J Vet Intern Med 2007;21:133–140
Expression of Receptor Activator of Nuclear Factor k-B Ligand (RANKL) in Neoplasms of Dogs and Cats Anne M. Barger, Timothy M. Fan, Louis-Philippe de Lorimier, Ian T. Sprandel, and Kristen O’Dell-Anderson Background: Receptor activator of nuclear factor k-B (RANK), RANK-ligand (RANKL), and the soluble decoy receptor osteoprotegerin (OPG) form a key axis modulating osteoclastogenesis. In health, RANKL-expressing bone stromal cells and osteoblasts activate osteoclasts through RANK ligation, resulting in homeostatic bone resorption. Skeletal tumors of dogs and cats, whether primary or metastatic, may express RANKL and directly induce malignant osteolysis. Hypothesis: Bone malignancies of dogs and cats may express RANKL, thereby contributing to pathologic bone resorption and pain. Furthermore, relative RANKL expression in bone tumors may correlate with radiographic characteristics of bone pathology. Animals: Forty-two dogs and 6 cats with spontaneously-occurring tumors involving bones or soft tissues were evaluated. Methods: A polyclonal anti-human RANKL antibody was validated for use in canine and feline cells by flow cytometry and immunocytochemistry. Fifty cytologic specimens were collected from bone and soft tissue tumors of 48 tumor-bearing animals and assessed for RANKL expression. In 15 canine osteosarcoma (OSA) samples, relative RANKL expression was correlated with radiographic characteristics of bone pathology. Results: Expression of RANKL by neoplastic cells was identified in 32/44 canine and 5/6 feline tumor samples. In 15 dogs with OSA, relative RANKL expression did not correlate with either radiographic osteolysis or bone mineral density as assessed by dual energy x-ray absorptiometry. Conclusions and Clinical Importance: In dogs and cats, tumors classically involving bone and causing pain, often may express RANKL. Confirming RANKL expression in tumors is a necessary step toward the rational institution of novel therapies targeting malignant osteolysis via RANKL antagonism. Key words: Bone pain; Cancer; Companion animals; Dual energy x-ray absorptiometry; Immunocytochemistry; Malignant osteolysis.
umor-bearing companion animals initially may present for clinical signs attributable to pain. Although various types of cancer may cause discomfort, chronic or breakthrough pain is frequently observed in patients with bone malignancies, whether primary or metastatic. Although the exact mechanisms are not completely understood, pain associated with neoplastic bone lesions likely is mediated by 2 basic mechanisms. First, the invasion of cancer cells into normal bone may induce pain directly by stimulating the nociceptor-rich endosteum and periosteum. Second, cancer cells may directly and indirectly activate osteoclasts, leading to painful pathologic osteolysis and facilitating further local tumor expansion.1 These mechanisms responsible for osteolytic pain are the result of altered osteoclastic and osteoblastic activities, and thereby represent dysregulated bone turnover. Although both processes of bone destruction and bone formation are operative during malignant osteolysis, bone tumors that induce excessive and uncompensated bone resorption predispose patients to pain and pathologic fracture.2
T
From the Department of Pathobiology (Barger), and the Department of Veterinary Clinical Medicine (Fan, Lorimier, Sprandel, O’Dell-Anderson), University of Illinois at UrbanaChampaign, Urbana, IL. Findings of this study were presented in part at the 25th Annual Veterinary Cancer Society Conference, Huntington Beach, California, October 23, 2005. Reprint requests: Anne M. Barger, DVM, MS, Department of Pathobiology, 1008 West Hazelwood Drive, Urbana, IL 61802-4714; e-mail:
[email protected]. Submitted May 2, 2006; Revised July 14, 2006; Accepted August 31, 2006. Copyright E 2007 by the American College of Veterinary Internal Medicine 0891-6640/07/2101-0019/$3.00/0
