the association of k-ras gene mutation and vascular endothelial growth factor gene expression in...

The Association of K-ras Gene Mutation and VascularEndothelial Growth Factor Gene Expression inPancreatic Carcinoma

Naoya Ikeda, M.D.1,2

Yoshiyuki Nakajima, M.D.2

Masayuki Sho, M.D.2

Masashi Adachi, M.D.1

Cheng-long Huang, M.D.1

Katsumichi Iki, M.D.3

Hiromichi Kanehiro, M.D.2

Michiyoshi Hisanaga, M.D.2

Hiroshige Nakano, M.D.2

Masayuki Miyake, M.D.1

1 Department of Thoracic Surgery and DepartmentV of Oncology, Kitano Hospital, Tazuke KofukaiMedical Research Institute, Osaka, Japan.

2 First Department of Surgery, Nara Medical Uni-versity, Nara, Japan.

3 Department of Surgery, Kawasaki Medical Uni-versity, Kurashiki, Okayama, Japan.

Supported in part by Grants-in-Aid from the Min-istry of Education, Science Sports and Culture ofJapan to Masayuki Miyake (grant 12470280) andfrom the Vehicle Racing Commemorative Founda-tion to Masayuki Miyake.

The authors thank Dr. Nobuo Shirahashi for sta-tistical analysis.

Address for reprints: Masayuki Miyake, M.D., De-partment of Thoracic Surgery and Department V ofOncology, Kitano Hospital, Tazuke Kofukai MedicalResearch Institute, 13-3, Kamiyama-cho, Kita-ku,Osaka 530-8480, Japan; Fax: 181-6-6312-8816;E-mail: [email protected]

Received October 18, 2000; revision receivedMarch 26, 2001; accepted May 2, 2001.

BACKGROUND. Previously, the authors reported the role of the vascular endothelial

growth factor (VEGF) as an angiogenic factor in 40 patients with pancreatic carci-

noma. In this study, they investigated the mechanism underlying the regulation of

VEGF gene expression and evaluated VEGF expression and K-ras gene status in 48

patients with pancreatic carcinoma.

METHODS. The authors used quantitative reverse transcriptase-polymerase chain

reaction analysis and direct sequencing techniques for a retrospective study of

VEGF gene expression and K-ras gene status in tumor tissue samples from 48

patients with pancreatic carcinoma. Immunohistochemistry also was used to in-

vestigate VEGF protein expression.

RESULTS. Thirty-one tumors (64.6%) were evaluated with high VEGF expression,

and 17 tumors (35.4%) were evaluated with low VEGF expression. Of the 48 primary

pancreatic tumors studied, 33 tumors (68.8%) contained mutations of the K-ras

gene. There was a significant correlation between VEGF expression and K-ras

status. Twenty-five of 33 tumors (75.8%) with mutant K-ras genes showed high

VEGF expression, whereas only 6 of 15 tumors with the wild type K-ras (40.0%)

showed high VEGF expression (P 5 0.038). The mean (6 standard error) VEGF

conservation rate for the 33 tumors with mutant K-ras was 1.839 6 1.241, and that

for the 15 tumors with wild type K-ras was 1.057 6 0.983 (P 5 0.037). Furthermore,

the median survival for patients with mutant K-ras was shorter than for those with

wild type K-ras (10.6 months vs. 27.6 months, respectively; P 5 0.026), whereas the

median survival for patients with high VEGF expression was shorter compared with

that for patients with low VEGF expression (9.5 months vs. 26.4 months, respec-

tively; P 5 0.002). Cox regression model analysis indicated that only the VEGF

status was a significant factor for prognosis (P 5 0.024). Other variables, i.e., K-ras

status, histopathologic tumor grade, tumor status, lymph node status, metastatic

status, gender, and age at surgery, were not significant.

CONCLUSIONS. The results of this study suggest that K-ras oncogene mutation may

be associated with VEGF expression and that patients with pancreatic carcinoma

who have high VEGF expression are associated with a poor prognosis. Cancer 2001;

92:488 –99. © 2001 American Cancer Society.

KEYWORDS: K-ras, vascular endothelial growth factor, angiogenesis, pancreaticcarcinoma.

The prognosis for patients with pancreatic carcinoma remainspoor. Surgical resection offers the only chance for cure or long-

term survival, provided the tumor is localized and resectable at thetime of diagnosis.1,2 However, only 8% of patients are eligible toundergo resection at the time of diagnosis,3 and the overall 5-yearsurvival rate is , 10%.1,2 Although the reasons for the aggressive

488

© 2001 American Cancer Society

growth and metastatic behavior of pancreatic carci-noma remain to be classified, various molecular alter-ations have been identified in this malignancy thatmay confer a growth advantage on pancreatic carci-noma cells. Among the many molecular alterations,K-ras mutations are characterized by higher frequencyin pancreatic carcinoma.4,5

More recently, several studies have shown thatangiogenesis is one of the important factors in thegrowth, progression, and metastasis of solid tumors.6,7

Many angiogenic factors in pancreatic carcinomahave been identified, such as transforming growthfactor-a (TGF-a),8 TGF-b,9 acidic fibroblast growthfactor, basic fibroblast growth factor,10 angiogenin,11

and vascular endothelial growth factor (VEGF).12–14 Ina previous study, we reported that an increase in theexpression of VEGF and PD-ECGF was associated with apoor prognosis for patients with pancreatic carcinoma.15

VEGF originally was discovered by Senger et al.16

as a vascular permeability factor and is now believedto be one of the most important angiogenic factorsand a potent mitogen specific for endothelial cells andassociated with angiogenesis.17 Because no other an-giogenic growth factors are known to induce vesselpermeability and subsequent fibrin deposition, VEGFappears to be the most potent of known or putativeangiogenic growth factors.18 VEGF has four isoforms(121, 165, 189, and 206 amino acids) as a result ofalternative splicing of the primary transcript.19 Thetwo shorter forms of VEGF (VEGF121 and VEGF165)are secreted proteins, and the two longer forms(VEGF189 and VEGF206) are bound to heparin-con-taining proteoglycans in the cell surface or in thebasement membrane.20 VEGF165 was seen producingdose dependent growth stimulation of bovine capil-lary endothelial cells, and this effect essentially wasduplicated by VEGF121.21 However, not much isknown about the regulation of VEGF itself, and it wasthe objective of the current study to shed some lighton this area.

