Proinflammatory functions of vascular endothelial growth factor in alloimmunity

Vascular endothelial growth factor (VEGF), an established angiogenesis factor, is expressed in allografts undergoing rejection, but its function in the rejection process has not been defined. Here, we initially determined that VEGF is functional in the trafficking of human T cells into skin allografts in vivo in the humanized SCID mouse. In vitro, we found that VEGF enhanced endothelial cell expression of the chemokines monocyte chemoattractant protein 1 and IL-8, and in combination with IFN-γ synergistically induced endothelial cell production of the potent T cell chemoattractant IFN-inducible protein-10 (IP-10). Treatment of BALB/c (H-2^d) recipients of fully MHC-mismatched C57BL/6 (H-2^b) donor hearts with anti-VEGF markedly inhibited T cell infiltration of allografts and acute rejection. Anti-VEGF failed to inhibit T cell activation responses in vivo, but inhibited intragraft expression of several endothelial cell adhesion molecules and chemokines, including IP-10. In addition, whereas VEGF expression was increased, neovascularization was not associated with acute rejection, and treatment of allograft recipients with the angiogenesis inhibitor endostatin failed to inhibit leukocyte infiltration of the grafts. Thus, VEGF appears to be functional in acute allograft rejection via its effects on leukocyte trafficking. Together, these observations provide mechanistic insight into the proinflammatory function of VEGF in immunity.

It is well established that leukocyte recruitment into sites of inflammation is associated with angiogenesis, and there is considerable evidence to suggest that angiogenesis and inflammation occur in an interactive and overlapping manner (1–4). Vascular endothelial growth factor (VEGF) is a potent angiogenesis factor that has important roles in both normal physiological as well as pathological vasculogenesis and angiogenesis (5–7). VEGF stimulates endothelial cell proliferation and migration in vitro, and has also been found to reprogram endothelial cell gene expression and to prevent endothelial cells from undergoing apoptosis and senescence (7, 8). In addition, VEGF acts as a proinflammatory cytokine by increasing endothelial cell permeability, by inducing the expression of endothelial cell adhesion molecules, and via its ability to act as a monocyte chemoattractant (9–11). Thus, VEGF is likely a key intermediary between cell-mediated immune inflammation and the associated angiogenesis reaction.

VEGF is produced by endothelial cells, macrophages, activated T cells, and a variety of other cell types (7, 12, 13). It is a heparin-binding, homodimeric glycoprotein with several protein variants resulting from alternative mRNA splicing (5). VEGF binds to three high-affinity tyrosine kinase receptors, Flt-1 (VEGFR-1), KDR (VEGFR-2) (7) and neuropilin (14), expressed almost exclusively by endothelial cells, although Flt-1 is also expressed by monocytes (15). The major stimulus for VEGF expression is hypoxia (16), but other factors that can upregulate VEGF expression include the degree of cell differentiation; local concentrations of glucose and serum; hormones; prostaglandins; modulators of protein kinase C; nitric oxide; and stimulators of adenylate cyclase (7, 17–20). In addition, we recently demonstrated that ligation of CD40 on endothelial cells and monocytes by CD40 ligand (expressed on platelets and activated T cells) is potent for the induction of VEGF, thereby linking immune inflammation with angiogenesis (12, 21).

Little is known about the function of VEGF in alloimmunity. However, VEGF has been observed in human renal and cardiac allografts, and its expression is associated with acute and chronic rejection (22–25). Here, we provide evidence that VEGF functions in the local trafficking of allogeneic leukocytes in vivo and that anti-VEGF is potent in inhibiting the development of acute allograft rejection. These observations provide insight into a general role for VEGF in inflammation and provide compelling evidence that VEGF is of importance in alloimmunity.

Function of VEGF in allogeneic human leukocyte recruitment. We first used the huSCID mouse to determine the effect of VEGF on leukocyte trafficking into human skin. This model is unique, as circulating human leukocytes in the huSCID mouse selectively interact with human endothelial cells in the skin allograft (4, 36, 44), allowing for the analysis of molecules involved in endothelial cell–dependent mechanisms of human leukocyte trafficking. In our model, human foreskin is transplanted onto SCID mice and is allowed to heal for 4–6 weeks. After adoptive transfer of human PBLs into the mouse by intraperitoneal injection, CD4⁺ and CD8⁺ T cells as well as CD68⁺ monocytes/macrophages infiltrate the human skin over 7–14 days (4). As previously described (36), healed human skin in the huSCID mouse is vascularized by both human and mouse vasculature, but human alloreactive leukocytes selectively interact and mediate vasculitis of only the human vessels. Thus, skin allografts do not slough in this model, as the murine vasculature maintains blood flow and integrity of the skin despite marked leukocyte infiltration and marked vasculitis of the human vessels (4, 44).

