Expression of vascular endothelial growth factor in tumors and plasma ...

Expression of vascular endothelial growth factor in tumors and plasma from dogs with primary intracranial neoplasms John H. Rossmeisl Jr, DVM, MS; Robert B. Duncan†, DVM, PhD; William R. Huckle, PhD; Gregory C. Troy, DVM, MS

Objective—To quantitatively evaluate expression of vascular endothelial growth factor (VEGF) in intracranial tumors in dogs and determine whether relationships exist between circulating and intratumoral VEGF concentrations and tumor type and grade. Animals—27 dogs with primary intracranial neoplasms and 4 unaffected control dogs. Procedures—Plasma and brain tumor samples were obtained from each dog, and plasma and intratumoral concentrations of VEGF were measured by use of an ELISA. Results—Dogs with meningiomas (n = 11) were significantly older than dogs with oligodendrogliomas (7) or astrocytomas (9). Measurable VEGF was detected in all tumors, and a significant negative correlation between age and intratumoral VEGF concentration was detected. Age-adjusted comparisons identified significant differences in intratumoral VEGF concentrations among all tumor types; the highest VEGF concentrations were associated with astrocytomas. Within each tumor type, increasing tumor grade was significantly associated with increasing VEGF expression. Plasma VEGF concentrations were detectable in 9 of 27 dogs; the proportion of dogs with astrocytomas and a detectable circulating VEGF concentration (7/9 dogs) was significantly higher than the proportion of dogs with meningiomas (1/11 dogs) or oligodendrogliomas (1/7 dogs) with a detectable circulating VEGF concentration. Conclusions and Clinical Relevance—Overexpression of VEGF appears common in canine astrocytomas, oligodendrogliomas, and meningiomas. In the neoplasms examined, intratumoral VEGF concentrations correlated well with tumor malignancy. The VEGF expression patterns paralleled those of analogous human tumors, providing evidence that dogs are a suitable species in which to study angiogenesis and intracranial neoplasia for human application. (Am J Vet Res 2007;68:1239–1245)

S

tudies1,2 in humans with a variety of naturally occurring neoplasms and in laboratory animals with experimentally induced cancer have revealed that angiogenesis is crucial for the growth, progression, and metastasis of solid tumors. Vascular endothelial growth factor is a hypoxia-induced proangiogenic factor that is a potent promoter of endothelial cell growth, proliferation, permeability, and survival in states of health and disease.3,4 Tumor neovascularization is also known to be dependent on VEGF expression.2 Results of a study4 in tumor-bearing dogs suggest that the molecular biological function of canine VEGF strongly resembles that of human VEGF. Received December 7, 2006. Accepted May 1, 2007. From the Departments of Small Animal Clinical Sciences (Rossmeisl, Troy) and Biomedical Sciences and Pathobiology (Duncan, Huckle), Virginia-Maryland Regional College of Veterinary Medicine, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061. Supported by Virginia Tech New Initiative and Virginia-Maryland Regional College of Veterinary Medicine Quick Response grants. The authors thank Dr. Stephen Werre for assistance with statistical analyses. Address correspondence to Dr. Rossmeisl. †Deceased.

VEGF VEGFR WHO

Abbreviations Vascular endothelial growth factor Vascular endothelial growth factor tyro - sine kinase receptor World Health Organization

Several investigations1–8 have determined that overexpression of VEGF occurs in humans, rodents, dogs, and cats with a variety of spontaneous cancers including primary intracranial tumors. In humans and dogs with intracranial meningiomas, VEGF expression has been linked to prognostic factors such as patient survival following multimodality treatment8 and tumor recurrence following surgical resection.9 Upregulation of VEGF and its endothelial receptors has also been reported to be a feature essential for anaplastic progression of human gliomas, and VEGF expression appears to be closely correlated with tumor grade and vascular indices in human gliomas.10,11 Considering the reliance of solid tumors on angiogenic factors for growth and metastasis, preliminary evidence suggests that blockade of proangiogenic activity may hold promise as a novel therapeutic intervention for patients with a variety of naturally occuring neoplasms in the CNS.1–3,9,11