It has been reported recently that mutant ras on-cogenes are associated with the up-regulation ofVEGF.22,23 This study was conducted to determinewhether the levels of VEGF gene expression are asso-ciated with the K-ras genes and whether these levelsare related to prognosis for patients with pancreaticcarcinoma.

MATERIALS AND METHODSPatients and Surgical SpecimensIn a previous study, we reported on 40 patients withpancreatic carcinoma. For this study, to which weadded 8 patients who underwent surgery at the FirstDepartment of Surgery, Nara Medical University, be-

tween October 1997 and August 1998), we examined48 patients, including the preceding 40 patients withpancreatic carcinoma who underwent surgery at theFirst Department of Surgery, Nara Medical University,between December 1992 and August 1998. The me-dian age of the patients was 63.7 years, with a range of47– 80 years. Eight patients with distant metastasis,which was both solitary and resectable, underwentpancreatectomy. After undergoing pancreatectomy,almost all patients received intraoperative radiationtherapy with 20 grays of radiation to the retroperito-neal area, including the origins of the portal vein andthe celiac and superior mesenteric arteries. Moreover,postoperative, systemic chemotherapy was adminis-tered to all patients. All of the tissues were obtainedfrom resected specimens and quickly stored at 280 °Cuntil they were used. The residual specimens werefixed in 10% phosphate-buffered formalin and embed-ded in paraffin. To ascertain the presence of tumorcells, half of each fresh tumor tissue specimen wasembedded immediately in optimum cutting tempera-ture (OCT) compound (Miles Laboratories, Kankakee,IL), whereas frozen sections were cut on a cryostat ata thickness of 6 mm, and hematoxylin and eosin-stained slides of tumor specimens were reviewed im-mediately. After the connective tissues had beentrimmed off from the other half of the fresh tumorspecimens, those with tumor cells that accounted for. 80% of all tissue cells were selected for reversetranscriptase-polymerase chain reaction (RT-PCR)and sequencing analyses. The tumors were classifiedaccording to the tumor-node-metastasis (TNM) stag-ing system,24 and the clinical characteristics of thepatients are presented in Table 1. The median follow-up for all patients was 33.8 months, with a range of5.6 –74.4 months.

RT-PCR AnalysisThe total cellular RNA was extracted from the cell linesand frozen tumor tissues with the acid guanidiniumthiocyanate procedure.25 In a preliminary study, weconfirmed that the human colon carcinoma cell linecolo 201 had stabler VEGF mRNA expression than anyother pancreatic carcinoma cell line; therefore, thecolo 201 cell line was used as a positive control. First-strand cDNA synthesis was performed with 5 mg oftotal RNA using a cDNA synthesis kit (Pharmacia, Pis-cataway, NJ) according to the manufacturer’s proto-col. PCR amplification used a 1-mL aliquot of the re-action mixture. To obtain a reproducible, quantitativeperformance for the RT-PCR assays for VEGF, we ti-trated the amount of initial cDNA against the numberof amplification cycles. On the basis of the nucleotidesequence of VEGF, 59-GAAGTGGTGAAGTTCATGGAT-

K-ras Oncogene and VEGF in Pancreatic Carcinoma/Ikeda et al. 489

GTC-39 was used as the sense primer, and 59-CGATCGTTCTGTATCAGTCTTTCC-39 was used as theantisense primer.26 Alternative splicing of the mRNAproduced four variants with 121, 165, 189, and 206amino acids. The primer pair amplified a 408 base pair(bp) fragment as VEGF121, a 541-bp fragment asVEGF165, a 615-bp fragment as VEGF189, and a666-bp fragment as VEGF206. The reaction mixturewas subjected to 28 PCR amplification cycles of 40seconds at 94 °C, 40 seconds at 60 °C, and 90 secondsat 72 °C. b-actin DNA coamplification was used as theinternal PCR control; the sense primer was 59-GAGAGATGACCCAGATCATGT-39, and the antisenseprimer, was 59-ACTCCATGCCCAGGAAGGAAGG-39.27

The same PCR conditions with only 20 cycles wereused to amplify the b-actin DNA. Tubes containing allof the ingredients except the templates were includedin all runs and served as negative reaction controls.

Preparations of the VEGF positive human colon carci-noma cell line colo 201 were used as positive controls.The amplified DNA samples were run on a 1% agarosegel, and the bands were visualized with ethidium bro-mide and photographed with a Polaroid camera. Den-sitometric analysis of the photographic negatives wasused for band quantification.

Specimen Classification Based on RT-PCR ResultsThe densitometric values obtained for a VEGF band ina given tumor tissue sample was divided by the valuesfor the internal control, and the resultant ratio wasreferred to as the gene expression ratio. The expressionratio for a given tumor was then divided by the ex-pression ratio for the human colon carcinoma cell linecolo 201 to obtain the gene conservation rate. Becausemost prognostic factors usually are considered to bedichotomized, discontinuous variables, a cut-off pointwas selected to yield the optimal separation betweenthe low-risk and high-risk levels of overall survival, asdescribed previously.28 Consequently, the most signif-icant P value for survival was found at a relative cut-offpoint for a high VEGF expression of 1.1. If the conser-vation rate value of a given specimen was . 1.1, thenit was considered to indicate high VEGF expression,and, if the value was # 1.1, then it was considered toindicate low VEGF expression.

