M1-like macrophage polarization prevails in young children with classic Hodgkin Lymphoma from Argentina

The microenvironment in classical Hodgkin lymphoma (cHL) comprises a mixture of different types of cells, which are responsible for lymphoma pathogenesis and progression. Even though microenvironment composition in adult cHL has been largely studied, only few groups studied pediatric cHL, in which both Epstein Barr virus (EBV) infection and age may display a role in their pathogenesis. Furthermore, our group described that EBV is significantly associated with cHL in Argentina in patients under the age of 10 years old. For that reason, our aim was to describe the microenvironment composition in 46 pediatric cHL patients. M1-like polarization status prevailed in the whole series independently of EBV association. On the other hand, in children older than 10 years, a tolerogenic environment illustrated by higher FOXP3 expression was proved, accompanied by a macrophage polarization status towards M2. In contrast, in children younger than 10 years, M1-like was prevalent, along with an increase in cytotoxic GrB+ cells. This study supports the notion that pediatric cHL exhibits a particular tumor microenvironment composition.

We performed this study in order to study microenvironment composition in pediatric cHL, including an independent and geographically different series of cases from a single institution, in a region where EBV association is significant in patients under 10 years of age, and cHL subtype distribution is different from previous studies 18 .
Given that older children showed an immunoregulatory-rich tumor microenvironment, we evaluated if this characteristic could affect the macrophage composition. For this purpose, a ratio of FOXP3/GrB >1.5 as indicative of an immunoregulatory-rich microenvironment was defined, as published previously 8 . An immunoregulatory-rich tumor microenvironment was associated with lower mean of CD68+ pSTAT1+ cell numbers (mean 163 cells/mm 2 for FOXP3+/GrB ratio >1.5 vs. mean 270 cells/mm 2 for FOXP3+/GrB ratio ratio <1.5, p = 0.028, Mann-Whitney test). In fact, M1-like polarization was the exclusive pattern observed in patients Finally, hierarchical cluster analysis was conducted to identify underlying patterns of tumor microenvironment cell subsets, in relation to both EBV status and age group (Fig. 2). In this analysis, 2 distinctive clusters were observed: cluster I included the EBV− negative cases, those with high M2-like macrophages cell count, as demonstrated by CD68+ CMAF+ and CD163+ CMAF+ above the 50th percentile, and predominance of FOXP3+ over GrB+ cells, given that the ratio Foxp3+: GrB+ >1.5 prevails in this cluster. In contrast, the cluster II is mainly formed by EBV-associated cases and those with increased numbers of M1-like macrophages, confirmed by CD68+ pSTAT1+ and CD163+ pSTAT1+ cases above the 50th percentile, and predominance of GrB+ over FOXP3+ cells, since the ratio GrB+: Foxp3+ is higher than 1.5.

