Blood vessels, some fibroblasts and inflammatory cells in the underlying stroma .... Acknowledgements. We thank The Scottish Office Department of Health, The.
The Histochemical Journal 33: 287–294, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.
Expression of vascular endothelial growth factor in normal and tumour oral tissues assessed with different antibodies R. Baillie, K. Harada, J. Carlile, M. Macluskey, S.L. Schor & A.M. Schor∗ Oral Diseases Group, Dental School, University of Dundee, Park Place, Dundee DD1 4HR, Scotland ∗ Author for correspondence Received 20 March 2000 and in revised form 18 May 2001
Summary Expression of vascular endothelial growth factor (VEGF) in oral tissues was assessed using different antibodies. Quantitative and topographical differences were observed between paraffin and cryostat sections. Two polyclonal antibodies (PC36, PC37) differing in their cross-reactivity with VEGF121 (not recognized by PC36), were used to stain serial cryostat sections of normal oral mucosa (n = 8) and squamous cell carcinoma (n = 7). The expression of VEGF in the epithelium was overall higher with PC37 than with PC36, the difference being significant in normal oral mucosa (p = 0.001) but not in squamous cell carcinoma samples (p = 0.094). With PC36, VEGF expression was significantly higher in squamous cell carcinoma than in normal oral mucosa specimens, whereas the opposite was true with PC37. Our results suggest that the relative levels of isoform 121 to that of 165 (and possibly others) may be different in the tissues examined, with VEGF121 preferentially expressed in normal oral mucosa. Previously published conflicting results may, therefore, be due to the presence of variable ratios of VEGF isoforms in the tissues examined, combined with differences in the cross-reactivity of the antibodies used. VEGF isoforms 121, 165 and (for the first time) 189 were detected in two frozen oral tissues by polymerase chain reaction amplification. Quantification of specific VEGF isoforms will be required in future studies concerned with the clinical value of VEGF expression. Introduction Vascular endothelial growth factor (VEGF), also called vascular permeability factor (VPF), belongs to a family of related proteins which at present includes five distinct members: VEGF-A, VEGF-B, VEGF-C, VEGF-D and PIGF. VEGF-A (VEGF) is the best characterized, with alternative splicing of the gene transcript generating five isoforms of 121, 145, 165, 189 and 206 amino acids, respectively (Neufeld et al. 1999). The smaller three isoforms (121, 145 and 165) are more soluble and biologically active, while the two larger isoforms (189 and 206) are predominantly heparin bound. Isoforms 121 and 165 are the most abundant and widely distributed, whereas expression of 145 appears to be restricted to reproductive organs (Houck et al. 1992, Neufeld et al. 1999). VEGF is a potent inducer of angiogenesis, and is believed to stimulate tumour growth indirectly by this mechanism. Experimental animal studies have demonstrated that inhibition of VEGF-induced angiogenesis suppresses tumour growth (Kim et al. 1993) and neutralization of VEGF in established tumours results in apoptosis of immature blood vessels (Benjamin et al. 1999). VEGF mRNA and protein have been found in a variety of human tumours, as well as normal tissues. When assessed quantitatively, VEGF levels have been reported to be higher in the tumour in some cases, but not in others (Weninger et al. 1996, Denhart et al. 1997, Salven et al. 1997, Baillie et al. 2001, Harada et al. 2001). With specific reference to head
and neck cancer, tumour progression and VEGF expression have been variously reported to be directly correlated (Eisma et al. 1997, Sauter et al. 1999, Mineta et al. 2000), inversely correlated (Pammer et al. 1998, Tae et al. 2000, Carlile et al. 2001) or not correlated (Pammer et al. 1998, Carlile et al. 2001). In a previous study, we found that VEGF expression in paraffin-embedded oral tissues was similar in normal mucosa and dysplasia, whilst a comparison of normal mucosa and squamous cell carcinoma led to conflicting results depending on the antibody used (Carlile et al. 2001). Paraffin-embedding and routine histological processing may result in the loss of certain antigens as well as crosslinking and masking of epitopes of