histiocytic sarcomas and a canine histiocytic sarcoma xenograft model
Adnan Fayyad1, Stefanie Lapp1, Engy Risha1, Karl Rohn2, Yvonne Barthel1, Dirk Schaudien3, Wolfgang Baumgärtner1,*, Christina Puff1
1 Department of Pathology, University of Veterinary Medicine Hannover, Bünteweg 17, 30559 Hannover, Germany
2 Institute for Biometry, Epidemiology and Information Processing, University of Veterinary Medicine Hannover, Bünteweg 2, 30559 Hannover, Germany
3 Fraunhofer Institute for Toxicology and Experimental Medicine, Nikolai-Fuchs-Straße 1, 30625 Hannover, Germany
* corresponding author:
Prof. Dr. Wolfgang Baumgärtner, Ph.D.
Department of Pathology
University of Veterinary Medicine Hannover Bünteweg 17
30559 Hannover Germany
Tel: +49-511-953-8620 Fax: +49-511-953-8675
e-mail: Wolfgang.Baumgaertner@tiho-hannover.de
30
Abstract
Canine histiocytic sarcoma (HS) represents a malignant neoplastic disorder with a rapid and progressive clinical course. Due to its invasiveness, high rate of relapses as well as the common occurrence of metastases especially the disseminated form of the disease carries a grave prognosis. A better understanding of the interaction between tumor cells and the local microenvironment may provide new insights into mechanisms of tumor growth and metastases. The influence of matrix metalloproteinases (MMPs) and their inhibitors (TIMPs) on tumor angiogenesis, invasion and metastasis has been detailed in previous studies. In addition, inflammatory cells infiltrating neoplasms especially tumor associated macrophages (TAM) may contribute significantly to tumor progression. Due to the high variability of canine HS, standardized models are highly required to investigate tumor progression and interaction with the surrounding microenvironment. Therefore, the present study comparatively characterizes the intratumoral immune response as well as the expression of MMP-2, MMP-9, MMP-14 and TIMP-1 in spontaneous canine HS and its murine xenotransplantation model. In spontaneous canine HS, an intratumoral predominance of macrophages and T lymphocytes was observed while B lymphocytes were only found in low numbers. Interestingly, morphometric results demonstrated that MMP-2, MMP-9, MMP-14 and TIMP-1 were mainly expressed at the invasive front of the neoplasms while the tumor centers exhibited a significantly smaller immunoreactive area. Similar findings were obtained for the MMP and TIMP-1 distribution in subcutaneously xenotransplanted HS, rendering this model highly suitable to investigate HS under standardized conditions. Moreover, tumor associated macrophages (TAM) strongly express MMPs and TIMP-1 and MMP-14 showed a strong correlation with microvessel density in xenotransplanted HS. These results indicate that MMP expression contributed to tumor progression and invasion and TAM seem to be major players in the interaction between neoplastic cells and the local microenvironment rendering therapeutic approaches modulating these molecules and cells promising treatment options for increasing life expectancy and/or quality of affected individuals.
Key words: canine histiocytic sarcoma; MMP; murine xenotransplantation model;
spontaneous neoplasm; tumor associated macrophages, microvessel density; TIMP
31
1. Introduction
Canine proliferative histiocytic diseases represent a range of well-defined disorders with marked differences in their clinical behavior and pathologic features [1,2]. This group of diseases, which are common in dogs and less frequently observed in cats, include cutaneous histiocytoma, the histiocytic sarcoma (HS) complex, reactive histiocytoses (cutaneous and systemic forms) and hemophagocytic histiocytic sarcomas [3-6]. The histiocytic sarcoma (HS) complex of dogs comprises two distinct types of malignant proliferative disorders, characterized by infiltration of neoplastic cells arising from interstitial dendritic cells [7]. The localized HS most often involves a single tissue or organ, usually the skin, whereas the disseminated HS spreads beyond the local draining lymph nodes to distant sites, usually lungs, lymph nodes, liver, spleen and central nervous system [2,8]. HS are typically composed of poorly demarcated sheets of large, pleomorphic cells with one or multiple nuclei, marked cellular atypia and a high mitotic index with variable numbers of inflammatory cell infiltrates consisting mainly of T lymphocytes [2,8,9]. HS is a highly aggressive neoplasm in dogs with a rapid and progressive clinical course leading to a poor prognosis [5].