We initially evaluated the expression of VEGF in noninfiltrated and infiltrated human skin specimens from nonhumanized or huSCID mice, respectively, and found enhanced VEGF mRNA and protein expression in association with infiltrates (Figure 2). By immunohistochemistry, VEGF protein was expressed by both inflammatory infiltrates and by vascular endothelial cells (Figure 2, d and e). In contrast, in skin specimens without infiltrates, VEGF was only found on keratinocytes and endothelial cells with minimal expression overall (Figure 2f). We treated huSCID mice with a neutralizing antihuman VEGF, human IgG, or saline. Consistent with its known function as a monocyte chemoattractant (11) anti-VEGF had a significant inhibitory effect on the recruitment of CD68⁺ monocytes/macrophages (Figure 3). Few CD68⁺ cells were evident in skins harvested from animals treated with anti-VEGF (Figure 3i). Moreover, we unexpectedly found that anti-VEGF was most potent at inhibiting CD3⁺ T cell trafficking into human skin (Figure 3, d–f). By quantitative grid counting analysis, there was a statistically significant decrease in CD3⁺ T cell infiltrates in anti-VEGF-treated huSCID animals both at day 7 and at day 14 (Figure 3j). Therefore, VEGF is functional for peripheral recruitment and for the trafficking of human allogeneic lymphocytes and monocytes in vivo.

Figure 2

Expression of VEGF in association with human leukocytic infiltration of skin. SCID mice bearing healed human skin transplants received 3 × 10⁸ human PBLs or saline by intraperitoneal injection. Skin grafts were harvested after 14 days. (a and b) The expression of VEGF mRNA in normal noninfiltrated skins (N; nonhumanized SCID) or in infiltrated skin specimens (I; huSCID) evaluated by RT-PCR (a) and by RNase protection assay (b). (c) The relative expression of VEGF mRNA versus GAPDH evaluated by RNase protection quantified by densitometry. Bar graphs illustrate mean VEGF expression (±1 SD) for three noninfiltrated (N) or infiltrated (I) skins. (d and e) Expression of VEGF (rose-red) by immunohistochemistry in an infiltrated skin specimen. (f) Expression of VEGF (rose-red) in a normal noninfiltrated skin. Note that there is enhanced VEGF expression in association with leukocytic infiltration. Representative of at least ten experiments. Magnification of d–f, ×400.

Figure 3

Function of VEGF in human leukocyte recruitment. Saline, IgG, or anti–human VEGF were received by huSCID mice bearing human skin transplants. (a–i) Skin grafts were harvested 14 days after humanization and infiltrates were identified by H&E staining (a–c), by immunostaining with anti–human CD3 (d–f), and by immunostaining with anti–human CD68 (g–i). Treatment of huSCID mice with anti–human VEGF inhibited both CD3⁺ T cell (f) and CD68⁺ monocyte macrophage (i) infiltration of skin. (j) Quantitative assessment of CD3⁺ T cell infiltrates was performed by calibrated grid counting at a magnification of ×400 in skin specimens harvested at either day 7 or day 14 following humanization. The mean CD3 count per calibrated field is illustrated in skins harvested from animals treated with saline (n = 7 at day 7; n = 10 at day 14), anti–human VEGF (n = 5 at day 7; n = 10 at day 14), or control IgG (n = 4 at day 14).

We also assessed the effect of anti-VEGF on the angiogenesis reaction, which we have previously reported to be associated with mononuclear cell infiltration in this model (4). Our findings were that anti-VEGF markedly inhibited angiogenesis (data not shown), which is consistent with the interpretation that locally expressed VEGF is also of functional importance in leukocyte-associated angiogenesis.

In vitro analysis of VEGF-dependent chemokine production in endothelial cells. The function of VEGF as a monocyte chemoattractant is established in vitro (11). However, our finding that VEGF promotes T cell recruitment has not been previously reported in vitro or in vivo, to our knowledge. To study the mechanism of function of VEGF on T cell recruitment, we next assessed the effect of VEGF on human endothelial cell activation in vitro. Treatment of confluent cultures of endothelial cells with increasing concentrations of VEGF resulted in a dose-dependent increase in the expression of the endothelial cell adhesion molecules E-selectin, ICAM-1, and VCAM-1 (not shown, and as reported by others [refs.9, 10]). In addition, after treatment with VEGF, we found that the endothelial cell expression of the chemokines monocyte chemoattractant protein 1 (MCP-1) and interleukin-8 (IL-8) was also increased (Figure 4a). MCP-1 peaked in expression between 1 and 5 hours following treatment, and expression persisted for up to 24 hours (Figure 4b). In contrast, IL-8 peaked in expression at approximately 2 hours following treatment, and decreased in expression by 24 hours. But there was no increase in the expression of the T cell–chemoattractant chemokines lymphotactin, RANTES, IP-10 (Figure 4), or Mig (not shown) after VEGF treatment.

Figure 4

Effect of VEGF on endothelial cell chemokine expression. (a–d) Confluent cultures of human endothelial cells were treated for 4 hours (a and c) or as a time course (b and d) with VEGF alone or VEGF in combination with IFN-γ as indicated. Total RNA was harvested from endothelial cells and the expression of chemokines was analyzed by RNase protection assay. Concentration- (a) and time-dependent (b) effects of VEGF on chemokine expression. (c and d) Effect of IFN-γ alone or in combination with VEGF on chemokine expression. Note that treatment with IFN-γ alone resulted in IP-10 expression (c, and d lane 7), and a combination of IFN-γ with VEGF resulted in a synergistic induction of IP-10 (c, and d lanes 2–6). Bar graphs to the right of each blot represent the quantitative analysis of IP-10 mRNA expression in three representative RNase protection assays as illustrated in c and d. (e) The production of IP-10 by ELISA in culture supernatants of endothelial cells treated with IFN-γ (1,000 U/ml; black bars) or with VEGF (10 ng/ml) and IFN-γ (1,000 U/ml) (white bars) for different times as indicated. Representative of three similar experiments performed in triplicate (mean ± 1 SD).

