介白質－6與第二型環氧酵素增益胃癌之血管內皮生長因子及腫瘤血管新生

血管新生（angiogenesis）是許多實體腫瘤甚至血液科腫瘤發生、生長、侵犯、及轉移所不可或缺的致病機轉。近年來，慢性發炎與其相關的因子在癌症致病機轉中所扮演的角色逐漸被了解，然而發炎反應相關因子在腫瘤血管新生中的作用目前所知仍屬有限。本研究有兩部分，分別探討與發炎有關的介白質－6（interleukin-6, IL-6）及第二型環氧酵素（cyclooxygenase-2, COX-2）在胃癌中與血管新生及血管內皮生長因子（vascular endothelial growth factor, VEGF）的相關性及作用機制。第一部分、介白質-6在胃癌腫瘤血管新生的角色介白質－6是一個發炎前驅性細胞激素（pro-inflammatory cytokine）。過去研究報告顯示，介白質－6與幽門螺旋桿菌（Helicobacter pylori, H. pylori）相關的胃炎及胃癌（gastric carcinoma, GC）疾病病程有關，但其對於胃癌之致癌機轉仍未完全了解。血管內皮生長因子則是一個過去研究已知的胃癌重要血管新生因子。本研究目的在釐清介白質-6是否可以調節胃癌血管內皮生長因子及血管新生機制。臨床研究部分蒐集54個取自臺大醫院胃癌手術切除標本，進行介白質－6、血管內皮生長因子、及腫瘤血管的組織免疫染色檢查。結果顯示介白質－6免疫活性與血管內皮生長因子免疫活性及腫瘤細微血管密度（microvessel density, MVD）有顯著正相關。基礎研究部分，本研究先篩檢四株胃癌細胞株對介白質－6刺激的反應，以酵素連結免疫吸收測定法（enzyme-linked immunosorbent assay, ELISA）測定其血管內皮生長因子表現量，發現其中兩株細胞株可以隨介白質－6刺激的劑量及時間增加而增加血管內皮生長因子分泌量。接下來，本研究利用火蠅發光酵素報導基因分析法（luciferase reporter gene assay）探討介白質－6的三種細胞訊息傳遞路徑中，何者與血管內皮生長因子的調控有關。結果顯示JAK/STAT路徑（JAK/STAT pathway）可以媒介介白質－6刺激血管內皮生長因子的效應。確認介白質－6可以經由特定訊息傳遞路徑刺激胃癌細胞株產生血管內皮生長因子後，本研究利用人類臍靜脈內皮細胞（human umbilical vein endothelial cell, HUVEC）進一步探討介白質－6是否可以有效增加體外血管新生（angiogenesis in vitro）。結果顯示介白質－6的確可以增進人類臍靜脈內皮細胞的生長及管狀網絡形成。另外，以Matrigel栓子試驗（Matrigel plug assay）進行小鼠的體內血管新生（angiogenesis in vivo）分析，亦顯示介白質－6也可以增進體內血管新生。上述的介白質－6體外及體內血管新生效應，經由中和性抗血管內皮生長因子抗體拮抗血管內皮生長因子後，均可以有效抑制介白質－6引起的血管新生效應。此結果顯示介白質－6對血管新生增進的效應，主要來自於其誘導之血管內皮生長因子分泌的作用。第二部分、第二型環氧酵素在胃癌腫瘤血管新生的角色第二型環氧酵素（cyclooxygenase-2, COX-2）是在發炎中一個可被誘導的酵素，同時也被發現在多種惡性腫瘤中過量表現。過去研究發現，第二型環氧酵素能促進癌症腫瘤逃脫細胞凋亡（apoptosis）機制、抑制宿主免疫攻擊、和增益癌細胞侵襲力。但第二型環氧酵素與腫瘤血管新生的研究仍屬有限。本研究臨床部分探討第二型環氧酵素與胃癌血管新生的相關性是否存在。54個取自臺大醫院胃癌手術切除標本的組織免疫染色，顯示了胃癌的第二型環氧酵素、血管內皮生長因子、和腫瘤血管密度的顯著正相關性。同時，也發現此種正相關性主要存在於Lauren氏瀰漫型胃癌、非賁門胃癌、以及幽門螺旋桿菌感染陽性胃癌。此結果一方面證實了第二型環氧酵素在胃癌血管新生扮演角色，另一方面也指出了特定胃癌亞型有此相關性。基礎研究部分，目的建立解釋上述臨床現象的實驗模式並探索其機制。在轉殖表現第二型環氧酵素的載體後，胃癌細胞株AGS可以增加第二型環氧酵素及前列腺素E2（PGE2）表現，並增益了人類臍靜脈血管內皮細胞的增殖和管狀網絡形成。第二型環氧酵素在人類臍靜脈內皮細胞的體外生成血管效應，可經由阻斷血管內皮生長因子而被抑制，也可被第二型環氧酵素抑制劑NS-398抑制。為了闡明第二型環氧酵素增加血管新生的機制，我們建立了一個穩定表現第二型環氧酵素的胃癌細胞株AGS/COX-2和它的載體控制組細胞株AGS/pcDNA3（不表達第二型環氧酵素），轉殖效果經測定第二型環氧酵素及前列腺素E2表現量加以確認。在 6個血管新生相關的因子中，血管內皮生長因子在 AGS/COX-2顯著地表現。本研究接著轉殖反義（anti-sense）缺氧可誘導因子－1α（hypoxia-inducible factor-1α, HIF-1α）以減少細胞的缺氧可誘導因子－1α蛋白，此使血管內皮生長因子在AGS/COX-2的生成顯著減少。我們發現原型AGS細胞以前列腺素E2刺激後，缺氧可誘導因子－1α及血管內皮生長因子同時增加，但此效應可以被前列腺素E2 EP1受體拮抗劑 SC19220阻抑。此外，用NS-398先處理以減少前列腺素E2生合成後，可以抑制缺氧可誘導因子－1α的累積，同時也如同轉殖反義缺氧可誘導因子－1α般地抑制了血管內皮生長因子的製造。這部分研究成果支持「第二型環氧酵素/前列腺素E2/缺氧可誘導因子－1α/血管內皮生長因子路徑」為胃癌的腫瘤血管新生的機制之一。結論本研究呈現了臨床上介白質－6及第二型環氧酵素與胃癌血管新生相關的證據，並分別指出該相關存在於那些胃癌亞型。同時基礎實驗也分別闡明了關於介白質－6及第二型環氧酵素在胃癌血管新生中作用的機制。簡言之，即介白質－6循「JAK/STAT路徑」增加血管內皮生長因子生合成來增益胃癌血管新生；第二型環氧酵素循「前列腺素E2/缺氧可誘導因子－1α路徑」增加血管內皮生長因子生合成來增益胃癌血管新生。這些成果增進了慢性發炎及其相關因子對於胃癌血管新生及相關因子的了解，亦提供了未來以抗血管新生療法治療胃癌的可能方向及理論基礎。