interest. Consequently, it is commonly found that paraffin-embedded tissues require pre-treatment of the sections for optimal antigen retrieval, and the pre-treatment used may significantly alter the results obtained (Schoret al. 1998). In contrast, frozen tissues rarely require pre-treatment for antigen retrieval and are therefore believed to be less likely to yield artefactual results. For these reasons, the aim of this study has been to identify those experimental parameters responsible for the reported variable pattern of VEGF expression in normal oral mucosa and oral squamous cell carcinoma. This has involved: (i) a comparison of paraffin-embedded and frozen tissues using five different antibodies, (ii) confirmation of the presence of isoforms 121 and 165 in oral tissues, (iii) demonstration of differential cross-reactivity of two antibodies (PC36 and PC37) for these isoforms, (iv) comparison
288 of VEGF expression using antibodies PC36 and PC37 in serial sections of frozen specimens and (v) comparison of methods to evaluate VEGF expression. Materials and methods Specimens Archival oral tissues were obtained from the Oral Pathology Laboratory, Unit of Oral Surgery and Medicine, University of Dundee. The specimens included frozen normal oral mucosa (NOM; n = 8) and squamous cell carcinoma (SCC; n = 7), as well as paraffin-embedded NOM (n = 5) and SCC (n = 7). Local Ethical Committee approval had been granted for the use of surgical trimming (i.e. redundant normal mucosa). Frozen and paraffin-embedded blocks were available from six additional tumour specimens which had been divided in half at collection, with one half being frozen and the other paraffin-embedded; five of these specimens contained abundant non-tumour epithelium adjacent to the lesion. Immunostaining VEGF immunostaining was evaluated on paraffin-embedded and frozen specimens using the following five antibodies: (1) PC36: rabbit polyclonal raised against a peptide from the C-terminus region of VEGF165 (VEGF Ab-1, cat # PC36 from Calbiochem, Nottingham, England); (2) PC37: rabbit polyclonal raised against a peptide from the N-terminus region of VEGF165 (VEGF Ab-2, cat # PC37 from Calbiochem); (3) A-20: rabbit polyclonal raised against amino acids 1–20 of VEGF165 amino terminus (sc-152 from Santa-Cruz Biotechnology Inc., Santa Cruz, California); (4) R&D: goat polyclonal raised against human rhVEGF165 (Ab-293-NA, from R&D Systems, Oxon UK); (5) GF25: mouse monoclonal raised against a peptide from the N-terminus region of VEGF165 (VEGF Ab-3, cat # GF25 from Calbiochem). Specimens were sectioned (6 µm thick) and mounted on silane-coated slides. Staining was carried out according to standard immunohistochemistry procedures (Baillie et al. 2001, Carlile et al. 2001). Briefly, paraffin sections were de-paraffinized in xylene, then rehydrated through a graded series of ethanols to water. Endogenous peroxidase activity was quenched in 3% hydrogen peroxide in phosphatebuffered solution (PBS) for 30 min. In the case of cryostat sections, this was followed by 2.5% periodic acid for 5 min. Various pre-treatments were tested for optimal antigen retrieval, prior to addition of the primary antibody (Schor et al. 1998). Non-specific binding was blocked by incubating the sections for 30 min in 100% normal serum from the species in which the secondary antibody was raised. The sections were then incubated overnight at 4 ◦ C with the appropriate primary antibody and detection was carried out using the relevant biotinylated secondary antibody,
R. Baillie et al. followed by avidin/biotin peroxidase ABC complex for 30 min. All stages were separated by PBS washes, and final visualization was by 3,3 -diaminobenzidine (DAB) using a DAB substrate kit for 2 min (all reagents from Vector Labs, Peterborough, England). Mayer’s haemalum was used as a light counterstain. Negative controls were incubated with normal rabbit IgG (for PC36, PC37 and A-20), mouse IgG (for GF25) or goat IgG (for R&D antibody). The latter was isolated from normal goat serum before use (Baillie et al. 2001) (all normal serum and IgG from Dako). As a further negative control, A-20 antibody was incubated with 10–100-fold of blocking peptide (Santa Cruz: sc-152P) for 20 min at 37 ◦ C and this solution tested for staining in parallel with the primary antibody. Quantification