Since IFN-γ-induced chemokines play a major role in the local recruitment of allogeneic lymphocytes, we also determined an effect of VEGF on IFN-γ-mediated activation of the potent T cell–chemoattractant chemokines IP-10 and Mig. We treated endothelial cells with IFN-γ and found induced expression of IP-10 and Mig, as reported (45, 46). Moreover, we observed that a combination of VEGF with IFN-γ resulted in a synergistic induction of endothelial cell IP-10 mRNA and protein expression, but not Mig expression (Figure 4, c–e, and not shown). In addition, we observed that IFN-γ inhibited the expression of IL-8, and that combinations of VEGF and IFN-γ restored IL-8 expression (Figure 4, c and d). Thus, VEGF may promote human allogeneic lymphocyte trafficking in vivo via regulation of endothelial cell production of chemokines including the potent T cell chemokine IP-10.

Blockade of VEGF in a fully MHC-mismatched murine model of acute cardiac allograft rejection. Our results also raise the possibility that VEGF is of importance in allograft rejection. Allograft rejection is a complex multicellular process that is critically dependent upon T cell recognition of alloantigen processed and presented to T cells by either donor or recipient antigen-presenting cells (47). The subsequent recruitment of activated T cells and effector cells into an allograft has been shown to be mediated in part by locally expressed leukocyte-endothelial adhesion molecules (48) and chemokines (49). We used a physiological in vivo model of acute allograft rejection in which fully MHC-mismatched C57BL/6 (H-2^b) donor hearts were transplanted into BALB/c (H-2^d) recipients. Similar to our findings in the huSCID mouse (above), VEGF was expressed in association with graft infiltrates and acute rejection (Figure 5, a and b). Administration of our neutralizing anti-VEGF antiserum resulted in a significant prolongation of allograft survival up to 25 days after transplant versus 11 days in control serum–treated animals (Figure 5c, P < 0.01). Severe cellular rejection with extensive mononuclear cell infiltration and myocyte necrosis was evident in sections of rejecting grafts at day 7 in recipients treated with control rabbit serum (Figure 5, d–g). In contrast, we observed only a mild to moderate infiltrate, reduced numbers of CD3⁺ T cells and monocytes/macrophages, less myocyte damage, and minimal vasculitis in allografts from recipients treated with anti-VEGF antiserum (Figure 5, i–l). Moreover, the T cell chemoattractant IP-10 was observed in abundance on multiple cell types in rejecting grafts (Figure 5h), but there was minimal expression in grafts harvested from animals treated with anti-VEGF (Figure 5m). Therefore, the intense expression of VEGF in allografts in vivo is of mechanistic significance in promoting the development of rejection.

Figure 5

Blockade of VEGF in a fully MHC-mismatched murine model of acute cardiac allograft rejection. Fully MHC-mismatched C57BL/6 (H-2^b) donor hearts were transplanted into BALB/c (H-2^d) mice as recipients. Untreated recipients develop marked leukocytic infiltrates and rejection by day 7. (a and b) Immunohistochemical analysis of VEGF in a normal nontransplanted heart (a) and in a rejecting allograft at day 7 (b), showing intense VEGF expression in association with allograft rejection. (c) Graft survival curves for recipients treated with anti-VEGF antiserum (filled squares; n = 6) or normal rabbit serum as a control (open squares; n = 5). Anti-VEGF–treated recipients show prolonged heart allograft survival (P < 0.001). (d–m) Histological analysis of cardiac allografts harvested at day 7 after transplantation from rejecting control serum–treated animals (d–h) or from animals treated with anti-VEGF (i–m). Control serum–treated animals had evidence of severe cellular rejection with extensive mononuclear cell infiltration (d), including CD45⁺, (e) CD3⁺ (f), and macrophage (Mac) (g) infiltrates. In contrast, recipients treated with anti-VEGF had minimal infiltrates and no evidence of vasculitis (i–l). (h and m) IP-10 protein expression was diffuse and intense within allografts in association with rejection (h) but was of low intensity and sparse in anti-VEGF–treated recipients (m). Representative immunostaining of five animals from each group; magnification, ×400.

Mechanism of function of VEGF in allograft rejection in vivo. The effect of anti-VEGF on leukocyte trafficking and allograft rejection in our in vivo studies could be explained by an effect of VEGF on T cell activation responses. However, the addition of our antibodies against VEGF to cocultures in human or mouse MLRs failed to inhibit T cell activation responses (Figure 6, a and b). Also, the in vivo priming of T cells to alloantigen as manifested by ELISPOT analysis of IFN-γ in murine recipients of allografts (above) was unaltered by anti-VEGF treatment (Figure 6c). Thus, VEGF does not function in alloimmune T cell activation in vitro or in vivo.