關鍵字

介白質－6 ；第二型環氧酵素；血管內皮生長因子；胃癌；血管新生

並列摘要

INTRODUCTION Gastric carcinoma (GC) remains one of the leading causes of cancer-related death around the world[1]. Although considerable studies have addressed the roles of genetic and environmental factors in the pathogenesis of GC, evidences are growing that angiogenesis is crucial in determining the initiation[143], progression, and metastasis of cancer, as well as overall survival rates[9, 14]. Tumor angiogenesis, defined as the process in which tumor neovasculature develops from a pre-existing vascular network[144], involves several key steps: degradation of the extracellular matrix, proliferation of vascular cells, chemotaxic migration of cells, formation of vascular lumen, and change in endothelial permeability[143]. This process is controlled by balancing various angiogenic and anti-angiogenic factors[143]. However, the molecular mechanisms involved in developing tumor neovasculature may vary greatly, depending on the type of cancer[9]. Among the vast angiogenic factors, vascular endothelial growth factor (VEGF) is one of the most potent mitogens specific to endothelial cells and is central to the angiogenic process[145, 146]. VEGF can promote the proliferation and migration of endothelial cells, the migration of vascular smooth muscle cells, the formation of vascular tubes, and the permeability of blood vessels[29, 30]. It is over-expressed in a broad diversity of solid and hematological malignancies, including cancers of the colorectum[108], breast[109], liver[110], lung[111], oropharynx[112], uterine cervix[113], ovary[114], urinary bladder[115], and multiple myeloma[54]. VEGF was also shown to increase tumor vasculature, invasion, metastasis, and hematogenous recurrence in GC, and was reported as an indicator of prognosis and a predictor of the mode of recurrence of GC[45-47, 147]. During recent decades, the mechanism of tumor angiogenesis and the regulation of angiogenesis-related factors have attracted much interest. Although VEGF is well known to participate importantly in angiogenic switching during carcinogenesis[143], few types of extracellular molecules have been found to be able to modulate tumor VEGF expression, including transforming growth factor-β in folliculostellate pituitary cells[117], insulin-like growth factor 1 in colorectal cancer cells[118], and epidermal growth factor in malignant glioma cells[119], Hypoxia is well known to be the most common stimulus of VEGF induction via hypoxia-inducible factor1-α[116, 148]. The extracellular factors other than tissue hypoxia that regulate VEGF expression during gastric carcinogenesis, however, remain unknown. Evidence is rapidly growing that chronic inflammation may contribute to carcinogenesis through up-regulating growth, angiogenesis, and metastasis in a number of neoplasms, including gastric carcinoma (GC)[7]. Interleukin-6 (IL-6) and cyclooxygenase-II (COX-2) are both factors involved in the gastritis and are shown to be associated with GC status. The present study comprises two parts. The first part investigated the relationship among interleukin-6 (IL-6), VEGF, and angiogenesis in GC. The second part investigated the relationship among cyclooxygenase-II (COX-2), VEGF, and angiogenesis in GC. Both parts studied the clinical association and then conducted the experiments to elucidate the causal mechanism. Part I-IL-6, VEGF, and angiogenesis in GC IL-6 has been demonstrated to be associated with the disease status and outcome of GC[49]. However, the cause of this association remains unknown. IL-6 is a multi-functional cytokine secreted by a wide variety of cells to regulate immune responses, hematopoiesis, acute phase reaction, and fever response[48]. Upon binding of IL-6 to the binding molecule, IL-6Rα, a receptor complex is formed by recruiting the signal-transducing subunit gp130[88], which in turn mediates signals into the nucleus via the Janus kinase (JAK)/signal transducers and activators of transcription (STAT) pathway[48], the Ras/mitogen-activated protein kinase (MAPK) pathway[89], and the phosphatidylinositol 3-kinase (PI3K)/Akt pathway[90]. Various mechanistic strategies for cancer progress are well documented, and involve self-sufficiency of growth signals, evasion of apoptosis, invasion of tissue, and sustained angiogenesis[6]. IL-6 has been shown to exhibit an anti-apoptotic effect against transforming growth factor-β in human hepatoma cell line Hep3B through concomitant activation of the PI3K/Akt and the JAK/STAT pathways[90]. With respect to tumor angiogenesis, studies of the interaction between IL-6 and VEGF in malignancies have been reported controversial results. Dankbar et al. studied the involvement of VEGF and IL-6 in paracrine tumor-stromal cell interactions in multiple myeloma, finding that IL-6 induced VEGF expression in myeloma cells and then VEGF reciprocally stimulated IL-6 secretion from stromal and microvascular endothelial cells[54]. Wei et al. reported that concordant expression of IL-6 and VEGF occurred in both cancerous and non-cancerous tissues in early-stage cervical cancers and showed that IL-6 induced VEGF production in cervical cancer cell C33A[55]. In contrast, Hatzi et al. found that IL-6 inhibited VEGF-induced rabbit cornea angiogenesis and demonstrated that IL-6-expressing neuroblastoma xenograft tumors in mice exhibited reduced neovascularization and suppressed growth[56]. These reports imply that the effect of IL-6 on tumor VEGF expression and angiogenesis varies greatly, probably being influenced by the type of cancer. In gastrointestinal malignancies, the relationship between IL-6 and VEGF as well as the role of IL-6 on angiogenesis remains unknown until now. The authors’ earlier research has found an association between serum IL-6 and VEGF levels in GC patients[149]. Accordingly, we hypothesized that IL-6 might interact with VEGF and angiogenesis in GC. The results of this work provide clinical evidence that IL-6 activity is highly correlated with VEGF expression and tumor vasculature in GC, as determined by immunohistochemistry, and show that IL-6 increases angiogenesis in vitro and in vivo by activating VEGF gene via the JAK/STAT pathway