[1576] The Presence of Langerhans Cells Is Positively Correlated with the Extent of Eosinophilia in Classical Hodgkin Lymphoma

Emina E Torlakovic, Gregory Young, Carol Cheung, Stephan Dirnhofer, Alexander Tzankov. University Health Network/University of Toronto, Toronto, Canada; Institute for Pathology/University of Basel, Basel, Switzerland

Background: In Classical Hodgkin Lymphoma (cHL), Reed-Sternberg cells constitute only a minor component of the tumor, whereas the majority of the tumor is composed of a mixed inflammatory infiltrate. Tissue eosinophilia, while present to varying degrees, is often a feature in mixed cellularity (MC) as well in some cases of nodular sclerosis (NS). Since Langerhans cells secrete CCL5 and IL-16, both of which are chemoattractant for eosinophils, we evaluated whether the presence of Langerhans cells correlated with the degree of eosinophilia in cHL.
Design: A tissue microarray (TMA) was created from cHL cases (n=243) diagnosed between 1974 and 2001 from the archives of the Institutes of Pathology at the University Hospitals of Basel and Innsbruck and the Unit of Hematopathology at the University of Bologna. Absolute numbers of eosinophils as well as CD1a, Langerin, CD68, and PU.1 positive cells within the benign inflammatory infiltrate were counted for each tissue core. Spearman correlation test and One-Way ANOVA were used as appropriate. Two-Step Cluster Analysis was perfromed by using six variables (diagnosis, esonophilia, CD1a, Langerin, CD68, and PU.1), which produced good cluster quality.
Results: 44% of cores had no CD1a+ cells. 25% of cores had no eosinophils or CD1a+ cells. The number of CD1a+ cells positively correlated with Langerin expression (r=.296, p<0.0001), and showed a strong positive correlation with the number of eosinophils/hpf (r=.507, p<0.0001). This correlation remained statistically significant for all levels of CD1a expression (low, intermediate, high). Cluster analysis revealed that NS could be stratified into two subgroups based on the number of eosinophils and CD1a cells: i) a larger cluster (84%, NS_NoEo) with low numbers of eosinophils and Langerhans cells, and intermediate numbers of CD68+ and PU.1+ cells, and ii) a smaller cluster (16%, NS_Eo) with high numbers of eosinophils, Langerhans cells and PU.1+ cells, and intermediate numbers of CD68+ cells. MC appeared as a separate homogeneous cluster that could not be further sub-divided irrespective of the number of eosinophils.
Conclusions: Two thirds of cHL is contains CD1a+ dendritic cells that show a strong positive correlation to tissue eosinophilia. Our results suggest a novel pathophysiological mechanism of tissue eosinophilia in a subset of cHL. Cluster analysis revealed two types of NS cHL based on eosinophil count and Langerhans cell count, while MC did not show additional clusters/subtypes.
Category: Hematopathology

Wednesday, March 21, 2012 9:30 AM

Poster Session V # 204, Wednesday Morning


Close Window