Whether in health or disease, bone resorption is mediated by activated osteoclasts, and factors promoting osteoclastic activity may be either soluble or membrane bound.2,3 Receptor activator of nuclear factor k-B ligand (RANKL) is a membrane-bound protein expressed primarily on the surface of osteoblasts and bone marrow stromal cells and can be secreted in a soluble form by activated T lymphocytes.4,5 Through its binding to receptor activator of nuclear factor k-B (RANK) present on osteoclasts, RANKL controls osteoclast differentiation, activation, and survival.3,5 Given that uncontrolled osteoclastic activity can produce deleterious effects, the biologic activity of RANKL is tightly counter-regulated by osteoprotegerin (OPG), a soluble decoy receptor, normally produced and secreted by bone marrow stromal cells and osteoblasts.6 Thus, the rate, magnitude, and direction of normal bone turnover are dictated by the regulatory and counterregulatory activities exerted by RANK, RANKL, and OPG. In disease states such as malignant osteolysis, excessive bone resorption may be caused by increased RANKL expression, decreased OPG production, or a combination of both.6,7 In human cancer patients, various tumor types capable of inducing pathologic osteolysis have been associated with RANKL expression and include osteosarcoma, prostatic carcinoma, breast carcinoma, multiple myeloma, and squamous cell carcinoma.2,7–9 The expression of RANKL by neoplastic cells potentially offers them a survival advantage, because the degradation of bone matrix results in the local release of growth factors, cytokines, and nutrients conducive for tumorcell mitogenesis and invasion in the bone microenvironment.10 Novel therapies blocking RANKL activity have proven effective recently for reducing pathologic bone
134
Barger et al
loss in human cancer patients,11 underscoring the biologic importance of RANKL in the process of malignant osteolysis. Similar to human cancer patients, various tumor types are associated with bone invasion in companion animals. In dogs, primary skeletal tumors are most commonly associated with malignant osteolysis and include osteosarcoma (OSA), chondrosarcoma, fibrosarcoma, multiple myeloma, and multilobular tumor of bone. Likewise in cats, oral squamous cell carcinoma (OSCC) is a locally infiltrative tumor often resulting in mandibular or maxillary osteolysis from direct tumor invasion. In addition, metastatic carcinomas can produce multifocal osteolytic or osteoblastic skeletal lesions in both species. Despite the similar tumor histology associated with malignant osteolysis in both people and companion animals, the characterization of RANKL expression in bone-involving neoplasms has yet to be reported in dogs and cats. The objectives of this prospective study were (1) to validate a polyclonal anti-human RANKL antibody for use in canine and feline cells; (2) to evaluate the expression of RANKL by immunocytochemistry in tumors of dogs and cats associated with bone involvement; and (3) to determine if correlations exist between relative RANKL expression and radiographic characteristics of bone pathology in 15 dogs with OSA.
Materials and Methods Study Population Forty-two dogs and 6 cats were prospectively evaluated between January 2005 and March 2006. In addition, 2 dogs with appendicular OSA had 2 sites (primary and metastatic) evaluated for RANKL expression, for a total of 50 samples. All patients had a diagnosis of neoplasia by either cytopathology alone (n 5 3), or both cytopathology and histopathology (n 5 47). The study population of dogs consisted of 23 neutered males, 4 intact males, 13 spayed females, and 2 intact females. The median age and weight were 9.1 years (range, 4.1–15.4 years) and 37 kg (range, 13.6–75.9 kg), respectively. Breeds included mixed breed (n 5 11), Labrador Retriever (n 5 7), Rottweiler (n 5 4), Golden Retriever (n 5 2), German Shorthaired Pointer (n 5 2), Saint Bernard (n 5 2), and 14 additional breeds (n 5 1 each). In 33 dogs diagnosed with OSA, affected skeletal sites included distal radius (n 5 11), proximal humerus (n 5 8), proximal tibia (n 5 4), proximal femur (n 5 2), and 8 additional sites (n 5 1 each). In the remaining 9 dogs diagnosed with bone-involving neoplasms, tumor types included chondrosarcoma (n 5 4), prostatic carcinoma (n 5 2), malignant plasma cell tumor (n 5 2), and histiocytic sarcoma (n 5 1). The study population of cats comprised 6 patients, being either neutered male (n 5 5) or spayed female (n 5 1) domestic shorthaired cats. The median age and weight were 14.5 years (range, 9–17 years) and 4.9 kg (range, 3.7–5.9 kg), respectively. All cats were diagnosed with OSCC involving the maxilla (n 5 2) or mandible (n 5 4). Pet owners were informed of available treatment options, including surgery with adjuvant systemic chemotherapy, and traditional or investigational palliative therapies. Patients were treated in accordance with the animal care guidelines of the University of Illinois Institutional Animal Care and Use Committee.