Figure 6

Function of VEGF in alloimmune T cell activation and allograft rejection. (a) Anti–human VEGF or anti–murine VEGF antiserum was added into the human or the mouse MLR, respectively. Proliferation was assessed by [³H]thymidine incorporation for the last 18 hours of coculture. (b) The production of IFN-γ and IL-2 was assessed by ELISA in coculture supernatants from a human MLR. As illustrated, blockade of VEGF had no effect on proliferation or cytokine production in the MLR. Bars indicate the mean ± 1 SD for triplicate wells. Data are representative of three experiments with similar results. S, stimulators alone; R, responders alone. (c) Frequency of IFN-γ-producing cells in murine recipients of cardiac transplants as assessed by ELISPOT. Illust production of IFN-γ from a syngeneic, an untreated, and an anti-VEGF-treated animal. Representative of three such experiments performed in triplicate.

We next assessed whether the intense intragraft expression of VEGF in acute rejection could mediate leukocyte recruitment in part via neovascularization. By immunohistochemistry, we found that endothelial CD31 was expressed diffusely on large vessels as well as microvessels throughout all grafts examined (Figure 7a). Moreover, the staining pattern of CD31 as well as the vascularization of isografts was very similar to that found in anti-VEGF-treated grafts. In contrast, we found fewer CD31-expressing vessels in control Ig–treated rejecting grafts; and some focal areas of rejecting grafts were devoid of any vascular staining (Figure 7a, and not shown). To assess the possibility of a role for angiogenesis in acute rejection, we also treated animals with endostatin, a known endothelial cell–specific inhibitor of angiogenesis (29) (Figure 7, b–d). Our findings were that endostatin had a minimal effect on allograft survival, prolonging survival in three of eight animals; and failed to have any effect on leukocyte trafficking into the graft (Figure 7, b–d). Because the effect of endostatin was so markedly different than that following treatment with anti-VEGF (Figure 5), we interpret these data as suggesting that VEGF mediates leukocyte trafficking into an allograft independent of an effect on angiogenesis.

Figure 7

Function of angiogenesis in acute rejection. Fully MHC-mismatched C57BL/6 (H-2^b) donor hearts were transplanted into BALB/c (H-2^d) mice, and recipients were treated with control Ig, anti-VEGF, or endostatin, as described in Methods. (a) Immunohistochemical analysis of CD31 in isografts or allografts harvested at day 7 from animals treated with control Ig or anti-VEGF. (b) H&E staining of allografts harvested from untreated or endostatin-treated animals at day 7 showing notable infiltrates in both untreated and treated grafts at low magnification (left; magnification, ×200) and at high magnification (right; magnification, ×400). Representative of eight animals. (c) Immunostaining for CD3-expressing T cells in a representative allograft from an endostatin-treated animal. (d) Graft survival curves for untreated recipients (dotted line; n = 10), or recipients treated with endostatin (solid line; n = 8). Note that the endostatin used in these studies inhibited angiogenesis in control animals with tumors (not shown and as described in ref.30), but only minimally prolonged graft survival in three of eight animals.

To further define the role of VEGF in T cell recruitment into cardiac allografts, we examined the effect of VEGF in primary cultures murine myocardial endothelial cells in vitro. In contrast to our findings in human endothelial cells, we found that VEGF alone induced the expression of IP-10 in murine endothelial cells (Figure 8a). Thus, it is possible that VEGF may regulate T cell trafficking primarily via IP-10-dependent mechanisms. Indeed, local intragraft endothelial cell expression of IP-10 has been previously reported to be of critical importance in the recruitment of T cells into allografts in rejection (39). Transplanted animals were next treated with neutralizing anti–IP-10 alone or in combination with anti-VEGF (Figure 8b). Our findings were that anti–IP-10 alone prolonged graft survival, and that combined treatment with anti–IP-10 and anti-VEGF significantly prolonged allograft survival versus treatment with anti–IP-10 alone or anti-VEGF alone (Figure 8b, P < 0.005 and P < 0.03, respectively). This is suggestive that a major effect of anti-VEGF in vivo is independent of IP-10. To further confirm this interpretation, we also transplanted allografts from IP-10^–/– donors into wild-type mice and again found that anti-VEGF prolonged survival in the absence of local IP-10 production (Figure 8c, P < 0.04).

Figure 8

Function of VEGF-dependent regulation of IP-10 in allograft rejection. (a) Confluent cultures of murine myocardial endothelial cells were treated for 4 hours with recombinant murine VEGF or IFN-γ alone or VEGF in combination with IFN-γ as indicated. Total RNA was harvested from endothelial cells and the expression of IP-10 was analyzed by RNase protection assay. (b and c) The ability of VEGF to mediate IP-10-dependent trafficking and rejection was evaluated using anti-VEGF and anti–IP-10 in fully MHC-mismatched C57BL/6 (H-2^b) or IP-10^–/– (H-2^b) donor hearts transplanted into BALB/c (H-2^d) mice. Recipients of wild-type grafts were treated with anti-VEGF alone, or with anti-VEGF in combination with anti–IP-10. Recipients of IP-10^–/– donor grafts were treated with anti-VEGF. Both anti–IP-10 and anti-VEGF were administered according to the schedule outlined in Methods. As illustrated in b, we found that addition of anti-VEGF with anti–IP-10 significantly prolonged allograft survival in wild-type combinations (P < 0.005); and in c, anti-VEGF prolonged survival in mice that received IP-10^–/– donor hearts (P < 0.04). The survival of control untreated wild-type grafts are illustrated by the dotted line.