Several studies have provided evidence that increased expression of matrix metalloproteinases (MMPs) is associated with invasion, metastasis and poor prognosis in numerous human and animal malignancies, including lung and oral neoplasms, breast carcinoma and esophageal squamous cell carcinoma [10-15]. An increased MMP expression is often accompanied by an enhanced invasion and metastasis rate and therefore a poor prognosis [16,17]. However, little is known about the importance and distribution of MMPs and their inhibitors in HS.
MMPs are a family of zinc-dependent endoproteinases whose enzymatic activity is directed mainly against components of the extracellular matrix (ECM) [18]. These proteinases are linked by a core of common domain structures and by their relationship to a family of proteinase inhibitors called tissue inhibitors of metalloproteinases (TIMPs) [19]. MMPs play important roles in tumor angiogenesis, invasion and metastasis through degradation of the stromal connective tissue and basement membrane components which permit the migration of tumor cells and secretion of growth factors, cytokines and vascular growth factors necessary for tumor development [20-24]. In addition, MMPs are also able to influence the tumor microenvironment by regulating innate and acquired immunity through modulating the function of cytokines and chemokines and increase the infiltration of inflammatory cells [25,26]. Notably, MMPs and their inhibitors are secreted by different cell types, including stromal fibroblasts in the vicinity of the neoplasm, tumor infiltrating macrophages or from the tumor cells themselves [27-30]. Inflammatory cells infiltrating tumors include macrophages,
32
dendritic cells, myeloid-derived suppressor cells and T-cells which contribute either positively or negatively to tumor invasion, growth and metastasis [31]. Analysis of cellular phenotypes of inflammatory cells infiltrating the neoplasm seems to be of critical importance for predicting tumor outcome [31]. Tumor-associated macrophages (TAM) are considered as the major players of tumor-related inflammation and are one important source of MMPs [32,33].
It’s suggested that tumor cells use MMPs produced by these macrophages, fibroblasts and other stromal cells for invasion [34] and tumor progression, including metastasis, which is only possible through close interaction between neoplastic and stromal cells [35]. However, the functional interactions between tumor and surrounding stromal cells are not completely understood [36] and are often unpredictable in spontaneously arising neoplasms. These difficulties highlight the importance of a well characterized, standardized model for detailed investigation of specific questions and analysis of the effectiveness of novel treatment options especially in relatively rare tumor types. To overcome these limitations and allow detailed investigations of the tumor microenvironment in histiocytic sarcomas, the aim of the present study was to comparatively analyze spontaneous canine histiocytic sarcomas and xenotransplanted neoplasms in a murine model to verify the usability of this model. Therefore the intratumoral immune response in spontaneous canine histiocytic sarcomas in combination with analysis of selected MMPs and TIMPs was characterized and obtained data were compared to xenotransplanted canine histiocytic sarcomas in a murine model. Furthermore, a detailed investigation of the local microenvironment, especially a correlation of the MMP expression with microvessel density was performed in xenotransplanted histiocytic sarcoma neoplasms.
2. Materials and Methods
2.1 Samples of spontaneous canine histiocytic sarcomas
Spontaneous histiocytic sarcomas were obtained from 16 dogs either by biopsy or necropsy.
Breed, sex, age and investigated tumor location are detailed in Table 1. After excision of biopsies and necropsy, respectively, obtained samples were fixed in 10% neutral-buffered formalin, routinely embedded in paraffin wax and stained with hematoxylin and eosin.