Immunohistochemical Staining of VEGFA rabbit polyclonal antibody (Santa Cruz Biotechnol-ogy, Inc., Santa Cruz, CA) at a 1:200 dilution was usedfor VEGF. After incubation of the sections with normalgoat serum to block nonspecific binding, they weretreated with anti-VEGF antibodies at room tempera-ture. Immunostaining was performed with thestreptavidin-biotin (SAB) method using a HistofineSAB-POt kit (Nichirei Corporation, Tokyo, Japan). Thestaining was visualized with 3,39-diaminobenzidinetetrahydrochloride, and the slides were counter-stained with hematoxylin. Sections that were incu-bated with mouse myeloma SP2 supernatant served asnegative reaction controls. All of the immunostainedsections were reviewed by two pathologists who hadno knowledge of the patients’ clinical status. Slideswere examined under low power (34 objective) toidentify regions containing low-staining, invasive tu-mor cells. In cases of multiple areas of low intensity,five randomly selected areas were scored, and, in sec-tions in which all of the staining appeared intense, onefield was selected at random. The proportion of tumorcells that showed high and low staining in each se-lected field was determined by counting individualtumor cells at high magnification. At least 200 tumorcells were scored per 3400 field. Positive tumor cells

TABLE 1Correlation between Vascular Endothelial Growth Factor or K-rasStatus and Known Prognostic Factors

CharacteristicsTotal(no.)

VEGF statusPvalue

K-ras status

PvalueHigh Low Mutant

Wildtype

Age at surgery (yrs), 60 15 7 8 NSa 9 6 NSa

$ 60 33 24 9 — 24 9 —Gender

Male 37 23 14 NSa 23 14 NSa

Female 11 8 3 — 10 1 —Tumor status

T1 4 1 3 NSb 1 3 NSb

T2 4 2 2 — 1 3 —T3 12 8 4 — 10 2 —T4 28 20 8 — 21 7 —

Lymph node statusN0 24 13 11 NSa 14 10 NSa

N1 24 18 6 — 19 5 —Metastatic status

M0 40 24 16 NSa 27 13 NSa

M1 8 7 1 — 6 2 —Pathologic stage

I 7 2 5 0.038b 2 5 NSb

II 5 3 2 — 4 1 —III 6 4 2 — 5 1 —IVA 23 16 7 — 17 6 —IVB 7 6 1 — 5 2 —

Histopathologic grade1 15 5 10 0.005b 7 8 0.041b

2 28 22 6 — 22 6 —3 5 4 1 — 4 1 —

Total number of patients 48 31 17 — 33 15 —

VEGF: vascular endothelial growth factor; NS: not significant.a Chi-square test.b Mann-Whitney U test.

490 CANCER August 1, 2001 / Volume 92 / Number 3

were stained at an intensity equivalent to that of nor-mal fibroblasts. When . 10% of the carcinoma cells ina given specimen were stained positively, the samplewas classified as VEGF positive, and staining of # 10%of the cells resulted in a negative classification.

DNA SequencingGenomic DNA was extracted from the specimens byusing proteinase K digestion followed by phenol/chlo-roform extraction and ethanol precipitation.29 The oli-gonucleotides synthesized as partially intron-basedPCR primers of K-ras, as described previously,30 were59-CATGTTCTAATATAGTCACA-39 and 59-CTCTATT-GTTGGATCATAT TCGTCC-39. Forty cycles of PCR am-plification were then performed, consisting of dena-turation at 94 °C for 1 minute, annealing at 50 °C forexon 1 of K-ras, and extension at 72 °C for 1.5 minutes,followed by a final extension at 72 °C for 7 minutes.For detection of an individual base, 30 cycles of PCRamplification of exon 1 of K-ras, using the fmol™ DNAsequencing system (Promega, Madison, WI), were per-formed with biotinylated primers for end labeling.Next, 3 mL of the PCR product for each detected basewere applied to each lane of an 8% polyacrylamide gel,and electrophoresis was performed at 30 Watts for 2hours. After electrophoresis, the bands on the nylonmembrane were visualized with the aid of a chemilu-minescent DNA sequencing kit (NRS-203; Toyobo,Osaka, Japan) according to the manufacturer’s proto-col. Briefly, the transfer of PCR products bands to thenylon membrane was done for 6 hours. The nylonmembrane was then stained with streptavidin for 5minutes, with biotinylated alkaline phosphatase for 5minutes, and finally stained with PPD. The membranewas reacted with enhanced chemiluminescence de-tection reagent (Amersham, Buckinghamshire, United

TABLE 2Expression Rate of Vascular Endothelial Growth Factor and Mutationof K-ras in 48 Patients with Pancreatic Carcinoma