Cell Lines One canine OSA line (HMPOS)a and 1 OSCC line of cat (SCCF1)b were evaluated for RANKL expression. The SaOs-2 human OSA cell line,c known to express RANKL,12 was purchased and used as a positive control. Both OSA cell lines were cultured in Dulbecco’s modified eagle’s medium (DMEM)d supplemented with 2 mM l-glutamine,e and 10% fetal bovine serum (FBS). The SCCF1 cell line was grown in Williams E mediaf supplemented with 2 mM l-glutamine, 0.05 mg/mL gentamicin,g 10 ng/mL epidermal growth factor,h 0.01 nM cholera toxin,i and 10% FBS.
Primary and Secondary Antibodies A rabbit polyclonal anti-human RANKL antibodyj was evaluated for cross-reactivity with canine cells and cells of cats A corresponding rabbit immunoglobulin (IgG1)k was used as an isotype control for flow cytometric analysis. Secondary antibodies used for flow cytometry and immunocytochemistry were a goat anti-rabbit IgG:fluorescein isothiocyanate (FITC) conjugatel and a goat anti-rabbit antibody,m respectively.
Flow Cytometry (Antibody Validation) The cells were harvested and resuspended (1 3 106 cells/mL) as a stock concentration in cold phosphate-buffered saline (PBS). Cell stock solutions (200 mL) were added to flow tubes, washed with cold PBS, centrifuged at 1000 3 g for 5 minutes, resuspended in 1000 mL of cold fixation/permeabilization buffer,n pulse-vortexed, and incubated at 4uC for 12 hours. After incubation, cells were washed twice with 2000 mL of cold permeabilization buffer.m Permeabilized cells were resuspended with 200 mL PBS/2% bovine serum albumin (BSA) (wt/vol) for 15 minutes at room temperature, and then incubated with primary antibody (anti-RANKL or isotype control, 1:50) at 4uC for 30 minutes. After a wash step with the permeabilization buffer, the secondary antibody (antirabbit FITC conjugated, 1:10) was incubated with cells at 4uC in the dark for 30 minutes. The cells were washed twice, resuspended in 500 mL of PBS, and analyzed on an Epics-XL flow cytometern based on their forward and side scatter properties and FITC fluorescence.
Immunocytochemistry (Antibody Validation) Cytospin preparations of all cell lines (HMPOS, SCCF1, and SaOs-2) were incubated in acetone for 10 minutes, allowed to air dry, and loaded on the DAKO autostainer.o To minimize nonspecific peroxidase background staining, all preparations were blocked with 10% hydrogen peroxide for 10 minutes. Cytospin preparations were incubated with the polyclonal anti-human RANKL antibody (1:50, 60 minutes). Next, samples were incubated with the secondary goat anti-rabbit immunoglobin (1:200, 30 minutes) followed by 3,39-diaminobenzidine chromagen solution (7 minutes), and then counterstained with Mayer’s hematoxylinl (5 minutes). The SaOs-2 cell line was used as a RANKLpositive control. For each respective cell line, staining only with the secondary antibody was used as a negative control. Positive RANKL expression was characterized by membranous staining with or without cytoplasmic staining.
Immunocytochemistry (Patient Samples) Fine-needle aspiration (n 5 3) or touch preps of tissue biopsies (n 5 47) were collected for RANKL analysis. An initial sample was evaluated after staining with Wright-Giemsa to evaluate the cellularity of the sample and make a cytologic diagnosis. Unstained slides were prepared as previously described for immunocytochemistry antibody validation. Cytospin preparations of SaOs-2 cell line
RANKL-Expressing Tumors were used as a RANKL-positive control, and patient samples incubated with secondary antibody only were used as a negative control. Sample staining for RANKL was evaluated by 1 pathologist (AMB), who was blinded to the radiographic characteristics of each bone-involving tumor. Cytology samples were graded based on percentage of positive staining cells as follows: negative (0–10% of cells stain positive), weak (10–20% of cells stain positive), moderate (21–50% of cells stain positive), and strong (.50% of cells stain positive).
Radiographic Characterization of Primary Tumor in 15 Dogs with Appendicular OSA At presentation, plain film radiographs of 15 appendicular OSAs were evaluated by 1 investigator (KOA), blinded to RANKL staining results. Radiographs from all primary tumors were subjectively characterized for their degree of osteolysis and sequentially ranked in order of increasing bone lysis (1–15).
