Refined Prognostic Role of CD68-Positive Tumor Macrophages in the Context of the Cellular Micromilieu of Classical Hodgkin Lymphoma

Objective: Classical Hodgkin lymphoma (HL) consists of neoplastic Hodgkin and Reed-Sternberg cells (HRSC) and a nonneoplastic micromilieu that greatly outnumbers the HRSC. Studies on HRSC-related prognostic biomarkers have been unsuccessful, but the microenvironmental composition is of prognostic importance. Recently, the number of CD68-positive macrophages was correlated with adverse survival in HL, and there was a call to validate these results. Methods: We utilized immunohistochemistry to analyze the prognostic importance of the CD68-positive macrophage number compared to other cellular environmental components in an unselected series of 105 HLs in tissue microarrays. Results: Applying a cutoff score of 1 0.82% tumor macrophages, cases with increased numbers showed worse overall survival (mean 185 months, median 192) compared to cases with lower numbers (mean 285 months, median not reached). Eleven of 62 patients with ̂ 0.82% tumor macrophages died, compared to 19 of 43 with 1 0.82% (p ! 0.001). The number of macrophages correlated with a low FOXP3-/high granzyme B-/high PD-1-positive micromilieu and patient age, but did not have independent prognostic significance. A combination background score taking into consideration all negaReceived: August 23, 2010 Accepted after revision: September 16, 2010 Dr. A. Tzankov Department of Pathology University of Basel, Schoenbeinstrasse 40 CH–4031 Basel (Switzerland) Tel. +41 61 328 68 80, Fax +41 61 265 31 94, E-Mail atzankov @ uhbs.ch © 2011 S. Karger AG, Basel 1015–2008/10/0776–0301$26.00/0 Accessible online at: www.karger.com/pat D ow nl oa de d by : 54 .7 0. 40 .1 1 1/ 1/ 20 18 1 1: 55 :2 3 P M Tzankov/Matter/Dirnhofer Pathobiology 2010;77:301–308 302 diovascular diseases [4] . Thus, it is important to be able to predict/identify patients who are treatment resistant and/ or have a higher risk of relapse and find alternative therapeutic options. In addition, patients who might be overtreated by the current treatment regimens must be identified to reduce treatment-associated toxicity. Prior to the introduction of modern therapy protocols, histological type and nodular sclerosis grading were predictive for disease-specific outcome in cHL [5–8] . Today, disease stage, presence of B symptoms, patient clinical history, including risk factors such as coincidental HIV infection, and laboratory parameters are the most important predictive factors [9, 10] . Studies searching for a meaningful HRSC-related phenotypic prognostic biomarker in cHL are contradictory and poorly reproducible [1, 11, 12] . Contrary to HRSC-related biomarkers, reactive cellular environment characteristics seem to be of greater prognostic significance in cHL [13–19] . Importantly, cHL lymphomagenesis is more closely related to the cellular background infiltrate than in other lymphomas. HRSC produce many cytokines and chemokines, resulting in the influx and activation of CD4+ cells and histiocytes; conversely, HRSC respond to the chemokines and growth factors produced by these surrounding cells, the latter providing essential feedback signals to stimulate proliferation and inhibit apoptosis in HRSC [20, 21] . A possible prognostic role of macrophages and histiocytes in cHL was suggested decades ago [22–24] . A recent gene expression profiling study on an unselected cHL collective with subsequent immunohistochemical verification on a collective enriched for treatment failures showed an important correlation between the number of CD68-positive macrophages and adverse disease-specific survival (DSS) and worse outcome after secondary high-intensity therapy with autologous stem cell support, but results have thus far not been validated in an independent collective [25] . Therefore, we systematically analyzed the prognostic importance of the number of CD68-positive macrophages in an unselected series of 105 clinically and histopathologically well-characterized cHLs in tissue microarrays (TMAs). Materials and Methods We used a patient set of our previously reported TMA collective [19, 26] that met the following criteria: complete clinical documentation, presence of at least 5 unequivocal HRSC/case on the examined slides and presence of at least 1 CD68-positive cell/case on the examined slides. Only 105 of the initial 259 cases could be considered based on these criteria. Patients cHL cases (n = 105) diagnosed between 1974 and 2000 and reclassified according to the World Health Organization 2008 criteria were collected from the archives of the Institutes of Pathology at the University Hospitals of Basel, Switzerland, and Innsbruck, Austria. Paraffin blocks were selected based on availabil ity and preservation. Clinical and follow-up data (summarized