University of ZurichZurich Open Repository and Archive

Winterthurerstr. 190

CH-8057 Zurich

http://www.zora.uzh.ch

Year: 2006

The influence of fractionated radiation therapy on plasmavascular endothelial growth factor (VEGF) concentration in dogs

with spontaneous tumors and its impact on outcome

Wergin, M C; Roos, M; Inteeworn, N; Laluhovà, D; Allemann, K; Kaser-Hotz, B

Wergin, M C; Roos, M; Inteeworn, N; Laluhovà, D; Allemann, K; Kaser-Hotz, B (2006). The influence offractionated radiation therapy on plasma vascular endothelial growth factor (VEGF) concentration in dogs withspontaneous tumors and its impact on outcome. Radiotherapy & Oncology, 79(2):239-44.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:Radiotherapy & Oncology 2006, 79(2):239-44.

Wergin, M C; Roos, M; Inteeworn, N; Laluhovà, D; Allemann, K; Kaser-Hotz, B (2006). The influence offractionated radiation therapy on plasma vascular endothelial growth factor (VEGF) concentration in dogs withspontaneous tumors and its impact on outcome. Radiotherapy & Oncology, 79(2):239-44.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:Radiotherapy & Oncology 2006, 79(2):239-44.

The influence of fractionated radiation therapy on plasmavascular endothelial growth factor (VEGF) concentration in dogs

with spontaneous tumors and its impact on outcome

Abstract

BACK GROUND AND PURPOSE: Vascular endothelial growth factor (VEGF), a specificpro-angiogenic factor is proposed to be involved in cancer progression and resistance to radiationtherapy by promoting angiogenesis and by protecting endothelial cells from radiation induced apoptosis.The aim of this study, was first to assess the influence of ionizing radiation on plasma VEGFconcentration in spontaneous canine tumors during fractionated radiation therapy with curative orpalliative intent and second to analyze plasma VEGF concentration as predictor for treatment outcome.PATIENTS AND METHODS: For plasma VEGF analysis a human VEGF enzyme linkedimmunosorbent assay was used. Sixty dogs with various tumor types were included in this study. Dogswere irradiated with either low dose per fx (3-3.5 Gy per fraction, total dose: 42-49 Gy, group A:curative intent) or high dose per fx (6-8 Gy per fraction, total dose: 24-30 Gy, group B: palliativeintent). Blood samples were taken before and after dose application at certain time points duringtherapy. Follow-up evaluation was performed for analysis of time to treatment failure and survival.RESULTS: Repeated measures analysis showed no increase of plasma VEGF in dogs treated withfractionated radiation therapy (group A and B). Dichotomizing baseline plasma VEGF into two groupswith high and low plasma VEGF, resulted in shorter time to treatment failure in dogs with high plasmaVEGF levels (TTF, group A: P=0.038, group B: P=0.041). CONCLUSIONS: This study demonstratedthat dogs with a plasma VEGF level higher than 5 pg/ml had a poorer outcome after radiation therapy. Itis therefore, suggested, to use plasma VEGF as predictor for treatment outcome in radiation therapy.

Radiotherapy and Oncology 79 (2006) 239–244www.thegreenjournal.com

Clinical radiobiology

The influence of fractionated radiation therapy on plasma vascularendothelial growth factor (VEGF) concentration in dogs with

spontaneous tumors and its impact on outcome

Melanie C. Wergina,*, Malgorzata Roosb, Nathalie Inteeworna, Dagmar Laluhovaa,Katrin Allemanna, Barbara Kaser-Hotza

aDiagnostic Imaging and Radio-Oncology, Vetsuisse Faculty, and bBiostatistics, ISPM,University of Zurich, Switzerland

Abstract

Back ground and purpose: Vascular endothelial growth factor (VEGF), a specific pro-angiogenic factor is proposed to beinvolved in cancer progression and resistance to radiation therapy by promoting angiogenesis and by protectingendothelial cells from radiation induced apoptosis. The aim of this study, was first to assess the influence of ionizingradiation on plasma VEGF concentration in spontaneous canine tumors during fractionated radiation therapy withcurative or palliative intent and second to analyze plasma VEGF concentration as predictor for treatment outcome.

