Several of the endothelial cell polypeptide mitogens that have been described probably play a role in blood vessel homeostasis. Two overlapping complementary DNA clones encoding human endothelial cell growth factor (ECGF) were isolated from a human brain stem complementary DNA library. Southern blot analysis suggested that there is a single copy of the ECGF gene and that it maps to human chromosome 5 at bands 5q31.3 to 33.2 A 4.8-kilobase messenger RNA was present in human brain stem messenger RNA. The complete amino acid sequence of human ECGF was deduced from the nucleic acid sequence of these clones; it encompasses all the well-characterized acidic endothelial cell polypeptide mitogens described by several laboratories. The ECGF-encoding open reading frame is flanked by translation stop codons and provides no signal peptide or internal hydrophobic domain for the secretion of ECGF. This property is shared by human interleukin-1, which is approximately 30 percent homologous to ECGF.

Hypoxia inducible factors HIF1alpha and HIF2alpha are important proteins involved in the regulation of the transcription of a variety of genes related to erythropoiesis, glycolysis and angiogenesis. Hypoxic stimulation results in rapid increase of the HIF1alpha and 2alpha protein levels, as a consequence of a redox-sensitive stabilization. The HIFalphas enter the nucleus, heterodimerize with the HIF1beta protein, and bind to DNA at the hypoxia response elements (HREs) of target genes. In this study we evaluated the immunohistochemical expression of these proteins in 108 tissue samples from non-small-cell lung cancer (NSCLC) and in normal lung tissues. Both proteins showed a mixed cytoplasmic/nuclear pattern of expression in cancer cells, tumoural vessels and tumour-infiltrating macrophages, as well as in areas of metaplasia, while normal lung components showed negative or very weak cytoplasmic staining. Positive HIF1alpha and HIF2alpha expression was noted in 68/108 (62%) and in 54/108 (50%) of cases respectively. Correlation analysis of HIF2alpha expression with HIF1alpha expression showed a significant association (P< 0.0001, r = 0.44). A strong association of the expression of both proteins with the angiogenic factors VEGF (P< 0.004), PD-ECGF (P< 0.003) and bFGF (P< 0.04) was noted. HIF1alpha correlated with the expression of bek-bFGF receptor expression (P = 0.01), while HIF2alpha was associated with intense VEGF/KDR-activated vascularization (P = 0.002). HIF2alpha protein was less frequently expressed in cases with a medium microvessel density (MVD); a high rate of expression was noted in cases with both low and high MVD (P = 0.006). Analysis of overall survival showed that HIF2alpha expression was related to poor outcome (P = 0.008), even in the group of patients with low MVD (P = 0.009). HIF1alpha expression was marginally associated with poor prognosis (P = 0.08). In multivariate analysis HIF2alpha expression was an independent prognostic indicator (P = 0.006, t-ratio 2.7). We conclude that HIF1alpha and HIF2alpha overexpression is a common event in NSCLC, which is related to the up-regulation of various angiogenic factors and with poor prognosis. Targeting the HIF pathway may prove of importance in the treatment of NSCLC.

The prognosis of hepatocellular carcinoma (HCC) still remains dismal, although many advances in its clinical study have been made. It is important for tumor control to identify the factors that predispose patients to death. With new discoveries in cancer biology, the pathological and biological prognostic factors of HCC have been studied quite extensively. Analyzing molecular markers (biomarkers) with prognostic significance is a complementary method. A large number of molecular factors have been shown to associate with the invasiveness of HCC, and have potential prognostic significance. One important aspect is the analysis of molecular markers for the cellular malignancy phenotype. These include alterations in DNA ploidy, cellular proliferation markers (PCNA, Ki-67, Mcm2, MIB1, MIA, and CSE1L/CAS protein), nuclear morphology, the p53 gene and its related molecule MD M2, other cell cycle regulators (cyclin A, cyclin D, cyclin E, cdc2, p27, p73), oncogenes and their receptors (such as ras, c-myc, c-fms, HGF, c-met, and erb-B receptor family members), apoptosis related factors (Fas and FasL), as well as telomerase activity. Another important aspect is the analysis of molecular markers involved in the process of cancer invasion and metastasis. Adhesion molecules (E-cadherin, catenins, serum intercellular adhesion molecule-1, CD44 variants), proteinases involved in the degradation of extracellular matrix (MMP-2, MMP-9, uPA, uPAR, PAI), as well as other molecules have been regarded as biomarkers for the malignant phenotype of HCC, and are related to prognosis and therapeutic outcomes. Tumor angiogenesis is critical to both the growth and metastasis of cancers including HCC, and has drawn much attention in recent years. Many angiogenesis-related markers, such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), platelet-derived endothelial cell growth factor (PD-ECGF), thrombospondin (TSP), angiogenin, pleiotrophin, and endostatin (ES) levels, as well as intratumor microvessel density (MVD) have been evaluated and found to be of prognostic significance. Body fluid (particularly blood and urinary) testing for biomarkers is easily accessible and useful in clinical patients. The prognostic significance of circulating DNA in plasma or serum, and its genetic alterations in HCC are other important trends. More attention should be paid to these two areas in future. As the progress of the human genome project advances, so does a clearer understanding of tumor biology, and more and more new prognostic markers with high sensitivity and specificity will be found and used in clinical assays. However, the combination of some items, i.e., the pathological features and some biomarkers mentioned above, seems to be more practical for now.