of VEGF expression in frozen tissues and statistical analyses Serial sections from 15 frozen specimens were stained with antibodies PC36 and PC37. These specimens included 8 NOM (from buccal mucosa) and 7 SCC (from gingiva, tongue and floor of the mouth). The expression of VEGF in the epithelium of these specimens was evaluated according to three indices: (i) percentage of cells stained, (ii) intensity of staining and (iii) product of area and intensity (final score). The intensity of staining was graded by comparison with pre-selected calibration slides as: 1 = weak, 2 = moderate and 3 = strong. Some samples were considered intermediate (e.g. 1.5) or consisted of distinct scores (e.g. 80% intensity 1 and 20% intensity 3); this was taken into account to obtain the final score. The assessments were carried out independently by two or three observers, based on the same calibration slides, and the final results were obtained by consensus. For each antibody and VEGF index, data for normal and tumour specimens were pooled and the two types of tissue were compared by two-tailed Mann–Whitney tests. For VEGF-intensity index, an additional non-parametric test (Mood’s Median) was also used. The same results were obtained by either test and, therefore, only Mann–Whitney’s results are presented. Immunoblotting Immunoblotting was used to confirm the cross-reactivity of VEGF antibodies PC36, PC37 and A-20 with VEGF 121 and 165, i.e. the two most abundant isoforms. Recombinant human VEGF121 (R&D, Abingdon Oxfordshire, UK) and VEGF165 (Oncogene Research Products, Cambridge, MA, USA) (250 ng each) were run on 12% SDS-polyacrylamide gel electrophoresis at 100 V. Proteins were blotted onto a nylon membrane (Bio-Rad Laboratories, Hemel Hempstead, Herts, UK) using a Trans-Blot semi-dry electrophoresis cell (Bio-Rad Laboratories, Hemel Hempstead, Herts, UK). The membrane was then blocked in 5% dried milk in TBS-Tween (25 mM Tris, 2.7 mM KCl, 140 mM NaCl, 2% Tween 20) either overnight (for PC37 and A-20) or 2 h (for PC36). The membrane was next incubated with the primary VEGF
VEGF in oral tissues antibody (2 mg/ml A-20 or 5 mg/ml PC37 for 1 h; 100 mg/ml PC36 overnight), followed by horseradish peroxidaseconjugated goat-anti-rabbit IgG (Dako) (0.125 mg/ml) for 1 h. Finally, the membrane was incubated in Super Signal West-Dura (Pierce & Warriner, Chester, UK) and exposed to X-ray film (Sigma, Poole, UK). All steps were separated by rinsing in TBS-Tween. RNA extraction and detection of VEGF isoforms by PCR Total RNA was purified from frozen normal oral mucosa and oral cancer specimens (one of each type) using the RNeasy minikit (Qiagen, Crawley West Sussex, UK), according to the manufacturer’s protocol. V-12 cells were used as a positive control. V-12 are MCF-7 breast carcinoma cells which have been transfected with a full length complementary DNA for VEGF121 (Zhang et al. 1995). The cells were cultured in MEM supplemented with 15% donor calf serum and 20 mM glutamine. Cells were lysed when confluent using the Qiagen kit lysis buffer and total RNA was purified in parallel with the oral tissues. Reverse transcription was carried out in a total volume of 20 µl containing 500 ng total RNA, 0.5 mM dNTPs, 1 mM oligo-dT primer, 10 units/reaction RNAse inhibitor, 4 units/reaction Omniscript reverse transcriptase for 1 h at 37 ◦ C. (All reverse transcription reagents from Qiagen.) The resulting cDNA (in 1 µl of reaction mixture) was used to amplify the VEGF splice isoforms 121, 165 and 189 using two primer sets: (i) Set A (for isoforms 121, 165 and 189) (Cheung et al. 1998): – Forward AGCTACTGCCATCCAATCG; – Reverse GGCGAATCCAATTCCAAGAG; (ii) Set B (for isoforms 165, 189 and 206) (R&D matched primer set, cat no. RDP-33-025; R&D, Abingdon, Oxfordshire, UK). This set included a positive RNA control. Samples were amplified in a reaction mix containing 2 mM dNTPs (Hybaid, Ashford, Middlesex, UK), primers at a concentration of 15 pmol/reaction, Expand PCR buffer with MgCl2 included and Expand Polymerase mix (Roche, Lewes, East Sussex, UK). Reaction conditions were 94 ◦ C for 4 min followed by 35 cycles of 94 ◦ C for 45 sec, 55 ◦ C for 45 sec and 72 ◦ C for 45 sec concluding with 72 ◦ C for 10 min. Reaction products were visualized using ethidium bromide-containing agarose gels. RNA was omitted for negative controls.