in GC cells. STUDY FINDINGS IL-6 was positively associated with VEGF expression and tumor vasculature in GC A total of 54 patients (male : female = 28 : 26; age 59.9 ± 13.5 years) were enrolled in this study. Table 1 summarized demographic data, IL-6 immunoreactivity, VEGF immunoreactivity, and tumor vasculature expressed as MVD on GC tissues. Forty-two patients had GC with high IL-6 expression and 12 had GC with low IL-6 expression. These two study groups had similar age and gender distribution (p = 0.539 and 0.332, respectively). Generally, the high IL-6 group had higher VEGF expression (90 % vs. 50 %, p = 0.005) and tumor vascular density (42.5 ± 32.0 vs. 21.5 ± 23.3, p = 0.040) than the low IL-6 group. Both IL-6 and VEGF immunoreactivities were found mainly localized in cancer cells, rather than stromal cells. The relationship between IL-6 and VEGF immunoreactivities in each subgroup of GC was further investigated according to various pathologic characteristics (Table 2). High IL-6 expression was associated with increased VEGF expression in the subgroups of advanced (92% vs. 60%, p = 0.027), intestinal type (96% vs. 50%, p = 0.009), non-cardiac (89% vs. 45%, p = 0.006), and H. pylori-positive GCs (85% vs. 20%, p = 0.009). All cardiac GCs in this series had high VEGF expression in both high and low IL-6 groups. However, the subgroups of early, diffuse type, and H. pylori-negative GCs did not exhibit an association between the expression of IL-6 and that of VEGF. Collectively, the results that high IL-6 expression was associated with high VEGF and tumor angiogenesis in GC supported the hypothesis that IL-6 might play a role for VEGF production and angiogenesis in GC. IL-6 induces VEGF production in various GC cell lines. Four serum-starved GC cell lines were stimulated with IL-6 to evaluate the effect of IL-6 on VEGF induction in GC cells. A dose-dependent increase in VEGF production was found in AGS and NCI-N87 cell lines but not in KATO-III and SNU-1 cell lines (Fig. 2A). VEGF concentrations considerably increased from a dose of 50 ng/ml IL-6 in AGS cell (p = 0.032) and 100 ng/ml IL-6 in NCI-N87 cells (p = 0.047). Moreover, AGS cell also exhibited a time-dependent increase in VEGF production after IL-6 stimulation at 24th h (Fig. 2B). The data provided the in vitro evidence that IL-6 promotes VEGF expression in GC cell lines. JAK/STAT pathway mediates signaling of IL-6 on VEGF induction in GC cell. The luciferase reporter gene assay was performed using three inhibitors in the signaling pathways of IL-6, JAK/STAT, Ras/MAPK and PI3K/Akt pathways, to elucidate further the pathway that mediates the signaling of IL-6 on VEGF induction in GC (Fig. 3). After IL-6 stimulation (positive control), the activation of the VEGF promoter represented by luciferase activity increased 2.4 times that associated with AGS cells without IL-6 stimulation (negative control). Luciferase activity in AGS cells pretreated with AG490 was reduced to near the level of the negative control (p = 0.292). However, the two groups pretreated with PD98059 and LY294002 still exhibited enhanced luciferase activity, approximating the level of the positive control (p = 0.511 and 0.411). The results showed that signaling through JAK/STAT pathway is the primary mechanism that mediates IL-6 on VEGF induction in GC. IL-6 stimulated the proliferation and tube formation of HUVEC primarily via VEGF in GC. The effect of IL-6 on angiogenesis in vitro was examined by assessing the proliferation (Fig. 4) and tube formation (Fig. 5A-D) of the HUVECs cultured with CM. Neutralizing anti-VEGF antibody effectively blocked the enhancement of proliferation of HUVECs by IL-6-stimulated AGS cells, which was reduced to a level slightly lower than that of HUVECs cultured with CM from AGS cells without IL-6 treatment, whereas non-specific IgG only minimally reduced HUVEC proliferation. The results further supported that IL-6 can promote HUVEC proliferation by inducing VEGF. The HUVEC tube formation assay was then performed to verify the effect of IL-6 on promoting HUVECs to form a network-like tubular structure, which process involves the integration of migration, invasion, and differentiation, but not proliferation of HUVECs[74]. This assay yielded results agreed with those of the proliferation assay. The formation of HUVEC tubes became significant after culturing with CM from IL-6-stimulated AGS cells (Fig. 5A,B). However, neutralizing anti-VEGF antibody completely blocked HUVEC tube formation (Fig. 5C). In contrast, non-specific IgG did not at all inhibit tube formation (Fig. 5D). Taken together, in vitro experiments on HUVEC demonstrated that IL-6 considerably increases the ability of AGS cells to promote the functions of the endothelial cells, mainly by inducing VEGF production. IL-6 increased in vivo angiogenesis primarily via VEGF in GC. The effect of IL-6 on the induction of angiogenesis in vivo was assessed using Matrigel plug assay. Matrigel plugs were harvested on the 7th day. They were examined by gross inspection and by measuring the density of hemoglobin, as an indicator of vascularization. Grossly, the Matrigel plugs that embedded CM from IL-6-stimulated AGS cells (positive control) developed substantial vasculature, even with hemorrhage in plugs, as compared to negative control (Fig. 6A,B). However, the plugs that contained CM neutralized by anti-VEGF antibody exhibited considerably less vascularization (Fig. 6D). The plugs blocked by non-specific IgG (Fig. 6C) developed similar vasculature and marked intra-plug hemorrhage as positive control. Hemoglobin concentrations were determined and normalized to the weight of the plugs to represent the plug vascular densities and thus further the effect of IL-6 (Fig. 6E). Plugs that embedded CM neutralized by anti-VEGF antibody had hemoglobin densities as low as those of the negative control, whereas both plugs that embedded CM from IL-6-stimulated AGS cells (positive control) and those blockaded by non-specific IgG had substantially higher hemoglobin densities. Collectively, the Matrigel plug assay demonstrated that IL-6 markedly potentiates the AGS cells to enhance angiogenesis in vivo, primarily by inducing