To evaluate whether angiogenic factors are of clinical relevance to actual human pancreatic cancers, we studied the intratumoral microvessel density (IMD), and PD-ECGF, VEGF protein expression in 40 pancreatic cancers using immunohistochemistry. We also investigated PD-ECGF and VEGF gene expression using reverse transcriptase-PCR (RT-PCR). Of the 40 pancreatic cancers studied, 30 carcinomas (75.0%) were evaluated to be PD-ECGF-positive and 10 carcinomas (25.0%) were determined to be PD-ECGF-negative. In contrast, 27 carcinomas (67.5%) were evaluated to be VEGF-positive, whereas 13 carcinomas (32.5%) were VEGF-negative. VEGF gene expression was moderately associated with an increase in the IMD (r2 = 0.181, P = 0.006), but no significant relationship was found between PD-ECGF gene expression and the IMD (r2 = 0.093, P = 0.059). However, tumours with positive expression for both PD-ECGF and VEGF had a higher IMD (P = 0.027). The results of the immunohistochemistry agreed well with the results of the quantitative RT-PCR. The median survival time of the hypervascular group was significantly shorter than that of the hypovascular group (P < 0.0001). In comparing the survival according to PD-ECGF and VEGF gene expression, the median survival time of the patients with positive PD-ECGF expression was significantly shorter than those with negative PD-ECGF expression (P = 0.040). Furthermore, the median survival time of the patients with positive VEGF expression was significantly shorter than those with negative VEGF expression (P = 0.048). However, the Cox multivariate analysis indicated that the IMD and VEGF expression were independent prognostic factors of the various clinicopathologic variables in pancreatic cancer patients (P = 0.0021 and P = 0.0443, respectively).

OBJECTIVE

To investigate the prognostic values of putative hepatic stem/progenitor cell (HSC/HPC) biomarkers in patients with hepatocellular carcinoma (HCC).

METHODS

Fourteen biomarkers related to HSCs/HPCs or tumour angiogenesis were assessed by qRT-PCR and then validated by tissue microarrays (TMAs) in three independent cohorts of patients with HCC undergoing curative resection (n=67, 314 and 73).

RESULTS

Most of the biomarkers were found to be overexpressed in patients with recurrent HCC by quantitative reverse transcription-PCR (qRT-PCR). The HSC/HPC biomarkers cytokeratin 19, ATP-binding cassette subfamily G member 2 (ABCG2), CD133, Nestin and CD44, and the markers of angiogenesis microvessel density (MVD, determined by CD34 immunostaining), vascular endothelial growth factor (VEGF) and platelet-derived endothelial cell growth factor (PD-ECGF) were confirmed as significant predictors for overall survival (OS) and/or relapse-free survival (RFS) in TMA analysis. As compared with the low HSC/HPC profile group, patients with a high HSC/HPC profile who had higher VEGF levels (p=0.012) and MVD (p=0.030) in tumours had significantly lower OS and RFS (p<0.0001). Based on Cox regression, a simplified model including CD133, CD44, Nestin and MVD was constructed and confirmed as an independent predictor for OS (p<0.0001) and RFS (p<0.0001), regardless of alpha-fetoprotein level, tumour stage and recurrence time (p<0.0001 for all).

CONCLUSIONS

High expression levels of HSC/HPC biomarkers are related to tumour angiogenesis and poor prognosis of HCC. The simplified model based on the HSC/HPC and tumour angiogenesis profile can be used to classify patients with HCC with a high risk of tumour recurrence after surgery.

Endothelial cell growth factor (ECGF), an anionic polypeptide mitogen, binds to immobilized heparin. The interaction between the acidic polypeptide and the anionic carbohydrate suggests a mechanism that is independent of ion exchange. Monoclonal antibodies to purified bovine ECGF inhibited the biological activity of ECGF in crude preparations of bovine brain. These data indicate that ECGF is the principal mitogen for endothelial cells from bovine brain, that heparin affinity chromatography may be used to purify and concentrate ECGF, and that the affinity of ECGF for heparin may have structural and perhaps biological significance.