Relative Bone Mineral Density of Primary Tumor in 15 Dogs with Appendicular OSA At presentation, dual energy x-ray absorptiometry (DEXA) scans of 15 appendicular OSA were performed to measure bone mineral density (BMD) of the primary tumor and equivalent regions of the normal contralateral limb by 1 investigator (LPdL), blinded to the RANKL staining results. Three representative regions of tumor and normal limb were acquired and designated (T1, T2, and T3) and (N1, N2, and N3), respectively. Relative bone mineral density (rBMD) of the primary tumor was calculated by the following formula and sequentially ranked in order of decreasing rBMD (1–15): rBMD ~ ½average (T1 z T2 z T3 ) = average (N1 z N2 z N3 ) ð1Þ
Statistical Analysis In 15 canine appendicular OSA patients, where both plain film radiographs and DEXA scans of the affected site were available, relative RANKL expression of each tumor was quantified by enumerating RANKL-positive cells identified per 100 cells examined, and sequentially ranked in order of increasing positive RANKL expression (1–15). Spearman rank correlation test was used to evaluate the correlations between (1) relative RANKL expression and radiographic osteolysis, and (2) relative RANKL expression and relative bone mineral density. Statistical analysis was performed using commercial computer software. Statistical significance was defined as P , .05.
Results Polyclonal Anti-Human RANKL Antibody Validation Human (SaOs-2), canine (HMPOS), and cat (SCCF1) cell lines were all positive for RANKL protein expression as assessed by flow cytometry and immunocytochemistry. Flow cytometric analysis demonstrated that a majority of cells (.85%), regardless of species, expressed RANKL protein as identified by the distinct population with positive FITC staining (Fig 1). Immunocytochemistry identified predominant membranous staining for RANKL in human, canine, and cat cell lines (Fig 2). Based on the protein detection methodologies employed in this study, the findings support cross-
135
Fig 1. Flow cytometric detection of RANKL. Expression of RANKL in (A) human OSA SaOs-2 cells, (B) canine OSA HMPOS cells, and (C) OSCC SCCF1 cells of cats. Open line represents respective cell lines with isotype control, and filled shaded line represents respective cell lines with polyclonal anti-human RANKL antibody. Positive RANKL expression demonstrated in the majority (.85%) of cells evaluated (R2 region). RANKL, receptor activator of nuclear factor k-B ligand; OSA, osteosarcoma; OSCC, oral squamous cell carcinoma.
reactivity of the polyclonal anti-human RANKL antibody for both canine cells and cells of cats.
RANKL Expression in Spontaneous Tumors A total of 50 cytologic specimens derived from 48 tumor-bearing patients (canine, n 5 42 and cat, n 5 6) were examined. Of the 50 specimens, 46 were derived from bone, whereas the remaining 4 samples were collected from affected soft tissue (OSA skin metastases, n 5 2; primary prostatic carcinoma, n 5 1; and OSCC soft tissue involvement, n 5 1). Tumor types represented in this study included OSA (appendicular, n 5 29; axial, n 5 4; skin metastases, n 5 2), OSCC (mandible, n 5 4 and maxilla, n 5 2), chondrosarcoma (n 5 4), prostatic carcinoma (n 5 2), malignant plasma cell tumor (n 5 2), and histiocytic sarcoma (n 5 1). Tumor types staining positive for RANKL included primary OSA (23/33), skin metastases of OSA (2/2), OSCC (5/6), chondrosarcoma (1/4), prostatic carcinoma (2/2), malignant plasma cell tumors (2/2), and histiocytic sarcoma (1/1) (Fig 3A– D; Table 1).
Correlation of Relative RANKL Expression and Radiographic Characteristics of Bone Pathology In 15 canine appendicular OSA patients, where both plain film radiographs and DEXA scans of the affected site were available, an attempt was made to correlate relative RANKL expression with subjective (degree of osteolysis on plain radiographs) and objective (rBMD as assessed by DEXA) radiographic characteristics of bone pathology. It was hypothesized that increasing RANKL expression by OSAs would positively correlate with increasing radiographic osteolysis, decreasing rBMD of the primary tumor, or both. No significant correlation between relative RANKL expression and radiographic osteolysis (rs 5 0.16, P . .05) or rBMD (rs 5 0.35, P . .05) could be identified in these 15 dogs.