AJVR, Vol 68, No. 11, November 2007

1239

Compared with human medicine, relatively little is known about VEGF expression in tumors in animals. To date, most of the studies in animals performed to investigate VEGF in specific tumor types have qualitatively or semiquantitatively assessed VEGF expression by use of immunohistochemical methods5,7,8 or quantitatively measured soluble VEGF isoforms in blood or body fluids.6,12–14 However, a recent study by Dickinson et al15 revealed that the expression patterns of VEGFRs (assessed by use of quantitative PCR assays) in gliomas and meningiomas in dogs are similar to those in humans.15 In addition, results of an immunohistochemical study8 indicated that survival was significantly and negatively associated with several qualitative measures of VEGF expression in dogs with intracranial meningiomas. The purpose of the study reported here was to quantitatively evaluate expression of VEGF in canine intracranial tumors and determine whether relationships exist between circulating and intratumoral VEGF concentrations and tumor type and grade. It was hypothesized that VEGF expression would be correlated with the degree of tumor malignancy in dogs with a variety of intracranial neoplasms. Materials and Methods Source of tissue samples—Plasma and brain tumor samples were collected from 27 client-owned dogs that had histologically confirmed, supratentorial primary intracranial neoplasms with no clinical or necropsy evidence of metastatic disease. As control specimens, brain tissue was collected from 4 dogs with no clinical or histopathologic evidence of brain disease that were euthanatized for reasons unrelated to the study. All study procedures were approved by the institutional animal care and use committee, and clients provided written, informed consent prior to inclusion of their dog in the study. Samples of blood were obtained from dogs during antemortem diagnostic investigations or immediately prior to euthanasia. Via jugular venipuncture, 5 to 10 mL of blood was collected from each dog and placed into tubes containing EDTA. Plasma was removed from whole blood samples within 30 minutes after collection. Following centrifugation (by use of a fixed-rotor centrifuge) at 3,000 X g at 10°C for 7 minutes, plasma samples were separated into 0.5-mL aliquots, then frozen and kept at –80°C until assayed. Brain tumor tissue was collected from client-owned dogs during surgery for treatment of the neoplasm or at the time of necropsy. Brain tumor tissue samples were collected during surgery in 10 of 27 dogs (9 dogs with meningiomas and 1 dog with a low-grade astrocytoma). When required, dogs were euthanatized by an overdose of barbiturate administered IV (16/17 dogs); 1 dog died during hospitalization. Brain tumor tissue samples were collected at the time of necropsy (ie, within 1 hour after death or euthanasia) in 17 dogs. Prior to surgery, all dogs included in the study were evaluated for distant metastases via thoracic radiography (3 views) and abdominal ultrasonography. Tumor and control tissue specimens were macroscopically separated from surrounding brain parenchyma via sharp dissection under magnification, divided into approximately 2 equivalent 1240

volumes, and then fixed in neutral-buffered 10% formalin or snap-frozen in liquid nitrogen and stored at –80°C within 15 minutes after collection. Histologic evaluation—Formalin-fixed, paraffinembedded brain tumor or brain tissue specimens were sectioned (5 µm in thickness), routinely processed, and stained with H&E. For the purposes of the study, a veterinary pathologist (RBD) who was unaware of the original diagnosis reviewed all slides and classified all neoplasms according to standard WHO criteria.16,17 Plasma VEGF ELISA assay—Vascular endothelial growth factor concentrations were assayed by use of a commercially available ELISA kit designed to detect human VEGF,a which has been previously validated by our laboratory for use in dogs.12 Assessment of circulating VEGF concentrations in the study dogs was performed in triplicate by use of 100 µL of plasma/replicate. Colorimetric endpoints were measured on a microplate reader,b and standard curves were generated according to manufacturer’s instructions. On the basis of the lowest reproducible absorbance change (0.004) relative to reagent blanks, the limit of detection of the assay was defined as 2.5 pg/mL. The mean intra-assay and interassay coefficients of variation obtained in the study for this kit were 3.8% and 5.7%, respectively, which were both within limits specified by the manufacturer. Brain tumor tissue VEGF assay—Protein extracts were prepared from frozen brain tumor or brain tissue specimens (approx 0.5 cm3 of tissue/sample). Four extracts were prepared from brain tissue collected from control dogs, specifically pooled white matter samples from the corpus callosum and internal capsule; pooled gray matter samples from the frontal and temporal cerebral cortices; a composite white and gray matter extract prepared from the internal capsule and surrounding basal ganglia; and a section of meninges overlying the left cerebral hemisphere that was bounded rostrally by the cruciate sulcus, medially by the longitudinal fissure, laterally by the ectosylvian sulcus, and caudally by the osseous tentorium. The anatomic regions from which control dog extracts were prepared were chosen because they approximated the neuroanatomic structures of origin of the tumors evaluated. Extracts were homogenized in an extraction buffer (20mM Tris [pH, 7.5], 0.5M NaCl, 0.1mM phenylmethyl sulfonyl fluoride, and 10 µg of leupeptin/mL; 3 mL of extraction buffer/g of tissue), sonicated, and clarified via microcentrifugation (14,000 X g for 10 minutes). Precipitates were then reextracted in the same buffer with a high saline content (1M NaCl) to extract soluble factors bound to heparin sulfate proteoglycans that were present in the matrix and on cells. The 2 supernatants were pooled and dialyzed against Tris-buffered saline solution (20mM Tris [pH, 7.5] and 150mM NaCl). Vascular endothelial growth factor concentration was then assayed in triplicate in 100-µL tumor tissue extracts by use of a commercially available ELISA kit designed to detect human VEGF,a as described. Extract protein concentrations were determined by use of a bicinchoninic acid protein assay kit.c Concentrations of VEGF that were determined via ELISA were then standardized to AJVR, Vol 68, No. 11, November 2007