Human umbilical vein (HUV) endothelial cells were grown for 15 to 21 passages at a split ratio of 1:5 (at least 27 population doublings) on a human fibronectin (HFN) matrix in Medium 199 supplemented with fetal bovine serum (FBS) and endothelial-cell growth factor (ECGF). This system also permitted the growth of HUV endothelial cells at cell densities as low as 1.25 cells/cm2. In addition to delaying the premature senescence of HUV endothelial cells, ECGF also reduced the serum requirement for low-density HUV endothelial-cell growth; 2.5% serum and ECGF yields half-maximum growth as compared to high serum controls. Significant HUV endothelial-cell growth was also observed in medium supplemented with either ovine hypophysectomized (HYPOX) serum, plasma-derived serum (PDS), or HYPOX-PDS in the presence of ECGF, suggesting that neither the pituitary nor the platelet contributes to HUV endothelial-cell growth.

BACKGROUND

Platelet-derived endothelial cell growth factor (PD-ECGF) is known to promote the development of new blood vessels, which are fundamental to tumor growth and metastasis. We previously found that thymidine phosphorylase (dThdPase) and PD-ECGF are the same protein.

OBJECTIVE

We retrospectively examined the expression of dThdPase in primary colorectal carcinomas, its association with angiogenesis and clinicopathologic findings, and its prognostic value.

METHODS

Tissues were obtained from the tumors of 163 patients whose colorectal carcinomas were completely removed by surgery. Microvessels assessed by immunostaining endothelial cells for factor VIII were counted on a 400x field in the most active areas of neovascularization within the tumor. We purified the monoclonal antibody against dThdPase and studied the expression of dThdPase in the same serial sections used for the detection of factor VIII. Those who carried out microvessel counting and dThdPase expression assessment had no knowledge of clinicopathologic findings. The significance of dThdPase in the prognosis of patients with colorectal carcinomas was also examined in the survival analysis of mortality follow-up data covering the period between 1984 through 1991. Reported P values are from two-sided tests of statistical significance.

RESULTS

The mean microvessel count (+/- standard deviation) in dThdPase-positive colorectal carcinoma specimens (17.5 +/- 7.2) was higher (P < .001) than that in dThdPase-negative carcinoma specimens (9.3 +/- 5.5). The dThdPase positivity was in accordance with the microvessel count. dThdPase positivity showed highly significant statistical associations with tumor size, extent of invasion, lymph node metastasis, lymphatic invasion, and venous invasion. Cox regression analysis revealed that dThdPase expression was prognostic for poor disease outcome after adjustment for Dukes' stage and microvessel count.

CONCLUSIONS

These findings suggest that higher levels of dThdPase expression in colorectal carcinomas are associated with more extensive angiogenesis, poor clinical and laboratory findings, and unfavorable clinical outcome.

CONCLUSIONS

Inhibition of dThdPase in human colorectal carcinomas might improve prognosis for some patients.

High microvessel density, an indirect measure of angiogenesis, has been shown to correlate with increased tumour size, lymph node involvement and poor prognosis in non-small-cell lung cancer (NSCLC). Tumour cell vascular endothelial growth factor (VEGF) and platelet-derived endothelial cell growth factor (PD-ECGF) expression correlate with angiogenesis and a poor outcome in this disease. In a retrospective study VEGF and PD-ECGF expression and microvessel density were evaluated immunohistochemically in surgically resected specimens (T1-3, N0-2) from 223 patients with operable NSCLC using the VG1, P-GF.44C and JC70 monoclonal antibodies respectively. High VEGF immunoreactivity was seen in 104 (46.6%) and PD-ECGF in 72 (32.3%) cases and both were associated with high vascular grade tumours (P= 0.009 and P= 0.05 respectively). Linear regression analysis revealed a weak positive correlation between VEGF and PD-ECGF expression in cancer cells (r= 0.21; P = 0.002). Co-expression of VEGF and PD-ECGF was not associated with a higher microvessel density than VEGF or PD-ECGF only expressing tumours. Furthermore a proportion of high vascular grade tumours expressed neither growth factor. Univariate analysis revealed tumour size, nodal status, microvessel density and VEGF and PD-ECGF expression as significant prognostic factors. Tumour size (P < 0.02) and microvessel density (P < 0.04) remained significant on multivariate analysis. In conclusion, VEGF and PD-ECGF are important angiogenic growth factors and have prognostic significance in NSCLC. Furthermore the study underlines the prognostic significance of microvessel density in operable NSCLC.