Discussion Malignant bone pain is a common and important complication associated with advanced metastatic can-
136
Barger et al
Fig 2. Immunocytochemical staining of RANKL. Cytoplasmic and membranous staining of RANKL of (A) human OSA SaOs-2 cells (original magnification 3 1000), (B) canine OSA HMPOS cells (original magnification 31000), (C) OSCC SCCF1 cells of cats (original magnification 3500), (D) negative control for canine OSA HMPOS cells (original magnification 31000), and (E) negative control for OSCC SCCF1 cells of cats (original magnification 3500). Arrows indicate representative cells. RANKL, receptor activator of nuclear factor k-B ligand; OSA, osteosarcoma; OSCC, oral squamous cell carcinoma.
cers in human patients. Likewise as a consequence of pathologic skeletal destruction, primary and metastatic bone cancers in companion animals often compromise quality of life by induction of chronic or breakthrough pain. The effective management of malignant osteolysis and associated pain syndromes requires an understand-
ing of the mechanisms used by cancer cells to invade bone. Therefore, in dogs and cats suffering from painful bone tumors, identifying operative pathways involved in malignant bone destruction is a necessary and fundamental step toward the development and institution of appropriate therapies.
RANKL-Expressing Tumors
137
Fig 3. RANKL expression in cytologic preparations of patient samples. (A) OSA strongly positive for RANKL (original magnification 3500), (B) OSA negative for RANKL expression (original magnification 31000), (C) OSCC with moderate staining for RANKL expression (original magnification 31000), (D) Prostatic carcinoma with moderate staining for RANKL expression (original magnification 31000), and (E) Negative control for representative OSA (original magnification 3500). Arrows indicate representative cells. RANKL, receptor activator of nuclear factor k-B ligand; OSA, osteosarcoma; OSCC, oral squamous cell carcinoma.
Although the dynamics of bone metabolism are complex, recent investigations have characterized the functional roles of RANK, RANKL, and OPG, a triad of cellular proteins that regulate the rate, magnitude, and direction of bone turnover. During skeletal homeostasis, a balance exists among RANK, the agonist
RANKL, and the decoy receptor OPG, thereby ensuring stable and normal bone metabolism. However, during malignant osteolysis, dysregulated osteoclast activation occurs and potentially leads to excessive bone resorption. Mechanistically, tumor cells can promote pathologic bone resorption either by stimulating osteoclasts
138
Barger et al
Table 1. Immunocytochemical staining for RANKL for individual tumors. Strong Moderate Weak Negative OSA (n 5 35) Chondrosarcoma (n 5 4) Plasma cell (n 5 2) Prostatic Carcinoma (n 5 2) Histiocytic sarcoma (n 5 1) OSCC (n 5 6)
13 -
8 1 2 4
4 2 1 1 1
10 2 1
RANKL, receptor activator of nuclear factor k-B ligand; OSA, osteosarcoma; OSCC, oral squamous cell carcinoma. ‘‘-’’ 5 no tumors staining in this category.
directly through production of RANKL or indirectly by prompting resident stromal cells or osteoblasts to express RANKL.1,7 Tumor types demonstrated to express RANKL in people include carcinomas of the breast, prostate, thyroid, kidney, and lung, as well as osteosarcoma, squamous cell carcinoma, multiple myeloma, and malignant melanoma.2,3,13 Despite the similar tumor types associated with malignant osteolysis in both people and companion animals, the mechanisms involved in pathologic osteoclast activation, such as the direct expression of RANKL by tumor cells have yet to be reported for dogs and cats and were the main premise for this study. In order to accurately identify RANKL expression in spontaneously arising bone tumors derived from companion animals, it was necessary to validate a commercially available rabbit polyclonal anti-human RANKL antibody for use on canine cells and cells of cats. In order to confirm the specific recognition of cellular RANKL, 2 different protein detection methodologies were used for antibody validation, including flow cytometry and immunocytochemistry. Through the use of a known RANKL-expressing human OSA cell line, SaOs-2, the true identification of cellular RANKL was ensured in both protein methodologies evaluated in this report. Based on the results of this study, the employed rabbit polyclonal anti-human RANKL antibody demonstrated cross-reactivity for both canine cells and cells of cats. Despite the utility of detecting RANKL by flow cytometry and immunocytochemistry, the suitability of the tested antibody for RANKL identification by means of immunohistochemistry was not determined. The investigators favored immunocytochemistry on fresh cytologic preparations over immunohistochemistry on formalin-fixed, paraffin-embedded tissues for various reasons, including the avoidance of antigen retrieval and decalcification steps, both of which are known potentially to lead to altered antibody binding. In evaluating samples derived from tumor-bearing companion animals, the immunostaining results demonstrate that the majority of osteolytic tumors in dogs and cats express RANKL. This finding may support 1 mechanism by which canine neoplastic cells and neoplastic cells of cats may induce bone destruction, because the direct expression of RANKL by tumor cells