extract protein content and expressed as nanograms of VEGF per milligram of sample protein. Statistical analysis—To stabilize the variance and obtain a symmetric Gaussian distribution, a natural logarithmic transformation was applied to measured VEGF concentrations. Differences in age, sex, and body weight among dogs with different histologic tumor types were assessed by use of an ANOVA. Correlation between age and intratumoral VEGF concentration was assessed by use of the Spearman correlation coefficient. To test the hypotheses (null) that histologic tumor type does not have an effect on intratumoral VEGF concentration and that within tumor type, tumor grade does not have an effect on intratumoral VEGF concentration, data were modeled by use of an ANOVA. The model for betweentumor type comparisons included age as a covariate. Because plasma VEGF concentrations were less than the ELISA assay detection limit in 18 of 27 dogs, a categoric variable (VEGF detectable in plasma vs VEGF not detectable in plasma) was created and used in some of the analyses. To test the hypothesis (null) that tumor type does not have an effect on the proportion of dogs with detectable plasma VEGF concentrations, data were modeled by use of logistic regression. To compare the proportions of dogs with detectable plasma VEGF concentrations among tumor grades within each tumor type (for dogs with meningiomas and oligodendrogliomas only), a Fisher exact χ2 test was applied. Because 7 of 9 dogs with astrocytomas had detectable plasma VEGF concentrations, a repeated-measures ANOVA was used to assess the effect of tumor grade for outcome in this group. Significance was set at an α value of 0.05. A statistical software packaged was used for all analyses. Results

ing 4 WHO grade II (low grade) astrocytomas, 3 WHO grade III (anaplastic) astrocytomas, and 2 WHO grade IV (glioblastoma multiforme) astrocytomas. Of the 27 dogs, the 11 dogs with meningiomas included 5 mixed-breed dogs, 2 Labrador Retrievers, 2 Boxers, 1 German Shepherd Dog, and 1 Golden Retriever. Breeds of dogs with oligodendrogliomas (7/27 dogs) included 2 Boston Terriers, 2 mixed, 1 Labrador Retriever, 1 French Bulldog, and 1 Mastiff. Breeds of dogs with astrocytomas (9/27 dogs) included 2 Boxers, 2 Boston Terriers, 1 mixed, 1 Poodle, 1 Corgi, 1 Shetland Sheepdog, and 1 Jack Russell Terrier. The mean body weight for all dogs with tumors was 20.4 ± 13.3 kg. Of the 27 dogs, 15 were spayed females and 12 were neutered males. There was no difference (P > 0.05) in sex or body weight among dogs with meningiomas, oligodendrogliomas, or astrocytomas. The mean ± SD age of all dogs with tumors was 9.7 ± 2.2 years (range, 5 to 14 years). Dogs with meningiomas were significantly (P < 0.001) older (mean age, 11.7 ± 1.7 years) than dogs with oligodendrogliomas (mean age, 8.3 ± 1.1 years) or astrocytomas (mean age, 8.2 ± 1.2 years). Mean ages of dogs with oligodendroglial and astrocytic tumors were not significantly (P > 0.05) different. Brain tumor tissue VEGF concentration—Vascular endothelial growth factor was undetectable in all brain tissue extracts prepared from control dogs. Vascular endothelial growth factor was quantifiable via ELISA in all brain tumor homogenates, and concentrations ranged from 0.07 to 6.38 ng/mg of tumor tissue. Mean ± SD intratumoral VEGF concentrations were 0.24 ± 0.28 ng/mg for meningiomas, 0.93 ± 0.3 ng/mg for oligodendrogliomas, and 2.84 ± 1.81 ng/mg for astrocytomas (Figure 1). A significant (P < 0.001) negative correlation between age and intratumoral VEGF concentration was identified. Analysis of covariance with a repeated-measures component revealed a significant (P < 0.001) interaction between age of dog and tumor type. The slope (regression coefficients for age) for dogs with oligodendrogliomas was not significantly (P = 0.92) different from that for dogs with astrocytomas. As a result, these 2 groups were combined into 1 group for the slope