Patients and methods: For plasma VEGF analysis a human VEGF enzyme linked immunosorbent assay was used. Sixtydogs with various tumor types were included in this study. Dogs were irradiated with either low dose per fx (3–3.5 Gy perfraction, total dose: 42–49 Gy, group A: curative intent) or high dose per fx (6–8 Gy per fraction, total dose: 24–30 Gy,group B: palliative intent). Blood samples were taken before and after dose application at certain time points duringtherapy. Follow-up evaluation was performed for analysis of time to treatment failure and survival.

Results: Repeated measures analysis showed no increase of plasma VEGF in dogs treated with fractionated radiationtherapy (group A and B). Dichotomizing baseline plasma VEGF into two groups with high and low plasma VEGF, resulted inshorter time to treatment failure in dogs with high plasma VEGF levels (TTF, group A: PZ0.038, group B: PZ0.041).

Conclusions: This study demonstrated that dogs with a plasma VEGF level higher than 5 pg/ml had a poorer outcomeafter radiation therapy. It is therefore, suggested, to use plasma VEGF as predictor for treatment outcome in radiationtherapy.q 2006 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 79 (2006) 239–244.

Keywords: Canine; Plasma VEGF; Time to treatment failure; Survival; Radiation

Tumor angiogenesis is an important factor in diseaseprogression and formation of metastasis [12]. One majorpro-angiogenic factor is vascular endothelial growth factor(VEGF). VEGF promotes tumor angiogenesis, endothelialcell survival and vessel maintenance of immature vessels[12,13]. VEGF is involved in development and growth of awide variety of different tumors [15,17]. VEGF is activelysecreted from tumor cells and its soluble form (VEGF165) isdetectable in the blood compartment [9].

Exposure of tumor cells to ionizing radiation can increaseVEGF mRNA [2] and protein concentration [14]. Ionizingradiation can lead to phosphorylation of the PI-3kinase viathe epidermal growth factor receptor (EGFR) in tumor cells.This is the first step in activating the ERK pathway, amitogen activated protein kinase (MAPK) pathway. Thispathway is mediating cell survival and upregulation of VEGF

0167-8140/$ - see front matter q 2006 Elsevier Ireland Ltd. All rights rese

expression in tumor cells [4,9]. In human squamouscarcinoma cells the activation was dependent on theradiation dose, with low doses (1 Gy) causing prolongedactivation of the ERK pathway and higher doses (6 Gy)having a weaker activation [4].

Cells incubated under hypoxic conditions had increasedVEGF expression [22,25]. This increase was mediated by thehypoxia inducible factor-1 (HIF-1). HIF-1 consists of twosubunits, HIF-1a and HIF-1b. Under normoxic conditions HIF-1a is rapidly degraded [22,25], but in hypoxic cells theheterodimer HIF-1 can bind to DNA at specific regions, calledhypoxia responsive elements (HRES) which includes a 28-bpelement that is sufficient to mediate upregulation of VEGFtranscription [19]. Several studies have associated HIF-1expression with human cancer progression, such as head andneck cancer, ovarian cancer and esophageal cancer [11,33].

rved. doi:10.1016/j.radonc.2006.03.021

Plasma VEGF in dogs during fractionated radiation therapy240

Recently, an association between tumor hypoxia undelevated systemic levels of VEGF in head and neck cancer hasbeen shown [3]. In dogs with spontaneous tumors lowhemoglobin levels correlated with high plasma VEGF levels[27]. In this study, a significant difference of plasma VEGFconcentration in different tumor types was found with highlevels in carcinoma, osteosarcoma, and melanoma [27].

Low pre-operative serum VEGF levels (!575 pg/ml) wereshown to correlate with increased disease free survival, inhuman patients with colorectal cancer indicating that serumVEGF was a good predictor for outcome [29]. When serumand plasma VEGF in patients with primary colorectalcarcinoma were analyzed separately, both, high serum andhigh plasma VEGF concentrations predicted for shortersurvival, but significance was only reached for serum VEGFvalues [29]. In contrast, high plasma VEGF was shown to be asignificant predictor of reduced overall survival and ofearlier onset of local recurrence for patients with breastcancer [23,30].

