Association between Absolute Tumor Burden and Serum Bone-Specific Alkaline Phosphatase in Canine Appendicular Osteosarcoma

Sternberg, R.A., Pondenis, H.C., Yang, X., Mitchell, M.A., O'Brien, R.T., Garrett, L.D., Helferich, W.G., Hoffmann, W.E. and Fan, T.M. (2013), Association between Absolute Tumor Burden and Serum Bone-Specific

Alkaline Phosphatase in Canine Appendicular

Osteosarcoma. Journal of Veterinary Internal Medicine,

27: 955–963. doi: 10.1111/jvim.12121

Abstract

Background

In dogs with appendicular osteosarcoma (OSA), increased pretreatment serum bone-specific alkaline phosphatase (BALP) activity is a negative prognostic factor, associated with shorter disease-free intervals and survival times, but a biologic basis for observed differential serum BALP activities in canine OSA patients remains incompletely defined.

Objective

Serum BALP activity will correlate with absolute tumor burden in dogs with OSA.

Animals

This study included 96 client-owned dogs with appendicular OSA.

Methods

In canine OSA cell lines, the expression and membranous release of BALP was evaluated in vitro. The correlation between serum BALP activity and radiographic primary tumor size was evaluated in OSA-bearing dogs. In dogs developing visceral OSA metastases, serial changes in serum BALP activities were evaluated in relation to progression of macroscopic metastases, and visceral metastatic OSA cells were evaluated for BALP expression.

Results

In vitro, BALP expression was not associated with either tumorigenic or metastatic phenotype, rather the quantity of membranous BALP released was proportional with cell density. In dogs devoid of macroscopic metastases, there was a positive correlation between serum BALP activity and absolute primary tumor size. In dogs with progressive OSA metastases, serum BALP activity increased and coincided with the development of macroscopic metastases. OSA cells derived from visceral metastatic lesions retained BALP expression.

Conclusions and Clinical Importance

Tumor burden is a determinant of serum BALP activity in dogs with appendicular OSA. The association between increased pretreatment BALP activity and negative clinical prognosis may simply be attributed to greater initial tumor burden, and consequently more advanced tumor stage.

Abbreviations

ALP: alkaline phosphatase
BALP: bone-specific alkaline phosphatase
NTx: N-telopeptide
OSA: osteosarcoma
PI-PLC: phosphatidylinositol-specific phospholipase C
rBMD: relative bone mineral density
TALP: total alkaline phosphatase

Alkaline phosphatase (ALP) consists of a group of metalloenzymes that hydrolyze monophosphate esters at alkaline pH. There are 2 different isoenzymes of ALP, produced from different genes, and several different isoforms, derived from posttranslational modification in isoenzymes. The liver and bone isoforms, products of the tissue nonspecific ALP gene, comprise the majority of serum total ALP (TALP), whereas corticosteroid ALP, an intestinal ALP gene product, also contributes to TALP in dogs.[1] Bone ALP (BALP) is produced by osteoblasts and is anchored to the outer membrane surface by a hydrophobic glycosylphosphatidylinositol (GPI) linkage.[2] This type of membrane insertion permits the release of linked proteins, including BALP, from the outer membrane by the action of 1 or more GPI-specific phospholipases.[3]

In both dogs and humans with appendicular osteosarcoma (OSA), increased pretreatment serum TALP and BALP activities are negative prognostic factors, associated with shorter disease-free intervals and survival times.[4-11] Because of the restricted cellular source of BALP, direct assessment of BALP activity might serve as a more accurate biomarker than TALP in the evaluation and monitoring of patients with OSA. Several biologically based theories have been proposed to explain the variability in BALP activities among individuals with OSA, including (1) differences in rate of enzyme production or release by malignant osteoblasts, (2) differences in degree and extent of host reparative osteoblastic responses, and (3) variability in malignant osteoblast tumor burden.

In human OSA patients, a positive correlation between serum BALP activity and absolute tumor volume has been identified, with absolute tumor volume being a predictor of metastasis-free survival and overall survival.[12-14] In addition, patients with metastatic OSA have significantly higher serum BALP activity than those with nonmetastatic OSA, suggesting that overall tumor burden contributes to serum BALP.[12, 15] Despite the definitive relationships observed for serum BALP activity, OSA burden, and clinical outcome in people, the biologic basis for the correlation between BALP activities and prognosis in OSA-bearing dogs remains speculative and incompletely defined.

The purpose of this study was to investigate plausible biologic foundations underlying the prognostic relevance of serum BALP activities in canine appendicular OSA. The objectives of this study were as follows: (1) to evaluate the contributions of inherent biologic phenotype and absolute cell number on BALP expression in canine OSA cell lines in vitro, and (2) to evaluate the relationship between serum BALP activities and various tumor parameters in dogs with appendicular OSA. We hypothesized that serum BALP is a surrogate of malignant osteoblast numbers and would correlate with absolute tumor burden in OSA-bearing dogs.