Histopathologic diagnoses and descriptive data— Brain tissue was collected from 4 control dogs with no clinical or histopathologic evidence of brain disease. Among these 4 control dogs, there were 2 mixed-breed dogs, 1 Boxer, and 1 Labrador Retriever. The group included 2 spayed females and 2 neutered males. The mean ± SD body weight of the control dogs was 22.6 ± 8.1 kg, and the mean age was 8.3 ± 1.5 years (range, 5 to 12 years). Plasma and brain tumor samples from 27 dogs with primary intracranial neoplasms were included in the study. Among the 27 dogs, 11 had meningiomas; tumor samples from 9 of those dogs were consistent with benign (transitional [n = 5], psammomatous [3], and fibrous [1]) WHO grade I meningiomas. Two of the 11 dogs had tumors that were categorized as anaplastic WHO grade III meningiomas. Oligodendroglial tumors were identified in 7 of 27 dogs, of which 5 were classified Figure 1—Box-and-whisker plot of intratumoral VEGF concentrations (ng of VEGF/mg as WHO grade II (low grade) tumors of sample protein) in dogs with meningiomas (n = 11 dogs), oligodendrogliomas (7), or (9). Each box represents the 25th to 75th percentile (interquartile range), and 2 were classified as WHO grade III astrocytomas and the bars (whiskers) represent extreme observations that are < 1.5 X interquartile (high grade) tumors. The remaining 9 range from the upper and lower limits of the interquartile range. Within each box, the tumors were astrocytic gliomas, includ- horizontal line represents the median value; the mean value (+) is identified. AJVR, Vol 68, No. 11, November 2007

1241

Table 1—Age-controlled comparisons of intratumoral VEGF concentrations* between tumor types in dogs with meningiomas (n = 11 dogs), oligodendrogliomas (7), or astrocytomas (9). Age (y)

Tumor type comparison

Mean intratumoral VEGF concentration*

Difference (95% CI)

P value†

8 Meningioma vs oligodendroglioma –3.04 vs –0.06 Meningioma vs astrocytoma –3.04 vs 0.91 Oligodendroglioma vs astrocytoma –0.06 vs 0.91

–2.99 (–4.04 to –1.99) –3.96 (–4.96 to –1.96) –3.96 (–4.96 to –1.96)

 0.001

9.7 Meningioma vs oligodendroglioma –2.51 vs –0.41 Meningioma vs astrocytoma –2.51 vs 0.56 Oligodendroglioma vs astrocytoma –0.41 vs 0.56

–2.09 (–2.91 to –1.28) –3.06 (–3.86 to –2.28) –0.97 (–1.55 to –0.39)

 0.001

11 Meningioma vs oligodendroglioma –2.09 vs –0.68 Meningioma vs astrocytoma –2.09 vs 0.29 Oligodendroglioma vs astrocytoma –0.68 vs 0.29

–1.42 (–2.38 to –0.46) –2.38 (–3.32 to –1.44) –2.38 (–3.32 to –1.44)

0.007

 0.001  0.001

 0.001 0.004

 0.001  0.001

The ages represented the mean age of all dogs in the study (9.7 y) as well as the upper (11 y) and lower (8 y) limits of the interquartile age range (the 25th to 75th percentile) of the studied population. *Comparisons of least square means of natural log values of intratumoral VEGF concentrations (ng of VEGF/mg sample protein). †A value of P  0.05 was considered significant. CI = Confidence interval.

