Increased Glioma Collagen Is Associated with Greater Blood Flow and Peri-Tumoral Fluid Flux, but Less Infiltration
Nagaraja TN, Elmghirbi R, Brown SL, Lee IY, Cabral G, Panda S, Knight RA, and Ewing JR. Increased Glioma Collagen Is Associated with Greater Blood Flow and Peri-Tumoral Fluid Flux, but Less Infiltration. Cancer Res 2019; 79(13).
Introduction: Composition of the glioma microenvironment (TME) determines tumor growth rate, invasion, metastasis and resistance to treatment. Compared to normal brain, tumor extracellular matrix (ECM) has a higher concentration of structural proteins including collagen, laminin, tenascin, and vinculin. We compared collagen expression in two rat models of glioma, human U251 grown in immune-compromised rats and 9L in syngeneic Fischer-344 rats, representing primary and recurrent glioma features, respectively. Experimental procedures: The U251 and 9L cells were implanted in the right brain hemisphere of female RNU rats (n=10) and Fischer-344 rats (n=7), respectively. Rats were imaged 2 to 3 weeks after implantation by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Tumor size, blood flow, apparent diffusion coefficient (ADC), blood-to-tumor forward volumetric transfer constant (Ktrans) , peri-tumoral contrast flux (Flux), hydraulic conductivity (K), and extracellular volume fraction (VD ) in the tumor (VD-tumor), tumor rim (VD-rim) and its periphery (VD-peri) along with tumor interstitial fluid pressure (TIFP) were measured. Immediately after imaging, brains were processed for paraffin embedding and histopathology. Brain sections (6 |jm) containing the tumor were stained for collagen using Picrosirius red and adjacent brain sections with hematoxylin and eosin (H&E). MRI data were compared by t-tests and significance inferred at p≤0.05. Digital images of Picrosirius red staining were analyzed using ImageJ for collagen expression expressed as % fraction of total tumor area. It was compared to the MRI biomarkers, and to the patterns of tumor cell dispersion into normal brain observed on H&E stained images. Results: Tumor diameters averaged 4 mm for both models at the time of imaging. The U251 tumors showed greater Flux (p=0.03), higher blood flow (p=0.02), smaller VD-tumor (p=0.01) and VD r im (p=0.02) values than the 9L tumors. TIFP also tended to be higher in the U251 model (p=0.06). Collagen expression in U251 tumors was significantly higher than in 9L tumors (p=0.002). On H&E stained sections, the U251 tumors showed a well delineated tumor margin with very few cells invading the host tissue. In contrast, the 9L tumors showed cancer cells infiltrating singly and in clusters, along white matter tracts as well as in perivascular spaces. Conclusions: Increased collagen content was associated with elevated blood flow, Flux and TIFP, but less tumor cell infiltration in the U251 glioma. Observations of low collagen content in conjunction with increased tumor cell invasion in the 9L model suggest that treatment with collagenase, while increasing drug penetration may also make the tumors more infiltrative. A closer examination of TME to determine the composition of ECM proteins that regulate glioma aggressiveness and modulate its response to treatments is warranted.