Interactions Proangiogenic Function of CD40 Ligand-CD40

The Journal of Immunology

Proangiogenic Function of CD40 Ligand-CD40 Interactions1,2

Marlies E. J. Reinders, Masayuki Sho, Stuart W. Robertson, Christopher S. Geehan, and David M. Briscoe

Angiogenesis is a characteristic component of cell-mediated immune inflammation. However, little is known of the immunologic mediators of angiogenesis factor production. Interactions between CD40 ligand (CD40L) and CD40 have been shown to have pluripotent functions in inflammation, including the production of cytokines, chemokines, as well as the angiogenesis factor, vascular endothelial growth factor (VEGF), by endothelial cells. In this study we found that treatment of cultured human endothelial cells with an anti-CD40 Ab (to ligate CD40) resulted in the expression of several other angiogenesis factors, including fibroblast growth factor-2 and the receptors Flt-1 and Flt-4. To determine the proangiogenic effect of CD40L in vivo, human skin was allowed to engraft on SCID mice for 6 wk. These healed human skins express CD40 on resident endothelial cells and monocyte/macrophages, but not on CD20-expressing B cells. Skins were injected with saline, untransfected murine fibroblasts, or murine fibroblasts stably transfected with human CD40L. We found that the injection of CD40L-expressing cells, but not control cells, resulted in the in vivo expression of several angiogenesis factors (including VEGF and fibroblast growth factor) and a marked angiogenesis reaction. Mice treated with anti-VEGF failed to elicit an angiogenesis reaction in response to injection of CD40L- expressing cells, suggesting that the proangiogenic effect of CD40L in vivo is VEGF dependent. These observations imply that ligation of CD40 at a peripheral inflammatory site is of pathophysiological importance as a mediator of both angiogenesis and inflammation. The Journal of Immunology, 2003, 171: 1534–1541.

ngiogenesis, the generation of new blood vessels from 164, 145, and 121 aa, which arise from the alternative splicing of pre-existing ones, is a complex process involving endo- a single gene (12). Despite significant physical differences, the A thelial cell (EC)3 division, degradation of the vascular several VEGF isoforms bind VEGF receptors 1 and 2 (called Flt-1 basement membrane and surrounding extracellular matrix, and EC and KDR, respectively) and function as growth and survival fac- migration (1–3). It is a characteristic component of many normal tors for EC (2, 12, 13). However, only the VEGF164 isoform binds physiologic processes and is pathologic in various diseases, in- VEGF receptor 3 (also called neuopilin) (14), and recent studies cluding cancer, arthritis, atherosclerosis, and some forms of blind- suggests that there may be isoform-specific differences in function, ness (3, 4). Angiogenesis is regulated by a balance between the especially with regard to developmental angiogenesis (15–17). relative expression and function of pro- vs antiangiogenic media- VEGF may act synergistically with other factors, such as fibroblast tors. To this end, it is proposed that pathologic angiogenesis results growth factor (FGF), in vitro to induce endothelial growth and either from an overproduction of proangiogenic factors or an un- capillary formation (18–20). VEGF and FGF also act as survival derproduction or lack of function of antiangiogenic factors. factors for endothelium, and their withdrawal causes apoptosis in Multiple factors have been identified to possess angiogenic ef- vitro and regression of newly formed vessels in vivo (21). fects, but vascular endothelial growth factor (VEGF) is recognized It has been established for over 30 years that lymphocytes and as the most potent angiogenesis factor described to date. VEGF is monocytes produce angiogenesis factors and stimulate angiogen- a 45-kDa protein produced by EC, monocyte/macrophages, and a esis in the course of inflammatory responses (7, 22–24). For ex- variety of other cell types, including activated T cells (2, 5–10). As ample, activated T cells are a source of angiogenesis factors, in- its name suggests, VEGF stimulates EC proliferation in vitro and cluding VEGF, and can stimulate angiogenesis in association with angiogenesis in vivo, and it has important roles in both normal early lymphocyte recruitment (25). Products of activated macro- physiological as well as pathological vasculogenesis and angio- phages, including VEGF, TNF-␣, TGF-␤, NO, and FGF-2, are genesis (11). It exists in at least five different isoforms of 206, 189, most potent to induce angiogenesis (26–29). Thus, it is not sur- prising that angiogenesis is characteristic of delayed-type hyper- sensitivity (1) and that excessive production of angiogenesis fac- Division of Nephrology, Department of Medicine, Children’s Hospital, and Depart- ment of Pediatrics, Harvard Medical School, Boston, MA 02115 tors occurs in many chronic inflammatory disorders (11). Received for publication November 19, 2002. Accepted for publication May Interactions between CD40 ligand (CD40L; also called CD154) 23, 2003. and its counter-receptor CD40 are well established to promote T The costs of publication of this article were defrayed in part by the payment of page cell activation, T cell-B cell interactions, T cell-dependent macro- charges. This article must therefore be hereby marked advertisement in accordance phage activation, as well as endothelial activation (30–32). We with 18 U.S.C. Section 1734 solely to indicate this fact. recently demonstrated that interactions between CD40L and CD40 1 This work was supported by National Institutes of Health Grants AI46756 and AI50157 (to D.M.B.). M.R. was supported in part by a grant from Stichting Dr. were the most potent for the expression of VEGF in EC and mono- Catharine van Tussenbroek. cytes in vitro (33). Moreover, we found that injection of soluble 2 Address correspondence and reprint requests to Dr. David M. Briscoe, Division of CD40L into skin resulted in enhanced VEGF expression as well as Nephrology, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115. E-mail angiogenesis in the athymic nude mouse. CD40 signals also reg- address: [email protected] ulate VEGF expression in fibroblasts (34) and other cells (35) and 3 Abbreviations used in this paper: EC, endothelial cell; CD40L, CD40 ligand; FGF, fibroblast growth factor; PCNA, proliferating cell nuclear Ag; ; VEGF, vascular en- have been demonstrated to induce the expression of matrix met- dothelial growth factor; vWf, von Willebrand factor. alloproteinases of functional importance in EC tube formation