Therefore, we were interested whether (1) the influenceof a low dose per fx (3–3.5 Gy per fraction, total dose: 42–49 Gy, group A: curative intent) or high dose per fx (6–8 Gyper fraction, total dose: 24–30 Gy, group B: palliative intent)on the VEGF release from tumor cells can be measured inplasma of dogs with spontaneous tumors, and (2) if baselineplasma VEGF as well as the course of plasma VEGF duringtherapy can predict time to treatment failure (TTF) andsurvival.

Material and methodsPatient selection

Sixty tumor bearing dogs were included in this study.There were 38 male dogs and 22 female dogs (12 neuteredmale and nine neutered female dogs). The median age ofpatients was 9.0 years (range 3–16 years). The study group

Table 1Tumor related parameters and treatment response

Tumor histology Stage of disease (N)

I II III IV

Group ASarcomaFibrosarcoma 4 1 2Osteosarcoma 1Myxosarcoma

Carcinoma 3 3 1Epulis 2 1

Group BSarcomaFibrosarcoma 2 1Osteosarcoma 1 4Sarcoma, other 2 2 4 1

Histiocytosis, malig.Melanoma 1 1

a No response to therapy.b Partial response.c Complete response.

included 35 pure breed dogs of various breeds and 25 mixedbreed dogs. The median weight of these dogs was 29.7 kg(range 5.1–66 kg). The routine work up included in allanimals a complete physical exam, blood count, serumchemistry, thoracic radiographs, biopsies of the primarytumor using a needle punch biopsy (Tru-cut, Baxter GeneralHealth Care, Deerfield, Illinois) or an incisional biopsy forroutine histopathological diagnosis. Fine needle aspirationof enlarged lymph nodes was performed using a 22-gaugeneedle and a 5 ml syringe. Further diagnostic work up wasdone as indicated. Patients with sarcomas of soft tissue(nZ28) or bone origin (nZ14), carcinomas (nZ9), oralmelanomas (nZ6) and three dogs with acanthomatous epuliswere included (Table 1). All patients were staged based onthe world health organization (WHO) system [24] (Table 1).For 17 patients no staging system was applicable. Thelength, width and depth of tumors were measured and thevolume was calculated using the rotation ellipsoid formula(pabc/6). The median tumor volume was 35.7 cm3 (range0.1–392.5 cm3). The initial tumor response at the end offractionated radiation therapy was defined as no response(unchanged tumor volume), partial response (decreasedtumor volume), or complete response (no macroscopictumor disease) (Table 1). Acute side effects of normal tissuewere recorded by using the toxicity criteria suggested by theVRTOG [18]. Informed consent of owners was obtained.

Anesthesia protocolAll dogs were anesthetized for radiation therapy.

Midazolam (Dormicumw, Roche Pharma AG, Reinach,Switzerland) was injected at a dosage of 0.125 mg/kgintravenously immediately followed by propofol (Propofolw,Fresenius Kabi AG, Stans, Switzerland), slowly administeredto effect. Anesthesia was maintained after intubation withisoflurane and oxygen (Forenew, Abott AG, Baar, Switzer-land). During anesthesia dogs were monitored with pulseoxymetry and ECG.

Tumour response Plasma VEGF(pg/ml)

N (total)

NRa PRb CRc

2 4 1 4.7 71 12.2 11 0.0 1

3 1 4 11.7 93 3.2 3

2 1 2.8 37 6 9.8 144 4 1 4.5 121 2 1 0.8 41 2 2 12.5 6

M.C. Wergin et al. / Radiotherapy and Oncology 79 (2006) 239–244 241

Treatment regimensAll dogs were treated with radiation therapy. The

treatment was chosen based on tumor volume, tumorstage, tumor histology and overall health of the dog. To beincluded all dogs had to have a measurable tumor mass.Radiation therapy was given with a linear accelerator (BBCDinary 20) using 6 MV photons or 5- to 16-MeV electrons. Ifindicated, computer treatment planning was performedwith CadPlan 6.0. A total of 21 dogs (group A) wereirradiated with a low dose per fraction (fx) with a totaldose of 42–49 Gy (curative intent). Dogs were treated eitherwith 3.5 Gy fractions on a Monday, Tuesday, Thursday andFriday schedule or a daily fractionation scheme with 3 Gyfractions was used. Thirty-nine dogs (group B) were treatedwith a high dose per fx with a total dose of 30 Gy delivered infive fractions, or a total dose of 24 Gy delivered in three orfour fractions (palliative intent). Detailed data are given inTable 2.