Discussion
In pediatric cHL, only two studies were performed in children from developed socioeconomic populations 5,7 , while one group has extensively studied a pediatric population from a developing country, Brazil 6,8,9 . Gupta et al. 5 and Englund et al. 7 performed tumor microenvironment analysis in pediatric cHL cases from developed population (Canada and Sweden, respectively), in which cHL incidence, rate of EBV-association and subtype distribution is different. In developed countries, nodular sclerosis (NS) subtype is prevalent over mixed cellularity (MC). In fact, NS subtype was described in 91.6% of cases demonstrated by Gupta et al. 5 , and in 69% of cases reported by Englund et al. 7 . In line with this, Barros et al. 6,8,9 described in their pediatric series that 69% of cases were NS subtype, even though their study was performed in a developing population. In contrast, in our study 58.7% of cases were MC subtype. EBV was associated with 25% of pediatric cases from a developed population published Englund et al., particularly in patients younger than 10 years old 7 . Meanwhile 44.8% of cases were EBV-associated by Barros et al. in pediatric cHL cases from a developing country, without correlation with age-groups 6,8,9 . In contrast, we observed an increase in EBV-association in our series from a developing population along with a correlation with young age-group.
Barros et al. suggested a shift from a cytotoxic to a regulatory profile in the tumor microenvironment of pediatric cHL with the increase of the age 8 . In line with this, we observed a regulatory profile in children older than 10 years old, confirmed by a higher FOXP3+ cell counts. Nevertheless, an increment on the numbers of GrB+ cells was not observed in younger patients, in spite of EBV prevalence on this specific group, which would be expected to trigger an antiviral cytotoxic profile. However it is remarkable that EBV-associated cases exhibited higher means of CD8+ cells/mm 2 , as described previously 8 . The reasons why older children develop a tumor microenvironment rich in Treg cells, similar to adult cases 19 , are not clear yet. One possible explanation could be the increase of Treg cells numbers and suppressive function with the aging of the immune system 20 .
Polarized M1 macrophages are generally pro-inflammatory and anti-tumoral. In contrast, M2 macrophages were described as an immunoregulatory phenotype, which mediate tumor promotion 21 . In cHL, the immunomodulatory phenotype adopted by macrophages may be the potential mechanism by mean of which they promote immune evasion 22 . In this study we observed an increase in M1-like macrophages cell count and M1-like polarized tumor microenvironment in younger children. Also, this kind of polarization was linked to cytotoxic tissue signature, as evidenced by the increase in GrB+ cells in M1-like polarized cases. In contrast, higher numbers of M2-like macrophages and M2-like polarized tumor microenvironment was observed in older patients and related to an immunoregulatory-rich environment, in line with previous results 9 . Together, our results support previous hypothesis that the tumor microenvironment in pediatric cHL is influenced by age, and a possible cross-talk between cells of adaptive and innate immune system in this microenvironment may contribute to the macrophage polarization 6,8,9 . In summary, a predominance of M1-like polarized macrophages in young children, as well as M2-like polarized macrophages and Treg cells in old children was observed, supporting the notion that pediatric cHL exhibits a particular tumor microenvironment composition when compared with adult cases 6,8,9,23-25 . www.nature.com/scientificreports www.nature.com/scientificreports/ Methods Patients and samples. Formalin-fixed paraffin-embedded (FFPE) biopsy samples from 46 patients were collected retrospectively, on the basis of the availability of sufficient material, from the archives at Pathology Division, Ricardo Gutierrez Children's Hospital in Buenos Aires, Argentina, from 1990 through 2012, and a tissue microarray (TMA) was constructed as described 26 . From each case, two 3-mm-diameter cores, selected from two different representative tumor areas rich in HRS cells selected by the pathologists EDM and MHB, were included 26 .
The Ricardo Gutierrez Children Hospital Ethics Committee on Research (CEI) approved the study, and the subjects gave informed consent to the study. Diagnosis was revised according to the updated WHO scheme for lymphomas 27 . All methods were performed in accordance with the relevant guidelines and regulations.
In addition, macrophage polarization was assessed by double immunohistochemistry, as described previously 10 . pSTAT1 or CMAF (Santa Cruz Biotechnology, Dallas, USA) were used as first primary antibodies and the detection of bound antibodies was performed using ZytoChem Plus HRP polymer kit (Zytomed Systems, Berlin, Germany), employing diaminobenzidine (DAB) as chromogen 10 . Then, CD68 (Dako) or CD163 (Novocastra, Wetzlar, Germany) antibodies were incubated as second antibodies, followed by detection with AP Polymer System (Zytomed Systems, Berlin, Germany), wiht Blue Alkaline Phosphatase substrate kit (Vector Laboratories, Burlingame, USA) as substrate. The sections were not counterstained 10 . Co-expression of pSTAT1 together with CD68 or CD163 was used to identify M1-polarized macrophages, while co-expression of CMAF in conjunction with CD68 or CD163, was performed to identify M2-polarized macrophages 20 . The ratio M1: M2 per case was assessed for each marker combination, and a polarized response (M1 > M2 or M2 > M1) was defined as one cell population 1.5x higher than the other 10 .
Tumor microenvironment analyzes. For lymphocyte and macrophage markers evaluation, the slides were scanned with Aperio LV1 digital pathology slide scanner (Leica Biosystems, CA, USA) at a 200× magnification. Using Aperio Image Scope software (Leica Biosystems, CA, USA), each core was extracted as separated JPG file. After, the numbers of labeled lymphocytes were determined per 1 mm 2 using the image analysis free software Image J 28 . The cells were counted optically without the use of plug-in.