Introduction
Human angiosarcoma and canine hemangiosarcoma are aggressive malignancies of vascular tissue or vascular forming cells [
1,
2], and their morphology and pathological progression are virtually indistinct. Although doxorubicin remains the foundation of chemotherapy for humans and dogs, it provides minimal tumor control or survival benefit for angiosarcoma patients as only 16-36% of patients respond to treatment and over 50% of affected dogs still die within four to six months of their diagnosis [
1,
2]. In addition to their poor response rates, tumors in both species can be highly drug resistant. Thus, the outcomes of people and dogs are not likely to improve without an increased understanding of the biology and pathogenesis of these tumors. To complicate matters further, angiosarcomas occur infrequently restricting the study of their basic biology and clinical treatment [
3]. However, the frequent occurrence of hemangiosarcomas in dogs (up to 7% of all canine malignancies) [
4] provides a readily available resource to investigate tumor biology and explore related treatment options.
Identifying more successful treatment approaches may lie in addressing the cellular origin of these tumors rather than their histological appearance. Canine hemangiosarcoma and human angiosarcoma have been classified historically as tumors of malignant endothelium due to their histology and the expression of endothelial cell surface markers [
2,
5]. Based on the expression of early hematopoietic and endothelial progenitor markers, other studies have challenged this idea, suggesting instead that hemangiosarcomas might arise from bone marrow-derived angioblastic progenitors [
6‐
8]. A similar analogy of progenitor cell origin has been drawn for human angiosarcoma [
9].
In a more recent analysis, we identified and characterized a myeloid subpopulation from hemangiosarcoma cell lines that showed the expression of markers associated with bone marrow-derived myeloid progenitors (CD14 and CD115 or the colony stimulating factor 1 receptor, CSF-1R) co-expressed with surface markers associated with endothelial progenitor (CD34 and CD133) and endothelial cell differentiation (CD105, CD146, and α
vβ
3) [
8]. These cells also possessed phagocytic activity, and the co-expression of endothelial markers suggests a role in angiogenesis. Yoder
et al. described a similar population of human myeloid cells that express a variety of hematopoietic (CD14, CSF-1R, and CD45) and endothelial markers (CD133, CD34, VEGFR2) and participate in blood vessel formation [
10]. These cells possessed a myeloid progenitor cell activity and differentiated into phagocytic macrophages, but failed to contribute to the capillary endothelial layer
in vivo. These similarities suggest that the myeloid-endothelial cell phenotype in hemangiosarcoma may represent a viable target for therapeutic intervention, and more specifically, targeting of CSF-1R.
CSF-1R and its ligands, CSF-1 and IL-34, are commonly associated with the survival, proliferation, differentiation, and activation of mononuclear phagocytes [
11‐
15]. CSF-1R expressed by tumor associated macrophages (TAMs) can have therapeutic implications since TAMs impact tumor growth by promoting myeloid cell-mediated angiogenesis, chemoresistance, and metastatic spread [
16‐
20]. However, expression of CSF-1R by tumor cells also indicates a non-macrophage functional role for the receptor. In this regard, Cioce
et al. reported increased expression of CSF-1R mRNA in mesothelioma versus normal tissue specimens and demonstrated that CSF-1R expression identified chemoresistant cells in both primary cultures and mesothelioma cell lines [
21]. Thus, CSF-1R expression may serve as a marker to identify drug resistant populations in some cancers.
For this study, we demonstrate that both hemangiosarcoma and angiosarcoma cells with high expression of CSF-1R are more drug resistant than their CSF-1R low-expressing counterparts, indicating a shared mechanism for the observed treatment failures and subsequent drug resistance. Our data also suggest that part of this resistance may be achieved through drug sequestration within cellular lysosomes. Intriguingly, drug resistance in canine hemangiosarcoma is associated with CD133 expression, suggesting that resistance may be associated with a stem or progenitor cell phenotype and may be related to the degree of cellular differentiation. Further characterization of these cells and utilization of approaches to disrupt lysosomal drug trapping could improve drug responses as well as treatment outcomes.
Materials and methods
Cell culture
The DD-1 cell line was derived from a splenic hemangiosarcoma [
22], and the COSB line was derived from a xenograft of the original cell line, SB-HSA [
23]. The AS5 human angiosarcoma cell line was derived from a primary angiosarcoma of the thigh [
24]. All cell lines were cultured as described previously [
6,
22,
25]. Cells were maintained in culture for up to 8 weeks before new vials were thawed to ensure similar passage numbers were used for all experiments.