Table 2—Comparisons of intratumoral VEGF concentrations* by WHO tumor grade in dogs with meningiomas (n = 11 dogs), oligodendrogliomas (7), or astrocytomas (9). Tumor grade Tumor type comparison

Mean intratumoral VEGF Difference concentration* (95% CI) P value†

Meningioma I vs III –2.23 vs –0.23 Oligodendroglioma II vs III –0.28 vs 0.30 Astrocytoma II vs III 0.32 vs 0.97 III vs IV 0.97 vs 1.78 II vs IV 0.32 vs 1.78

–1.99 (–2.63 to –1.34) –0.58 (–0.96 to –0.20) –0.65 (–1.02 to –0.28) –0.80 (–1.24 to –0.36) –1.45 (–1.87 to –1.04)

 0.001 0.011 0.005 0.004 0.001

See Table 1 for key.

component of the model before comparing mean intratumoral VEGF concentrations between tumor types at ages 8, 9.7 (the mean age), and 11 years (Table 1). At each age, comparisons between VEGF concentrations of the various tumor types involved every dog within each respective tumor type, while controlling for the effect of age. At all ages, all 2-way comparisons of mean intratumoral VEGF concentrations between tumor types were significant (all P values < 0.007). At all ages, intratumoral VEGF concentrations were significantly (P = 0.004) higher in astrocytomas than in oligodendroglial or meningeal tumors, and intratumoral VEGF concentrations were significantly (P = 0.007) higher in oligodendrogliomas than meningiomas. Within each histologic subtype of tumor, increasing tumor grade was significantly (all P values < 0.011) associated with increasing VEGF concentration (Table 2). Plasma VEGF concentration—Concentrations of VEGF were detectable in 9 of 27 plasma samples from dogs with intracranial neoplasms; values ranged 1242

from 9.2 to 145.1 pg/mL. The proportion of dogs with detectable plasma VEGF concentrations was significantly (all P values < 0.027) greater among dogs with astrocytomas than among dogs with meningiomas or oligodendrogliomas (Table 3). The proportion of dogs with detectable plasma VEGF concentrations was not different (P = 0.73) between dogs with meningiomas and dogs with oligodendrogliomas. Among dogs with astrocytomas, plasma concentrations of VEGF were significantly greater in dogs with grade IV astrocytomas than those in dogs with grade III (P = 0.008) or grade II (P = 0.004) tumors (Table 4). There was no significant (P = 0.17) difference in plasma VEGF concentrations between dogs with grade II and grade III astrocytomas. Discussion In the present study of dogs with primary intracranial neoplasms, measurable VEGF concentrations were detected in all tumors evaluated, and VEGF expression was undetectable in brain tissue samples obAJVR, Vol 68, No. 11, November 2007

Table 3—Associations between detectable plasma VEGF concentration and tumor type in dogs with meningiomas (n = 11 dogs), oligodendrogliomas (7), or astrocytomas (9). Tumor type comparison Astrocytoma vs meningioma Oligodendroglioma vs meningioma Astrocytoma vs oligodendroglioma

Rate (proportion of dogs) of detectable plasma VEGF concentration

Odds ratio (95% CI)

P value†

77% (7/9) vs 9% (1/11) 14% (1/7) vs 9% (1/11) 77% (7/9) vs 14% (1/7)

35 (2.63 to 465.4) 1.67 (0.09 to 31.87) 0.05 (0.003 to 0.665)

0.007 0.73 0.027

See Table 1 for key. Table 4—Comparisons of plasma VEGF concentrations* in 9 dogs with astrocytomas by WHO tumor grade. Astrocytoma grade comparison II vs III II vs IV III vs IV

Mean plasma VEGF concentration*

Difference (95% CI)

P value†

2.51 vs 3.03 2.51 vs 4.59 3.03 vs 4.59

–0.53 (–1.41 to 0.35) –2.09 (–3.05 to –1.12) –1.56 (–2.44 to –0.68)

0.17 0.004 0.008

Among the astrocytomas, 4 were grade II, 3 were grade III, and 2 were grade IV. *Comparisons of least square means of natural log values of plasma VEGF concentrations (pg/mL). See Table 1 for remainder of key.