PatientExpression rateof VEGF gene

Mutationof K-ras Patient

Expression rateof VEGF gene

Mutationof K-ras

1 0.64 AGT 34 0.02 Wild type2 1.15 CGT 35 0.22 Wild type3 1.21 CGT 36 0.43 Wild type4 1.32 CGT 37 0.45 Wild type5 0.64 GAT 38 0.51 Wild type6 1.20 GAT 39 0.52 Wild type7 1.30 GAT 40 0.61 Wild type8 1.38 GAT 41 0.64 Wild type9 1.64 GAT 42 0.95 Wild type10 2.40 GAT 43 1.34 Wild type11 2.80 GAT 44 1.39 Wild type12 2.90 GAT 45 1.42 Wild type13 1.41 GCT 46 1.59 Wild type14 1.16 GGA 47 1.72 Wild type15 0.50 GGC 48 4.05 Wild type16 0.21 GTT — — —17 0.36 GTT — — —18 0.37 GTT — — —19 0.39 GTT — — —20 1.05 GTT — — —21 1.21 GTT — — —22 1.52 GTT — — —23 1.58 GTT — — —24 1.70 GTT — — —25 1.71 GTT — — —26 2.60 GTT — — —27 2.70 GTT — — —28 3.50 GTT — — —29 3.60 GTT — — —30 3.85 GTT — — —31 3.99 GTT — — —32 4.98 GTT — — —33 3.70 TGT — — —

FIGURE 1. Agarose gel electrophore-

sis of the reverse transcriptase-poly-

merase chain reaction-amplified, 408-

base-pair (bp) fragment as vascular

endothelial growth factor (VEGF) isoform

121 (VEGF121) (A) and the amplified

internal control b-actin (B). Lane 1, size

marker; lane 2, colon carcinoma cell line

colo 201 with positive VEGF; lanes 3–5,

pancreatic adenocarcinomas with high

VEGF expression; lanes 6–8, primary

pancreatic adenocarcinomas with low

VEGF expression.


Kingdom) and exposed to X-ray film at room temper-ature.

Statistical AnalysisOverall disease specific survival was calculated fromthe date of surgery to the date of death due to disease.The statistical significance of the correlation betweenthe incidence of high VEGF gene expression and sev-eral clinical parameters or K-ras gene status was as-

sessed with the chi-square test or the Mann–WhitneyU test. For VEGF expression, we assessed the accuracyof the RT-PCR method and the immunohistochemicalmethod using a McNemar test. The Kaplan–Meiermethod was used to estimate the probability of overallsurvival as a function of time, and the two factors werecompared with the log-rank test.31 Multivariate anal-ysis was performed with the Cox regression model,32

and eight factors (VEGF status, K-ras status, his-

FIGURE 2. Immunohistochemical

staining of vascular endothelial growth

factor (VEGF). VEGF was expressed

mainly in the cytoplasm and nuclei of

tumor cells. A: VEGF positive; B: VEGF

negative (original magnification, 3100).


topathologic grade, tumor status, lymph node status,metastatic status, gender, and age at surgery) werestudied; scores also were assigned to each variable forthe regression analysis. All P values were obtainedwith a two-tailed statistical analysis, and P values, 0.05 were considered statistically significant.

RESULTSRT-PCR Analysis of VEGF Expression in PancreaticCarcinoma TissuesThere are four structural variants of VEGF (VEGF121,VEGF165, VEGF189, and VEGF206) resulting from al-ternative mRNA splicing in the regions encoding thecytoplasmic domains.19,21 Figure 1A shows that ampli-fication of VEGF121 and VEGF165 cDNA was detected,and some tumors had amplification of VEGF189 andVEGF206. These results indicate that, in pancreaticcarcinoma tissues, the expression of VEGF121 andVEGF165 corresponds to VEGF expression, and wedecided to focus on VEGF121 expression. Of the 48pancreatic carcinoma samples examined, 31 (64.6%)were evaluated with high VEGF gene expression, and17 (35.4%) were evaluated with low VEGF gene expres-sion. The VEGF gene expression rate ranged from 0.02to 4.98, with a mean value of 1.59 (Table 2). Thecorrelations between VEGF gene expression and vari-ous prognostic factors are listed in Table 1, whichshows that VEGF gene expression was associated withhistopathologic grade (P 5 0.005) and pathologic stage(P 5 0.038). High gene expression was seen in 80% ofpatients with Grade 3 tumors compared with 33.3% ofpatients with Grade 1 tumors. In addition, 85.7% ofpatients with Stage IVB disease patients showed highgene expression compared with 28.6% of patients withStage I disease. There was no significant correlationbetween VEGF gene expression and patient age atsurgery, gender, tumor status, lymph node status, ormetastatic status.

Immunohistochemical Analysis of VEGF ProteinExpressionThe VEGF protein positive ratio determined by immu-nohistochemistry was 58.3% (28 of 48 samples). TheVEGF protein was found mainly in the cytoplasm andislets, and the nuclei of the tumor and stromal cells,including macrophages and fibroblasts, also occasion-ally were stained (Fig. 2A). The result of the immuno-histochemical analysis approximated the results of theRT-PCR analysis (McNemar test; see Table 3). Overall,the results of the VEGF immunohistochemistry anal-ysis showed good agreement with the RT-PCR assaydata, and 81.2% of the samples coincided exactly.When there was a discrepancy, the results from the

RT-PCR analysis were used for the specimen classifi-cation.

Detection of Mutations of K-ras in Pancreatic CarcinomaTissuesOf the 48 primary pancreatic carcinoma tissues stud-ied, 33 tumors (68.8%) had K-ras mutations (Table 1,Fig. 3). The correlations between mutations of K-rasand various prognostic factors are listed in Table 1,which shows that K-ras was associated significantlywith histopathologic grading (P 5 0.041): 80.0% ofpatients with Grade 3 tumors had K-ras mutationscompared with 46.7% of patients with Grade 1 tumors.However, there was no statistically significant correla-tionship between K-ras and the other factors.

