Nodal histiocytic sarcoma with prominent eosinophilic infiltration: expression of eotaxin-2 on tumor cells

Histiocytic sarcoma (HS) is a rare hematopoietic neoplasm derived from non-Langerhans histiocytes that accounts for less than 1% of all hematolymphoid neoplasms [1] and shows morphological and immunophenotypic features of mature tissue histiocytes. This tumor occurs in the skin and connective tissue (35.8%), followed by the lymph nodes (17%), respiratory system (8.2%), and nervous system (6.9%) [2‐4]. The neoplastic cells are usually large and round-to-oval in shape [2] and are immunophenotypically positive for histiocytic markers, such as CD4, CD68, CD163, and lysozyme. Somewhat frequently, cases of HS arise from other hematopoietic malignancies, such as follicular lymphoma, chronic lymphocytic leukemia, diffuse large B-cell lymphoma, and B- or T-lymphoblastic leukemia, which have been shown to include secondary HS [4‐6]. The prognosis of patients who have concomitant hematological neoplasms (non-Hodgkin lymphoma and acute myeloid leukemia) is relatively poor, and HS has an increased risk of subsequent development in patients with non-Hodgkin lymphoma [4]. Indeed, clonal IgH or TCR rearrangements are found in these secondary HS cases [7]. Recently, the variation rate of the BRAF V600E mutation was reported to be low (12%), and HS cases in which complete remission is achieved via molecular targeted therapy are rare [5, 8].

HS tumors are usually accompanied by a variable number of reactive cells, including small lymphocytes, plasma cells, neutrophils, and scant eosinophils [2, 9, 10]. Infiltrating eosinophils are usually not predominant, so no previous reports on HS have paid close attention to the significance of this type of inflammatory cell. Here, we first report a patient with nodal HS exhibiting marked reactive eosinophilic infiltration. This is the first case showing eotaxin-2 (CCL24) expression on HS neoplastic cells themselves, which may induce eosinophil migration in the tumor tissue.

HS is a rare neoplasm of mature histiocytes that often has an aggressive clinical course and can arise de novo or from a low-grade B-cell lymphoma [1, 9]. HS usually shows a diffuse architecture involving either nodal or extranodal tissue. Neoplastic cells are large and round to oval with abundant eosinophilic cytoplasm. Neoplastic giant cells or spindle cells can be seen. Hemophagocytosis or emperipolesis by neoplastic cells can be present. There is usually a prominent inflammatory background consisting of neutrophils and lymphocytes. Tumor cells are immunohistochemically positive for CD4, CD11c, CD14, CD45, CD45RO, CD68, CD163, and lysozyme and often or rarely positive for CD15, the S100 protein, and CD56 but negative for markers associated with the B cell lineage, the T cell lineage, follicular dendritic cells (CD21, CD23, and CD35) and Langerhans cells (langerin, and CD1a) as well as CD30, CD34, HMB45, myeloperoxidase, EMA, and cytokeratins [1, 9]. The Ki-67 index varies from 10 to 90% of tumor cells. Our case was morphologically and immunohistochemically compatible with HS.

In HS, a variable number of reactive cells, including small lymphocytes, plasma cells, benign histiocytes, neutrophils, and frequently fewer/scant eosinophils, may be seen [2, 9]. In this case, especially on cytological and histological findings, the maximum number of eosinophils infiltrating the tumor was 108 cells/HPF (hot spot) (Fig. 2f). Therefore, we investigated the mechanism of eosinophilic infiltration of this tumor by utilizing anti-eotaxin-2 and anti-eotaxin-3 antibodies. Eotaxins comprise three types: eotaxin-1 (CCL11), eotaxin-2, and eotaxin-3. Eotaxin-2 and eotaxin-3 have been identified in the human genome and are known to have eosinophil-selective chemoattractant activity [13]. In humans, eotaxin-2 is secreted by various types of cells, including most infiltrating eosinophils as well as other inflammatory cells, epithelial cells, endothelial cells in nasal polyps, and cytokine-stimulated human lung alveolar epithelial cells [14, 15]. Eotaxin-3, and not the other eotaxins, seems to play a major and specific role in sustained severe eosinophil infiltration in asthmatic tissues [16]. Eotaxins are unusual (but not unique) in signaling via a single receptor: C-C motif chemokine receptor 3 (CCR3). CCR3 is found on eosinophils, basophils, mast cells and a subpopulation of Th2 lymphocytes [17]. The eosinophilic infiltration in neoplastic tissue in our case was induced by some chemotactic factors, including eotaxin-2 and eotaxin-3, which are secreted by neoplastic cells and eosinophils, respectively. Namely, this phenomenon may include the secretion of eotaxin-2 from neoplastic cells combined with CCR3 expression on eosinophils. Furthermore, activated eosinophils secrete eotaxin-3. Subsequently, eotaxin-3 may combine with CCR3 expressed on eosinophils and induce the selective migration of numerous eosinophils into tumor tissues.