tained from clinically normal adult dogs; these data support findings of other investigations8,15 and have suggested that VEGF or VEGFR overexpression occurs commonly in canine astrocytomas, oligodendrogliomas, and meningiomas. In our study, after correction for age, VEGF expression was significantly different among all tumor types examined and was quantitatively highest in astrocytic gliomas, intermediate in oligodendroglial tumors, and lowest in meningiomas. In addition, quantitative VEGF expression was significantly correlated with tumor grade. Within each subtype of tumor studied, higher WHO tumor grades (which is a graded representation of malignant potential based on histomorphologic criteria) were associated with significant increases in VEGF expression. Angiogenesis is a complex process balanced by a number of pro- and antiangiogenic factors. Previous research has revealed that VEGF is a potent endothelial cell mitogen and chemotactic factor, as well as a regulator of vascular permeability.2,3,12 Results of several studies4,7,10 have indicated that VEGF is a major promoter of angiogenesis in solid neoplasms. Expression of VEGF has been extensively studied in high-grade gliomas because those tumors represent some of the most highly vascularized solid neoplasms in humans.10,11,18 Hypoxia of endothelial cells and other cell types is considered the primary stimulus for upregulation of VEGF expression within neoplastic cells, although other cytokines and growth factors also participate.1–3 In both human and canine gliomas, overexpression of epidermal growth factor receptor-1 and platelet-derived growth factor has also been associated with increased VEGF secretion.15,18 Microenvironmental hypoxia is known to develop in high-grade glial neoplasms and is partially responsible for some of the neuropathologic features that characterize anaplastic gliomas (ie, WHO grade III and IV astrocytomas and grade III oligodendrogliomas) such as necrosis, vascular hypertrophy, and microvascular proliferation.10,11,15–17 Expression of VEGF mRNA within human high-grade gliomas is spatially restricted

to areas of palisading cells that border areas of necrosis, further suggesting that hypoxia is a potent inducer of VEGF expression.1,11 Vascular endothelial growth factor expression also correlates well with microvessel density and indices of microvascular proliferation in human meningiomas and gliomas.3,8,10 Therefore, although microvascular indices were not specifically assessed in the dogs of the present study, it is not surprising that VEGF expression was quantitatively higher in high-grade tumors that were in part defined by the presence of necrosis and highly vascular phenotypes, compared with benign, less vascularized tumors such as meningiomas.16,17 Investigations1,10,11 in humans have clearly identified the importance of VEGF and VEGFR expression in the neovascularization, growth, and malignant progression of gliomas. Hypoxia-induced overexpression of VEGF is considered to be a crucial step in the angiogenic switch that accompanies progression of low-grade gliomas to more anaplastic phenotypes.11,19 Although comparatively little is known about the in vivo progression of naturally occurring intracranial tumors in dogs, the results of the study reported here are similar to findings of other investigations19,20 to evaluate intratumoral VEGF expression in humans, in which VEGF expression was quantitatively highest in patients with glioblastoma multiforme (WHO grade IV) or anaplastic (WHO grade III) astrocytomas and typically low in meningeal tumors. Our findings also resemble those of several studies10,11,19–23 in human medicine, which have indicated that VEGF expression is positively correlated with tumor grade, and thus with degree of tumor malignancy, in humans with astrocytomas, oligodendrogliomas, and meningiomas. Results of the present study also indicated that intratumoral VEGF concentrations in canine astrocytomas are quantitatively within a similar range as those in human astrocytomas determined by use of an ELISA by Lamszus et al.19 In healthy adults of several species, differential expressions of VEGF and VEGFRs occur within many


1243

organs. In the vasculature of clinically normal adult rodent and human brains, VEGF and VEGFR expression are low.1,3,15 In healthy adult rodent brains, expression of VEGF is primarily localized to the choroid plexi and pituitary gland.3,20 The organ-specific patterns of VEGF expression within the brains of clinically normal dogs are not currently known. Although additional studies are required to characterize VEGF expression in clinically normal canine brains, results obtained from the control dogs in the present study have suggested that VEGF expression is low in normal adult canine meninges and cerebrocortical and subcortical neural parenchyma. The signalment of dogs in our study was similar to that of dogs with naturally occurring intracranial tumors in a previous report.24 Our data support findings that dogs with meningiomas are significantly older than dogs with other intracranial tumor types at the time of evaluation because of signs of neurologic disease referable to brain dysfunction, and that brachycephalic breeds are predisposed to development of glia-origin intracranial tumors.24 Therefore, it is not unexpected that there was a significant and negative correlation between intratumoral VEGF content and age in the present study. In a previous study12 conducted in our laboratory, 24 of 54 (44%) tumor-bearing dogs had plasma VEGF concentrations (≥ 2.5 pg/mL) detectable via ELISA. Similarly, circulating concentrations of VEGF were detectable in 9 of 27 (33%) dogs in the present study. The range of plasma VEGF concentrations in tumorbearing dogs determined in our study (9.2 to 145.1 pg/mL) was also similar to that reported previously.12 Although plasma VEGF concentrations were detectable in a significantly higher proportion of dogs with astrocytomas, compared with the proportions of dogs with other tumor types in the present study, and dogs with glioblastomas had a significantly higher plasma VEGF concentration than dogs with other types of astrocytomas, the reasons for these findings are not clear. Results of studies14,25,26 of humans with intracranial neoplasms and humans and dogs with neoplastic body cavitary effusions suggest that circulating VEGF concentrations do not necessarily correlate with matched VEGF concentrations in tumors or malignant effusions. The results of the present study are in agreement with those of previous investigations14,25,26 in which circulating VEGF concentrations were frequently undetectable in patients with localized cancers; none of the dogs included in our study had clinical or necropsy evidence of metastases. Several potential explanations exist for these findings. First, little is known about the expression of the various VEGF isoforms in canine disease states, but all isoforms are not necessarily secreted in a coordinated fashion. Differential expression of soluble and cell-associated isoforms of VEGF by tumors may explain the failure to detect VEGF in the plasma of patients with localized tumors. Results of an investigation10 in humans have suggested that the diffusible VEGF isoforms (VEGF121 and VEGF165) are primarily overexpressed in humans with brain tumors, but a few tumors can express the cell-associated VEGF isoform (VEGF189), which remains tightly bound to the 1244