289 We found that optimal staining of paraffin sections required pre-treatment, such as incubation with 0.01% protease XXIV (PC36, PC37, R&D), incubation with 0.4% trypsin (PC36, PC37) or microwaving (A-20) (Schor et al. 1998; Carlile et al. 2001). No pre-treatment was necessary for cryostat sections. In our hands, antibodies A-20 and R&D stained paraffin-embedded and cryostat sections satisfactorily. PC36 and PC37 stained frozen tissues satisfactorily, but were less reliable for paraffin-embedded sections, as the background was often high. Antibody GF25 was found unsatisfactory for both types of tissue with the protocols tested. No staining occurred in any of the negative controls. An initial comparison of paraffin-embedded and frozen tissues was carried out using a total of 11 specimens of NOM (5 paraffin, 6 frozen), and 13 of SCC (7 paraffin, 6 frozen) from different patients. Assessment of the staining clearly indicated quantitative and qualitative differences, as a function of both the antibody used and tissue preservation. Regarding the latter, preferential staining was often observed in the suprabasal layers of the non-tumour epithelium in paraffin sections. In contrast, VEGF expression in cryostat sections occurred either throughout the epithelium or preferentially in the basal area. A direct comparison of frozen and paraffin-embedded sections from the same specimen was possible in six tumours which had been divided in half, one half being frozen and the other paraffin-embedded. In five cases, the sections included non-tumour tissue (either histologically normal or mild dysplasia) adjacent to the tumour. Serial sections of these specimens were stained with VEGF antibodies PC36, PC37, A-20 and R&D. Besides showing antibody-related differences, the results confirmed that there are topographical differences between paraffin-embedded and cryostat sections, and these are most apparent in the epithelium of the normal oral mucosa/dysplasia (Table 1 and Figures 1A and 1B). Detection of VEGF splice forms by immunoblotting Most commonly used VEGF antibodies and probes recognise all the main splice variants of VEGF. According to the manufacturers, antibody PC36 does not cross-react with VEGF121 , whereas PC37 and A-20 do so (Table 2). To verify this assertion, immunoblotting was carried out using recombinant Table 1. VEGF localization in non-tumour epithelium adjacent to oral carcinoma. Comparison of paraffin-embedded and cryostat sections from the same specimens (n = 5), stained with different antibodies. VEGF antibody
Results
Type of section and preferential staining Paraffin
Cryostat
Suprabasal Basal None
Suprabasal Basal None
4 3 3 1
0 0 0 0
Comparison of paraffin-embedded and cryostat specimens using various VEGF antibodies
PC36 PC37 A-20 R&D
0 0 0 0
1 2 2 4
0 1 0 0
5 4 5 5
Optimization of staining was assessed as previously reported (Schor et al. 1998) for five different antibodies (PC36, PC37, A-20, R&D, GF25) in both paraffin and frozen specimens.
Results indicate the number of specimens that show preferential staining in either the suprabasal or basal layer of the epithelium; ‘none’ indicates staining throughout the epithelium, with no preference for either layer.
290
R. Baillie et al. VEGF165 and VEGF121 . The results confirmed that both isoforms were detected by antibodies PC37 and A-20, while PC36 only detected VEGF165 (Figure 2 and Table 2). Expression of VEGF mRNA in oral tissues VEGF isoforms 121 and 165 have been previously identified in a variety of tumours (including head and neck tumours) and normal tissues (Weninger et al. 1996, Cheung et al. 1998, Mineta et al. 2000). In order to confirm this finding, we examined the presence of VEGF mRNA in two oral tissues (NOM and SCC) by PCR amplification. The V-12 cell line was used as a positive control (Zhang et al. 1995). The results, although not quantitative, demonstrated amplification products for VEGF121 , VEGF165 and VEGF189 (but not VEGF206 ) in both the cells and the oral tissues examined (Figure 3). Comparison of VEGF expression in normal and tumour frozen specimens Having confirmed the differential recognition of VEGF121 by antibodies PC36 and PC37, we examined VEGF expression in serial cryostat sections of normal oral mucosa (NOM; n = 8) and oral squamous cell carcinoma (SCC; n = 7) stained with these two antibodies. VEGF expression in the epithelium was quantified by three indices (percentage, intensity and final score) as described in the section Materials and methods. Comparison between antibodies: As observed with other antibodies (Denhart et al. 1997, Carlile et al. 2001), PC36 and PC37 stained both epithelial and stromal cells in NOM and SCC. However, quantitative (see below) and qualitative differences between the antibodies were observed. For example, in NOM, preferential staining of the basal layer was more apparent with PC37 than with PC36. Blood vessels, some fibroblasts and inflammatory cells in the underlying stroma were highlighted more strongly by PC36 than PC37. PC36 usually stained tumour and stromal cells with similar intensity, although occasionally blood vessels were more intensely stained than tumour cells (Figure 1C). In contrast, PC37 usually stained tumour cells more strongly than the surrounding stroma. Quantitative results indicated that, for most specimens, VEGF indices were higher with PC37 than with PC36, the only exceptions being NOM3 (for intensity only) and SCC6 (Table 3). VEGF-percentage was the most informative index. Combining the tissues according to their histological type
Figure 1. Examples of VEGF immunolocalization in paraffin-embedded and frozen oral tissues. A,B: Staining in non-tumour oral mucosa with antibody PC37. Paraffin-embedded section (A) shows localization of VEGF in the suprabasal layer of the epithelium. In contrast, frozen section (B) from the same specimen shows preferential localization in the basal layer, and weak staining in the rest of the epithelium. C: Frozen carcinoma stained with antibody PC36 shows no staining in the tumour cells, in contrast with strong staining in the vasculature. Bar = 50 µm.