Materials and Methods

In Vitro Studies

Cell Lines

Five canine OSA cell lines including the HMPOS (provided by Dr James Farese, University of Florida) and the KOS series 001-004 (provided by Dr Chand Khanna, National Cancer Institute) were evaluated for BALP expression. All cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Cell cultures were maintained at 37°C in 5% CO₂ and passaged twice weekly.

Primary Osteoblast Isolation and Culture

Bone fragments were obtained from the proximal humeri of 4 skeletally normal dogs submitted to the University of Illinois Veterinary Diagnostic Laboratory for necropsy in good postmortem condition. Care was taken to obtain samples from the metaphyseal region predominantly composed of metabolically active trabecular bone. Briefly, bone fragments of about 2 mm³ were washed with PBS 3 times and digested in 2 mg/mL type IV collagenase. The enzymatic reaction was stopped by adding an equal volume of DMEM with 10% FBS. Bone fragments were cultured in 100 mm culture dishes in DMEM with 10% FBS. After migration of cells off bone fragments and onto the dish, cells were maintained in modified osteoblast differentiation medium[16] (DMEM with 10% FBS, 1% Pen/Strep, 0.2 mM ascorbic acid, 10 mM beta-glycerol phosphate, and 1 μM dexamethasone), which was changed 3 times weekly until cells were harvested.

PI-PLC Induced Release of BALP from Cell Membranes

HMPOS cells were harvested and pelleted in microcentrifuge tubes at counts of 4 × 10⁶, 2 × 10⁶, 1 × 10⁶, and 5 × 10⁵ per tube. Cells were washed with phosphate buffered saline (PBS) and resuspended in 0.5 mL PBS with or without 0.5 U phosphatidylinositol-specific phospholipase C (PI-PLC) isolated from Bacillus cereus1 for 30 minutes at 37°C. After incubation, cells were pelleted, the supernatant was collected for Western blot, and the cells were resuspended in media, plated on chamber well slides, and allowed to adhere for ALP staining.

Cytologic Detection of ALP Activity

Unstained slides of PI-PLC treated and untreated HMPOS cells were stained with nitroblue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate toluidine salt2 (NBT/BCIP), an ALP substrate, as previously reported.[17] Briefly, slides were incubated for 10 minutes at room temperature with sufficient amounts of NBT/BCIP to coat the slide. Slides were rinsed with water, blotted dry, and examined microscopically. Positive staining was indicated by brown to black staining of the cell surface.

Western Blot Analysis

Cellular proteins from OSA cell lines and canine osteoblasts were extracted with a commercial reagent3 and concentrations quantified with a standard assay kit.4 For the detection of total cellular BALP, 50 μg of protein was collected from each respective OSA and osteoblast cell line. For assessing the amount of membranous cleaved BALP released into cell culture media, a 20 μL aliquot of supernatant derived from varying HMPOS cell densities after PI-PLC incubation was used. Protein was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using 10% polyacrylamide gels and transferred onto a nitrocellulose membrane. After 1 hour blocking at room temperature in 5% nonfat dry milk PBS-Tween, the membrane was incubated with a rabbit monoclonal anti-human/mouse BALP antibody5 (1 : 1,000) in 5% nonfat dry milk PBS-Tween for 1 hour at room temperature. The membrane was washed 3 times and then incubated for 1 hour at room temperature with a horseradish peroxidase conjugated anti-rabbit secondary antibody (1 : 1,000) in 5% nonfat dry milk PBS-Tween and developed using a standard chemiluminescence detection kit.6 For subjective comparison of BALP expression among canine OSA and normal osteoblast cell lines, rabbit anti-β-actin polyclonal antibody5 (1 : 5,000) was used to ensure equivalent loading of protein in separate wells.

Canine Studies

Study Population

The study included dogs with OSA that were evaluated at the University of Illinois Cancer Care Clinic between 2003 and 2012. All dogs had a diagnosis of appendicular OSA confirmed by either histopathology or cytology with concurrent positive ALP staining.[17] All dogs included in the study had orthogonal radiographs of the primary tumor, 3-view thoracic radiographs, DEXA scan, and urine and serum collected at the time of diagnosis. Study candidates were excluded if they had previously received other forms of treatment including radiation treatment, cytotoxic chemotherapy, or IV bisphosphonates and if they had radiographically apparent pulmonary metastases or metastases to other distant sites at the time of diagnosis. After their initial visits, all dogs were treated palliatively for OSA, receiving coarse-fraction radiation treatment with or without an IV aminobisphosphonate. Dogs were re-evaluated at 28-day intervals, at which time serum and urine were collected. When owner consent was given, necropsy was performed at the time of euthanasia.