cell surface or in the extracellular matrix immediately adjacent to the tumor. The ELISA used in the study of this report primarily detected the soluble isoforms of VEGF.12 Second, locally secreted, soluble isoforms of VEGF may never gain access to the circulation because of a short biological half-life or as a result of local binding to vascular endothelium or soluble VEGFRs.25 Although mechanisms regulating expression of VEGFRs are poorly understood, some evidence suggests that the stimuli that induce expression of VEGF may have the same effect on VEGFRs or that VEGF itself may stimulate the expression of its own receptors in a paracrine fashion.1,3 In the dogs with neoplasms in the present study, plasma VEGF concentrations may also have been influenced by variables independent of tumor type that were not evaluated. It is known that expression of VEGF is increased within WBCs and platelets of humans with cancer (compared with VEGF expression in unaffected humans),27 and increases in plasma VEGF concentrations in some dogs could perhaps be explained by the presence of leukocytosis or thrombocytosis. Results of studies12,13 to examine the effects of WBC and platelet counts on plasma VEGF concentrations in dogs have provided inconsistent results; however, we have previously determined that VEGF concentrations may be positively correlated with WBC and platelets counts.12 In addition, anemia of chronic disease is also a common clinicopathologic abnormality in dogs with neoplasia, and it has been reported12 that plasma VEGF concentration may be inversely correlated with PCV or hemoglobin concentrations, as a result of hypoxia-induced VEGF expression. On the basis of our findings, it is evident that intratumoral VEGF concentrations in dogs can be measured by use of a commercially available ELISA assay and that VEGF overexpression had developed in all of the naturally occurring neoplasms examined. In dogs of the present study, intratumoral VEGF expression was significantly higher in high-grade vascular and necrotic glia-origin tumors than in meningiomas and correlated well with tumor grade. In dogs with localized intracranial tumors, plasma VEGF concentrations were often undetectable by use of the ELISA. The patterns of intratumoral VEGF expression determined in dogs paralleled those in analogous human tumors. Our findings and those of other investigations have reinforced the potential value of dogs for the study of alterations in VEGF expression in naturally occurring intracranial neoplasia in humans.8,15 In addition, novel molecular antiglioma treatments that are specifically targeted against the VEGF-VEGFR pathway, which have entered into clinical trials in humans,18 may also have a role in the treatment of canine intracranial tumors. a. b. c. d.

VEGF assay DVE00, R & D Systems, Minneapolis, Minn. Spectramax 250, Molecular Devices, Sunnyvale, Calif. BCA-1, Sigma-Aldrich, St Louis, Mo. SAS, version 9.1.3, SAS Institute Inc, Cary, NC.

References 1.

Berkman RA, Merrill MJ, Reinhold WC, et al. Expression of the vascular permeability factor/vascular endothelial growth factor gene in central nervous system neoplasms. J Clin Invest 1993;91:153–159. AJVR, Vol 68, No. 11, November 2007

2. 3. 4. 5. 6.

7. 8.

9. 10. 11.

12.

13. 14. 15.