Blood samples and VEGF assayThree milliliters of blood were collected into sterile CTAD

tubes (Beckton and Dickinson Vacutainer System, France) andplaced on ice. CTAD tubes contained sodium citrate,theophyllin, adenosine and dipyridamine allowing maximalplatelet stabilization. The tubes were centrifuged within15 min at 2500!g for 30 min at 4 8C [31]. The resulting plasmawas separated and stored immediately at K80 8C. Thetechnique of blood sampling and handling of probes has beenpublished previously [10,31]. Thrombocyte count in CTADplasma was measured and the amount was at the detectionlimit (!5000). A blood sample, taken before therapy, servedas baseline plasma VEGF. In group A, blood samples werealways taken before the 1st and after the 3rd, 6th, 8th, 10th,12th and 14th fraction. Blood samples were obtained beforethe 1st and after every following fraction in group B.

VEGF concentration was determined by using the HumanVEGF enzyme linked immunosorbent assay (ELISA, R&DSystem, Inc., Abingdon, United Kingdom) designed fordetection of VEGF165. This ELISA has already been provenreliable for plasma VEGF evaluations in dogs in recentstudies [6,7,27]. The VEGF ELISA was accomplished using theprotocol from the manufacturer. The standard curve was

Table 2Fractionation scheme for curative (group A) and palliative (groupB) treatment

Dose perfx (Gy)

Fx perweek

Total fx Total dose (Gy) Dogs (N)

Group Aa

3.5 4 12 42 33.5 4 14 49 113.0 5 14 49 43.0 5 16 52 3

Group Bb

6.0 2 5 30 256.0 1 4 24 88.0 1 3 24 6

a Dogs receiving a low dose per fx, with a high total dose.b Dogs receiving a high dose per fx, with low total dose.

adapted for low concentrations and the detection limit wastested up to 1 pg/ml. A standard curve was assayed for eachmicrotiter plate. After thawing, every aliquot was assayedtwice and the mean value was taken for statistical analysis.Inter- and intra-donor variations have been tested in healthydogs [28].

Outcome measuresTime to treatment failure (TTF) was defined to be the

interval from the date of completion of radiation therapy tothe date of disease recurrence or progression and/or theobservation of metastasis. The survival time was defined asthe time interval from start of therapy to death/euthanasia.Follow-up evaluation was performed in our clinic or by thereferring veterinarian 3 and 6 months after completion oftherapy and then every year thereafter.

Statistical analysisPlasma VEGF levels showed a skewed distribution,

therefore plasma VEGF levels were transformed logarithmi-cally before analysis (plasma ln VEGF). Statistical analysis(StatView Version 4.0 statistical software application,Abacus concept) was performed using the descriptivestatistics and for correlation of plasma VEGF and tumorresponse the Kruskal–Wallis test was used. Estimates ofprobabilities for time to treatment failure (TTF) werecalculated by the Kaplan–Meier method, and differencesbetween groups were tested with the log rank statistic. Theplasma VEGF concentration was scored as low if plasma VEGFwas less than the 50th percentile of baseline plasma VEGF ofall tumor bearing dogs. The median plasma VEGF for all dogswas 5.2 pg/ml. Therefore, groups were dichotomized inpatients having a plasma VEGF concentration lower or higher5 pg/ml. To analyze the effect of ionizing radiation on VEGFrelease over time the repeated measures analysis withGreenhouse–Geisser correction was used (SPSS program).This test was distinguished from MANOVA to test withinsubject effects. For this analysis plasma VEGF levels afterdose application were subtracted from baseline plasmaVEGF. P values%0.05 were considered significant.

ResultsBaseline analysis

Dogs receiving a low dose per fx, high total dose(group A)

Median plasma VEGF was 6.2 pg/ml. Dogs having noresponse (NR) to radiation therapy had median plasma VEGFof 10.1 pg/ml, dogs with a partial response (PR) had medianplasma VEGF of 7.5 pg/ml and dogs with a completeresponse (CR) had median plasma VEGF of 5.7 pg/ml. Butdifferences between groups were not statistically significant(PZ0.67).