Correlation between VEGF Gene Expression and K-rasStatusThere was a significant association between VEGF ex-pression and K-ras status. Of the 33 tumors with mu-tant K-ras, 25 (75.8%) showed high VEGF expression,whereas only 6 of 15 (40.0%) tumors with wild typeK-ras showed high VEGF expression (P 5 0.038). Themean (6 standard error) VEGF conservation rate forthe 33 tumors with mutant K-ras was 1.839 6 1.241,and that for the 15 tumors with wild type K-ras was1.057 6 0.983 (P 5 0.037) (Fig. 4).

Association of VEGF Expression and K-ras Mutationswith Survival of Patients with Pancreatic CarcinomaThe prognosis for patients with pancreatic carcinomawas very poor, and the overall survival rate for almostall of the factors was 0%. Of the 48 patients withpancreatic carcinoma, the median survival of patientswith high VEGF gene expression was shorter com-pared with the median survival of patients with lowVEGF expression, as demonstrated in Table 4 (9.5months vs. 26.4 months; P 5 0.002) (Fig. 5A). In addi-tion, significant differences were noted between pa-

TABLE 3The Results of Immunohistochemistry and Reverse-Transcriptase-Polymerase Chain Reaction Analysisa

RT-PCR

Immunohistochemistry

Positive Negative Total

Positive 24 7 31Negative 4 13 17Total 28 20 48

RT-PCR: reverse transcriptase-polymerase chain reaction.a Differential, 20.0625; standard error, 0.313148; 95% confidence interval, 20.375648 – 0.250648; Mc-

Nemar test (P value), 0.5465 (not significant).


tients with high and low VEGF expression for the fol-lowing variables: younger patients (6.1 months vs. 27.6months; P 5 0.044), elderly patients (10.6 months vs.14.0 months; P 5 0.012), male gender (7.5 months vs.26.4 months; P 5 0.005), T4 tumor status (8.3 monthsvs. 14.0 months; P 5 0.020), N1 lymph node status (6.8months vs. 13.2 months; P 5 0.013), M0 metastaticstatus (10.6 months vs. 26.4 months; P 5 0.004), StageIVA disease (9.5 months vs. 26.4 months; P 5 0.026),and Grade 2 histopathologic grade (7.5 months vs.27.6 months; P 5 0.030).

Conversely, the median survival of patients withmutant K-ras was shorter compared with the mediansurvival of patients with wild type K-ras (10.6 monthsvs. 27.6 months; P 5 0.026) (Table 4, Fig. 5B). This

difference was particularly evident for male patientsand patients with M0 metastatic status (P 5 0.018 forboth groups). Furthermore, there was a significantdifference for patients with N0 lymph node status(14.0 months vs. not reached; P 5 0.045).

Prognostic Value of VEGF Status and K-ras StatusWe determined the independent prognostic value foreach variable used in the Cox regression analysis, asshown in Table 5. It was found that only VEGF statuswas a significant prognostic factor (P 5 0.024), and theother variables (K-ras status, histopathologic grade,tumor status, lymph node status, metastatic status,gender, and age at surgery) did not reach significance.

DISCUSSIONVEGF is a potent and widely distributed angiogenicpeptide33 and, as a growth factor, consists of adimeric, 34 – 42 kDa, glycosylated, basic protein.19 Itstwo shorter isoforms (VEGF12l and VEGF165) are sec-retary proteins, and the two longer isoforms (VEGF189and VEGF206) are bound to heparin-containing pro-teoglycans or to the cell surface or basement mem-brane.20 The former isoforms induce their angiogeniceffects by binding to the specific transmembrane ty-rosine kinase receptors KDR/flk-l and flt-l, respec-tively, which are expressed selectively expressed onvascular endothelial cells.33 In the current study,VEGF121 and VEGF165 mRNAs were strongly ex-pressed. Because high VEGF121 expression was iden-tified in the tumors with elevated vascularization,20 weselected this VEGF121 expression as the subject of ourinvestigation. Although many environmental stimuli,such as cytokines, growth factors,34 and hypoxia,35 areknown to transiently up-regulate VEGF, the ras onco-gene regulation of VEGF in actual patients with pan-creatic carcinoma still has not been studied.

FIGURE 4. The correlation between vascular endothelial growth factor (VEGF)

gene expression and K-ras gene status.

FIGURE 3. Exon 1 of K-ras DNA se-

quencing analysis. Each sequence is

shown from 59 (top) to 39 (bottom). A,

wild type K-ras; B, mutant type K-ras.


Ras proteins play a pivotal role in relaying signalsfrom the cell surface (through receptor protein ty-rosine kinases) to a cascade of intracellular serinethreonine kinases, resulting in the activation by phos-phorylation of various transcription factors, such asthe AP1 (jun/fos) complex.36,37 Because the VEGF gene(in humans) contains four potential AP1 sites for thejun/fos heterodimeric transcription factor,19 it clearlyis possible that mutations in ras proto-oncogenes mayinduce or enhance the expression of VEGF.23 Of theras family, K-ras has been widely identified in humantumors.5 The K-ras gene encodes a 21-kDa guanosinetriphosphate (GTP)-binding protein (G-protein),which has intrinsic GTPase activity and regulates cellproliferation by signal transduction across the cellmembrane.5 The mechanism that turns a normal ras

gene into an oncogene is primarily a point mutationlocated in the GTP-binding region of this protein,which results in loss of the intrinsic GTPase activityand, thus, an inability of the G-protein to switch backto its inactive state.38 Mutations of this oncogene re-portedly have been detected in 64 –90% of pancreaticcarcinomas.5,39,40 Most point mutations of K-ras inhuman pancreatic carcinoma are guanine to thymine(GGT3GTT) or guanine to adenine (GGT3GAT) tran-sitions in codon 12 of exon 1.5,41 The current resultsgenerally agree with these findings (Table 2) andindicate that 25 of 33 tumors (75.8%) with mutantK-ras had high VEGF expression, whereas only 6 of15 tumors (40.0%) with wild type K-ras had highVEGF expression (P 5 0.038). These results furthersuggest that activation of the K-ras oncogene by a

TABLE 4Association of Vascular Endothelial Growth Factor Status and K-ras Status with the Survival of Patients with Pancreatic Carcinoma

Characteristics Total (no.)