Serum BALP

Serum was collected at the time of initial presentation, and samples were stored at −80°C until analysis. Serum BALP was evaluated with a commercial ELISA7 test kit utilizing a murine monoclonal anti-BALP antibody, previously validated for use in dogs.[18] Reference range for dogs ≥2 years of age is 0–14 U/L (reflects mean activity ±2 standard deviations).[18]

Urine NTx Excretion

Urine was collected (voided or cystocentesis) at the time of initial presentation. Urine samples were immediately centrifuged at 4°C, 400 × g for 10 minutes, and the supernatant was collected and stored at −80°C until analysis. Urine NTx concentrations were measured with a commercial ELISA8 test kit, previously validated for use in dogs[19, 20] and expressed as normalized nanomolar (nM) bone collagen equivalents (BCE) per millimolar (mM) urinary concentration of creatinine.

Radiography

Anterior-posterior (AP) and lateromedial radiographic views of affected limbs were evaluated and interpreted separately for each dog by 1 investigator (RAS). Radiographic assessment of tumor involvement was determined based on the following characteristic changes in bone: lysis of cortical or medullary bone, sclerosis, and periosteal new bone formation.[21]

Calculation of Tumor Size

The following size parameters similar to those defined by Bieling et al[14] were determined from each dog's radiographic images:

Absolute tumor length (ATL): the greater longitudinal extension of the tumor measured on either AP or lateral view (cm).
Absolute tumor width (ATW): horizontal tumor extension measured on the AP view.
Absolute tumor depth (ATD): horizontal tumor extension measured on the lateral view.
Absolute tumor plane (ATP): the maximal cross-sectional area calculated from the tumor length and the larger of the 2 diameters (width or depth), designated Max [ATW, ATD], by the ellipse formula:
- $display math$
Relative tumor plane (RTP): ATP divided by body surface area (BSA, cm²/m²).
Absolute tumor volume (ATV): calculated from the 3 parameters length, width, and depth, by the ellipsoid formula:
- $display math$
Relative tumor volume (RTV): ATV divided by BSA (cm³/m²).
Absolute tumor area (ATA): the cross-sectional area calculated from the tumor width and depth by the ellipse formula:
- $display math$
Relative tumor area (RTA): ATA divided by BSA (cm²/m²).
Absolute tumor surface area (ATSA): the surface area calculated from the tumor width, depth, and length by the approximate ellipsoid surface area formula[22, 23] (Knud Thomsen's formula):
- $display math$
where p = 1.6075 and a, b, and c represent the semiaxes with a < b < c.
Relative tumor surface area (RTSA): ATSA divided by BSA (cm²/m²).

Relative Bone Mineral Density (rBMD) of Primary Tumor

At initial presentation, DEXA9 scans were performed to measure rBMD of the primary tumor and an equivalent anatomic area of the unaffected contralateral limb. Dogs were sedated and positioned in sternal or lateral recumbency for forelimb or hind limb lesions, respectively. Bone mineral density of the primary tumor was divided by BMD of the contralateral normal limb to determine rBMD as previously described.[24]

BALP Immunohistochemistry Antibody Validation

Normal canine kidney was used as a positive control for antibody validation and staining protocol optimization. Tissue was fixed in 10% neutral buffered formalin, embedded in paraffin, and sectioned at 5 μm. The protocol consisted of deparaffinization and rehydration of the sections through graded alcohol baths, endogenous peroxidase quenching with a methanol–hydrogen peroxidase solution, microwave antigen retrieval with citrate buffer (pH 6.0), and incubation in the primary antibody. Anti-BALP monoclonal antibody5 was diluted 1 : 500, incubated for 30 minutes, and detected with a non avidin-biotin immunoperoxidase- diaminobenzidine (DAB) detection method.10 Primary antibody was omitted as a negative control.

BALP Immunohistochemistry of OSA Samples

Complete necropsy was performed on 8 dogs with not only primary bone OSA but also advanced macroscopic metastases involving distant visceral organs. Tissue specimens from the primary tumor and all major organs were submitted for histologic examination. All tissues were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (HE). Based on the HE-stained slides, additional paraffin sections were obtained from the primary tumors and selected soft tissue metastatic sites for BALP immunohistochemistry.

Statistical Analysis

The distribution of the continuous variable data was evaluated using the Shapiro–Wilk test, skewness, kurtosis, and q–q plots. Data that were not normally distributed were log transformed to meet the assumption of normality. A Pearson product-moment correlation coefficient was used to assess the relationship between pretreatment serum BALP and tumor parameters (measurements of tumor size, urine NTx, and rBMD). Changes in BALP and urine NTx over time were evaluated with repeated measures ANOVA followed by posthoc paired 2-tailed t-tests for normally distributed data (BALP) or by Wilcoxon matched-pairs signed-ranks test for nonparametric data (urine NTx). Statistical calculations were performed using commercial software programs.11^,12 P < .05 was considered statistically significant for all analyses.