Dogs receiving a high dose per fx, low total dose(group B)

Median plasma VEGF in this group was 4.8 pg/ml. For non-responders the median plasma VEGF was 9.7 pg/ml, dogs

Plasma VEGF in dogs during fractionated radiation therapy242

with PR had median plasma VEGF of 3.2 pg/ml and dogswithout macroscopic disease had median plasma VEGF of0.4 pg/ml. But differences between groups were notstatistically significant (PZ0.2).

Plasma VEGF measured over the courseof radiation therapy

Dogs receiving a low dose per fx, high total dose(group A)

Change of plasma VEGF during radiation therapy was notstatistically different (PZ0.7). Plasma VEGF during radiationtherapy in dogs with sarcomas was stable (sarcoma: PZ0.57). The mean plasma VEGF level of dogs with carcinomaincreased slightly over the course of radiation therapyalthough this was not significant (PZ0.3). Plasma VEGFlevels directly before and after dose application did notdiffer. The repeated measures analysis with the covariables:tumor volume, tumor stage and radiation reaction werecomputed separately and none of the covariables influencedplasma VEGF during therapy (tumor volume: PZ0.2, tumorstage: PZ0.3, radiation reaction: PZ0.4).

Dogs receiving a high dose per fx, low total dose(group B)

Plasma VEGF in dogs treated with a high dose per fractionremained unchanged (PZ0.8). The same was seen whenpatients were grouped according to tumor histology, such asmelanoma (PZ0.6), sarcoma (PZ0.5) and carcinoma (PZ0.8). Plasma VEGF levels directly before and after doseapplication did not differ. None of the covariables influencedthe course of plasma VEGF (tumor volume: PZ0.7, tumorstage: PZ0.4, radiation reaction: PZ0.5) (Fig.1).

Treatment outcomeComparison of dogs irradiated with a curative intent (low

dose per fx, high total dose) to dogs irradiated with apalliative intent (high dose per fx, low total dose) resulted in

Fig. 1. TTF curve for dogs treated with a high total dose: Kaplan–Meier time to treatment failure curves for baseline plasma VEGFconcentration in dogs treated with a high total dose (group A). Themedian plasma VEGF for all dogs was 5.2 pg/ml. Therefore, groupswere dichotomized in patients having a plasma VEGF concentrationlower or higher 5 pg/ml. Dogs with a plasma VEGF level higher5 pg/ml had a significantly shorter TTF (PZ0.038).

significantly longer survival in dogs irradiated with a lowdose per fraction (high total dose) (data not shown).Therefore, the survival analysis was computed for bothgroups separately. Dogs remaining alive at last follow-up ordogs that died of unrelated reasons were censored.

Dogs receiving a low dose per fx, high total dose(group A)

The median follow-up time was 524 days (mean:490 days). At the time of analysis 14 out of 21 dogs haddied because of their cancer disease. Seven dogs werecensored, five dogs were still alive at the time of analysis andtwo dogs died of unrelated reasons. The median TTF of the14 dogs was 153 days (31–440 days). When baseline plasmaVEGF was dichotomized into two groups with plasma VEGFlevels of %5 pg/ml (nZ7), respectively O5 pg/ml (nZ7),then dogs with a plasma VEGF level higher 5 pg/ml had asignificantly shorter TTF (PZ0.04) (Fig. 2). The median TTFof dogs with a plasma VEGF %5 pg/ml was 243 days and fordogs with plasma VEGF O5 pg/ml the median TTF was136 days. This difference in pretreatment plasma VEGF alsoresulted in a shorter survival time (PZ0.04).

Further, we looked at the course of plasma VEGF duringradiation therapy and its influence on TTF and survival. Dogswere separated into two groups. One group had increasingthe other had decreasing or unchanged plasma VEGF levels.However, TTF and survival were not influenced by increasingor decreasing plasma VEGF levels.

