd activityirrespective of tumor architecture, led us to hypothesize that solublefactors might have been playing a role. Fibroblast- and endotheliumconditionedmedia induced a slight inhibition of F512Pr activityin SB2 melanoma cells. In clear contrast, both conditioned mediastrongly enhanced F512Pr activity in pancreatic MIA PaCa-2 cancercells. In addition, fi broblast-conditioned media enhanced viral lyticactivity on SB2 cells and other melanoma cells as well and bothconditioned media enhanced at a different extent CRAd activity onMIA PaCa-2 cells. Surprisingly, preinfection of stromal cells with theCRAd completely obliterated the enhancement in CRAd lytic activityinduced by the conditioned media on SB2 melanoma cells. But inclear contrast, conditioned media obtained from pre-infected stromalcells dramatically enhanced CRAd lytic activity on MIA PaCa-2 cells.Previous evidence demonstrated that oncolytic viruses increasedthe amount of cells in S-phase and that E1B mutant viruses, such asthe present CRAds, exhibited better cytopathic effect on cells in Sphase. To dissect CRAd’s effects, we fi rst assessed whether arrestedmalignant cells infected with the CRAd showed an accelerated exitfrom quiescence compared to non-infected cells. Indeed, infectedMIA PaCa-2 cells exhibited a higher cell number in S-phase at 20 hrcompared to non-infected cells. Surprisingly, SB2 melanoma cellsexhibited a retarded exit from G0/G1 compared to MIA PaCa-2 cells,regardless of whether SB2 cells were infected or not with the CRAd.Next, we assessed whether stromal cells-conditioned media mightalso accelerate MIA PaCa-2 cells exit from quiescence. Twenty fourhours after MIA PaCa-2 cells release from G0/G1, we observed a clearincrease in the amount of cells in S-phase when they were exposed toconditioned media obtained from fi broblast cells pre-infected with theCRAd, compared to MIA PaCa-2-own conditioned media obtainedfrom cells infected also with the CRAd or the control media. Thisdata suggest that soluble factors produced by stromal cells can playa dramatic and differential role in defining the therapeutic effi cacyof a CRAd on specific tumor types.