in table 1) were obtained from chart reviews. Retrieval of tissue and clinical data was performed according to the regulations of the local institutional review boards and data safety laws. Treatment was either standard or consistent with the investigational protocols active during the time the patients were diagnosed. Patients were staged surgically as stage I or II when radiotherapy (RT) was administered as the sole treatment. Otherwise, patients were staged clinically and treated with risk-adapted treatment strategies according to the protocols of the German Hodgkin Study Group valid at the time of diagnosis. In advanced stages, RT was applied for treatment of residual disease or primary bulky disease [27, 28] . Disease remission was defined as absence of disease for at least 1 month as determined by laboratory and imaging studies as well as physical examination. Disease relapse was defined as disease progression occurring at least 1 month after achieving disease remission. Treatment failure was defined as disease relapse or primary treatment resistance; the latter was not observed in the present collective. Overall survival (OS) was defined as the time from initial diagnosis to death or the last follow-up examination or other patient contact. DSS was defined as the time from initial diagnosis to death from lymphoma or the last follow-up examination or other patient contact. Relapse-free survival (RFS) was defined as the time from initial diagnosis to disease relapse or the last follow-up examination or other patient contact without imaging, clinical or anamnestic evidence of disease recurrence. TMA Construction The TMAs were constructed and validated as described elsewhere [19, 26] . Two cores from every sample were arrayed. The construction of the TMAs was approved by the institutional review boards of the participating centers at the time of case submission, and cases have been kept anonymous to meet the legal and ethical requirements for their use in the present study. Immunohistochemical Evaluation The slides were immunohistochemically stained for CD68 using an automated immunostainer (Nexes, Ventana, USA), utilizing the PGM1 antibody from Dako (M0876; 1: 200; antigen retrieval with citrate buffer, pH 6, in a microwave oven at 100 ° C for 15 min). The total number of cells and CD68-stained cells on each array spot was counted at ! 200 magnification. The percentage of CD68-positive cells for each spot was subsequently extracted. The mean percentage over both spots, provided that both contained HRSC, was considered for final statistical analysis, except for cases in which the difference between both spots exceeded 1% or would lead to reclassification of a case as either negative or positive when considering the cutoff score for CD68-positive tumor macrophages. In both of the latter instances, the HRSC-richer array spot was further considered (see Results). Details on FOXP3, granzyme B and PD-1 staining have been reported previously [18, 19] . To assess the reproducibility of the CD68 data, comparison between the results of two observers (A.T. and M.S.M.) in one D ow nl oa de d by : 54 .7 0. 40 .1 1 1/ 1/ 20 18 1 1: 55 :2 3 P M Macrophages in Hodgkin Lymphoma Pathobiology 2010;77:301–308 303 third of the cases was performed in a blinded fashion and assessed using interclass correlation coefficients. Statistics Statistical analysis, including data description, was performed using the Statistical Package of Social Sciences 15.0 software (Chicago, Ill., USA). The degree of agreement between observers was evaluated by interclass correlation coefficients, using Cronbach’s alpha reliability analysis, with alpha 1 0.75 indicating very good reproducibility. Spearman rank correlation coefficients were used to analyze relationships between markers and/or clinical variables. The Mann-Whitney U and Kruskal-Wallis tests were applied, where appropriate, to assess mean differences between groups. The prognostic performance of the variables and determination of optimal cutoff values of continuous variables were established by receiver operating characteristic (ROC) curves plotting sensitivity versus 1 – specificity. The optimal cutoff point was calculated using Youden’s index (Y), where Y = sensitivity + specificity – 1, since this method can be applied when there is no particular requirement for sensitivity and/or specificity [29] . ROC analysis was applied for OS, DSS and RFS. Survival was then analyzed by the Kaplan-Meier method applying the cutoff values calculated by ROC/Y. Survival results from the present study as well as prognostic factors already established in the study collective [18, 19] , for which the ROC or Kaplan-Meier methods suggested statistical (p ! 0.05) or borderline significance (p ! 0.1), were further considered in the multivariate analysis using a Cox regression model.