Endothelial cell growth factor (ECGF) is a potent polypeptide mitogen for endothelial cells and fibroblasts. The mitogenic effects of ECGF are inhibited by the lymphokine gamma-interferon (gamma-IFN) in a dose-dependent manner. Gamma-IFN also induces a unique change in endothelial cell morphology which is maximally expressed in the presence of ECGF. The antiproliferative and phenotypic modulatory effects of gamma-IFN on endothelial cells are reversible. Inhibition of ECGF-induced endothelial cell proliferation by gamma-IFN is accompanied by a concentration- and time-dependent decrease in binding of 125I-ECGF to the endothelial cell surface. Scatchard analyses of the binding data in the presence and absence of gamma-IFN demonstrate a decrease in the number of ECGF-binding sites rather than a decrease in ligand affinity for the receptor. Cross-linking experiments with disuccinimidyl suberate demonstrate a decrease in the 170,000 Mr cross-linked receptor-ligand complex. These data suggest that gamma-IFN inhibits endothelial cell proliferation by a mechanism which involves growth factor receptor modulation.

Tumour angiogenesis, as assayed by microvessel density (MVD), and the expression of vascular endothelial growth factor (VEGF) and platelet-derived endothelial cell growth factor (PD-ECGF) have become established as important prognostic indicators for many tumour types. In this study, MVD and the expression of VEGF and PD-ECGF were examined by immunohistochemical staining of 50 pancreatic cancer tissues, and the relationships between either MVD or the expression of these two angiogenic factors and the clinicopathological features, including survival, were analysed. The expression of VEGF and PD-ECGF proteins were confirmed by Western blot analysis and VEGF mRNA isoforms were determined by reverse transcriptase-polymerase chain reaction (RT-PCR) in five pancreatic cancer cell lines. Twenty-eight (56%) of 50 pancreatic cancers were positive for VEGF protein in cancer cells, and 16 (32%) showed strong PD-ECGF staining in cancer and infiltrating cells. VEGF121 and VEGF165 were identified as the predominant species produced in pancreatic cancer cells. The overexpression of VEGF and PD-ECGF protein significantly correlated with high MVD (P = 0.002, 0.044, respectively). Advanced stage of disease was significantly more frequent in patients with high MVD (P = 0.025). No significant association was found between the expression of VEGF or PD-ECGF and clinicopathological features, except for tumour histology. The expression of PD-ECGF correlated with poor survival (P = 0.011), but MVD and VEGF expression were not found to be useful for the prediction of overall survival. This study suggests that VEGF and PD-ECGF may play an important role in tumour angiogenesis, and that PD-ECGF expression seems to be useful for establishing prognoses for pancreatic cancer.

Human tumors can constitutively express cytokines and growth factors, but the extent of this expression has not been investigated. Using 44 different probes to cytokines, growth factors, and their receptors, we tested 21 melanoma and 5 melanocyte cultures for RNA transcript expression by reverse transcriptase-polymerase chain reaction. With 30 amplification cycles, expression of the cytokines interleukin (IL)-1 beta, IL-6, leukemia inhibitory factor (LIF), IL-7, gro alpha, IL-8 and the p35 chain of IL-12 was detected in more than 60% of melanomas. Concomitant receptors for IL-6 and IL-7 were also detected. IL-1 alpha, IL-5, Rantes, IL-10, interferon (IFN)-beta, tumor-necrosis factor (TNF)-alpha, G-colony-stimulating factor (CSF) and GM-CSF were expressed at lower levels. Melanocytes showed greatly reduced cytokine RNA transcripts, and only gro alpha was consistently detected. No expression of IL-2, IL-3, IL-4, IL-9, the p40 chain of IL-12, IFN-alpha or IFN-gamma RNA transcripts was detected in melanomas or melanocytes. The growth factors expressed by melanomas and, after further signal amplification, by melanocytes were transforming growth factor (TGF)-alpha, epidermal growth factor (EGF), TGF-beta, endothelial-cell growth factor (ECGF), basic-fibroblast growth factor (bFGF), nerve growth factor (NGF) and steel. The receptors EGFR, FGFR, NGFRp70 and c-kit were also expressed by melanomas and melanocytes. These results point to new possible autocrine and paracrine pathways in melanoma biology.