Dogs receiving a high dose per fx, low total dose(group B)

The median follow-up time was 93 days (mean: 125 days).Out of 39 dogs, 10 dogs were censored, of which three werestill alive at the time of analysis, five dogs died of unrelatedcauses and two dogs were lost to follow-up. Dogs with highplasma VEGF (O5 pg/ml) had a significant shorter TTF (PZ0.04) (Fig. 2). The median TTF for dogs with plasma VEGF%5 pg/ml (nZ14) was 102 days and dogs with O5 pg/ml

Fig. 2. TTF curve for dogs treated with a low total dose: Kaplan–Meier time to treatment failure curves for baseline plasma VEGFconcentration in dogs treated with a low total dose (group B). Themedian plasma VEGF for all dogs was 5.2 pg/ml. Therefore, groupswere dichotomized in patients having a plasma VEGF concentrationlower or higher 5 pg/ml. Dogs with a plasma VEGF level higher5 pg/ml had a significantly shorter TTF (PZ0.041).

M.C. Wergin et al. / Radiotherapy and Oncology 79 (2006) 239–244 243

(nZ15) plasma VEGF had a median TTF of 80 days. Thisdifference in pretreatment plasma VEGF also resulted in ashorter survival time (PZ0.03). The course of plasma VEGF,as mentioned above, was not influencing TTF or survival.Dogs that responded to therapy and had a plasma VEGF level%5 pg/ml had a significantly longer TTF (PZ0.004) andlonger life span (PZ0.005).

DiscussionVEGF is known to protect endothelial cells for radiation

induced apoptosis [21,26,32]. Therefore, we were inter-ested in changes of plasma VEGF during fractionatedradiation therapy with either high or low dose per fraction.However, our study did not reveal a significant increase ofplasma VEGF during fractionated radiation therapy in bothtreatment groups. In vitro, a 2.5-fold increased VEGFconcentration has been measured in lung squamous carci-noma cells exposed to a single fraction of 15 Gy [2]. Thisincrease was measurable up to 24 h and it was dependent onthe activation of the ERK cascade through activation of theEGFR [2,9]. The activation of the ERK pathway was dosedependent with increased activity after exposure to a lowradiation dose (1 Gy) [4]. Interestingly, we found that dogswith carcinoma irradiated with a low dose per fraction had aslight increase of plasma VEGF, although this was notsignificant.

The regulatory mechanism of VEGF expression is not onlydependent on the activation of ERK. Hypoxia is also a strongstimulus for VEGF expression in tumor cells. Recently, serialoxygen partial pressure measurements (pO2) were done inspontaneous canine tumors during fractionated radiationtherapy and the data suggest that tumors may behavedifferently with increasing or decreasing tumor oxygen levelsduring radiation therapy [1]. These findings may explainincreasing or decreasing plasma VEGF concentrations duringtherapy in individual dog patients in this study. Additionally,micro vessel density (MVD) seems to negatively correlatewith VEGF and positively correlate with tumor oxygenationstatus [8]. We were able to analyse MVD in seven patientsand we found a median MVD of 3.2 vessels per visual field.Dogs with a high plasma VEGF (O5 pg/ml, nZ2) had amedian MVD of 2.2 and dogs with a low plasma VEGF(%5 pg/ml, nZ5) had a median MVD of 4.8. This preliminaryresult might indicate that tumors with a high MVD are betteroxygenated resulting in a lower plasma VEGF. Due to smallsample size a definitive conclusion cannot be drawn.

In a previous study [28], we analyzed the course of VEGFduring fractionated radiation therapy in a smaller studygroup and there was a tendency for increased plasma VEGFlevel in curatively treated dogs. That this tendency did notreach significance in this larger patient group might be dueto several factors. First, a counterbalancing effect mightoccur between ERK pathway activation and changes in theoxygenation status of the tumor. Second, the existence ofmechanisms avoiding VEGF release from tumor tissue havebeen demonstrated. Possibly, binding of VEGF to its receptorand to extracellular matrix occurs and would explain theunchanged plasma VEGF levels. For example, after breastcancer surgery higher VEGF levels were found in wound fluid

than in the blood compartment [16]. The lack of significantdifferences in plasma VEGF levels during radiation therapymight be the simple fact that changes were too small to bedetected.

This study demonstrated that high pre-irradiation plasmaVEGF levels in dogs with spontaneous tumors resulted inshortened survival times. This was seen in both treatmentgroups. Interestingly, this prediction was not dependent ontumor histology or tumor stage. Additionally, we foundlowest median plasma VEGF levels in dogs with a completeresponse to therapy in both treatment groups. Correspond-ingly, palliatively treated dogs (group B) with a low plasmaVEGF level that responded to therapy had also a significantlylonger TTF and a longer live span. In a human colorectalcancer study, it was also shown that high plasma and serumVEGF concentration prior to surgery predicted decreasedoverall survival [5,29].