VEGF production, consistent with the results of the experiments on HUVECs. DISCUSSION Since its discovery in 1989[29], VEGF has been ranked among the most potent angiogenic factors during tumorigenesis in a wide variety of malignancies[54, 108-115]. Although great advances have been made in the understanding of VEGF function during the last decade, molecules reported to be able to up-regulate VEGF expression in cancers are still very few, including hypoxia-inducible factor 1-α[116] and extracellular factors such as transforming growth factor-β[117], insulin-like growth factor 1[118], and epidermal growth factor[119]. This work presents the clinical findings that VEGF expression and tumor vasculature increased with the multi-functional cytokine IL-6 activity in GC, and provides experimental data to show that IL-6 induces VEGF production via the JAK/STAT signaling pathway to promote angiogenesis in GC cells. Originally, a relationship was hypothesized to exist between IL-6 and VEGF in GC, mainly following two reports: one describing an association between serum IL-6 levels and the disease status and the outcome of GC[49]; the other, by the authors, showing a correlation between serum levels of IL-6 and VEGF in patients with GC[149]. However, the serum levels of IL-6 and VEGF may not fully reflect their protein activities in tumor tissues. Consequently, the first task of this study was to examine the expression patterns of IL-6 and VEGF proteins in GC tissues. Immunohistochemical analysis revealed concomitant increments of IL-6 and VEGF immunoreactivities in GC, especially in the subgroups of advanced, intestinal type, non-cardiac, and H. pylori-positive GCs. The results indicated that the positive correlation between IL-6 and VEGF expressions is not uncommon in GC. Different subgroups of GC may have different molecular mechanisms of carcinogenesis[87]. From the viewpoint of angiogenesis, the data imply that GCs with different pathologic characteristics might also follow differential mechanisms to control VEGF expression. Further studies must be undertaken to clarify the mechanism that governs the correlation in some pathologic subgroups but not others. Additionally, both IL-6 and VEGF immunoreactivities were found mainly localized in cancer cells. This indicated that an autocrine and/or paracrine interaction might exist between IL-6 and VEGF in GC cells, further supporting the claim that IL-6 might promote VEGF expression to increase tumor angiogenesis. Accordingly, the correlation between the extent of IL-6 expression and tumor angiogenesis was further examined by measuring the densities of anti-CD34-stained microvessels. In general, high IL-6 tumors developed more significant vasculature than low IL-6 tumors in the series studied here. Collectively, the immunohistochemical studies revealed trend of higher IL-6 activity toward higher VEGF expression and tumor angiogenesis in GC and further prompted us to investigate whether a causal mechanism exists, that is, whether IL-6 can modulate VEGF expression and hence increases tumor angiogenesis in GC. GC cell lines were screened for responses to VEGF induction after IL-6 stimulation and two out of four cell lines, AGS and NCI-N87, were found to produce VEGF in a significant dose-dependent response to IL-6. In the time-dependent experiment, AGS also exhibited a considerable response to IL-6 stimulation at 24 h. The findings confirmed that GC cells can respond to IL-6 stimulation and VEGF is among the target genes of IL-6. JAK/STAT, Ras/MAPK, and PI3K/Akt pathways mediate signals of IL-6 after gp130 dimerization and activation triggered by the binding of IL-6 to IL-6Rα[88]. All the three pathways have been reported to mediate VEGF production in response to various factors in a range of different cells. In cardiac myocytes, STAT is required by the gp130-mediated induction of VEGF[91]. In the glioblastoma cell line U87 MG, MAPK mediates acidic extracellular pH-induced VEGF expression[92]. In the prostate cancer cell line DU145, PI3K/Akt mediates vanadate-induced VEGF expression[93]. Moreover, in colonic cancer cells, insulin-like growth factor 1 induces hypoxia-inducible factor 1-mediated VEGF expression via both MAPK and PI3K/Akt pathways[94]. The luciferase reporter gene assay for the activation of VEGF promoter was employed to assess the effects of the inhibitors in the three IL-6 signaling pathways and thus clarify the signaling pathway responsible for IL-6-induced production of VEGF in GC. Only AG490, the specific inhibitor of JAK2, could effectively eliminate the activity of the VEGF promoter induced by IL-6, whereas neither PD98059 nor LY294002, the inhibitors in Ras/MAPK and PI3K/Akt pathways, had such an inhibitory effect. These data imply that JAK/STAT pathway is primarily involved in the signaling of IL-6-related activation of the VEGF gene in GC cells. Whether VEGF is the most important factor in IL-6-induced angiogenesis in vitro and in vivo was addressed here, since other genes activated by IL-6 may also affect GC angiogenesis. The assays of proliferation and tube formation of endothelial cells are currently accepted as in vitro indicators to assess angiogenic factors[95, 96]. This study used MTT assay[97] to measure the number of viable cells, and compared the results of those of the negative control to determine the relative proliferation of HUVECs. AGS cells were found to promote significantly HUVEC growth after IL-6 stimulation. Additionally, this effect could be completely eliminated by neutralizing anti-VEGF antibody but it was not affected substantially by non-specific IgG. During angiogenesis, proliferating endothelial cells migrate to form a net-like structure, an important characteristic in evaluating the angiogenic factor. In this study, the HUVEC tube formation assay yielded a result similar to that of the MTT assay. IL-6 significantly potentiated AGS cells to induce the tube formation of HUVECs, whereas anti-VEGF antibody effectively eliminated this effect. Accordingly, this study established that IL-6 increases endothelial growth and migration primarily through VEGF in GC. Other factors not tested in this study might also be induced