VEGF status

P value

K-ras status

P value

Mediansurvival VEGFstatus(months)

Overall survivalrate (%)

Median survival(months)

Overall survival rate(%)

High Low High Low Mutant Wild type Mutant Wild type

Age at surgery (yrs), 60 15 6.1 27.6 14.3 0.0 0.044 8.7 27.6 27.8 33.3 0.278$ 60 33 10.6 14.0 0.0 33.8 0.012 11.0 13.2 0.0 38.9 0.121

GenderMale 37 7.5 26.4 11.0 26.1 0.005 8.7 13.2 0.0 36.7 0.018Female 11 10.6 NR 0.0 100.0 ND 12.7 NR 0.0 100.0 ND

Tumor statusT1 4 NR NR 100.0 100.0 ND NR NR 100.0 100.0 NDT2 4 11.3 NR 50.0 100.0 ND NR 11.3 100.0 50.0 NDT3 12 6.3 13.0 0.0 50.0 0.174 12.7 4.6 0.0 0.0 0.164T4 28 8.3 14.0 5.3 0.0 0.020 10.6 13.2 0.0 0.0 0.125

Lymph node statusN0 24 13.2 27.6 16.0 34.1 0.096 14.0 NRa 0.0 51.4 0.045N1 24 6.8 13.2 0.0 44.4 0.013 8.7 11.3 0.0 0.0 0.628

Metastatic statusM0 40 10.6 26.4 8.4 31.7 0.004 13.0 27.6 0.0 43.5 0.018M1 8 6.4 8.7 0.0 0.0 0.591 6.4 4.6 0.0 0.0 0.487

Pathologic stageI 7 NR NR 100.0 100.0 ND NR NR 100.0 100.0 NDII 5 6.3 9.7 0.0 50.0 0.502 6.3 9.7 0.0 0.0 0.774III 6 11.3 13.0 25.0 50.0 0.144 13.0 11.3 30.0 0.0 0.247IVA 23 9.5 26.4 6.7 0.0 0.026 10.6 13.2 0.0 0.0 0.125IVB 7 5.0 8.7 0.0 0.0 0.569 6.4 4.6 0.0 0.0 0.549

Histopathologic grade1 15 11.3 26.4 40.0 45.0 0.333 26.4 NR 0.0 58.3 0.4952 28 7.5 27.6 0.0 0.0 0.030 10.6 9.7 0.0 0.0 0.1613 5 4.2 8.7 0.0 0.0 0.754 4.2 8.3 0.0 0.0 0.774

Total number ofpatients 48 9.5 26.4 6.8 29.7 0.002 10.6 27.6 0.0 37.7 0.026

VEGF: vascular endothelial growth factor; NR: not reached; ND: not done.


point mutation may be associated with VEGF ex-pression.

High levels of VEGF gene expression in patientswith pancreatic carcinoma were found to be associ-ated with a poor prognosis. The median survival of 31patients who had tumors with high VEGF expressionwas 9.5 months, whereas the median survival of 17patients who had tumors with low VEGF expressionwas 26.4 months (P 5 0.002). Moreover, mutant K-ras

was associated with a poor prognosis, because themedian survival of patients with mutant K-ras wasshorter compared with the median survival of patientswith wild type K-ras (10.6 months vs. 27.6 months; P5 0.026). With regard to survival, it is interesting tonote that the survival of patients with and withoutK-ras mutation seemed to be very similar up until10 –12 months postsurgery. However, at this sametime, the survival for patients who had tumors with

FIGURE 5. Overall survival of 48 pa-

tients with pancreatic carcinoma in re-

lation to vascular endothelial growth fac-

tor (VEGF) gene status (A) and K-ras

gene status (B).


high VEGF expression was much poorer comparedwith patients who had tumors with low VEGF expres-sion. We assumed that VEGF gene expression wasdirectly dependent on the presence or absence of K-ras mutation; however, data from the survival curvessuggest that there are several other factors that maycause tumors with high VEGF expression to becomevery aggressive tumors from the time of surgery. Ofcourse, all of this speculation regarding K-ras muta-tion driving VEGF expression was derived from indi-rect evidence, and we await confirmation in futuretissue culture experiments. In the Cox regression anal-ysis, only VEGF gene status was a significant prognos-tic indicator (P 5 0.024). These observations, togetherwith the biologic activities of VEGF described above,suggest that high expression of VEGF may contributeto the aggressiveness of pancreatic carcinoma. Even ifthe down-regulation of VEGF can hardly be expectedto result in a complete cure, it may improve a poorprognosis, inducing tumor dormancy in patients withpancreatic carcinoma.

Recently, several studies on the inhibition of an-giogenesis through VEGF have been reported. Anti-Flk-1 monoclonal antibody was found to inhibit tumor

development and angiogenesis in murine hepatocel-lular carcinoma cells.42 Recombinant humanizedmonoclonal antibody VEGF inhibits VEGF-inducedproliferation of endothelial cells in vitro and tumorgrowth in vivo.43 Considering that mutant ras onco-genes may be associated with the up-regulation ofVEGF, the inhibition of K-ras signaling pathways alsoshould be considered an effective strategy for thetreatment of patients with pancreatic carcinoma. Fur-thermore, disruption of the mutant ras function maycause an antitumor effect by means of suppression oftumor cell proliferation and also by increasing thelevels of tumor cell apoptosis.18 Therefore, we con-clude that the down-regulation of VEGF or the disrup-tion of mutant K-ras functions may have some poten-tially important consequences for the treatment ofpatients with pancreatic carcinoma.