Lastly, we examined adhesion molecule and chemokine expression in vivo in cardiac allografts harvested on day 7 from control or anti-VEGF-treated wild-type mice. By RNase protection assay, we found a significant decrease in the expression of E-selectin, ICAM-1, and VCAM-1 in anti-VEGF-treated animals compared with that of controls (Figure 9a). In addition, anti-VEGF markedly inhibited the intragraft expression of several T cell–chemoattractant chemokines (including lymphotactin, RANTES, and IP-10) as well as the monocyte chemoattractant MCP-1 (Figure 9b). Thus, blockade of VEGF in vivo inhibits the local expression of several chemokines that are not directly regulated by VEGF. We interpret our findings as suggesting that VEGF is of importance in allograft rejection at early times after transplantation and prior to the induced expression and function of T cell–derived chemokines. Intragraft VEGF expression appears to promote early T cell and monocyte trafficking, and subsequent activation events result in additional VEGF expression as well as the expression of chemokines (including IP-10); both of these effects are additive in promoting rejection.

Figure 9

Effect of Anti-VEGF on the intragraft expression of adhesion molecules and chemokines. Analysis of endothelial cell adhesion molecule expression (a) and chemokine expression (b) in cardiac syngeneic grafts (Syn) or allografts (Allo) at day 7. Recipients received either control normal rabbit serum (–) or anti-VEGF antiserum (+). The bar graphs represent quantitative analysis of mRNA expression as the relative expression of adhesion molecules (adh mol) or chemokines (chem) compared with the expression of GAPDH. Mean expression (± 1 SD) for three animals treated with normal rabbit serum (black bars) or anti-VEGF antiserum (white bars) is shown.

VEGF is a most potent and critical mediator of physiological as well as pathological angiogenesis. Its expression and function have been reported in several chronic inflammatory processes; however, its mechanism of function in immunity has been unclear. Recent reports have defined the angiogenesis reaction mediated by VEGF to be proinflammatory in chronic inflammation (7). Here, we provide evidence that VEGF promotes endothelial cell chemokine production in vitro and in vivo and functions in the recruitment of monocytes and T cells into allografts. VEGF is thus an important proinflammatory cytokine in transplant rejection.

The expression of VEGF has been reported to be associated with allograft rejection (especially chronic rejection) in both experimental models and following human transplantation (22, 24, 25). In addition, it has been reported that transplant recipients with polymorphisms in the VEGF gene, encoding high VEGF production, are at increased risk for the development of acute renal allograft rejection (23). Here, we extend upon those observations and demonstrate that intragraft VEGF has a biological effect in vivo in promoting rejection, as leukocyte infiltration is inhibited following treatment with neutralizing anti-VEGF. Moreover, we demonstrate that induced intragraft VEGF is unlikely to be functional for the development of angiogenesis, as neovascularization is not a component of the rejection process. Rather, we show that the mechanism of function of VEGF appears to involve its ability to mediate intragraft leukocyte trafficking.

It is well established that the process of recruitment involves coordinate interactions among adhesion molecules and chemokines (3, 49). Multiple individual molecules have been shown to be functional in the process of rejection (49–52), suggesting that there is some redundancy in the function of adhesion molecules and chemokines in vivo. It is known that VEGF induces the expression of adhesion molecules in human endothelial cells (9, 10), and two recent reports have pointed to an effect of VEGF on chemokine production (53, 54). Here, we extend upon those observations and show that VEGF regulates IP-10, MCP-1, and IL-8, as well as the endothelial cell adhesion molecules E-selectin, ICAM-1, and VCAM-1, in vitro and in vivo. All of these molecules are expressed in association with rejection, and blockade of several molecules or their ligands individually (e.g., IP-10, MCP-1, ICAM-1, and VCAM-1) (49–52) has been shown to prolong allograft survival in murine models. Our data in this study imply that the ability of VEGF to regulate several of these molecules is the basis for its function in leukocyte trafficking. However, we have not fully addressed this complex question at an individual molecular level, and thus this interpretation will require further study. Nevertheless, of all chemokine-chemokine receptor interactions studied to date, the most profound effects on graft survival are seen when the T cell chemoattractant IP-10 or its ligand CXCR3 are targeted (39, 50, 55). We found that VEGF-dependent T cell trafficking and rejection in vivo was not restricted to the function of IP-10. Our results therefore provide a framework for an analysis of the role of VEGF in the expression and function of individual chemokines and adhesion molecules in leukocyte trafficking, in rejection, and in other cell-mediated immune inflammatory diseases (7, 38).