by IL-6, but they do not substantially contribute to the effect of IL-6 on HUVECs. The success of demonstrating the effect of IL-6 on HUVEC functions in vitro enables further testing the postulate using the Matrigel plug assay to assess an in vivo angiogenesis[75]. Hemoglobin densities were consistent with gross observation, showing that angiogenesis is notably enhanced by IL-6 but effectively blocked by anti-VEGF antibody. The higher hemoglobin densities in the plugs blocked by non-specific IgG than in the positive control might be due to intra-plug hemorrhage. Systemic chemotherapy for advanced GC is often associated with significant treatment-related toxicity, especially in patients in poor general condition[131]. On the other hand, cancer cells may acquire resistance to cytotoxic chemotherapeutic agents through their genetic instability, rapidly accumulating mutations in drug’s target genes. Although tumor neovasculature is essential in supporting tumor development, its endothelial cells are genetically stable[138]. Therefore, anti-angiogenic therapy through targeting genetically stable endothelial cells has been developed as a novel therapeutic approach alternative to conventional cytotoxic chemotherapy, either alone or in combination, given its effectiveness and low toxicity[138]. A better appreciation of the mechanisms by which the angiogenic factors are modulated is crucial to developing anti-cancer agents that can more effectively antagonize tumor angiogenesis. VEGF is very important under a wide range of both physiologic and pathologic conditions[139]. It is now one of the most extensively studied targets in anti-angiogenesis therapy[138]. However, the wide distribution of VEGF in both cancerous and non-cancerous tissues in a single patient might conversely limit the role of VEGF as a therapeutic target, given the possibility of unexpected effects in non-cancerous tissues. The inhibition of tissue-specific factors capable of modulating local VEGF expression might be an alternative strategy. This study establishes that IL-6 up-regulates VEGF expression and angiogenesis in GC, implying that IL-6 might be among the potential therapeutic targets in the treatment of GC by blocking VEGF-induced angiogenesis. Although VEGF is released automatically or induced in response to hypoxia by cancer cells, other factors such as IL-6 might be also important in regulating VEGF. Clinical anti-angiogenic strategies for treating malignancies have been proposed to be perhaps most effective when multiple, rather than single, anti-angiogenic agents are used[140]. Therefore, this work may encourage further investigation into potent angiogenesis-modulating agents to improve the effectiveness of anti-angiogenic therapy. In conclusion, support for the hypothesis that IL-6 is among the few molecular factors currently known to be able to modulate VEGF expression in tumorigenesis comes from both the clinical and the experimental evidences presented here. To the authors’ knowledge, IL-6 is the first extracellular molecule reported to be capable of up-regulating VEGF expression in GC. This study might hint at a new strategy for studying organ-specific factors that control tumor angiogenesis in gastrointestinal malignancies. Part II-COX-2, VEGF, and angiogenesis in GC Clinical studies in patients with chronic inflammatory diseases demonstrated an effect on reducing colorectal cancer risk and mortality after long-term use of aspirin or non-steroidal anti-inflammatory drugs (NSAIDs)[57]. Cyclooxygenase-2 (COX-2), an inducible enzyme pivotal in the inflammatory response, converts arachidonic acid to the prostaglandins required in initiating and maintaining reactions during the inflammatory process[58]. COX-2 is blocked by NSAIDs and now is the most studied therapeutic target of NSAIDs[58]. Angiogenesis, the process by which new vascular network develops from the pre-existing vessels, plays a crucial role in tumor initiation, progression, invasion and metastasis and is generally accepted as an indicator of prognosis[9, 14]. Angiogenic process is finely controlled by balancing a vast of pro- and anti-angiogenic factors. Among them, vascular endothelial growth factor (VEGF) is one of the most potent tumor angiogenic factors, capable of promoting proliferation and migration of endothelial cells and increasing vascular permeability[150]. The main stimulus for VEGF expression comes from hypoxia, which activates the VEGF gene through increasing hypoxia-inducible factor-1 (HIF-1)[63]. HIF-1 is a heterodimeric transcriptional factor consisting of a regulated subunit HIF-1a and a constitutive subunit HIF-1β[64] and is important in activating multiple genes involved in angiogenesis, cell survival, and invasion[65]. HIF-1a is primarily regulated by oxygen concentration change[64] via degradation by the ubiquitin-proteasome pathway[66]. Interestingly, recent studies discovered that HIF-1a is also regulated by oxygen-independent mechanisms, mainly activation of oncogenes and mutation of tumor suppressor genes[65]. Earlier clinical studies have confirmed an association between COX-2 over-expression and GC occurrence[60]. An animal model of GC also showed that celecoxib, a selective COX-2 inhibitor, is effective for chemoprevention of tumor development[61]. The known mechanisms by which COX-2 promotes carcinogenesis include evasion from apoptosis, suppression of immunity, and promotion of invasiveness[59]. Recently, COX-2 was also found to play a role in tumor angiogenesis[62]. However, the mechanism of COX-2 for promoting angiogenesis in GC remains obscure, especially its effects on various angiogenic factors. Therefore, we conducted the present study to investigate the clinical relationship among COX-2, VEGF and tumor vasculature on surgical specimens of GC, to clarify the effect of COX-2 on growth and function of vascular endothelial cells, and to assess the role of COX-2 in the regulation of angiogenic factors in GC cells. STUDY FINDINGS COX-2 Positively Correlated with VEGF and Tumor Vasculature in GC This study enrolled a total of 54 GC patients (28 males and 26 females, aged 59.9 ± 13.5 yr). Forty-two patients had a high COX-2 tumor and 12 