Angiogenesis is a significant prognostic factor in melanoma, but the angiogenic factors controlling the neovascularization are not well defined. The purpose of this study was to investigate whether the angiogenesis and metastasis of melanoma are promoted by vascular endothelial growth factor (VEGF), interleukin 8 (IL-8), platelet-derived endothelial cell growth factor (PD-ECGF), and/or basic fibroblast growth factor (bFGF). Cells from human melanoma lines (A-07, D-12, R-18, and U-25) transplanted to BALB/c nu/nu mice were used as tumor models. Expression of angiogenic factors was studied by ELISA, Western blotting, and immunohistochemistry. Angiogenesis was assessed by using an intradermal angiogenesis assay. Lung colonization and spontaneous lung metastasis were determined after i.v. and intradermal inoculation of tumor cells, respectively. The specific roles of VEGF, IL-8, PD-ECGF, and bFGF in tumor angiogenesis, lung colonization, and spontaneous metastasis were assessed in mice treated with neutralizing antibody. The melanoma lines expressed multiple angiogenic factors, and each line showed a unique expression pattern. Multiple angiogenic factors promoted angiogenesis in the most angiogenic melanoma lines, whereas angiogenesis in the least angiogenic melanoma lines was possibly promoted solely by VEGF. Tumor growth, lung colonization, and spontaneous metastasis were controlled by the rate of angiogenesis and hence by the angiogenic factors promoting the angiogenesis. Lung colonization and spontaneous metastasis in A-07 were inhibited by treatment with neutralizing antibody against VEGF, IL-8, PD-ECGF, or bFGF. Each of these angiogenic factors may promote metastasis in melanoma, because inhibition of one of them could not be compensated for by the others. Our observations suggest that efficient antiangiogenic treatment of melanoma may require identification and blocking of common functional features of several angiogenic factors.

Culture conditions that favor rapid multiplication of human umbilical vein endothelial cells (HUV-EC) also support long-term serial propagation of the cells. This is routinely achieved when HUV-EC are grown in Medium 199 (M-199) supplemented with fetal bovine serum (FBS) and endothelial cell growth factor (ECGF), on a human fibronectin (HFN) matrix. The HUV-EC can shift from a proliferative to an organized state when the in vitro conditions are changed from those favoring low density proliferation to those supporting high density survival. When ECGF and HFN are omitted, cultures fail to achieve confluence beyond the first or second passage: the preconfluent cultures organize into tubular structures after 4-6 wk. Some tubes become grossly visible and float in the culture medium, remaining tethered to the plastic dish at either end of the tube. On an ultrastructural level, the tubes consist of cells, held together by junctional complexes, arranged so as to form a lumen. The smallest lumens are formed by one cell folding over to form a junction with itself. The cells contain Weibel-Palade bodies and factor VIII-related antigen. The lumens contain granular, fibrillar and amorphous debris. Predigesting the HFN matrix with trypsin (10 min, 37 degrees C) or plasmin significantly accelerates tube formation. Thrombin and plasminogen activator had no apparent effect. Disruption of the largest tubes with trypsin/EDTA permits the cells to revert to a proliferative state if plated on HFN, in M-199, FBS, and ECGF. These observations indicate that culture conditions that do not favor proliferation permit attainment of a state of nonterminal differentiation (organization) by the endothelial cell. Furthermore, proteolytic modification of the HFN matrix may play an important role in endothelial organization.

Mast cells accumulate at sites of angiogenesis. The factor(s) that control mast-cell recruitment at these sites have yet to be defined. We sought to determine if angiogenic factors result in mast-cell chemotaxis. In this study, we observed that platelet-derived growth factor-AB (PDGF-AB), vascular endothelial cell growth factor (VEGF), and basic fibroblast growth factor (bFGF) each cause directed migration of murine mast cells at picomolar concentrations, with a typical bell-shaped dose-response curve. Another potent angiogenic factor, platelet-derived endothelial cell growth factor (PD-ECGF), appears to promote chemokinesis of mast cells, whereas tumor necrosis factor-alpha, a weak angiogenic factor, is less robust but still functions as a mast cell chemotactic factor. Epidermal growth factor (EGF), a growth factor with minimal angiogenic properties, was ineffective as a mast cell chemotactic factor. A checkerboard analysis confirmed the directional chemotactic response of PDGF-AB, VEGF, and bFGF, while indicating the chemokinetic response induced by PD-ECGF. Cross-desensitization of growth-factor-induced directed migration was observed between PDGF-AB and bFGF, and also between PDGF-AB and PD-ECGF. Tyrosine kinase-inhibitor genistein effectively dampened the chemotactic responses, whereas pertussis toxin had no effect. In summary, our findings suggest that factors known to act on endothelial cells and stimulate neovascularization may simultaneously serve to recruit mast cells to these sites. The local accumulation of mast cells is believed to facilitate new vessel formation through complex cell:cell interactions.