Table 2. Cross-reactivity of anti-VEGF antibodies with different VEGF isoforms. Antibody PC36 PC37 A-20
Source Calbiochem Calbiochem Santa Cruz
Epitope
VEGF121
VEGF165
VEGF189
VEGF206
C terminus N terminus N terminus
− +∗ +∗
+ +∗ +∗
+ + +
+ + +
∗
∗
Cross-reactivity (+) or lack of cross-reactivity (−) is based on manufacturer’s data. ∗ represents data confirmed in our study (see Figure 2).
VEGF in oral tissues
291
Figure 2. Immunoblots of VEGF121 and VEGF165 with three antibodies. Antibodies tested were: Santa Cruz A-20, Calbiochem PC37 and Calbiochem PC36. For each antibody, isoform 121 was applied to track A and isoform 165 to track B (250 ng each). All three antibodies detected VEGF165 (tracks B, and some ‘spill-over’ in tracks A). VEGF121 was detected with antibodies A-20 and PC37, but not with PC36.
Figure 3. Identification of VEGF isoforms in oral tissues by RT-PCR amplification. (A): Amplification products obtained with primer set A. Tracks A–D contain negative control (A) and RNA from: V-12 cells (B), oral carcinoma (C) and normal oral mucosa (D). (B): Amplification products obtained with primer set B. Tracks A–E contain negative control (A) and RNA from: positive control provided by manufacturers (B), V-12 cells (C), oral carcinoma (D) and normal oral mucosa (E). Table 3. VEGF expression in oral tissues stained with two different antibodies. Serial cryostat sections of NOM and SCC were stained with VEGF antibodies PC36 and PC37. Results show VEGF staining in the epithelium assessed by two indices: percentage of cells stained and intensity. Specimen NOM1 NOM2 NOM3 NOM4 NOM5 NOM6 NOM7 NOM8 SCC1 SCC2 SCC3 SCC4 SCC5 SCC6 SCC7
PC36
PC37
Percentage
Intensity
Percentage
Intensity
0 5 30 10 0 10 15 0 20 30 15 20 30 80 20
0 1 1.5 1 0 1 1 0 1 1 1 1 1 2 1
100 100 100 100 100 100 100 10 95 75 40 35 50 15 60
1.15 1.2 1.3 1.3 1.1 1.3 1.1 1.5 1.4 1.8 2 2 2 1 2
(NOM or SCC), the difference between VEGF-percentage values obtained with PC36 and PC37 were statistically significant in NOM samples (p = 0.001), but not in SCC samples (p = 0.097) (Figure 4). In NOM, results obtained with PC37
Figure 4. VEGF expression (percentage of cells stained) in serial sections of normal oral mucosa and oral squamous cell carcinoma using antibodies PC36 and PC37. Box-plots show the median (horizontal line), interquartile range (box) and full range (vertical line). (A): In normal mucosa (n = 8), VEGF expression was significantly higher with PC37 than with PC36 (p = 0.001). (B): In squamous cell carcinoma (n = 7) the difference between PC37 and PC36 was not statistically significant (p = 0.094).
were also significantly higher than those with PC36 for both VEGF-intensity (p = 0.009) and final score (p = 0.001). In SCC, PC37 values were higher than PC36 values for VEGFintensity (p = 0.038), but not for final score (p = 0.128). Comparison between NOM and SCC: When PC36 was used, VEGF expression in SCC was significantly higher than in NOM (for both percentage and final score, but not intensity). In contrast, percentage staining with PC37 was significantly higher in NOM than in SCC, VEGF-intensity was significantly higher in SCC than in NOM and final score was similar in both types of tissue (Table 4).