REFERENCES1. Bramhall SR, Allum WH, Jones AG, Allwood A, Cummins C,

Neoptolemos JP. Treatment and survival in 13,560 patientswith pancreatic cancer, and incidence of the disease, in theWest Midlands: an epidemiological study. Br J Surg 1995;82:111–5.

TABLE 5Multivariate Cox Analysis of Overall Survival of 48 Patients with Pancreatic Carcinoma

Variable Assigned score b SE Chi-square Hazard ratio (95% CI) P value

VEGF statusLow 0 21.041 0.461 5.113 0.353 (0.143–0.871) 0.024High 1

K-ras statusWild type 0 20.483 0.482 1.003 0.617 (0.240–1.587) 0.317Mutant type 1

Histopathologic grade1 1 0.184 0.439 0.176 1.202 (0.509–2.841) 0.6752 23 3

Tumor statusT1 1 0.470 0.324 2.110 1.600 (0.849–3.017) 0.146T2 2T3 3T4 4

Lymph node statusN0 0 0.693 0.400 2.995 1.999 (0.912–4.382) 0.084N1 1

Metastatic statusM0 0 0.899 0.547 2.699 2.457 (0.841–7.181) 0.100M1 1

GenderMale 0 0.378 0.513 0.541 1.459 (0.533–3.989) 0.462Female 1

Age (yrs), 60 0 20.135 0.437 0.096 0.874 (0.371–2.057) 0.757$ 60 1

95% CI: 95% confidence interval; VEGF: vascular endothelial growth factor.


2. Nitecki SS, Sarr MG, Colby TV, van Heerden JA. Long-termsurvival after resection for ductal adenocarcinoma of thepancreas. Is it really improving? Ann Surg 1995;221:59 – 66.

3. Gastrointestinal Tumor Study Group. Further evidence ofeffective adjuvant combined radiation and chemotherapyfollowing curative resection of pancreatic cancer. Cancer1987;59:2006 –10.

4. Almoguera C, Shibata D, Forrester K, Martin J, Arnheim N,Perucho M. Most human carcinomas of the exocrine pan-creas contain mutant c-K-ras genes. Cell 1988;53:549 –54.

5. Bos JL. Ras oncogenes in human cancer: a review. CancerRes 1989;49:4682–9.

6. Folkman J. What is the evidence that tumors are angiogen-esis dependent? J Natl Cancer Inst 1990;82:4 – 6.

7. Liotta LA, Steeg PS, Stetler-Stevenson WG. Cancer metasta-sis and angiogenesis: an imbalance of positive and negativeregulation. Cell 1991;64:327–36.

8. Barton CM, Hall PA, Hughes CM, Gullick WJ, Lemoine NR.Transforming growth factor alpha and epidermal growthfactor in human pancreatic cancer. J Pathol 1991;163:111– 6.

9. Friess H, Yamanaka Y, Buchler M, Ebert M, Beger HG, GoldLI, et al. Enhanced expression of transforming growth factorbeta isoforms in pancreatic cancer correlates with de-creased survival. Gastroenterology 1993;105:1846 –56.

10. Yamanaka Y, Friess H, Buchler M, Beger HG, Uchida E,Onda M, et al. Overexpression of acidic and basic fibroblastgrowth factors in human pancreatic cancer correlates withadvanced tumor stage. Cancer Res 1993;53:5289 –96.

11. Shimoyama S, Gansauge F, Gansauge S, Negri G, Oohara T,Beger HG. Increased angiogenin expression in pancreaticcancer is related to cancer aggressiveness. Cancer Res 1996;56:2703– 6.

12. Itakura J, Ishiwata T, Friess H, Fujii H, Matsumoto Y,Buchler MW, et al. Enhanced expression of vascular endo-thelial growth factor in human pancreatic cancer correlateswith local disease progression. Clin Cancer Res 1997;3:1309 –16.

13. Ellis LM, Takahashi Y, Fenoglio CJ, Cleary KR, Bucana CD,Evans DB. Vessel counts and vascular endothelial growthfactor expression in pancreatic adenocarcinoma. Eur J Can-cer 1998;34:337– 40.

14. Seo Y, Baba H, Fukuda T, Takashima M, Sugimachi K. Highexpression of vascular endothelial growth factor is associ-ated with liver metastasis and a poor prognosis for patientswith ductal pancreatic adenocarcinoma. Cancer 2000;88:2239 – 45.

15. Ikeda N, Adachi M, Taki T, Huang C, Hashida H, Takaba-yashi A, et al. Prognostic significance of angiogenesis inhuman pancreatic cancer. Br J Cancer 1999;79:1553– 63.

16. Senger DR, Perruzzi CA, Feder J, Dvorak HF. A highly con-served vascular permeability factor secreted by a variety ofhuman and rodent tumor cell lines. Cancer Res 1986;46:5629 –32.

17. Ferrara N, Houck K, Jakeman L, Leung DW. Molecular andbiological properties of the vascular endothelial growth fac-tor family of proteins. Endocrine Rev 1992;13:18 –32.

18. Rak J, Filmus J, Finkenzeller G, Grugel S, Marme D, KerbelRS. Oncogenes as inducers of tumor angiogenesis. CancerMetast Rev 1995;14:263–77.