Folkman J. Tumor angiogenesis. Adv Cancer Res 1985;43:175– 203. Harrigan MR. Angiogenic factors in the central nervous system. Neurosurgery 2003;53:1362–1376. Scheidegger P, Weiglhofer W, Suarez S, et al. Vascular endothelial growth factor (VEGF) and its receptor in tumor-bearing dogs. Biol Chem 1999;380:1449–1454. Graham JC, Myers RK. The prognostic significance of angiogenesis in mammary tumors. J Vet Intern Med 1999;13:416–418. Wergin MC, Ballmer Hofer K, Roos M, et al. Preliminary study of plasma vascular endothelial growth factor (VEGF) during low- and high-dose radiation therapy in of dogs with spontaneous tumors. Vet Radiol Ultrasound 2004;45:247–254. Restucci B, Maiolino P, Paciello O, et al. Expression of vascular endothelial growth factor in canine mammary tumors. Vet Pathol 2002;39:488–493. Platt SR, Scase TJ, Adams V, et al. Vascular endothelial growth factor expression in canine intracranial meningiomas and association with patient survival. J Vet Intern Med 2006;20:663– 668. Pistolesi S, Boldrini L, Gisfredi S, et al. Angiogenesis in intracranial meningiomas: immunohistochemical and molecular study. Neuropathol Appl Neurobiol 2004;30:118–125. Huang H, Held-Feindt J, Buhl R, et al. Expression of VEGF and its receptors in different brain tumors. Neurol Res 2005;27:371– 377. Chaundry IH, O’Donovan DGO, Brenchley PEC, et al. Vascular endothelial growth factor expression correlated with tumor grade and vascularity in gliomas. Histopathology 2001;39:409– 415. Troy GC, Huckle WR, Rossmeisl JH, et al. Endostatin and vascular endothelial growth factor concentrations in healthy dogs, dogs with selected neoplasia, and dogs with non-neoplastic diseases. J Vet Intern Med 2006;20:144–150. Clifford CA, Hughes D, Beal MW, et al. Plasma vascular endothelial growth factor concentrations in healthy dogs and dogs with hemangiosarcoma. J Vet Intern Med 2001;15:131–135. Clifford CA, Hughes D, Beal MW, et al. Vascular endothelial growth factor concentrations in body cavity effusions in dogs. J Vet Intern Med 2002;16:164–168. Dickinson PJ, Roberts BN, Higgins RJ, et al. Expression of re-

16. 17.

18. 19.

20. 21. 22. 23.

24. 25. 26. 27.

ceptor tyrosine kinases VEGFR-1 (FLT-1), VEGFR-2 (KDR), EGFR-1, PDGFRα, and c-Met in canine primary brain tumors. Vet Comp Oncol 2006;4:132–140. Kleinheus P, Cavenee WK. WHO classification of tumors. Pathology and genetics—tumors of the nervous system. Lyon, France: IARC Press, 2000;176–184. Koestner A, Bilzer T, Schulman FY, et al. Histological classification of tumors of the nervous system of domestic animals. 2nd series, Vol 5. Washington, DC: Armed Forces Institute of Pathology, 1999. Morokoff AP, Novak U. Targeted therapy for malignant gliomas. J Clin Neurosci 2004;11:807–818. Lamszus K, Ulbricht U, Matschke J, et al. Levels of soluble vascular endothelial growth factor (VEGF) receptor 1 in astrocytic tumors and its relation to malignancy, vascularity, and VEGF-A. Clin Cancer Res 2003;9:1399–1405. Pietsch T, Valter MM, Wolf H, et al. Expression and distribution of vascular endothelial growth factor protein in human brain tumors. Acta Neuropathol 1997;93:109–117. Oehring RD, Miletic M, Valter MM, et al. Vascular endothelial growth factor in astrocytic gliomas—a prognostic factor? J Neurooncol 1999;45:117–125. Korshunov A, Golanov A, Sycheva R. Immunohistochemical markers for prognosis of oligodendroglial neoplasms. J Neurooncol 2002;58:237–253. Korshunov A, Golanov A. The prognostic significance of vascular endothelial growth factor (VEGF C-1) immunoexpression in oligodendrogliomas. An analysis of 91 cases. J Neurooncol 2000;48:13–19. Snyder JM, Shofer FS, Van Winkle TJ, et al. Canine intracranial primary neoplasia: 173 cases (1986–2003). J Vet Intern Med 2006;20:669–675. Takano S, Yoshii Y, Kondo S, et al. Concentration of vascular endothelial growth factor in the serum and tumor tissue of brain tumor patients. Cancer Res 1996;56:2185–2190. Kraft A, Weindel K, Ochs A, et al. Vascular endothelial growth factor in the sera and effusions of patients with malignant and non-malignant disease. Cancer 1999;85:178–187. Salven P, Orpana A, Joensuu H. Leukocytes and platelets with cancer contain high levels of vascular endothelial growth factor. Clin Cancer Res 1999;5:487–491.


1245