Flow cytometry and magnetic enrichment
The primary antibodies used were: anti-CSF-1R (CD115)-Cy5.5 (Bioss Inc., Woburn, MA), anti-CD34-Alexa Fluor 647, anti-CSF-1R-RPE (AbD Serotec, Raleigh, NC); anti-CD117(c-kit)-PE and APC, anti-CD34-PE, anti-CD243(ABCB1)-PE and APC, anti-CD338 (ABCG2)-PE and APC (eBioscience, San Diego, CA), anti-CD34-APC (human) (eBioscience), anti-CD34-PE and APC (canine) (eBioscience), and anti-CD133/AC133-PE and APC (Miltenyi Biotech, Auburn, CA). The detection of cell surface markers was carried out as described, previously [
8]. Data were collected on a BD FACS Calibur or a BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA) and then analyzed using FlowJo software. To exclude dead cells from analysis, 7-AAD was added to cells 10 minutes before acquisition, and compensation was performed to account for 7-AAD fluorescence in FL2. CSF-1R
low and CSF-1R
high populations were enriched from hemangiosarcoma and angiosarcoma cells by staining 2 × 10
7 to 4 × 10
7 cells with anti-CD115-RPE antibody followed by magnetic separation using the EasySep PE-Selection Kit (Stemcell Technologies, British Columbia, Canada) according to the manufacturer’s instructions. Cells were analyzed immediately after enrichment by flow cytometry to determine the percent enrichment from the original cell line culture.
Phagocytosis assay
Cells were plated in triplicate using 10,000 cells per well in 100 μL of culture medium. FITC-conjugated, rabbit-IgG-coated latex beads (Cayman Chemical Company, Ann Arbor, MI) were added to the cells. No beads were added to negative control wells. The emission at 535 nm was measured for each well after 24 hours using a Wallac Victor2 1420 Multilabel Counter. Relative phagocytosis for each cell line was determined by dividing the fluorescence of the wells with beads by the fluorescence of the respective negative controls.
Cytotoxicity assays
Cells were plated in triplicate at 2,500 (DD-1), 5,000 (COSB) or 10,000 (AS5) cells per well in 100 μL of culture medium and exposed to increasing concentrations of doxorubicin (Bedford Laboratories, Bedford, OH). Cell viability was determined 72 hours later using the colorimetric Cell Titer 96® Aqueous Non-Radioactive Cell Proliferation Assay (MTS Assay; Promega, Madison, WI) according to the manufacturer’s instructions. In order to compare the percent viability between the CSF-1Rlow and the CSF-1Rhigh populations and reduce artifacts due to potential differences in metabolic properties between cell populations, standard curves were generated for each CSF-1R population in each cell line, and the cell viability was determined based upon the total cell number from the standard curve. A separate standard curve for the CSF-1Rlow and the CSF-1Rhigh populations was generated for each experiment.
Side population analysis
Filtered DD-1 or COSB cells were incubated in the presence or absence of 10 μM verapamil for 15 minutes at 37°C. DyeCycle Violet (DCV) (Life Technologies, Eugene, OR) was added to a final concentration of 10 μM, and the cells were incubated for an additional 60 minutes at 37°C with intermittent mixing. Cells were washed, filtered, and maintained on ice until analysis. Propidium iodide was added to each sample immediately before collection to exclude dead cells from analysis. DCV emission was detected using a BD LSRII flow cytometer (BD Biosciences). Verapamil was used to determine the SP gates, and data were analyzed using FlowJo software (Tree Star Inc.).
Doxorubicin assays and detection
Cells were plated at 100,000 cells per well in 3 mL of culture medium in a 6-well plate. After 2 hours, cells were treated with 1 μM doxorubicin for 1 hour, washed with PBS, and stained for CSF-1R expression using a Cy5.5 labeled CSF-1R antibody at t = 0 and 24 hours. Intracellular doxorubicin was measured according to established methods [
26,
27]. To exclude dead cells from analysis, 7-AAD was added to filtered cells 10 minutes before data acquisition and compensation was used to account for the fluorescence of 7-AAD in FL2.
To determine the percent viable cells positive for CSF-1R expression after doxorubicin exposure, COSB cells were plated in triplicate as described for the cytotoxicity assays. Cells were exposed to increasing concentrations of doxorubicin (Bedford Laboratories) for 72 hours and then harvested. Cells were stained for CSF-1R expression using a Cy5.5 labeled CSF-1R antibody and 7-AAD was added to the cells before analysis.