The pathogenesis of rheumatoid arthritis (RA) and osteoarthritis (OA) remains obscure, although angiogenesis appears to play an important role. We recently confirmed an overexpression of two angiogenic factors, namely vascular endothelial growth factor (VEGF) and platelet-derived endothelial cell growth factor (PD-ECGF), by the lining and stromal cells of the synovium in both conditions. Because hypoxia inducible factor (HIF)-1alpha and HIF-2alpha are essential in regulating transcription of the VEGF gene, active participation of HIF-alpha molecules in the pathogenesis of these arthritides is anticipated. We investigated the immunohistochemical expression of HIF-1alpha and HIF-2alpha in the synovium of 22 patients with RA, 34 patients with OA and 22 'normal' nonarthritic individuals, in relation to VEGF, VEGF/KDR (kinase insert domain protein receptor) vascular activation, PD-ECGF and bcl-2. A significant cytoplasmic and nuclear overexpression of HIF-1alpha and HIF-2alpha was noted in the synovial lining and stromal cells of both diseases relative to normal. Overexpression of HIF-alphas was related to high microvessel density, high PD-ECGF expression and high VEGF/KDR receptor activation, suggesting HIF-alpha-dependent synovial angiogenesis in OA. By contrast, the activation of the angiogenic VEGF/KDR pathway was persistently increased in RA, as indeed was microvessel density and the expression of PD-ECGF, irrespective of the extent of HIF-alpha expression, indicating a cytokine-dependent angiogenesis. In all cases, the VEGF/KDR vascular activation was significantly lower in OA than in RA, suggesting a relative failure of the HIF-alpha pathway to effectively produce a viable vasculature for OA, which is consistent with the degenerative nature of the disease. The activation of the HIF-alpha pathway occurs in both RA and OA, although for unrelated reasons.

Endothelial cell growth factor (ECGF) binds specifically in vitro to membrane receptors present on the surface of several cell types, including murine and human endothelial cells and fibroblasts. Monoclonal antibodies prepared against ECGF that inhibit the mitogenic activity of the growth factor prevent receptor occupancy by the ligand. Heparin interacts structurally with ECGF [Maciag, T., Mehlman, T., Friesel, R. & Schreiber, A. B. (1984) Science 225, 932-935], potentiates the mitogenic activity of the polypeptide, restores the biological activity to inactivate ECGF, enhances the affinity of the ligand to cell surface receptors, and modifies antibody recognition of ECGF. These data suggest that the association between heparin and ECGF induces a conformational change in the polypeptide that increases or stabilizes the biological activity of the mitogen.

BACKGROUND

Development of new blood vessels is essential for tumor growth and metastasis and depends on the production of angiogenic factors by tumor and/or infiltrating cells. We previously showed that vascular endothelial growth factor (VEGF) expression and vessel count correlate with metastasis in human colon cancer. Although most tumors with high vessel counts express high levels of VEGF, some do not. Recently, platelet-derived endothelial cell growth factor (PD-ECGF), another potent angiogenic factor, has been reported to be expressed in colon cancer.

OBJECTIVE

In this study, we examined the role of PD-ECGF in colon cancer angiogenesis and whether PD-ECGF is derived from the tumor or infiltrating cells.

METHODS

Immunostaining for PD-ECGF was performed on 96 colon cancer specimens, some of which were previously stained for VEGF and factor VIII, a marker that is specific for endothelial cells. Double staining was done by using antibodies to PD-ECGF and to CD68 (macrophage specific) or CD3 (lymphocyte specific) to confirm which infiltrating cells produce PD-ECGF. Northern blot analysis for PD-ECGF messenger RNA (mRNA) was performed on four colon cancer specimens and corresponding normal colon mucosae (same patients) and four human colon cancer cell lines (KM12SM, SW620, HT29, and NCI-H508) to determine whether colon cancer epithelium expresses PD-ECGF.

RESULTS

Immunohistochemical analysis demonstrated that PD-ECGF was expressed in infiltrating cells in most of the colon cancer specimens (80 [83%] of 96) but rarely in tumor epithelium (five [5%] of 96). Double staining demonstrated that infiltrating cells staining positive for both PD-ECGF and CD68 were more predominant than those staining positive for both PD-ECGF and CD3. The intensity of staining for PD-ECGF in infiltrating cells correlated with vessel counts (Spearman's rank correlation coefficient (R) = .29; P = .004), but did not correlate with the intensity of VEGF staining (R = .176, P = .086) or metastasis (Mann-Whitney U test, P = .253). PD-ECGF staining intensity was higher in specimens with a high vessel count >> 50 at high magnification) and low VEGF-staining intensity (< or = 2+) than in specimens with a high vessel count (again,>> 50) and high VEGF-staining intensity (3+). Northern blot analysis revealed that colon cancer specimens and normal mucosae expressed relatively high levels of PD-ECGF mRNA, whereas PD-ECGF mRNA transcripts were not detectable in colon cancer cell lines.