may favor dysregulated osteoclast activation with subsequent pathologic osteolysis and bone pain. Given similar data in the human literature, it is not surprising or unexpected to observe that a majority of canine tumors and tumors of cats in this report expressed RANKL.2,3,13 Nonetheless, the findings of this study are novel and important, because they not only support the hypothesis that some canine tumors and tumors of cats may directly induce osteolysis via RANKL expression but they also strengthen the value of companion animals as comparative tumor models for studying the process of malignant osteolysis. In this investigation, of 35 (primary, n 5 33; metastatic, n 5 2) canine OSA samples evaluated, 10 primary tumors (29%) were negative for RANKL expression. This finding is not in total agreement with a study of humans in which all bone tumors (n 5 8) evaluated were positive for RANKL expression.2 Although, intuitively, it is expected that tumors that arise from or invade bone would express RANKL, the process of malignant osteolysis can be mediated by both direct and indirect mechanisms. For this reason, malignant canine OSA cells still could induce pathologic bone destruction via the promotion of RANKL expression by nonneoplastic, resident stromal cells or osteoblasts, thereby indirectly stimulating osteoclast differentiation and activation.7 Furthermore, in OSA samples lacking RANKL expression, it remains plausible that bone destruction could be mediated by other cellular products, because several proteins are known to induce osteolysis via osteoclast activation, including interleukin (IL)-1, IL-6, macrophage inflammatory protein 1a (MIP1a), and parathyroid hormone-related protein (PTHrP).3,6 This possibility is supported by the observation that bone metastases of human breast carcinoma and malignant melanoma have been shown to produce PTHrP, and the neoplastic cells in such cases often are negative for RANKL expression.2,14,15 In addition to PTHrP production, it is also proposed in human melanoma that the neoplastic cells may induce the production of IL-11 and tumor growth factor (TGF)b by osteoblasts, which in turn causes osteolysis.16,17 The relevance of these cytokine pathways, in relation to malignant osteolysis, has not been thoroughly evaluated in veterinary medicine but may account for the lack of direct RANKL expression by some tumor cells identified in the current study. In an attempt to correlate RANKL expression to detectable bone changes observed clinically with imaging modalities, we evaluated 15 cases of canine appendicular OSA for the degree of osteolysis, subjectively and objectively, via plain radiographs and DEXA scan, respectively. The relative RANKL expression did not correlate with the degree of osteolysis observed on radiographs or to the rBMD calculated with DEXA analysis. One potential explanation for this negative finding could stem from the small sample size (n 5 15), and perhaps the evaluation of a larger number of clinical specimens would demonstrate some linear relationship between RANKL expression and radiographic osteolysis. Another potential reason for the
RANKL-Expressing Tumors
absence of any correlation could be tumor heterogeneity, as OSA is a tumor type often showing a mixed radiographic pattern, with areas of severe lysis and areas of marked abnormal bone production. Therefore, the specific tumor area sampled may influence RANKLstaining results and not necessarily correlate with the radiographic appearance of the tumor as a whole. In support of this possibility, a recent study describes RANKL overexpression by stromal cells specifically at the interface of malignant plasma cells with normal marrow elements.8 In addition, other important factors capable of inducing malignant osteolysis, such as the previously described cytokines, also could play a primary role in the osteolytic process, thereby attenuating the possible correlation between 1 specific component, such as RANKL, and the global outcome as evaluated by imaging studies. Finally, imaging modalities reflect cumulative bone changes over time and may not accurately represent ongoing processes influenced by the expression of various cellular proteins, such as RANKL, in the actively evolving tumor microenvironment. Although this report offers new information, there are several inherent limitations that should be discussed. First, only a limited number of tumor types were evaluated for RANKL expression, with the majority of them being derived from dogs presented with appendicular OSA, simply based upon the high prevalence of this tumor type. Therefore, besides canine OSA, conclusive statements regarding RANKL-expression patterns in other tumor types is not possible without additional studies evaluating a larger and more diverse group of tumor types. In addition, the functional significance of RANKL expression in primary bone sarcomas, as predominantly