33
2.2 Canine histiocytic sarcoma xenotransplantation model
All animal experiments were approved and authorized by the local authorities (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES), Oldenburg, Germany, permission number: 33.9-42502-04-08/1515). All animal procedures were performed in accordance with the German regulations and legal requirements.
Canine histiocytic sarcoma cells (DH82 cells) were transplanted subcutaneously into severe combined immunodeficient mice (CB17/Icr-Prkdcscid/IcrIcoCrl mice) as described previously [37]. Briefly, mice were injected once subcutaneously into the left flank with of a total of 3.0 million DH82 cells. Subsequently tumor development was monitored every second to third day and calculated as [(shortest diameter2 x longest diameter)/2] [38]. Necropsies were performed and tumor tissue was harvested on day 7, 14, 21, 35 and 77 after transplantation.
Animals were preliminary sacrificed when tumors extended a volume of 1.7 cm³ (equals a diameter of approx. 1.5 cm).
2.3. Immunohistochemistry
Serial sections (2-4 µm) of formalin-fixed, paraffin-embedded tissue were used for immunohistochemistry (IHC) by applying the avidin-biotin-peroxidase complex (ABC) procedure as described [39]. Spontaneous canine HS diagnosis was confirmed using the immunohistochemical marker CD204 [40]. To study the expression of MMPs and TIMPs in spontaneous canine HS and the HS xenotransplantation model, antibodies directed against MMP-2, MMP-9, MMP-14 and TIMP-1 were used [41]. Antibodies applied for intratumoral inflammatory cell evaluation were directed against myeloid/histiocyte antigen, Pax-5 and CD3 in spontaneous canine HS and Mac3, CD3 and CD45R in the HS xenotransplantation model [42,43]. In addition, an antibody directed against CD31 was used for microvessel density (MVD) evaluation in both spontaneous and xenotransplanted HS [44; Table 2).
Positive controls included bone marrow from a stillborn puppy (TIMP-1), a cell pellet of DH82 cells (MMP-2, MMP-9, MMP-14, CD44 and CD204), granulation tissue from a dog (CD31), lymphoid tissue from immunocompetent SJL mice and immunocompromised SCID mice (Mac3), canine lymph nodes (myeloid/histiocyte antigen, CD3, Pax-5). Negative controls included substituting serum from non-immunized Balb/cJ mice for monoclonal antibodies (MMP-2, myeloid/histiocyte antigen, Pax-5 and CD204) substituting serum from non-immunized rabbits (for antibodies directed against MMP-9, MMP-14, TIMP-1, CD-31 and CD3, substituting serum from non-immunized rats for antibodies directed against Mac3 and CD45R and substituting DH82 cell culture supernatant for the antibody directed against
34
CD44. Control sera were used with protein concentrations identical to the ones of the specific antibodies.
The expression of MMP-2, MMP-9, MMP-14 and TIMP-1 was determined by using a microscope (Olympus BXS1) equipped with a camera (Olympus DP72) at a 400x magnification in both, spontaneous and xenotransplanted canine HS. Images were taken from 10 randomly selected fields in the tumor center and 10 randomly selected fields of the tumor periphery. Obtained pictures were analyzed using the positive pixel count algorithm of the Aperio ImageScope viewer (Version 12, Aperio Technologies). Only areas containing neoplastic tissue were analyzed. Tumor-free regions, necrotic areas and artifacts (e.g., tissue folding) were manually deselected and excluded from further analysis. For the positive pixel count algorithm, hue value of 0.1 and hue width of 0.5, color saturation threshold of 0.04 were used, and any intensity of staining was considered positive. The number of positive pixels was divided by the total number of pixels (negative and positive) in the analyzed area, and multiplied by 100 to receive the percentage of positive pixels. Values from each section were averaged.
Analysis of intratumoral inflammatory cell infiltration in spontaneous and xenotransplanted canine HS was performed by manual counting of immune-reactive cells in 10 randomly selected fields in the tumor center and 10 randomly selected fields in the tumor periphery (each 0.1734 mm²) at a 400x magnification. Then values obtained for tumor center and tumor periphery, respectively, were averaged. For microvessel density the counting was done as described for inflammatory cells with the exception that only immunolabeled structures with a definable lumen were counted.