LysoTracker detection
Cell lines were grown until they were approximately 80% confluent and then harvested for analysis. The cells were blocked with normal rat serum in cold PBS containing 2% fetal bovine serum (Atlanta Biologicals, Atlanta, GA) and 2 mmol/L EDTA (Sigma, St. Louis, MO) for 2 minutes, and then incubated with an anti-CSF-1R-RPE antibody (AbD Serotec) for 30 minutes on ice. After incubation, the cells were washed with cold PBS to remove serum and incubated with increasing concentrations of LysoTracker® Deep Red (Life Technologies) in serum free medium on ice for 10 minutes. The cells were centrifuged and washed with cold PBS. To exclude dead cells from analysis, 7-AAD or LIVE/DEAD® Cell Stain (Life Technologies) was added to the cells before acquisition on an Accuri or BD LSRII flow cytometer (BD Biosciences). The CSF-1Rhigh population (~1-2% of the total cell population) was identified using anti-CSF-1R RPE labeling. Approximately 3-5% of the dimmest cell population was used to represent the CSF-1Rlow population. The CSF-1R populations were normalized for comparison by subtracting the background (cells without LysoTracker, negative control) and then dividing the median fluorescence intensity of each concentration of LysoTracker by the median fluorescence intensity of the each cell population without LysoTracker (negative control). Values are presented as relative fluorescence levels for comparison.
Statistical analysis and data presentation
All EC50 calculations were made using Prism 5 Software (GraphPad Inc. San Diego, CA) using a 4-parameter curve fit. Bar graphs are presented as blank-subtracted (adjusted) means ± SD. Comparisons between groups were made using a two-sided Student’s t-test to evaluate statistical significance, and a p-value ≤ 0.05 was considered significant.
Discussion
Human angiosarcoma and canine hemangiosarcoma are aggressive vascular tumors where there are currently no effective treatments [
2,
3,
37]. Angiogenesis and inflammation are key features of hemangiosarcoma [
7,
8], and progenitor cell populations expressing markers of both endothelial and myeloid progenitors support this premise [
8]. Because these cells may participate in both the angiogenic and inflammatory responses, a better understanding of the cell biology would provide insight into new therapeutic opportunities. Here, we show that cell populations enriched for the myeloid marker, CSF-1R, are highly drug resistant in hemangiosarcoma and angiosarcoma, and this resistance may be mediated through drug sequestration within cellular lysosomes.
Lysosomes have been shown to be involved in the sequestration of amine-containing drugs such as doxorubicin [
38,
39]. The sequestration of drugs away from cellular target sites into cytoplasmic organelles prevents the drugs from reaching their targets of interest, leading to ineffective drug therapy. In support of this mechanism, our results suggest that a small population of cells identified by CSF-1R expression is more resistant due to doxorubicin sequestration into what are likely cellular lysosomes. While further studies are needed, the more myeloid-like character of the CSF-1R
high cells may play a role. Prior studies using human myeloid leukemia cell lines demonstrated that the more drug resistant variants of the lines showed increased drug sequestration within lysosomes [
33,
34]. As a consequence of the increased lysosomal drug trapping, nuclear accumulation of the drug was decreased leading to lower cytotoxicity [
34].
More recently, Sukhai
et al. showed that lysosomes isolated from primary human acute myeloid leukemia (AML) cells, CD34
+ AML progenitor cells, as well as human and mouse AML cell lines contained larger lysosomes when compared to lysosomes found in normal human CD34
+ hematopoietic cells [
40]. Although larger in size, the number of lysosomes per cell did not differ significantly between AML cells and normal cells. Alterations in lysosomal size may be indicative of altered metabolic processes, specifically alterations in fatty acid metabolism [
41], which may be important for altered energy needs of cancer cells. Furthermore, primary human AML and AML progenitor cells were more sensitive to the antimalarial agent mefloquine, a quinoline approved for the treatment of malaria and known to accumulate in the lysosomes of the malarial parasite [
42]. These observations provide the rationale for therapeutically targeting the lysosomal compartment in AML, and they also support the investigation of the effects of lysosomal disruption on hemangiosarcomas and angiosarcomas since this approach may target the more drug resistant CSF-1R
high cells.