had a low COX-2 tumor. Both ages and gender distributions were similar between these two groups (table 1). In general, high COX-2 tumors developed significant vasculature, compared with low COX-2 tumors (43.6 ± 32.2 vs. 17.8 ± 17.1 per HPF, p = 0.001). The trend of VEGF expression was the same with tumor angiogenesis, i.e. high COX-2 tumors exhibited significantly high VEGF activity (59%), compared with low COX-2 tumors (17%, p = 0.019). After confirming that both COX-2 and VEGF positively correlated MVD (p = 0.001 and 0.011, respectively), we performed stepwise linear regression and found that the significance of COX-2 activity on determining MVD was considerably influenced by VEGF expression (COX-2, p = 0.100; VEGF, p = 0.012). Next, we investigated the association of COX-2 with vasculature in different clinicopathological subsets of GC (table 2). GCs of Lauren’s diffuse type, non-cardia, or H. pylori infection exhibited a statistically positive correlation between COX-2 activity and tumor vasculature. In this series, all cardiac GCs had both high COX-2 activity and high MVD but all early GCs had low MVD. Taken as a whole, the investigation on surgical specimens supported the association of COX-2 with VEGF and angiogenesis in GC. Over-expression of COX-2 in GC Cells Promoted in Vitro Angiogenesis To understand the effect of COX-2 on angiogenesis in GC, we transiently transfected GC cell line AGS with COX-2-expressing vector and then harvested the culture medium supernatants for HUVEC growth and tube formation assays. Western blotting analysis confirmed the COX-2 expression increased in a dose-dependent manner after transient transfection (fig. 2A). Using the obtained CM to culture HUVEC, we found that endothelial cells proliferated considerably as the dose of COX-2 transfection increased from the lowest dose (0.5 mg) (p < 0.05), even greater than those cultured with HUVEC-specific culture medium (fig. 2B). Although HUVEC growth increased slightly in the vector control group, the extent was considerably less than positive control and higher doses of COX-2. Notably, neutralizing antibody against VEGF significantly blocked (p < 0.05) but did not completely abolish COX-2 effect, as compared with non-specific blocking IgG. Subsequently, we selected and obtained a stable COX-2 over-expressing clone AGS/COX-2 and a vector control clone AGS/pcDNA3, which were verified by Western blotting analysis of COX-2 protein expression and ELISA determination of PGE2 levels as enzymatic function assessment (fig. 3). In HUVEC tube formation assay, AGS/COX-2 effectively prompted network-like formation of HUVECs (fig. 4C), whereas wild type AGS and AGS/pcDNA3 still rendered many HUVECs distributed dispersedly (fig. 4A, B). Meanwhile, both neutralizing anti-VEGF antibody and the selective COX-2 inhibitor NS-398 also inhibited HUVEC tube formation (fig. 4D, F), as compared with non-specific blocking IgG (fig. 4E). Collectively, the results showed that both HUVEC growth and function are enhanced by COX-2 over-expression in GC cells and provided experimental evidence supporting the findings observed in our clinical study. Angiogenic Factor VEGF and HIF-1a Were Up-Regulated by COX-2 Over-expression through PGE2-dependent Pathway We wondered which pathway and factors are involved in GC angiogenesis influenced by COX-2. Six candidate angiogenic factors were screened, including VEGF, bFGF, PDGF, IL-6, IL-8, and p53, by RT-PCR of mRNAs from wild type AGS, AGS/pcDNA3, and AGS/COX-2 cells (fig. 5). We found that VEGF was markedly up-regulated after GC cells over-expressed COX-2, bFGF and IL-8 transcripts were modestly increased, whereas the other 3 genes were not changed. The result further supported the role of VEGF in angiogenesis caused by COX-2. We next intended to clarify whether the blockage of HIF-1a accumulation in COX-2 over-expressing GC cells can prohibit the effect of COX-2 on VEGF induction, since that VEGF is primarily regulated by HIF-1a, and whether HIF-1a is capable of being regulated independently of hypoxia stimulation. Our experiments confirmed that HIF-1a accumulation and VEGF induction caused by COX-2 were concomitantly reduced in a dose-dependent response to transfection of anti-sense HIF-1a (fig. 6A). PGE2 is a prostaglandin produced by COX-2 catalysis and is important in mediating COX-2 effects. We found that exogenous PGE2 stimulation effectively induced HIF-1a accumulation and increased VEGF production simultaneously in AGS cells and these effects can be reversed by SC51089, the antagonist of PGE2 EP1 receptor, in a dose-dependent manner (fig. 6B). NS-398 is an inhibitor to block PGE2 conversion from arachidonic acid by COX-2. Finally, we investigated whether the blockage of PGE2 production by NS-398 can suppress on VEGF production. The results were compatible with the above experiments, showing that VEGF expression and HIF-1a accumulation decreased concomitantly in a dose-dependent response to pretreatment of NS-398 (fig. 6C). Accordingly, COX-2 may increase VEGF by sequential signaling via PGE2 and then HIF-1a in GC cells. DISCUSSION An emerging issue in cancer research focuses on the mechanistic link between chronic inflammation and carcinogenesis, including tumor angiogenesis[7]. COX-2 is an important, inducible enzyme mediating inflammatory processes and is highly expressed in a diversity of cancers[58]. Recently, accumulating evidence indicated that COX-2 may play an important role in developing neovasculature[62] and therefore might be a potential therapeutic target against tumor angiogenesis[62]. However, the mechanism governing COX-2-induced angiogenesis remains largely unclear. Our initial task on investigating surgical specimens indeed revealed a positive association among COX-2 immunoreactivity, VEGF expression, and GC vasculature. The results were consistent with earlier reports[104-107]. Additionally, we found that the association between COX-2 and MVD was substantially influenced by VEGF in stepwise regression analysis. Although studies have discovered some upstream factors capable of stimulating COX-2 expression in cancer cells[120], the downstream effecter molecules by which COX-2 promotes