demonstrated in this study, remains poorly defined because most investigations conducted in human cancer patients address the participatory role of RANKL in the context of skeletal carcinoma metastases. However, the mechanisms responsible for malignant osteolysis may be universal and, regardless of tumor histology, common to both primary and metastatic skeletal malignancies. Another potential limitation of this investigation is the sole use of immunocytochemistry to evaluate RANKL expression in spontaneous tumor samples. Because the absolute number of cells retrieved by cytology tends to be limited, in circumstances of great tumor heterogeneity, cytology may not accurately reflect overall RANKL expression. The areas of greatest RANKL expression may occur at the periphery of the tumor interacting with normal surrounding bone, and samples obtained from a focal area of the tumor may not be representative of the tumor as a whole. Thus, the exclusive use of immunocytochemistry, in the absence of corroborative immunohistochemistry, has the potential for underestimating the participatory role of RANKLexpressing, nonneoplastic cells involved in the initiation and progression of malignant osteolysis. Finally, because it is only 1 component of a triad of cellular proteins that regulate bone metabolism, it is likely that the sole expression of RANKL may not be
139
the best reflection of ongoing pathologic osteolysis. Clearly, the quantification of OPG, the counter-regulatory decoy molecule of RANKL, would be important to globally assess the rate, magnitude, and direction of bone turnover. In this study, it would have been ideal to measure OPG concentrations either in the serum or in the local tumor microenvironment from all samples in which RANKL expression was assessed. However, commercially available reagents for assessing OPG have yet to be validated in companion animals. The investigations currently are evaluating OPG in cancerbearing companion animals and should hopefully complement the findings regarding RANKL expression reported in this study. Despite the limitations associated with this investigation, the validation of a polyclonal antibody and the identification of companion animal tumors that express RANKL are important steps for studying malignant osteolysis in dogs and cats. Already, novel therapeutic strategies are being investigated to halt malignant bone resorptive processes as demonstrated by clinical trials in humans evaluating anti-RANKL monoclonal antibodies.11 Furthermore, aminobisphosphonates, known to inhibit osteoclasts by other mechanisms, recently have been demonstrated to reduce RANKL expression in various experimental systems including a human OSA cell line and murine OSCC tumor model.18,19 Because the use of aminobisphosphonates for the treatment of malignant bone diseases in veterinary oncology recently has been described,20,21 investigating the dynamics of tumor RANKL expression before and after aminobisphosphonate therapy may warrant further investigation. In conclusion, this study demonstrates validation of a rabbit polyclonal anti-human RANKL antibody for use on both canine cells and cells of cats. In addition, it reports on the expression of RANKL in common veterinary tumors known to result in pathologic osteolysis and morbidity from bone pain. This study should lead to further investigations on the role of the RANK/RANKL/OPG axis in companion animal tumors associated with malignant osteolysis. Finally, it is hoped that these findings will facilitate the rational development of novel therapeutic strategies targeting direct or indirect inhibition of RANKL in bone malignancy, ultimately leading to improved quality of life scores in dogs and cats suffering from malignant osteolysis and associated bone pain syndromes.
Footnotes a
Canine OSA line (HMPOS) provided by James Farese, University of Florida, Gainesville, FL b OSCC line of cat (SCCF1) provided by Tom Rosol, Ohio State University, Columbus, OH c SaOs-2 human OSA cell line, ATCC, Manassas, VA d DMEM, Gibco, Grand Island NY e l-Glutamine, Biosource, Rockville, MD f Williams E media, Sigma, St Louis, MO g Gentamicin, Pepro Tech, Rocky Hill, NJ h Epidermal growth factor, Calbiochem, LaJolla, CA
140
Barger et al
i
Cholera toxin, Axxora, San Diego, CA Rabbit polyclonal anti-human RANKL antibody, Axxora, San Diego, CA k Rabbit immunoglobulin, SCB, Santa Cruz, CA l Goat anti-rabbit IgG:FITC conjugate, DAKO, Carpinteria, CA m Goat anti-rabbit antibody, eBiosciences, San Diego, CA n Fixation/permeabilization buffer, Beckman Coulter, Hialeah, FL o Dako Autostainer, DAKO, Carpinteria, CA j
Acknowledgments This study was conducted at the Cancer Care Clinic of the Veterinary Teaching Hospital, University of Illinois at Urbana-Champaign, Urbana, IL. This project was partly funded by the University of Illinois Companion Animal Memorial Fund. The authors thank Mrs Lisa Shipp of the Veterinary Diagnostic Laboratory for assistance with immunocytochemistry. The authors also thank Drs Hugues Lacoste, David Heller, Lorin Hillman, and Jackie Wypij of the Cancer Care Clinic for sample submissions.