292
R. Baillie et al. Table 4. Comparison between NOM and SCC using cryostat sections stained with VEGF antibodies PC36 and PC37. Antibody
Tissue
PC36
NOM
8
SCC
7
Difference (p) PC37
n
15
NOM
8
SCC
7
Difference (p)
15
VEGFpercentage
VEGFintensity
VEGFfinal score
8.7 ± 10.3 7.5 (13.7) 30.7 ± 22.4 20.0 (10.0) 0.006∗
0.7 ± 0.6 1.0 (1.0) 1.1 ± 0.4 1.0 (0.0) 0.281 ns
10.6 ± 15.0 7.5 (13.7) 42.1 ± 52.3 20.0 (10.0) 0.009*
88.7 ± 31.8 100.0 (0.0) 52.8 ± 26.6 50.0 (40.0) 0.000∗
1.2 ± 0.1 1.2 (0.2) 1.7 ± 0.4 2.0 (0.6) 0.029∗
107.5 ± 38.3 117.5 (20.0) 93.3 ± 42.7 100.0 (63.0) 0.613 ns
Results show the mean ± standard deviation, with median and inter-quartile range (within brackets) underneath. The number of specimens for each type of tissue (n) and p values for the difference between NOM and SCC (Mann–Whitney test) are also indicated. ∗ = significant difference. ns = not significant.
Discussion VEGF is considered to play an important role in tumour development due to its ability to induce angiogenesis and vessel permeability (Kim et al. 1993, Kumar-Singh et al. 1999, Harada et al. 2001). Consequently, a large number of studies have examined the expression of VEGF in histological sections of human tumours in order to determine the possible value of VEGF as a prognostic indicator and the association between VEGF expression and angiogenesis. Unfortunately, there is a lack of consensus in the literature regarding these issues (Baillie et al. 2000, Carlile et al. 2001, Harada et al. 2001), as well as the relative levels of VEGF in tumours and their normal tissue counterparts (see Introduction). Various antibodies to VEGF have been used in previous studies, but a direct comparison between antibodies has only rarely been presented. In this regard, we previously reported that when oral tissues were stained with two VEGF antibodies (Ab-293-NA from R&D and A-20 from Santa Cruz), a comparison between NOM and SCC led to opposite results, depending on the VEGF antibody used (Carlile et al. 2001). Paraffin-embedded tissues were used in our previous study; differences observed between paraffin-embedded and frozen tissues in this communication (Table 1) raise the additional question of possible artefacts due to antigen retrieval in paraffin sections. Consequently, we compared VEGF expression in frozen oral tissues by staining serial sections with two antibodies (PC36 and PC37). Following immunolocalisation, the expression of VEGF has been previously evaluated by different means, including the percentage of cells or area stained (Salven et al. 1997, Maeda et al. 1998, Pammer et al. 1998, Kumar-Singh et al. 1999), the intensity of staining (Inoue et al. 1997, Moriyama et al. 1997, Pammer et al. 1998, Sauter et al. 1999), and a combination of these (Mattern et al. 1999, Baillie et al. 2000, Tae et al. 2000, Carlile et al. 2001). Due to the lack of a standard protocol, we quantified VEGF expression by the consensus of at least two
independent observers using three VEGF indices (percentage of cells stained, intensity and final score). Most previous studies have used only one index, with intensity being most commonly used in oral tissues; however the percentage of cells stained has been found to be more reliable than intensity (this study and Lee et al. 1998). Our results demonstrate quantitative and qualitative differences between the antibodies (Tables 1, 3; Figure 4). Therefore, a comparison between NOM and SCC yielded conflicting results depending on both the antibody and the VEGF index used (Table 4). These data suggest that previous studies reporting contradictory results may be attributed to these two variables (antibody and quantification of VEGF expression), rather than to tissue preservation or staining protocol. Using RT-PCR, previous studies have demonstrated that VEGF isoforms 121 and 165 are present in normal skin (Weninger et al. 1996) and in a variety of tumour types and corresponding normal/adjacent tissues, including head and neck (Mineta et al. 2000), breast (Greb et al. 1999), colon (Tokunaga et al. 1998), lung (Cheung et al. 1998), ovary (Sowter et al. 1997, Shen et al. 2000) prostate (Latil et al. 2000) and thyroid (Katoh et al. 1999). Interestingly, isoforms 189 and 206 were not detected in head and neck