Plasma Cell Distribution within the Vascular Niche and Micro-Vessel Density Correlate with Gene Expression Profile
M E Salama, F Zhan, H Lange, J Kohan, N Landis, S Tripp, J Kruger, S Potts. University of Utah, Salt Lake CIty; Flagship Biosciences, Flagstaff
Background: Bone marrow (BM) angiogenesis is typically increased in multiple myeloma (MM) and microvessel density (MVD) is a known indicator of poor prognosis. However, the difficulty of detecting 3D vessels from 2D cut sections has previously limited the study of spatial distribution of plasma cells (PC) and their interaction with BM microenvironment. We utilized a novel whole tissue image (WSI) analysis approach (AngioMap) to identify vessels, then applied computational grown regions extended out from each vessel to identify the spatial distribution of PC. PC distribution was correlated with gene expression profile in a cohort of cases.
Design: After identifying vessels and drawing regions of 15 and 35 µm from vessels on CD34/CD138 double stained core biopsy slides, we identified both plasma and normal haematopoetic (NH) cells in these regions, and computed the following statistics across the entire hematopoietic regions of the slides (excluding bone and poor histology areas) (A & B): total PC/NH count, PC count within 15 & 35 µm of vessels borders, and MVD. We performed unsupervised hierarchic clustering of mRNA expression profiles (HG-U133A arrays, Affymetrix, Santa Clara, CA, USA) in CD138 enriched plasma cells from 16 patients' baseline diagnostic samples.
Results: AngioMap analysis of the 16 samples showed a spectrum of PC to NH at <15, 15-35 and >35 µm of (0.04-84), (0.04-131.41), and (0.02-21.76), respectively. Unsupervised cluster analysis produced two major dendrogram branches (C). Supervised hierarchic clustering performed on two subgroups based on low/high MVD and low/high PC density within 15 µm distance as designated by AngioMap findings identified 70 genes with differential expression between both groups based on statistically significant P value.
Conclusions: PC spatial distribution patterns are associated with distinct gene expression profiles. The use of dual staining immunohistochemistry in conjunction with WSI techniques is a powerful technique for better characterization of the plasma cell in core biopsies. This is the first study that profiles the distribution of PC around vessels across whole sections, and has widespread applications in other areas, including stem cell biology.
Tuesday, March 20, 2012 9:15 AM
Platform Session: Section C, Tuesday Morning