Abstract 222: Protein network mapping of bladder cancer: tumor compartment and microenvironment analysis

• Published in: Cancer research (Chicago, Ill.) Ill.), 2017-07, Vol.77 (13_Supplement), p.222-222
• Main Authors: Hodge, Kimberley A., Hau, Andrew, Pierobon, Mariaelena, Hansel, Donna, Petricoin, Emanuel
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2017-222

Background: Interactions between tumor and stroma are of intense interest in the field of oncology and provide new opportunities to understand the driving mechanisms behind tumor progression, metastases and responsiveness to therapy.  Laser capture microdissection (LCM) coupled with reverse phase protein microarrays (RPPA) are ideal technologies for isolating different cellular compartments from heterogeneous tissues and for exploring protein signaling networks. Material and Methods: We explored protein signaling networks of epithelium and matched surrounding stroma of 16 bladder cancers (BC) including papillary (n=8) and invasive (n=8) tumors.  Tumor epithelium and matched surrounding stroma were isolated with LCM for each case. Signaling network analysis of dissected material was performed using RPPA, allowing for the measurement of 111 proteins and phosphoproteins. Wilcoxon Rank Sum Test was used to compare the signaling network of papillary and invasive lesions. Spearman Rho correlation analysis (correlation coefficient > 0.90) was used to explore signaling networks of each cellular compartment across histotypes. Results: Mean comparison of tumor epithelium revealed increased PD-L1 and PDGFRβ expression in invasive tumors compared to papillary lesions (p= 0.015 and p=0.028 respectively). Correlation analysis of the epithelium compartments collected from papillary and invasive cancers revealed different signatures with invasive cases showing a phenotype of proliferation and survival in contrast to papillary tumors which were characterized by wound healing and metabolic network activation. Analysis of stroma showed increased expression of MMP9 in invasive compared to papillary stroma (p= 0.05). While stroma surrounding both invasive and papillary tumors showed high correlation between members of the PI3K/AKT pathway, only invasive BC presented with interconnections between immune response proteins. Finally, receptor tyrosine kinases (RTKs) in the invasive epithelium appeared to be highly correlated with RTKs in the stroma as well as their downstream targets suggesting cross-talk between the two compartments. Conclusions: LCM-RPPA workflow is a unique tool for exploring tumor and surrounding microenvironment interactions in BC. Signaling networks of epithelial cells and surrounding stroma of invasive BC has unique characteristics compared to papillary tumors including increased expression of immune checkpoint PD-L1 and PDGFRβ in the epithelium, two potent