tissues (Mineta et al. 2000) or in two oesophageal carcinoma cell lines (Inoue et al. 1997). In contrast, we detected isoforms 121, 165 and 189, but not 206, in two oral tissues examined (Figure 3). It should be noted that lack of detection may be due to poor efficiency of the primers used, as these tend to amplify smaller products preferentially. For example, the primer set A used in our study (able to detect VEGF 121, 165 and 189) revealed weak PCR-products for VEGF 121 and 165, but not for 189. Similarly the primer set B (able to detect VEGF 165, 189 and 206) amplified the first two, but not 206. The antibodies used in the present study were demonstrated to differ in their cross-reactivity with VEGF121 , which is recognized by PC37 but not by PC36 (Table 2, Figure 2). Although it is not possible to determine the relative
VEGF in oral tissues expression of VEGF isoforms using these or any other available antibodies, our results are consistent with the hypothesis that the levels of VEGF121 , relative to those of VEGF165 (and possibly VEGF189 ), may be different in the tissues examined, with VEGF121 being preferentially expressed in NOM. Most studies have relied on detection of VEGF with antibodies or probes that recognise all isoforms, or whose crossreactivity is unknown (Denhart et al. 1997, Salven et al. 1997, Mattern et al. 1999, Baillie et al. 2001, Carlile et al. 2001). Our results indicate that conflicting results may be obtained due to the presence of different ratios of VEGF isoforms in various tissues and to differences in the cross-reactivity of the antibodies used. Future studies concerned with VEGF expression in tissues and its possible association with clinical parameters require a quantitative investigation of the specific VEGF isoforms present. Acknowledgements We thank The Scottish Office Department of Health, The Medical Research Council, The Tayside Area Oncology Fund and The Anonymous ChariTable Trust for financial support; Drs. H.T. Zhang and R. Bicknell for providing the V-12 cells; and Profs. D.M. Chisholm and G.R. Ogden for providing the oral tissues. References Baillie R, Carlile J, Pendleton N, Schor AM (2001) Prognostic value of vascularity and VEGF expression in non-small cell lung cancer. J Clin Pathol 54: 116–120. Benjamin LE, Golijanin D, Itin A, Pode D, Keshet E (1999) Selective ablation of immature blood vessels in established human tumours follows vascular endothelial growth factor withdrawal. J Clin Invest 103: 159–165. Carlile J, Harada K, Baillie R, Macluskey M, Chisholm DM, Ogden GR, Schor SL, Schor AM (2001) The expression of VEGF in normal, dysplastic and cancerous oral tissues assessed by immunostaining with different antibodies. Possible relevance to angiogenesis, tumour progression and field cancerisation. J Oral Pathol Med. Cheung N, Wong MP, Yuen ST, Leung SY, Chung LP (1998) Tissuespecific expression pattern of vascular endothelial growth factor isoforms in the malignant transformation of lung and colon. Human Pathol 29: 910–914. Denhart BC, Guidi AJ, Tognazzi K, Dvorak HF, Brown LF (1997) Vascular permeability factor/vascular endothelial growth factor and its receptors in oral and laryngeal squamous cell carcinoma and dysplasia. Lab Invest 77: 659–664. Eisma RJ, Spiro JD, Kreutzer DL (1997) Vascular endothelial growth factor expression in head and neck squamous cell carcinoma. Am J Surg 174: 513–517. Greb RR, Maier I, Wallwiener D, Kiesel L (1999) Vascular endothelial growth factor A (VEGF-A) mRNA expression decrease after menopause in normal breast tissue but not in lesions. Br J Cancer 81: 225–231. Harada K, Baillie R, Lu S, Syrjanen S, Schor AM (2001) VEGF expression in skin warts. Relevance to angiogenesis and vasodilation. Arch Dermatol Res (in press). Houck KA, Leung DW, Rowland AM, Winer J, Ferrara N (1992). Dual recognition of vascular endothelial growth factor bioavailability by genetic and proteolytic means. J Biol Chem 267: 26031–26037. Inoue K, Ozeki Y, Suganuma T, Sugiura Y, Tanaka S (1997) Vascular endothelial growth factor expression in primary esophageal squamous