2.4. Double immunofluorescence staining
To determine whether MMPs and TIMPs were produced by neoplastic cells or tumor-infiltrating macrophages or both, double immunofluorescence was performed. The same antibodies against MMPs and TIMP-1 used for immunohistochemistry were used for immunofluorescence (Table 2). These antibodies were incubated together with a rat monoclonal antibody directed against CD44 for labeling of the transplanted DH82 cells or together with a rat monoclonal antibody directed against Mac3 for labelling of murine macrophages.
Positive and negative controls, same as used for immunohistochemistry, include a DH82 cell pellet for MMP-2, -9, -14, TIMP-1 and CD44. Lymphoid tissue from immunocompetent SJL
35
mice and immunocompromised SCID mice was used for Mac3. Negative controls included serum from non-immunized Balb/cJ mice mixed with DH82 cell culture supernatant for the combination of MMP-2and CD44 staining. Serum from non-immunized Balb/cJ mice mixed with rat serum was applied for the combination of MMP-2 and Mac3 and serum from non-immunized rabbits mixed with DH82 cell culture supernatant for the combinations of MMP-9 / CD44, MMP-14 / CD44 and TIMP-1 / CD44. Serum from non-immunized rabbits mixed with rat serum served as a negative control for the combinations of MMP-9 / Mac3, MMP-14 / Mac3 and TIMP-1 / Mac3.
Tissue sections were deparaffinized in Rotihistol (2 X 5 minutes), isopropanol (5 minutes), rehydrated through 96% ethanol (5 minutes) and rinsed in phosphate-buffered saline (PBS;
pH 7.4). After heat-induced epitope retrieval (800 Watt microwave treatment for 15 minutes in 10 mM sodium citrate buffer [pH 6.0]), slides were rinsed in PBS. Before incubation with the primary antibody, each section was blocked with 20% normal goat serum in PBS with 1%
bovine serum albumin (BSA) and 0,1% Triton X for 30 minutes. Afterwards, sections were incubated with a mixture of two primary antibodies diluted in PBS with 1% BSA and 0.1%
Triton X for 1.5 hours at room temperature, washed with PBS, and incubated with a mixture of two secondary antibodies; a Cy3-conjugated goat anti-rabbit antibody (dilution 1:200; Life Technologies, Darmstadt, Germany) and a Cy2-conjugated goat anti-rat antibody (dilution 1:200; Dianova, Hamburg, Germany) for 1 hour at room temperature. After washing with PBS, tissue sections were incubated with bisbenzimide H (dilution 1:100, Sigma, Taufkirchen, Germany) for 10 minutes at room temperature, mounted and analyzed using a fluorescence microscope (Biorevo BZ-9000; Keyence, Japan)
2.5. Statistical analysis
Statistical analysis was performed using the Statistical Analysis System (SAS) for Windows Software, version 9.1 (SAS Institute Inc., Cary, USA). The significance level was considered if (p ≤ 0.05) for comparison between different groups and between central and peripheral areas of the neoplasms.
All data were included into a descriptive analysis. For inflammatory cells normal distribution of the model residuals of MAC387, CD3 and Pax-5 were confirmed by Kolmogorov-Smirnov-Test and visual assessment of q-q – plots. Because three points was right skewed distributed, logarithmic transformation was performed prior to analysis; hence the results were tabulated on the original scale after exponential retransformation. For analysis, a two way ANOVA was calculated.
36
Analysis of the immunohistochemical data for MMPs, TIMP-1 and CD31 in spontaneous and xenotransplanted canine HS and MAC-3 in xenotransplanted neoplasms was performed using the Wilcoxon signed rank test. In addition the difference in MMP and TIMP-1 expression between spontaneous and xenotransplanted canine HS was performed using the Wilcoxon test.
