Low-DensityArray to Predict Outcome in Advanced Hodgkin ’ s Lymphoma Using Paraffin-Embedded Samples

Purpose:Despitemajoradvances inthe treatmentofclassicHodgkin’s lymphoma(cHL),f30%of patientsinadvancedstagesmayeventuallydieasresultof thedisease,andcurrentmethodstopredict prognosisare ratherunreliable.Thus, theapplicationof robust techniquesfor theidentificationofbiomarkers associatedwith treatment responseis essential ifnewpredictive tools are tobedeveloped. Experimental Design:We used gene expression data from advanced cHL patients to identify transcriptionalpatterns fromthe tumoralcells and theirnonneoplasticmicroenvironment, associatedwith lackofmaintained treatment response.Gene-SetEnrichmentAnalysiswasused to identify functionalpathwaysassociatedwithunfavorableoutcomethatwere significantlyenrichedineither the Hodgkin’s and Reed-Sternberg cells (regulationof the G2-M checkpoint, chaperones, histone modification,andsignalingpathways)or thereactivecellmicroenvironment(mainly representedby specificT-cellpopulations andmacrophageactivationmarkers). Results:Toexplore thepathways identifiedpreviously,weusedaseriesof52 formalin-fixedparaffin-embeddedadvancedcHL samples anddesigneda real-time PCR-based low-density array that included the most relevant genes. A large majority of the samples (82.7%) and all selected genes wereanalyzedsuccessfully with this approach. Conclusions:The resultsof this assaycanbe combinedina single risk score integrating thesebiologicalpathwaysassociatedwithtreatment responseandeventuallyusedinalarger series todevelopanewmolecularoutcomepredictor foradvancedcHL. Classic Hodgkin’s lymphoma (cHL) is considered to be a monoclonal proliferation of the characteristic Hodgkin’s and Reed-Sternberg (HRS) cells. It has a defective B-cell immunophenotype and a characteristic paucity of neoplastic cells within the tumor, diluted in a reactive inflammatory background composed of nonneoplastic B and T cells, macrophages, eosinophils, neutrophils, and plasma cells, which comprise the bulk of the infiltrate. The B-cell origin and monoclonality of the HRS cells have been clearly established in the last two decades (1–3). Likewise, progress has been made in recent years to clarify the particular composition of the enigmatic cell microenvironment (4, 5). This is commonly made up of a characteristic TH2 environment (6) that is involved in the production of survival signals. Although cHL is usually a curable tumor, f20% to 30% of patients relapse and eventually die due to progressive disease or complications of therapy (7, 8). Patients with advanced disease and clinical indicators of poor prognosis, and those with disease that persists despite optimized primary treatment, may need intensified treatment (9). In contrast, another fraction of patients could benefit from reduced treatment. Current predictive systems are based on clinical and analytic variables, such as the International Prognostic Score developed for advanced cHL (10), but this still fails to identify a significant fraction of patients with very short failure-free survival (11). In this context, the application of robust molecular techniques to identify molecular events and biological processes associated with treatment response is a necessary requisite for developing new predictive tools that enhance the Imaging, Diagnosis, Prognosis Authors’ Affiliations: The Lymphoma Group and Tumour Bank Network, Department of Molecular Pathology, Spanish National Cancer Centre ; Departments of Pathology and Hematology, M. D. Anderson International; Department of Internal Medicine and Pathology, Hospital Ramon y Cajal; Departments of Pathology and Hematology, Hospital ClI¤nico Universitario San Carlos ; Departments of Pathology and Oncology, Hospital Mo¤ stoles ; Departments of Pathology and Hematology, Hospital Severo Ochoa de Leganes, Madrid, Spain; Departments of Pathology and Hematology, Complejo Hospitalario Xeral-Cies,Vigo, Spain; Departments of Pathology andHematology, Hospital Central de Asturias, Oviedo, Spain; and Departments of Pathology and Oncology, Hospital Marque¤ s deValdecilla, Santander, Spain Received 5/6/08; revised 8/29/08; accepted11/7/08. Grant support: Ministerio de Sanidad y Consumo grants PI051623, PI052800, PI052327, RETICS, Accion¤ Transversal and Ministerio de Ciencia yTecnologI¤a grant SAF2005-00221. The costs of publication of this article were defrayed in part by the payment of page charges.This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section1734 solely to indicate this fact. Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). Requests for reprints: Miguel A. Piris, Centro Nacional de Investigaciones Oncolo¤ gicas, Melchor Ferna¤ ndez Almagro 3, Madrid 28029, Spain. Phone: 34-91224-69-00; Fax: 34-91-224-69-23; E-mail: mapiris@cnio.es. F2009 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-08-1119 www.aacrjournals.org Clin Cancer Res 2009;15(4) February15, 2009 1367 Cancer Research. on December 30, 2017. © 2009 American Association for clincancerres.aacrjournals.org Downloaded from accuracy of classic clinical variables. Reliable prognostic markers could allow subsets of patients to be identified who might benefit from alternative approaches. Several biological markers identified in tumor tissues, alone (12– 15) or in combination (16), have been associated with clinical outcome and treatment response. Not surprisingly, most of these variables reflect functional characteristics of the neoplastic cells in tumor tissues, revealed by proteins with deregulated expression in HRS cells. The HRS cells and the inflammatory infiltrate secrete cytokines, creating an elaborate cross-talk that contributes to the survival, proliferation, and immune evasion of the tumor cells in many interacting ways (17–20). Recent studies have indicated that the composition of this background is also associated with the clinical outcome of the patients (21–23). Indeed, previous work by our group identified specific gene signatures associated with treatment response that are attributable to the nonneoplastic component of the