angiogenesis are largely unknown. Our clinical data implicated that COX-2 may induce VEGF to promote GC angiogenesis. This hypothesis was further supported by our in vitro angiogenesis assessments on HUVECs. Firstly in the experimental study, we examined the effect of COX-2 over-expression on GC angiogenesis by assessing HUVEC growth and tube formation, both being generally accepted as indicators of in vitro angiogenesis. We found that GC cells increased HUVEC proliferation in a COX-2 dose-dependent manner. Meanwhile, the COX-2 over-expressing GC cells effectively promoted HUVEC tube formation arising from endothelial cell migration. Our results also showed that both proliferation and tube formation of HUVECs were prohibited either by the COX-2 inhibitor or by anti-VEGF antibody. The results support COX-2 as an angiogenic factor in GC. Secondly, we compared 6 angiogenesis-associated factors among wild type AGS, AGS/pcDNA3, and AGS/COX-2, revealing that the VEGF gene expression was markedly increased. This result was in consistency with the HUVEC experiments that blockage of VEGF effectively inhibited HUVEC proliferation and network formation induced by COX-2 over-expressing GC cells. In our study, increases of bFGF and IL-8 transcripts were modest and not as significant as seen in VEGF in AGS/COX-2. We do not exclude the possibility that bFGF, IL-8, and/or other factors not examined in our study may also play a role in COX-2-related angiogenesis but further studies are needed to address this issue. As regards subtle changes of PDGF, p53, and bFGF in AGS/pcDNA3 in comparison to wild type AGS cells, a clonal effect is considered. Because VEGF functions to promote the majority of the processes involved in angiogenesis, including endothelial proliferation and migration, vascular lumen formation, and vascular permeability[29, 30], the authors suggest that VEGF induction may be one of the main strategies by which COX-2 increases GC angiogenesis. In most conditions, hypoxia induces VEGF production through up-regulating HIF-1[121]. HIF-1a dimerizes with HIF-1β and then binds to the hypoxia response element of the target genes, including VEGF[122]. As the most important regulator for VEGF expression, however, HIF-1a is not only controlled by oxygen concentration but also regulated by several hypoxia-independent mechanisms, including transition metals, nitric oxide, reactive oxygen species, growth factors, and mechanical stress[77]. Jones et al reported that NSAID increased the expression of the von Hippel Lindau tumor suppressor, which targets proteins for ubiquitination and leads to reduced accumulation of HIF-1a, and resulted in suppression of VEGF expression[123]. Surprisingly, we found that concomitant up-regulation of HIF-1a and VEGF occurred in COX-2 over-expressing GC cells. We hypothesized that PGE2 may play a role in this process because PGE2 is well known in mediating a vast of effects by COX-2. Our additionally work further provided evidence supporting the link from PGE2, HIF-1a, to VEGF by showing that HIF-1a and VEGF were increased simultaneously by exogenous PGE2 stimulation and reversed by blocking PGE2 receptor. Moreover, VEGF production was markedly reduced in those cells treated either by the selective COX-2 inhibitor NS-398[151], which inhibits excess PGE2 production but does not affect the physiological level of PGE2[152], or by anti-sense HIF-1a, which blocks HIF-1a protein production, respectively. Earlier studies reported controversial results in regard to the relationship among COX-2, HIF-1 and VEGF among different cell types. A study in prostate cancer cells found that NSAIDs reduced HIF-1 protein levels but the inhibitory effect might be COX-2–independent[124]. Another study on a lung cancer cell line investigated the role of HIF-1a in interleukin-1β-induced inflammation , showing that an HIF-1a transcriptional inhibitor NSC-609699[125] suppresses VEGF expression induced by COX-2[126]. In benign cells, COX-2 was found to mediate the hypoxia effect on HIF-1 activation and VEGF production in retinal cells and hepatic stellate cells, respectively[127, 128]. Collectively, COX-2 may increase angiogenesis through oxygen-independent mechanisms in addition to its role as a hypoxia effecter to increase VEGF expression. Our work supports that COX-2/PGE2/HIF-1/VEGF pathway may play an important role in GC angiogenesis, by which chronic inflammation might be capable of contributing to GC development. The support for COX-2 as a molecular target for cancer prevention and treatment comes from multiple lines of evidence[120]. Epidemiological studies showed that non-steroidal anti-inflammatory drugs (NSAIDs) reduced the risk of developing malignancies, including colorectal cancer[129]. Animal studies also showed that NSAIDs protected the formation of tumors[130]. Currently, many clinical trails are under way to evaluate selective COX-2 inhibitors as the agents to prevent or treat human cancers[120]. Our clinical study demonstrated that the correlation between COX-2 and tumor angiogenesis predominated in certain subsets of GC, indicating that diverse mechanisms may govern the differential effect of COX-2 on angiogenesis and also that the selection of patient is mandatory to apply COX-2 inhibitors as an anti-GC agent in the future. One feature from our clinical study was that GCs with H. pylori infection exhibited a positive correlation between COX-2 and tumor angiogenesis but those without H. pylori infection did not. The result might indicate that inflammation associated with H. Pylori infection enhances GC neovasculature through COX-2 up-regulation. On the other hand, although our earlier research found that the multi-functional inflammatory cytokine IL-6 can promote VEGF and angiogenesis in GC[153], the present study showed that IL-6 does not mediate GC angiogenesis induced by COX-2. In summary, our work supports that the COX-2/PGE2/HIF-1/VEGF pathway may be important in GC angiogenesis. Further research to elucidate other important factors or mechanisms by which COX-2 promotes tumor angiogenesis are needed to shed more lights on the use of the COX-2 inhibitors in cancer chemoprevention and treatment.