293 cell carcinoma. Association with angiogenesis and tumor progression. Cancer 79: 206–213. Katoh R, Miyagi E, Hawaoi A, Hemmi A, Komiyama A, Oyama T, Shibuya M (1999) Expression of vascular endothelial growth factor (VEGF) in human thyroid neoplasms. Human Pathology 30: 891–897. Kim KJ, Li B, Armanini M, Gillet N, Phillips HS, Ferrara N (1993) Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature 362: 841–844. Kumar-Singh S, Weyler J, Martin MJH, Vermeulen PB, Vanmarck E (1999) Angiogenic cytokines in mesothelioma: A study of VEGF, FGF-1 and -2, and TGF Beta expression. J Pathol 189: 72–78. Latil A, Bieche I, Pesche S, Valeri A, Fournier G, Cussenot O, Lidereau R (2000) VEGF overexpression in clinically localized prostate tumors and neuropilin-1 overexpression in metastic forms. Int J Cancer 89: 167–171. Lee AHS, Dublin EA, Bobrow LG, Poulsom R (1998) Invasive lobular and invasive ductal carcinoma of the breast show distinct patterns of vascular endothelial growth factor expression and angiogenesis. J Pathol 185: 394–401. Maeda T, Matsumura S, Hiranuma H, Jikko A, Furukawa S, Ishida T, Fuchihata H (1998) Expression of vascular endothelial growth factor in human oral squamous cell carcinoma: its association with tumour progression and p53 gene status. J Clin Pathol 51: 771–775. Mattern J, Koomagi R, Volm M (1999) Biological characterisation of subgroups of squamous cell lung carcinomas. Clin Cancer Res 5: 1459–1463. Mineta H, Miura K, Ogino T, Takebayashi S, Misawa K, Ueda Y, Suzuki I, Dictor M, Borg A, Wennerberg J (2000) Prognostic value of vascular endothelial growth factor (VEGF) in head and neck squamous cell carcinomas. Br J Cancer 83: 775–781. Moriyama M, Kumagai S, Kawashiri S, Kojima K, Kakihara K, Yamamoto E (1997) Immunohistochemical study of tumour angiogenesis in oral squamous cell carcinoma. Oral Oncology 33: 369–374. Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z (1999). Vascular endothelial growth factor (VEGF) and its receptors. FASEB J 13: 9–22. Pammer J, Weninger W, Mildner M, Burian M, Wojta J, Tschachler E (1998) Vascular endothelial growth factor is constitutively expressed in normal human salivary glands and is secreted in saliva of healthy individuals. J Pathol 186: 186–191. Salven P, Heikkila P, Anttonen A, Kajanti M, Joensuu H (1997) Vascular endothelial growth factor in squamous cell head and neck carcinoma: expression and prognostic significance. Mod Pathol 10: 1128–1133. Sauter ER, Nesbit M, Watson JC, Klein-Szanto A, Litwin S, Herlyn M (1999) Vascular endothelial growth factor is a marker of tumour invasion and metastasis in squamous cell carcinoma of the head and neck. Clin Cancer Res 5: 775–782. Schor AM, Pendleton N, Pazouki S, Smither RL, Morris J, Lessan K, Heerkens E, Chandrachud LM, Carmichael G, Adi M, Chisholm DM, Stevenson H (1998) Assessment of vascularity in histological sections: effects of methodology and value as an index of angiogenesis in breast tumours. Histochem J 30: 849–856. Shen GH, Ghazizadeh M, Kawanami O, Shimizu H, Jin E, Araki T, Sugisaki Y (2000) Prognostic significance of vascular endothelial growth factor expression in human ovarian carcinoma. Br J Cancer 83: 196–203. Sowter HM, Corps AN, Evans AL, Clark DE, Charnock-Jones AS, Smith SK (1997) Expression and localization of the endothelial growth factor family in ovarian epithelial tumors. Lab Invest 77: 607–614. Tae K, El-Nagar AK, Yoo E, Lee JJ, Hong WK, Hittelman WN, Shin DM (2000) Expression of vascular endothelial growth factor and microvessel density in head and neck tumorigenesis. Clin Cancer Res 6: 2821–2828. Tokunaga T, Oshika Y, Abe Y, Ozeki Y, Sadahiro S, Kijima H, Tsuchida T, Yamazaki H, Ueyama Y, Tamaoki N, Nakamura M (1998) Vascular endothelial growth factor (VEGF) mRNA isoform expression pattern is correlated with liver metastases and poor prognosis in colon cancer. Br J Cancer 77: 998–1002.
294 Weninger W, Uthman A, Pammer J, Pichler A, Ballaun C, Lang IM, Plettenberg A, Bankl HC, Sturzl M, Tschachler E (1996) Vascular endothelial growth factor production in normal epidermis and in benign and malignant epithelial skin tumors. Lab Invest 75: 647–657.
R. Baillie et al. Zhang H-T, Craft P, Scott PAE, Ziche M, Weich HA, Harris AL, Bicknell R (1995) Enhancement of tumour growth and vascular density by transfection of vascular endothelial cell growth factor into MCF-7 human breast carcinoma cells. J Natl Cancer Inst 87: 213–219.