19. Tischer E, Mitchell R, Hartman T, Silva M, GospodarowiczD, Fiddes JC, et al. The human gene for vascular endothelialgrowth factor. Multiple protein forms are encoded throughalternative exon splicing. J Biol Chem 1991;266:11947–54.

20. Houck KA, Leung DW, Rowland AM, Winer J, Ferrara N.Dual regulation of vascular endothelial growth factor bio-availability by genetic and proteolytic mechanisms. J BiolChem 1992;267:26031–7.

21. Houck KA, Ferrara N, Winer J, Cachianes G, Li B, Leung DW.The vascular endothelial growth factor family: identificationof a fourth molecular species and characterization of alter-native splicing of RNA. Mol Endocrinol 1991;5:1806 –14.

22. Okada F, Rak JW, Croix BS, Lieubeau B, Kaya M, Roncari L,et al. Impact of oncogenes in tumor angiogenesis: mutantK-ras up-regulation of vascular endothelial growth factor/vascular permeability factor is necessary, but not sufficientfor tumorigenicity of human colorectal carcinoma cells.Proc Natl Acad Sci USA 1998;95:3609 –14.

23. Rak J, Mitsuhashi Y, Bayko L, Filmus J, Shirasawa S, SasazukiT, et al. Mutant ras oncogenes upregulate VEGF/VPF expres-sion: implications for induction and inhibition of tumorangiogenesis. Cancer Res 1995;55:4575– 80.

24. Sobin LH, Wittekind CH. TNM classification of malignanttumors. 5th ed. New York: John Wiley & Sons, Inc., 1997:87–90.

25. Chomczynski P, Sacchi N. Single-step method of RNA iso-lation by acid guanidinium thiocyanate-phenol-chloroformextraction. Anal Biochem 1987;162:156 –9.

26. Ohta Y, Endo Y, Tanaka M, Shimizu J, Oda M, Hayashi Y, etal. Significance of vascular endothelial growth factor mes-senger RNA expression in primary lung cancer. Clin CancerRes 1996;2:1411– 6.

27. Nakajima-Iijima S, Hamada H, Reddy P, Kakunaga T. Mo-lecular structure of the human cytoplasmic beta-actin gene:interspecies homology of sequences in the introns. ProcNatl Acad Sci USA 1985;82:6133–7.

28. Tandon AK, Clark GM, Chamness GC, Chirgwin JM,McGuire WL. Cathepsin D and prognosis in breast cancer.N Engl J Med 1990;322:297–302.

29. Perucho M, Goldfarb M, Shimizu K, Lama C, Fogh J, WiglerM. Human-tumor-derived cell lines contain common anddifferent transforming genes. Cell 1981;27:467–76.

30. Hongyo T, Buzard GS, Palli D, Weghorst CM, Amorosi A,Galli M, et al. Mutations of the K-ras and p53 genes in gastricadenocarcinomas from a high-incidence region around Flo-rence, Italy. Cancer Res 1995;55:2665–72.

31. Mantel N. Evaluation of survival data and two new rankorder statistics arising in its consideration. Cancer Che-mother Rep 1966;50:163–70.

32. Cox DR. Regression models and life-tables. J R Stat Soc B1972;34:187–220.

33. Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascularpermeability factor/vascular endothelial growth factor, mi-crovascular hyperpermeability, and angiogenesis. Am JPathol 1995;146:1029 –39.

34. Kolch W, Martiny-Baron G, Kieser A, Marme D. Regulationof the expression of the VEGF/VPS and its receptors: role intumor angiogenesis. Breast Cancer Res Treat 1995;36:139 –55.

35. Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelialgrowth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 1992;359:843–5.

36. Khosravi-Far R, Der CJ. The Ras signal transduction path-way. Cancer Metast Rev 1994;13:67– 89.

37. Pronk GJ, Bos JL. The role of p21ras in receptor tyrosinekinase signalling. Biochim Biophys Acta 1994;1198:131– 47.


38. Slebos RJ, Kibbelaar RE, Dalesio O, Kooistra A, Stam J, Mei-jer CJ, et al. K-ras oncogene activation as a prognosticmarker in adenocarcinoma of the lung. N Engl J Med 1990;323:561–5.

39. Castells A, Puig P, Mora J, Boadas J, Boix L, Urgell E, et al.K-ras mutations in DNA extracted from the plasma of pa-tients with pancreatic carcinoma: diagnostic utility andprognostic significance. J Clin Oncol 1999;17:578 – 84.

40. Rall CJ, Yan YX, Graeme-Cook F, Beauchamp R, Yandell DW,Povoski SP, et al. Ki-ras and p53 mutations in pancreaticductal adenocarcinoma. Pancreas 1996;12:10 –7.

41. Hruban RH, van Mansfeld AD, Offerhaus GJ, van WeeringDH, Allison DC, Goodman SN, et al. K-ras oncogene ac-tivation in adenocarcinoma of the human pancreas. A

study of 82 carcinomas using a combination of mutant-enriched polymerase chain reaction analysis and allele-specific oligonucleotide hybridization. Am J Pathol 1993;143:545–54.

42. Yoshiji H, Kuriyama S, Hicklin DJ, Huber J, Yoshii J, Miy-amoto Y, et al. KDR/Flk-1 is a major regulator of vascularendothelial growth factor-induced tumor development andangiogenesis in murine hepatocellular carcinoma cells.Hepatology 1999;30:1179 – 86.

43. Presta LG, Chen H, O’Connor SJ, Chisholm V, Meng YG,Krummen L, et al. Humanization of an anti-vascular endo-thelial growth factor monoclonal antibody for the therapy ofsolid tumors and other disorders. Cancer Res 1997;57:4593–9.


the association of k-ras gene mutation and vascular endothelial growth factor gene expression in...

Documents