3. Results
3.1 CD204 expression confirmed spontaneous canine histiocytic sarcomas
CD204 was expressed in neoplastic cells as well as tissue macrophages, including alveolar macrophages of the lung, Kupffer cells of the liver and resident macrophages of spleen and lymph nodes. All 16 cases of spontaneous HS expressed CD204.
3.2 Inflammatory cell infiltrates in spontaneous canine histiocytic sarcomas are dominated by macrophages and T lymphocytes
Histologically a moderate inflammatory infiltrate was found in spontaneous canine HS.
Inflammatory cells occurred as single cells randomly distributed within the tumor or as variably sized aggregates within tumor center and periphery. This inflammatory cell infiltrate was composed mainly of macrophages, T lymphocytes and lower numbers of B lymphocytes (Figure 1). All results are summarized in Table 3.
Macrophages showed an intense cytoplasmic immunoreactivity using an antibody directed against myeloid/histiocyte antigen (MAC387; Figure 1A). They were randomly distributed in tumor center and periphery, with median values of 23 cells / 0.1734 mm² and 33 cells / 0.1734 mm², respectively. Occasionally, macrophages were arranged around necrotic areas within the tumor.
T lymphocytes were determined by a membrane-bound CD3 expression and were also randomly distributed within tumor center and periphery. A median of 27 T lymphocytes / 0.1734 mm² were found within the tumor center and 33 T lymphocytes / 0.1734 mm² in the tumor periphery (Figure 1B).
B lymphocytes were characterized by a nuclear Pax-5 immunoreactivity and were randomly distributed within tumor center and periphery. A median of 0 B lymphocytes / 0.1734 mm²
37
were seen within the tumor center and 0 B lymphocytes / 0.1734 mm² in the tumor periphery (Figure 1C).
One case (No. 15) lacked infiltrating T and seven cases lacked infiltrating B lymphocytes only (No. 3, 4, 6, 8, 12, 14 and 16).
Statistically, the number of macrophages and T lymphocytes significantly outnumbered the one of B cells in the tumor periphery (p ≤ 0.05). Additionally, significantly more macrophages than B lymphocytes were present in the tumor periphery (p ≤ 0.05) whereas no significant difference was found between the number of macrophages and T lymphocytes in any localization (p > 0.05; Figure 2). Furthermore, there was no significant difference (p >
0.05) between tumor center and periphery in macrophage and lymphocyte distribution.
3.2 Inflammatory cell infiltrates in subcutaneous canine histiocytic sarcoma xenotransplants are mainly composed of macrophages
HS in a subcutaneous xenotransplantation model were accompanied by mild to moderate inflammatory cell infiltrates.
These inflammatory infiltrates were composed mainly of macrophages which were present as variably sized aggregates at the invasive front of the tumor and around necrotic areas. Within the tumor center a median of 23.6 macrophages / 0.1734 mm² was found (Figure 1D).
Similarly, the tumor periphery displayed a median number of 29.6 macrophages / 0.1734 mm²; Table 3). Statistically, the number of tumor associated macrophages did not differ between center and periphery.
Furthermore few B lymphocytes and some T lymphocytes were observed as single scattered cells or small aggregates within the tumor.
3.3 MMP-2, MMP-9, MMP-14 and TIMP-1 are mainly expressed at the invasive front of spontaneous canine histiocytic sarcomas
Serial sections of spontaneous canine histiocytic sarcomas were used to analyze the expression of MMP-2, MMP-9, MMP-14 and TIMP-1. All molecules examined were present in all tumor samples with a diffuse cytoplasmic staining but displayed a varying intensity and distribution.
The highest immunoreactivity was detected for MMP-14, where 40.87% (median) of the central tumor area and 61.46% (median) of the peripheral tumor area expressed this molecule.
38
MMP-2, MMP-9 and TIMP-1 displayed a similar but lower immunoreactivity than MMP-14 within central and peripheral tumor areas. The immunoreactivity of MMP-2 was 5.68%
(median) of tumor center and 23.97% (median) of tumor periphery, whereas the median for MMP-9 was 7.46% in central tumor areas and 21.23% (median) in the tumor periphery. The TIMP-1 immunoreactivity was 4.31% (median) and 14.75% (median) of tumor center and periphery, respectively.