In a mouse model study, it has been demonstrated thateven a minimal increase in plasma VEGF can cause growth ofmicrometastasis by tipping the balance towards angiogen-esis [20]. Therefore, we wanted to analyze if increasingconcentrations of plasma VEGF over the course of radiationtherapy in individual patients would result in shortened TTFand survival. However, increasing plasma VEGF over timehad no effect on survival or TTF.

In conclusion, plasma VEGF appeared to be a goodprognostic indicator for dog patients receiving radiationtherapy. First, dogs with a low plasma VEGF level are morelikely to respond to radiation therapy and second these dogsalso have a longer TTF and survival.

* Corresponding author. Melanie C. Wergin, Diagnostic Imagingand Radio-Oncology, Vetsuisse Faculty, University of Zurich,Winterthurerstr. 260, CH-8057 Zurich, Switzerland. E-mail address:mwergin@vetclinics.unizh.ch

Received 29 October 2004; received in revised form 19 January 2006;accepted 28 March 2006; Available online 4 May 2006

References[1] Achermann RE, Ohlerth SM, Rohrer-Bley C, et al. Oxygenation

of spontaneous canine tumors during fractionated radiationtherapy. Strahlenther Onkol 2004;180:1–9.

[2] Ando S, Nojima K, Majima H, et al. Evidence for mRNAexpression of vascular endothelial growth factor by X-rayirradiation in a lung squamous carcinoma cell line. Cancer Lett1998;132:75–80.

[3] Becker A, Stadler P, Krause U, et al. Association betweenelevated serum VEGF and polarographically measured tumorhypoxia in head and neck carcinomas. Strahlenther Onkol 2001;177:182–8.

[4] Carter S, Auer KL, Reardon DB, et al. Inhibition of the mitogenactivated protein (MAP) kinase cascade potentiates cell killingby low dose ionizing radiation in A431 human squamouscarcinoma cells. Oncogene 1998;16:2787–96.

[5] Chin KF, Greenman J, Reusch P, Gardiner E, Marme D,Monson JR. Vascular endothelial growth factor and solubleTie-2 receptor in colorectal cancer: associations with diseaserecurrence. Eur J Surg Oncol 2003;29:497–505.

[6] Clifford CA, Hughes D, Beal MW, et al. Plasma vascularendothelial growth factor concentrations in healthy dog anddogs with hemangiosarcoma. J Vet Intern Med 2001;14:131–5.

Plasma VEGF in dogs during fractionated radiation therapy244

[7] Clifford CA, Hughes D, Beal MW, et al. Vascular growth factorconcentrations in body cavity effusions in dogs. J Vet InternMed 2002;16:164–8.

[8] Couvelard A, O’Toole D, Turley H, et al. Microvessel density andhypoxia-inducible factor pathway in pancreatic endocrinetumors: negative correlation of microvascular density andVEGF expression with tumor progression. Br J Cancer 2005;92:94–101.

[9] Dent P, Reardon DB, Park JS, et al. Radiation-induced releaseof transforming growth factor a activates the epidermal growthfactor receptor and mitogen-activated protein kinase pathwayin carcinoma cells, leading to increased proliferation andprotection from radiation induced cell death. Mol Biol Cell1999;10:2493–506.

[10] Dittadi R, Meo S, Fabris F, et al. Validation of blood collectionprocedures for the determination of circulating vascularendothelial growth factor (VEGF) in different blood compart-ments. Int J Biol Markers 2001;16:87–96.

[11] Dunst J, Stadler P, Becker A, et al. Tumor hypoxia and systemiclevels of vascular endothelial growth factor (VEGF) in head andneck cancers. Strahlenther Onkol 2001;177:469–73.

[12] Ferrara N, Davis-Smyth T. The biology of vascular endothelialgrowth factor. Endocr Rev 1997;18:4–18.

[13] Goldman CK, Kendall RL, Cabrera G, et al. Paracrine expressionof a native soluble vascular endothelial growth factor receptorinhibits tumor growth, metastasis, and mortality rate. ProcNatl Acad Sci USA 1998;95:8795–800.