References 1. Clohisy DR, Mantyh PW. Bone cancer pain and the role of RANKL/OPG. J Musculoskel Neuron Interact 2004;4:293–300. 2. Good CR, O’Keefe RJ, Puzas JE, Schwarz EM, Rosier RN. Immunohistochemical study of receptor activator of nuclear factor kappa-B ligand (RANK-L) in human osteolytic bone tumors. J Surg Oncol 2002;79:174–179. 3. Tada T, Jimi E, Okamoto M, Ozeki S, Okabe K. Oral squamous cell carcinoma cells induce osteoclast differentiation by suppression of osteoprotegrin expression in osteoblasts. Int J Cancer 2005;116:253–262. 4. Takayanagi H. Inflammatory bone destruction and osteoimmunology. J Periodontal Res 2005;40:287–293. 5. Nishimura M, Yuasa K, Mori K, et al. Cytological properties of stromal cells derived from giant cell tumor of bone (GCTSC) which can induce osteoclast formation of human blood monocytes without cell to cell contact. J Orthop Res 2005;23:979–987. 6. Roodman GD. Mechanisms of bone metastasis. N Eng J Med 2004;360:1655–1664. 7. Wittrant Y, Theoleyre S, Chipoy C, et al. RANKL/RANK/ OPG: New therapeutic targets in bone tumors and associated osteolysis. Biochim Biophys Acta 2004;1704:49–57.
8. Roux S, Meignin V, Quillar J, et al. RANK (receptor activator of nuclear factor-kappa-B) and RANKL expression in multiple myeloma. Br J Haematol 2002;117:86–92. 9. Lee Y, Schwarz E, Davies M, et al. Differences in the cytokine profiles associated with prostate cancer cell induced osteoblastic and osteolytic lesions in bone. J Orthop Res 2003;21:62–72. 10. Clines GA, Guise TA. Hypercalcemia of malignancy and basic research on mechanisms responsible for osteolytic and osteoblastic metastasis to bone. Endocr Relat Cancer 2005;12: 549–583. 11. Body JJ, Fancon T, Coleman RE, et al. A study of the biological receptor activator of nuclear factor-kB ligand inhibitor, Denosumab, in patients with multiple myeloma or bone metastases from breast cancer. Clin Cancer Res 2006;12:1221–1228. 12. Kinpara K, Mogi M, Kuzushima M, Togari A. Osteoclast differentiation factor in human osteosarcoma cell line. J Immunoassay 2000;21:327–340. 13. Huang L, Cheng YY, Chow LTC, Zheng MH, Kumta SM. Tumour cells produce receptor activator of NF-kB ligand (RANKL) in skeletal metastases. J Clin Pathol 2002;55:876– 879. 14. Yeung JS, Eton O, Buton DW, et al. Hypercalcemia due to parathyroid hormone-related protein secretion by melanoma. Horm Res 1998;49:288–291. 15. Bundred NJ. Osteolytic factor in breast tumor cells: Invited commentary. World J Surg 1995;19:297–298. 16. Zhang Y, Naoya F, Tomoko O, et al. Production of interleukin-11 in bone-derived endothelial cells and its role in the formation of osteolytic bone metastasis. Oncogene 1998;16: 693–703. 17. Morinaga Y, Fujita N, Ohishi K, Tsuruo T. Stimulation of interleukin-11 production from osteoblast-like cells by transforming growth factor-b and tumor cell factors. Int J Cancer 1997;71:422–428. 18. Kobayashi A, Hirano F, Makino I. The inhibitory effect of bisphosphonates on glucocorticoid-induced RANKL expression in human cells. Scand J Rheumatol 2005;34:480–484. 19. Cui N, Numora T, Noma H, et al. Effect of YM529 on a model of mandibular invasion by oral squamous cell carcinoma in mice. Clin Cancer Res 2005;11:2713–2719. 20. Fan TM, de Lorimier LP, Charney SC, et al. Evaluation of intravenous pamidronate administration in 33 cancer-bearing dogs with primary or secondary bone involvement. J Vet Intern Med 2005;19:74–80. 21. Lacoste H, Fan TM, de Lorimier LP, et al. Urine Ntelopeptide excretion in dogs with appendicular osteosarcoma. J Vet Intern Med 2006;20:335–341.