Another interpretation of our findings is that the major proinflammatory function of VEGF is in the initiation of the inflammatory cascade, rather than in subsequent events resulting from T cell–mediated activation responses. Consistent with this possibility, it is important to note that hypoxia is the major stimulus for VEGF expression (16). This suggests that VEGF will be produced locally within allografts immediately following transplantation. Additional insults, including platelet and leukocyte recruitment into the graft, will further facilitate the expression of local VEGF (12) as well as other cytokines and chemokines (56) that are of importance in the rejection process (47, 49, 57). Here, we have provided evidence that early VEGF expression promotes T cell and monocyte recruitment. We suggest that the subsequent induced expression of chemokines (e.g., IP-10) and VEGF are additive in promoting the development of rejection. As it is well established that early insults to allografts (such as prolonged ischemia) have long-term consequences (58), dysregulation of VEGF at early times following transplantation may have beneficial effects in promoting graft survival. Indeed, in support of this possibility, it has been noted that early (but not late) blockade of VEGF can limit the development of chronic arthritis (38).

In summary, our results provide compelling evidence that VEGF is a proinflammatory cytokine in allograft rejection. In addition, they provide insight into the role of VEGF in cell-mediated inflammatory reactions in immunity. We suggest that VEGF warrants consideration as a target for future therapeutic intervention following transplantation.

This work was supported by NIH grants AI46756 to D.M. Briscoe and CA069212 to A.D. Luster. The authors wish to thank Don Senger, Beth Israel Deaconess Medical Center, for help with the generation of anti-VEGF antiserum; Kashi Javaherian and Judah Folkman for the gift of endostatin; and Stuart Robertson, Evelyn Flynn, and William Wong for assistance with these studies. We also thank the maternity staff at South Shore Hospital for the supply of umbilical cords used for the generation of endothelial cells.

Marlies E.J. Reinders and Masayuki Sho contributed equally to this work.

Conflict of interest: The authors have declared that no conflict of interest exists.

Nonstandard abbreviations used: IFN-inducible protein-10 (IP-10); human umbilical vein endothelial cell (HUVEC); peripheral blood lymphocyte (PBL); mixed lymphocyte reaction (MLR); humanized SCID (huSCID).