The neoplastic cells of HS have not been reported to express eotaxin-2. However, there are some lymphoma cases with eotaxin expression. Although a high level of eotaxin mRNA expression was detected in Hodgkin lymphoma, especially nodular sclerosis [18], eotaxin was expressed on fibroblasts rather than Hodgkin/Reed-Sternberg cells themselves [19]. Eotaxin-3 mRNA expression was detected in some cases of follicular lymphoma [20], and multiple reports on cancers associated with eotaxin-2 expression exist [21‐23]. In clear cell renal cell carcinoma, patients with high eotaxin-2 expression had poor survival rates [21]. In hepatocellular carcinoma, eotaxin-2 contributes to malignancy via the angiogenesis pathway and indicates a poor prognosis [22]. In colorectal cancer, eotaxin-2 was upregulated in cases with liver metastases [23]. Therefore, eotaxin-2 expression in neoplastic cells may be a poor prognostic factor related to angiogenesis and metastasis.

Via the secretion of various cytokines, tissue eosinophils are involved in diverse biological events, such as host protection against helminths; steady-state development in the mammary glands and intestines; metabolic homeostasis of adipose tissue in coordination with M2 macrophages; tissue remodeling (tissue regeneration and repair, angiogenesis, and fibrosis); cell-mediated immunity (T cell activation and polarization from Th0 to Th1, Th2, and Th17 cells and recruitment of dendritic cells and T cells); humoral immunity (maintenance of bone marrow plasma cells and secretory IgA in the intestines); and cell-cell interactions among dendritic cells, mast cells, and neurons [24]. In addition, eosinophils have ambivalent roles that are both anti- and protumorigenic [25]. In regards to their antitumorigenic role, eosinophils can induce tumor cell death by secreting cytotoxic proteins (major basic protein, eosinophil cationic protein, eosinophil-derived neurotoxin, and granzymes), suppress tumor metastasis by IL-10 and IL-12, and induce cytotoxicity towards tumor cells [25]. Furthermore, eosinophils indirectly promote antitumor immunity by releasing IFNγ and support antitumor immunity by controlling angiogenesis [25]. In regards to their protumorigenic activity, eosinophils inhibit effector T-cell responses and induce suppressive immunity via indoleamine 2,3-dioxygenase [25]. Furthermore, eosinophils may induce tumor cell growth and epithelial mesenchymal transition via epidermal growth factor and transforming growth factor-β1, respectively, and may facilitate metastases by promoting matrix remodeling via the secretion of matrix metalloproteinases 2 and 9 [25]. In our case, no fibrosis or lymphatic vessel proliferation was observed in the area with high eosinophilic infiltration, suggesting that areas with high eosinophil numbers existed in the stage prior to the start of remodeling. Conversely, fibrosis and lymphatic vessel proliferation were observed around areas with no or scant eosinophilic infiltration, suggesting that eosinophil infiltration had already subsided in the areas with extensive fibrosis and lymphatic proliferation. Namely, the increased eosinophilic infiltration in this case would be affected in both the protumorigenic and antitumorigenic activities phases in the tumor microenvironment.

LIGHT can be used to identify murine M2b macrophages but not human M2b macrophages [26]. In this case, neoplastic cells were identified as positive for LIGHT for the first time. This suggests that LIGHT may be useful as an HS marker in humans. The neoplastic cells were also positive for PD-L1 in this case. Although 8 of 10 HS cases showed neoplastic cell expression of PD-L1, these cases were not associated with gains of 9p24.1 [6], unlike classic Hodgkin lymphoma [27]. Furthermore, PD-L1 expression in HS may be associated with the macrophage phenotype and not associated with amplification of 9p24.1 because the staining intensity of PD-L1⁺ neoplastic cells was similar to that of normal macrophages [6]. Therefore, the effect of PD-L1/PD-1 blockade on HS is unknown [6].

In conclusion, we revealed that eotaxins induced the selective migration of eosinophils into tissues in this case. Although the reason why predominant eosinophilic infiltration occurred in this HS case is not clear, these eosinophils may be associated with tissue remodeling processes, such as fibrosis and lymphatic vessel proliferation.