[14] Gorski DH, Beckett MA, Jaskowiak NT, et al. Blockage of thevascular endothelial growth factor stress response increasesthe antitumor effects of ionizing radiation. Cancer Res 1999;59:3374–8.

[15] Hefler L, Tempfer C, Obermair A, et al. Serum concentrationsof vascular endothelial growth factor in vulvar cancer. ClinCancer Res 1999;5:2806–9.

[16] Hormbrey E, Han C, Roberts A, McGrouther DA, Harris AL. Therelationship of human wound vascular endothelial growthfactor (VEGF) after breast cancer surgery to circulating VEGFand angiogenesis. Clin Cancer Res 2003;9:4332–9.

[17] Hyodo I, Doi T, Endo H, et al. Clinical significance of plasmavascular endothelial growth factor in gastrointestinal cancer.Eur J Cancer 1998;34:2041–5.

[18] LaDue T, Klein MK. Toxicity criteria of the veterinary radiationtherapy oncology group. Vet Radiol Ultrasound 2001;42:475–6.

[19] Liu Y, Cox SR, Morita T, Kourembanas S. Hypoxia regulatesvascular endothelial growth factor gene expression in endo-thelial cells. Identification of a 5 0 enhancer. Circ Res 1995;77:638–43.

[20] Maniwa Y, Okada M, Ishii N, Kiyooka K. Vascular endothelialgrowth factor increased by pulmonary surgery accelerates thegrowth of micrometastases in metastatic lung cancer. Chest1998;114:1668–75.

[21] Mc Mahon G. VEGF-receptor signaling in tumor angiogenesis.Oncologist 2000;5:3–10.

[22] Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS,Sukhatme VP. Hypoxic induction of human vascular endothelialgrowth factor expression through c-Src activation. Nature1995;375:577–81.

[23] Nakamura Y, Yasuoka H, Tsujimoto M, et al. Prognosticsignificance of vascular endothelial growth factor D in breastcarcinoma with long-term follow-up. Clin Cancer Res 2003;9:716–21.

[24] Owen LN. The TNM classification of tumours in domesticanimals. Geneva: World Health Organisation; 1980. p. 49.

[25] Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelialgrowth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Lett Nat 1992;359:843–8.

[26] Veikkola T, Karkkainen M, Claesson-Welsh L, Alitalo K.Regulation of angiogenesis via vascular endothelial growthfactor receptors. Cancer Res 2000;60:203–12.

[27] Wergin M, Kaser-Hotz B. Plasma Vascular endothelial growthfactor (VEGF) in seventy dogs with spontaneously occuringtumors. Int J Vivo Res 2004;18:15–19.

[28] Wergin MC, Ballmer-Hofer K, Roos M, et al. Preliminary studyon plasma vascular endothelial growth factor during low- andhigh-dose radiation therapy of dogs with spontaneous tumors.Vet Radiol Ultrasound 2004;45:1–8.

[29] Werther K, Christensen IJ, Nielsen HJ. Prognostic impact ofmatched preoperative plasma and serum VEGF in patients withprimary colorectal carcinoma. Br J Cancer 2002;86:417–23.

[30] Wu Y, Saldana L, Chillar R, Vadgama JV. Plasma vascularendothelial growth factor is useful in assessing progression ofbreast cancer post surgery and during adjuvant treatment. IntJ Oncol 2002;20:509–16.

[31] Wynendaele W, Derua R, Hoylaerts MF, et al. Vascularendothelial growth factor measured in platelet poor plasmaallows optimal separation between cancer patients andvolunteers: a key to study an angiogenic marker in vivo? AnnOncol 1999;10:965–71.

[32] Zingg D, Riesterer O, Fabbro D, Glanzmann C, Bodis S,Pruschy M. Differential activation of the phosphatidylinositol3 0-kinase/Akt survival pathway by ionizing radiation in tumorand primary endothelial cells. Cancer Res 2004;64:5398–406.

[33] Zhong H, De Marzo AM, Laughner E, et al. Overexpression ofhypoxia-inducible factor-1a in common human cancers andtheir metastases. Cancer Res 1999;59:5830–5.