並列關鍵字

Cyclooxygenase-2 ； Gastric Carcinoma ； Angiogenesis ； Vascular Endothelial Growth Factor ； Interleukin-6

參考文獻

1. Fuchs, C. S. and Mayer, R. J. Gastric carcinoma. N Engl J Med, 333: 32-41, 1995.

2. Laur'En, P. The Two Histological Main Types of Gastric Carcinoma: Diffuse and So-Called Intestinal-Type Carcinoma. An Attempt at a Histo-Clinical Classification. Acta Pathol Microbiol Scand, 64: 31-49, 1965.

3. An international association between Helicobacter pylori infection and gastric cancer. The EUROGAST Study Group. Lancet, 341: 1359-1362, 1993.

4. Nomura, A., Stemmermann, G. N., Chyou, P. H., Kato, I., Perez-Perez, G. I., and Blaser, M. J. Helicobacter pylori infection and gastric carcinoma among Japanese Americans in Hawaii. N Engl J Med, 325: 1132-1136, 1991.

5. Tatsuta, M., Iishi, H., Okuda, S., Taniguchi, H., and Yokota, Y. The association of Helicobacter pylori with differentiated-type early gastric cancer. Cancer, 72: 1841-1845, 1993.

國際替代計量

介白質－6與第二型環氧酵素增益胃癌之血管內皮生長因子及腫瘤血管新生

全文下載

主題瀏覽