While it is likely that doxorubicin is sequestered in cellular lysosomes largely due to its lysosomotropic properties [
34], drug sequestration within lysosomes also may be due to ABC transporter activity. Chapuy
et al. determined that the ABCA3 transporter expression was localized to lysosome and multivesicular body membranes in AML, and that ABCA3 expression was associated with unfavorable treatment outcomes for AML patients [
43]. Further analysis in chronic myeloid leukemia (CML) showed that lysosomal storage capacity was increased with increases in ABCA3 expression, indicating that ABCA3 may contribute to drug resistance by facilitating lysosomal drug sequestration [
44]. ABCA3 expression appeared to be low across a population of cells with high CSF-1R expression enriched from hemangiosarcoma cell lines when expression was examined using microarray anlaysis (J-H. Kim, unpublished observation). Thus ABCA3 may not contribute to lysosomal sequestration in hemangiosarcomas. In contrast, ABCB1 expression on lysosomes has been identified and may contribute to doxorubicin sequestration since doxorubicin accumulation was inhibited in the presence of the ABCB1 inhibitor valspodar [
45]. Thus, further studies examining ABCB1 expression and localization in CSF-1R
low versus CSF-1R
high cells are warranted, and additional candidates may be identified through microarray analysis.
Although the hemangiosarcoma and angiosarcoma cell lines showed similar drug resistance mechanisms and these populations could be enriched by targeting CSF-1R expression, the cell surface marker expression profiles differed between species. Previous studies using hemangiosarcoma cell lines demonstrated expression of early hematopoietic (CD34, CD117, and CD113) and endothelial progenitor (CD34 and CD133) markers [
6]. While expression of CD34 and CD117 by both the CSF-1R
low and CSF-1R
high cells was observed in the hemangiosarcoma cell lines, expression of CD133 was exclusive to the CSF-1R
high cell subpopulation. The retention of CD133 expression by the CSF-1R
high cell population brings up the intriguing possibility in hemangiosarcomas that the CSF-1R
high cell population possesses the potential for differentiation into CSF-1R
low cells and may be responsible for maintaining tumor growth autocrine or paracrine loops since we previously determined that the ligands for CSF-1R, CSF-1 and IL-34, are expressed by these and other canine hemangiosarcoma cell lines (B. H. Gorden, unpublished data). Studies using cell sorting of lineage markers or single cell isolation followed by clonal expansion would address both hypotheses, but our group has not yet undertaken these approaches.
In contrast, expression of CD34 and CD117 were observed mainly in the monolayer and the CSF-1R
low cell population from the AS5 cell line, although some expression was observed in the CSF-1R
high cells. We did not detect expression of CD133. Our results are similar to those reported by Lui
et al. where the expression of both CD34 and CD117 was detected in angiosarcomas by immunohistochemistry. In keeping with our observations for CD133, detection of CD133 expression by immunohistochemistry was negative overall [
9]. In both cases, the expression levels of CD133 may have been below the detection limit by immunohistochemistry due to cellular differentiation or tumor heterogeneity and this may also be the case for flow cytometry [
9]. Regardless, the overall expression of these markers in the AS5 cell line reflects previous findings, and they also highlight that differences do exist between human angiosarcomas and canine hemangiosarcomas at the cellular and molecular level even though their overall pathologies appear to be virtually indistinct. Further studies are needed to characterize the potential differences as well as the noted similarities between these tumors.
Acknowledgements
We thank Drs. Aric M. Frantz, and Daisuke Ito for their technical assistance. We also thank Dr. Jaime F. Modiano for helpful comments and the critical reading of this manuscript.
This work was partially supported by Morris Animal Foundation, Grant # D13CA-062 (EBD) and Grant #D14CA-047 (EBD), the Office of the Vice President for Research, University of Minnesota, Grant #21873 (EBD), and with the assistance of the Flow Cytometry Core Facility of the Masonic Cancer Center (supported in part by NIH P30 CA77598).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
BHG conceived parts of the project, contributed to the writing of the manuscript, performed the experiments using the canine hemangiosarcoma cell lines, and analyzed and prepared data for publication. JS performed the flow cytometry and cytotoxicity experiments for the AS5 cell line, generated the LysoTracker data, and helped analyze and prepare date for publication. AK assisted with some of the flow cytometry experiments, data analyses, and interpretation, GKS conceived parts of the project and edited the manuscript, and EBD conceived parts of the project, contributed to the writing of the manuscript, analyzed and interpreted data, and coordinated the project. All authors read and approved the final manuscript.