CONCLUSIONS

PD-ECGF expression in human colon cancer specimens is associated with vessel count and may be responsible for tumor vascularity in those tumors with low VEGF expression. Infiltrating cells expressing PD-ECGF may contribute to angiogenesis, thus providing an additional mechanism for tumor neovascularization.

The response of human endothelial cell migration to various extracellular matrix components and growth factors has been assessed. Human endothelial cells demonstrate increased chemotaxis and chemokinesis when placed in a modified Boyden chamber with endothelial cell growth factor (ECGF) used at a concentration of 10(-9) M. Anti-ECGF antibody inhibits the chemotactic response. Heparin (10(-8) to 10(-10) M) was also chemotactic and was shown to potentiate the chemotactic activity of ECGF. Although laminin, fibronectin, the polypeptide (epidermal, fibroblast, and nerve) growth factors, and collagen types I, II, III, IV, and V demonstrate a chemotactic response, these activities were one third to one half less than observed with ECGF. These data suggest that ECGF and heparin may play a significant role as response modifiers of human endothelial cell migration which may be relevant to tumor metastasis, wound healing, and atherogenesis.

The purpose of this study was to investigate the role of interactions between tumor cells and macrophages during angiogenesis in human gastric carcinomas. Macrophage infiltration into tumors and monocyte chemoattractant protein-1 (MCP-1) expression was assessed in 72 archival specimens of gastric carcinoma for comparison with tumor vascularity. The mRNA expression of MCP-1 was examined by RT-PCR in 6 gastric carcinoma cell lines and in fresh biopsy specimens from 18 patients. Immunolocalization of representative angiogenic factors, vascular endothelial growth factor (VEGF), and platelet-derived endothelial cell growth factor (PD-ECGF) was also done. MCP-1 expression in tumor cells increased with the depth of tumor invasion (Tis 9.5%, T1 19.4%, T2-4 60.0%), as did microvessel density and macrophage infiltration. Macrophage counts correlated with vessel counts, and both were significantly higher in MCP-1-positive than in negative tumors. Of the 6 gastric carcinoma cell lines, 2 constitutively expressed MCP-1 mRNA. In 6 (33.3%) of 18 biopsy samples, MCP-1 mRNA was expressed at higher levels in tumor tissues than in normal mucosa. VEGF protein was expressed by gastric carcinoma cells, whereas PD-ECGF protein was expressed mainly by stromal mononuclear cells. MCP-1 expression correlated significantly with VEGF but not PD-ECGF expression in gastric carcinomas. These results suggest that MCP-1 produced by human gastric carcinoma cells plays a role in angiogenesis via macrophage recruitment and activation.

Thymidine phosphorylase (dThdPase) is identical to platelet-derived endothelial cell growth factor (PD-ECGF) and has angiogenic activity. Since dThdPase seems to have an important role in angiogenesis of tumours, we measured the activity and expression of dThdPase in various tumours and the adjacent non-neoplastic tissues. We assayed dThdPase activity by spectrophotometric means, and the expression of dThdPase was examined by immunoblotting and by immunohistochemical staining using a monoclonal antibody against dThdPase. In the oesophagus, stomach, colorectum, pancreas, and lung, dThdPase activity in carcinomas was significantly higher (P < 0.05) than that in the adjacent non-neoplastic tissues. The expression level of dThdPase detected by immunoblotting correlated well with the activity of dThdPase. In the oesophagus, stomach, colorectum, gall bladder, pancreas and lung, the proportion of dThdPase-positive tumours was significantly higher (P < 0.05 or 0.01) than that of the dThdPase-positive adjacent normal tissues. In oesophageal, gastric colorectal and lung carcinomas, the proportion of dThdPase positivity in advanced carcinomas was significantly higher (P < 0.01) than that in early carcinomas. Tumour-infiltrative macrophages or lymphocytes in the lymph node, alveolar macrophages and Kupffer cells expressed high levels of dThdPase. The results indicate that dThdPase activity and expression level in many tumours are higher than those in the adjacent non-neoplastic tissues, and that dThdPase may have an important role in the proliferation of these solid tumours.