1. Sidky, YA, Auerbach, R. Lymphocyte-induced angiogenesis: a quantitative and sensitive assay of the graft-vs.-host reaction. J. Exp. Med. 1975. 141:1084-1100.
   View this article via: PubMed Google Scholar
2. Auerbach, R, Sidky, YA. Nature of the stimulus leading to lymphocyte-induced angiogenesis. J. Immunol. 1979. 123:751-754.
   View this article via: PubMed Google Scholar
3. Cotran, R.S. 1994. Inflammation and repair. In Pathologic basis of disease. R.S. Cotran, V. Kumar, and S.L. Robbins, editors. WB Saunders, Philadelphia, USA. 51–92.
4. Moulton, KS, et al. Angiogenesis in the huPBL-SCID model of human transplant rejection. Transplantation. 1999. 67:1626-1631.
   View this article via: PubMed Google Scholar
5. Leung, DW, Cachianes, G, Kuang, WJ, Goeddel, DV, Ferrara, N. Vascular endothelial growth factor is a secreted angiogenic mitogen. Science. 1989. 246:1306-1309.
   View this article via: PubMed Google Scholar
6. Folkman, J. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat. Med. 1995. 1:27-31.
   View this article via: PubMed Google Scholar
7. Ferrara, N, Davis-Smyth, T. The biology of vascular endothelial growth factor. Endocr. Rev. 1997. 18:4-25.
   View this article via: PubMed Google Scholar
8. Gerber, HP, Dixit, V, Ferrara, N. Vascular endothelial growth factor induces expression of the antiapoptotic proteins Bcl-2 and A1 in vascular endothelial cells. J. Biol. Chem. 1998. 273:13313-13316.
   View this article via: PubMed Google Scholar
9. Melder, RJ, et al. During angiogenesis, vascular endothelial growth factor and basic fibroblast growth factor regulate natural killer cell adhesion to tumor endothelium. Nat. Med. 1996. 2:992-997.
   View this article via: PubMed Google Scholar
10. Kim, I, et al. Vascular endothelial growth factor expression of intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), and E-selectin through nuclear factor-κB activation in endothelial cells. J. Biol. Chem. 2001. 276:7614-7620.
    View this article via: PubMed Google Scholar
11. Barleon, B, et al. Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor flt-1. Blood. 1996. 87:3336-3343.
    View this article via: PubMed Google Scholar
12. Melter, M, et al. Ligation of CD40 induces the expression of vascular endothelial growth factor by endothelial cells and monocytes and promotes angiogenesis in vivo. Blood. 2000. 96:3801-3808.
    View this article via: PubMed Google Scholar
13. Freeman, MR, et al. Peripheral blood T lymphocytes and lymphocytes infiltrating human cancers express vascular endothelial growth factor: a potential role for T cells in angiogenesis. Cancer Res. 1995. 55:4140-4145.
    View this article via: PubMed Google Scholar
14. Soker, S, Fidder, H, Neufeld, G, Klagsbrun, M. Characterization of novel vascular endothelial growth factor (VEGF) receptors on tumor cells that bind VEGF165 via its exon 7-encoded domain. J. Biol. Chem. 1996. 271:5761-5767.
    View this article via: PubMed Google Scholar
15. Clauss, M, et al. The vascular endothelial growth factor receptor Flt-1 mediates biological activities. Implications for a functional role of placenta growth factor in monocyte activation and chemotaxis. J. Biol. Chem. 1996. 271:17629-17634.
    View this article via: PubMed Google Scholar
16. Shweiki, D, Itin, A, Soffer, D, Keshet, E. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature. 1992. 359:843-845.
    View this article via: PubMed Google Scholar
17. Sato, K, et al. Stimulation by thyroid-stimulating hormone and Grave’s immunoglobulin G of vascular endothelial growth factor mRNA expression in human thyroid follicles in vitro and flt mRNA expression in the rat thyroid in vivo. J. Clin. Invest. 1995. 96:1295-1302.
    View this article via: JCI PubMed Google Scholar
18. Shweiki, D, Neeman, M, Itin, A, Keshet, E. Induction of vascular endothelial growth factor expression by hypoxia and by glucose deficiency in multicell spheroids: implications for tumor angiogenesis. Proc. Natl. Acad. Sci. U. S. A. 1995. 92:768-772.
    View this article via: PubMed Google Scholar
19. Williams, B, Baker, AQ, Gallacher, B, Lodwick, D. Angiotensin II increases vascular permeability factor gene expression by human vascular smooth muscle cells. Hypertension. 1995. 25:913-917.
    View this article via: PubMed Google Scholar
20. Williams, B, Quinn-Baker, A, Gallacher, B. Serum and platelet-derived growth factor-induced expression of vascular permeability factor mRNA by human vascular smooth muscle cells in vitro. Clin. Sci. 1995. 88:141-147.
    View this article via: PubMed Google Scholar
21. Reinders, ME, Sho, M, Robertson, SW, Geehan, CS, Briscoe, DM. Proangiogenic function of CD40 ligand-CD40 interactions. J. Immunol. 2003. 171:1534-1541.
    View this article via: PubMed Google Scholar
22. Reinders, MEJ, Fang, JC, Wong, W, Ganz, P, Briscoe, DM. Expression patterns of vascular endothelial growth factor in human cardiac allografts: association with rejection. Transplantation. 2003. 76:224-230.
    View this article via: PubMed Google Scholar
23. Shahbazi, M, et al. Vascular endothelial growth factor gene polymorphisms are associated with acute renal allograft rejection. J. Am. Soc. Nephrol. 2002. 13:260-264.
    View this article via: PubMed Google Scholar
24. Pilmore, HL, Eris, JM, Painter, DM, Bishop, GA, McCaughan, GW. Vascular endothelial growth factor expression in human chronic renal allograft rejection. Transplantation. 1999. 67:929-933.
    View this article via: PubMed Google Scholar
25. Torry, RJ, Labarrere, CA, Torry, DS, Holt, VJ, Faulk, WP. Vascular endothelial growth factor expression in transplanted human hearts. Transplantation. 1995. 60:1451-1457.
    View this article via: PubMed Google Scholar
26. Kim, KJ, et al. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature. 1993. 362:841-844.
    View this article via: PubMed Google Scholar
27. Presta, LG, et al. Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer. Res. 1997. 57:4593-4599.
    View this article via: PubMed Google Scholar
28. Khan, IA, et al. IP-10 is critical for effector T cell trafficking and host survival in Toxoplasma gondii infection. Immunity. 2000. 12:483-494.
    View this article via: PubMed Google Scholar