The staining intensity for MMP-14 was moderate to strong in tumor cells, stromal cells surrounding the neoplasm and tumor infiltrating cells, whereas, weak to moderate staining was observed for MMP-2, MMP-9 and TIMP-1. Notably, all MMPs and TIMP-1 showed a significantly enhanced expression at the invasive front of the tumor (p ≤ 0.05; Figure 3 A-H).
Significant differences in immunohistochemistry expression were observed for all MMPs and TIMP-1 between tumor center and periphery (p ˂ 0.05; Figure 4). All results are summarized in Table 4.
3.4 MMP-9, MMP-14 and TIMP-1 are mainly produced by tumor-infiltrating macrophages in a canine histiocytic sarcoma xenotransplantation model
Serial sections of subcutaneously xenotransplanted HS were used to study the expression of MMP-2, MMP-9, MMP-14 and TIMP-1 (Table 4), which all displayed immunoreactivity in all neoplasms examined (Figure 3 I-S). Whereas MMP-9 and TIMP-1 were expressed diffusely within the tumors, MMP-2 and MMP-14 were found as a patchy, multifocal immunopositivity of neoplastic cells. Interestingly, compared to spontaneous canine histiocytic sarcomas, MMP-9, MMP-2 and TIMP-1 immunoreactivity was significantly higher in xenotransplanted neoplasms within the tumor center and MMP-14, MMP-9, MMP-2 and TIMP-1 within the tumor periphery (p ≤ 0.05; Figure 5).
When all time points from the HS xenotransplantation model were taken together, the MMP-14 expression was comparable (p > 0.05) to spontaneous canine HS in tumor central areas but significant differences (p ≤ 0.05) in tumor peripheral areas were observed. The latter displayed a median of 40.87% MMP-14 immunoreactive area in the tumor center and 61.46% at the tumor periphery compared to 50.65% (median) and 87.68% (median) in xenotransplanted HS, respectively.
In contrast, xenotransplanted neoplasms, exhibited a higher percentage of positive area for MMP-2 (median 17.81%), MMP-9 (median 28.38%) and TIMP-1 (median 22.10%) in the tumor center compared to spontaneous HS (see 3.3). Similar results were obtained for the tumor periphery, where a median of 49.75% (MMP-2), 57.05% (MMP-9) and 77.14%
(TIMP-39
1) of the cells showed immunoreactivity. Interestingly these values are all significantly (p ≤ 0.05) higher than in spontaneous canine HS.
Notably, the MMP-9 and MMP-14 immunopositive area increased over time in the tumor center of xenotransplanted histiocytic sarcomas whereas a high immunopositivity within the tumor peripherywas found at all time points. At all time points investigated, the MMP-2 and TIMP-1 immunopositive area was small within the tumor center and medium-sized in the tumor periphery without significant differences over time. In all cases, strong 9, MMP-14, TIMP-1 and weak MMP-2 immunoreactivity was detected in inflammatory cells infiltrating the neoplasms. Immunohistochemically, the presence of numerous macrophages infiltrating central tumor areas, peripheral areas at the invasive front of the tumor and also
Notably, the MMP-9 and MMP-14 immunopositive area increased over time in the tumor center of xenotransplanted histiocytic sarcomas whereas a high immunopositivity within the tumor peripherywas found at all time points. At all time points investigated, the MMP-2 and TIMP-1 immunopositive area was small within the tumor center and medium-sized in the tumor periphery without significant differences over time. In all cases, strong 9, MMP-14, TIMP-1 and weak MMP-2 immunoreactivity was detected in inflammatory cells infiltrating the neoplasms. Immunohistochemically, the presence of numerous macrophages infiltrating central tumor areas, peripheral areas at the invasive front of the tumor and also