Hypertension causes biochemical and morphological changes in the vessel wall by unknown mechanisms. Locally produced substances may have a role in mediating these vascular changes. We have studied the expression of platelet-derived growth factor (PDGF) B chain and PDGF A chain, insulin-like growth factor (IGF)-I and IGF-II, endothelial cell growth factor (ECGF), basic fibroblast growth factor (bFGF), and transforming growth factor-beta (TGF-beta) in aortic tissue from normotensive rats and rats made hypertensive by deoxycorticosterone (DOC)/salt treatment. Using Northern blotting, we found that genes for each of these growth factors were transcriptionally active in the aorta of both normotensive and hypertensive rats. TGF-beta aortic mRNA levels increased up to threefold as a result of DOC/salt hypertension. In contrast, no major changes in the expression of either PDGF chain, IGF-I or II, ECGF, or bFGF were detectable. The results indicate that at least seven genes coding for growth factors that were shown previously to influence growth and function of vascular cells in vitro, are expressed in rat aorta in vivo. These findings support the hypothesis that synthesis and release of growth factors in the arterial wall are involved in autocrine and/or paracrine regulatory mechanisms. In addition, the increased expression of TGF-beta in vivo may have a role in mediating the aortic changes induced by hypertension.

Cytokines are a heterogenous group of polypeptide mediators that have been associated with activation of numerous functions, including the immune system and inflammatory responses. The cytokine families include, but are not limited to, interleukins (IL-I alpha, IL-I beta, ILIra and IL-2-IL-15), chemokines (IL-8/ NAP-I, NAP-2, MIP-I alpha and beta, MCAF/MCP-1, MGSA and RANTES), tumor necrosis factors (TNF-alpha and TNF-beta), interferons (INF-alpha, beta and gamma), colony stimulating factors (G-CSF, M-CSF, GM-CSF, IL-3 and some of the other ILs), growth factors (EGF, FGF, PDGF, TGF alpha, TGF beta and ECGF), neuropoietins (LIF, CNTF, OM and IL-6), and neurotrophins (BDNF, NGF, NT-3-NT-6 and GDNF). The neurotrophins represent a family of survival and differentiation factors that exert profound effects in the central and peripheral nervous system (PNS). The neurotrophins are currently under investigation as therapeutic agents for the treatment of neurodegenerative disorders and nerve injury either individually or in combination with other trophic factors such as ciliary neurotrophic factor (CNTF) or fibroblast growth factor (FGF). Responsiveness of neurons to a given neurotrophin is governed by the expression of two classes of cell surface receptor. For nerve growth factor (NGF), these are p75NTR (p75) and p140trk (referred to as trk or trkA), which binds both BDNF and neurotrophin (NT)-4/5, and trkC receptor, which binds only NT-3. After binding ligand, the neurotrophin-receptor complex is internalized and retrogradely transported in the axon to the soma. Both receptors undergo ligand-induced dimerization, which activates multiple signal transduction pathways. These include the ras-dependent pathway utilized by trk to mediate neurotrophin effects such as survival and differentiation. Indeed, cellular diversity in the nervous system evolves from the concerted processes of cell proliferation, differentiation, migration, survival, and synapse formation. Neural adhesion and extracellular matrix molecules have been shown to play crucial roles in axonal migration, guidance, and growth cone targeting. Proinflammatory cytokines, released by activated macrophages and monocytes during infection, can act on neural targets that control thermogenesis, behavior, and mood. In addition to induction of fever, cytokines induce other biological functions associated with the acute phase response, including hypophagia and sleep. Cytokine production has been detected within the central nervous system as a result of brain injury, following stab wound to the brain, during viral and bacterial infections (AIDS and meningitis), and in neurodegenerative processes (multiple sclerosis and Alzheimer's disease). Novel cytokine therapies, such as anticytokine antibodies or specific receptor antagonists acting on the cytokine network may provide an optimistic feature for treatment of multiple sclerosis and other diseases in which cytokines have been implicated.

Human thymidine phosphorylase (dThdPase) has been reported to be identical with the platelet-derived endothelial cell growth factor (PD-ECGF). To investigate whether the dThdPase activity of PD-ECGF/dThdPase is indispensable to its angiogenic activity, three PD-ECGF/dThdPase mutants, K115E (Lys-115->>Glu), L148R (Leu-148->>Arg) and R202S (Arg-202->>Ser) were made by site-directed mutagenesis. Although the expression level of the three mutant PD-ECGF/dThdPases in the COS-7 cells was similar to that of wild-type PD-ECGF/dThdPase, dThdPase activity was not detected in the COS-7 cells transfected with the mutant PD-ECGF/dThdPase cDNA. The lysates of COS-7 cells transfected with the wild-type PD-ECGF/dThdPase cDNA had angiogenic activity, but those transfected with the mutant PD-ECGF/dThdPase cDNAs did not. An inhibitor of dThdPase, 6-amino-5-chlorouracil, inhibited the angiogenic activity of purified dThdPase. These findings indicate that dThdPase activity is indispensable to the angiogenic activity of PD-ECGF/dThdPase.