29. O’Reilly, MS, et al. Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell. 1997. 88:277-285.
    View this article via: PubMed Google Scholar
30. Boehm, T, Folkman, J, Browder, T, O’Reilly, MS. Antiangiogenic therapy of experimental cancer does not induce acquired drug resistance. Nature. 1997. 390:404-407.
    View this article via: PubMed Google Scholar
31. Bergers, G, Javaherian, K, Lo, KM, Folkman, J, Hanahan, D. Effects of angiogenesis inhibitors on multistage carcinogenesis in mice. Science. 1999. 284:808-812.
    View this article via: PubMed Google Scholar
32. Tilton, RG, et al. Vascular dysfunction induced by elevated glucose levels in rats is mediated by vascular endothelial growth factor. J. Clin. Invest. 1997. 99:2192-2202.
    View this article via: JCI PubMed Google Scholar
33. Basu, S, et al. The neurotransmitter dopamine inhibits angiogenesis induced by vascular permeability factor/vascular endothelial growth factor. Nat. Med. 2001. 7:569-574.
    View this article via: PubMed Google Scholar
34. Walter, DH, et al. The in vivo bioactivity of vascular endothelial growth factor/vascular permeability factor is independent of N-linked glycosylation. Lab. Invest. 1996. 74:546-556.
    View this article via: PubMed Google Scholar
35. Dufour, JH, et al. IFN-γ-inducible protein 10 (IP-10; CXCL10)-deficient mice reveal a role for IP-10 in effector T cell generation and trafficking. J. Immunol. 2002. 168:3195-3204.
    View this article via: PubMed Google Scholar
36. Briscoe, DM, et al. The allogeneic response to cultured human skin equivalent in the hu-PBL-SCID mouse model of skin rejection. Transplantation. 1999. 67:1590-1599.
    View this article via: PubMed Google Scholar
37. Corry, RJ, Winn, HJ, Russell, PS. Primarily vascularized allografts of hearts in mice. The role of H-2D, H-2K, and non-H-2 antigens in rejection. Transplantation. 1973. 16:343-350.
    View this article via: PubMed Google Scholar
38. Lu, J, et al. Vascular endothelial growth factor expression and regulation of murine collagen-induced arthritis. J. Immunol. 2000. 164:5922-5927.
    View this article via: PubMed Google Scholar
39. Hancock, WW, et al. Donor-derived IP-10 initiates development of acute allograft rejection. J. Exp. Med. 2001. 193:975-980.
    View this article via: PubMed Google Scholar
40. Yamada, A, et al. CD28-independent costimulation of T cells in alloimmune responses. J. Immunol. 2001. 167:140-146.
    View this article via: PubMed Google Scholar
41. Gimbrone, MA. Culture of vascular endothelium. Prog. Hemost. Thromb. 1976. 3:1-28.
    View this article via: PubMed Google Scholar
42. Marelli-Berg, FM, Peek, E, Lidington, EA, Stauss, HJ, Lechler, RI. Isolation of endothelial cells from murine tissue. J. Immunol. Meth. 2000. 244:205-215.
43. Denton, MD, Geehan, C, Alexander, SI, Sayegh, MH, Briscoe, DM. Endothelial cells modify the costimulatory capacity of transmigrating leukocytes and promote CD28-mediated CD4+ T cell alloactivation. J. Exp. Med. 1999. 190:555-566.
    View this article via: PubMed Google Scholar
44. Murray, AG, et al. Human T-cell-mediated destruction of allogeneic dermal microvessels in a severe combined immunodeficient mouse. Proc. Natl. Acad. Sci. U. S. A. 1994. 91:9146-9150.
    View this article via: PubMed Google Scholar
45. Luster, AD, Jhanwar, SC, Chaganti, RS, Kersey, JH, Ravetch, JV. Interferon-inducible gene maps to a chromosomal band associated with a (4;11) translocation in acute leukemia cells. Proc. Natl. Acad. Sci. U. S. A. 1987. 84:2868-2871.
    View this article via: PubMed Google Scholar
46. Farber, JM. A macrophage mRNA selectively induced by γ-interferon encodes a member of the platelet factor 4 family of cytokines. Proc. Natl. Acad. Sci. U. S. A. 1990. 87:5238-5242.
    View this article via: PubMed Google Scholar
47. Briscoe, DM, Sayegh, MH. A rendezvous before rejection: where do T cells meet transplant antigens? Nat. Med. 2002. 8:220-222.
    View this article via: PubMed Google Scholar
48. Briscoe, DM, et al. Predictive value of inducible endothelial cell adhesion molecule expression for acute rejection of human cardiac allografts. Transplantation. 1995. 59:204-211.
    View this article via: PubMed Google Scholar
49. Hancock, WW. Chemokines and transplant immunobiology. J. Am. Soc. Nephrol. 2002. 13:821-824.
    View this article via: PubMed Google Scholar
50. Melter, M, Exeni, A, Briscoe, DM. Chemokines and their receptors in human clinical solid organ transplantation. Curr. Opin. Organ Transplant. 2002. 7:77-84.
51. Winn, R, Vedder, N, Ramamoorthy, C, Sharar, S, Harlan, J. Endothelial and leukocyte adhesion molecules in inflammation and disease. Blood Coagul. Fibrinolysis. 1998. 9:S17-S23.
    View this article via: PubMed Google Scholar
52. Denton, MD, et al. The role of the graft endothelium in transplant rejection: evidence that endothelial activation may serve as a clinical marker for the development of chronic rejection. Pediatr. Transplant. 2000. 4:252-260.
    View this article via: PubMed Google Scholar
53. Marumo, T, Schini-Kerth, VB, Busse, R. Vascular endothelial growth factor activates nuclear factor-κB and induces monocyte chemoattractant protein-1 in bovine retinal endothelial cells. Diabetes. 1999. 48:1131-1137.
    View this article via: PubMed Google Scholar
54. Lee, TH, Avraham, H, Lee, SH, Avraham, S. Vascular endothelial growth factor modulates neutrophil transendothelial migration via up-regulation of interleukin-8 in human brain microvascular endothelial cells. J. Biol. Chem. 2002. 277:10445-10451.
    View this article via: PubMed Google Scholar
55. Hancock, WW, et al. Requirement of the chemokine receptor CXCR3 for acute allograft rejection. J. Exp. Med. 2000. 192:1515-1520.
    View this article via: PubMed Google Scholar
56. Denton, MD, et al. Central role for CD40/CD40 ligand (CD154) interactions in transplant rejection. Pediatr. Transplant. 1998. 2:6-15.
    View this article via: PubMed Google Scholar
57. Sayegh, MH. Why do we reject a graft? Role of indirect allorecognition in graft rejection. Kidney Int. 1999. 56:1967-1979.
    View this article via: PubMed Google Scholar
58. Gjertson, D.W. 2000. Impact of delayed graft function and acute rejection on kidney graft survival. Clin. Transpl. 467–480.