FIGURE 1. CD40 ligation induces FGF-2, Flt-4, and Flt-1 in EC in a time- and dose-dependent manner. Conﬂuent cultures of human EC were treated for 6 h with different concentrations of stimulating anti-CD40 (A, C, and D)or as a time course using 1 ␮g/ml anti- CD40 (B) as indicated. Total RNA was harvested, and the expression of angiogenesis factors was analyzed by RNase protection assay and PCR. Anti-CD40 induced the expression of FGF-2, Flt-4, and Flt-1 expression in EC (A and B), whereas several other angiogenesis factors and receptors had a consti- tutive level of expression and no detectable change after treatment with anti-CD40 (C and D). mRNA expression was quantiﬁed by densitometry as the relative expression of the angiogenesis factor compared with the expression of the housekeeping GAPDH signal. Bar graphs on the right of A and B illustrate the relative expression of each gene examined (mean Ϯ 1 SD of three blots, including the illustrated blots). The expression of GAPDH or ␤-actin served as internal housekeeping gene controls. All autoradiographs are representative of three experiments with similar results.

(36). In addition, some of the intracellular signaling pathways for Santa Cruz, CA). Humanized anti-human VEGF for in vivo studies was a CD40-dependent angiogenesis have recently been defined (35, 37). gift from Genentech (South San Francisco, CA). Human IgG was pur- Together, these in vitro studies point to a major function for chased from Sigma-Aldrich (St. Louis, MO). The proliferating cell nuclear Ag (PCNA) kit was purchased from Novocastra (Burlingame, CA). Murine CD40L-CD40 interactions in angiogenesis. fibroblasts stably transfected with human CD40L (CD40L cells) or un- In this study we extended upon these observations in an in vivo transfected cells (mock cells) were a gift from C. van Kooten (38). model and tested the mechanism of function of CD40L for angiogenesis in vivo. Our findings clearly demonstrate that the provision Cell culture of CD40L to tissues in vivo is a physiologic stimulus for the pro- Single-donor human umbilical vein EC were purchased from Clonetics duction of several angiogenesis factors, but that the CD40L-in- (Walkersville, MD) and were cultured in complete endothelial medium duced angiogenesis reaction is VEGF dependent. (EGM Bulletkit; Clonetics) as supplied and according to the recommended instructions. EC were subcultured and used at passages four to six.

Materials and Methods In vivo assessment of angiogenesis Reagents CB.17 SCID mice were purchased from Taconic Farms (Germantown, The following Abs were used in these studies: anti-human CD31, anti-von NY) and were used at 6–8 wk of age. The SCID mice used in these studies Willebrand factor (anti-vWf), anti-CD20, anti-CD68, and anti-mouse were occasionally monitored for “leakiness” by assessment of circulating T CD31 (DAKO, Carpenteria, CA); anti-human CD40 (G28.5) and CD40L cells, and in all cases were found to be immunodeﬁcient and stable. Full- (106; gifts from D. Hollenbaugh, Bristol Myers Squibb, Princeton, NJ); thickness human neonatal foreskin grafts were transplanted onto CB.17 and anti-human CD40L, VEGF, and FGF-2 (Santa Cruz Biotechnology, (SCID) mice as previously described (25, 39). Following engraftment for 1536 CD40-DEPENDENT ANGIOGENESIS

FIGURE 2. Ligation of CD40 induces angiogenesis in vivo. We used murine fibroblasts stably transfected with human CD40L to test the function of CD40L-CD40 interactions for the development of angiogenesis in vivo. A and B, Flow cytometry illustrating the expression of human CD40L on CD40L-transfected cells, but not on control-transfected mock cells. C–E, SCID mice bearing human skin transplants received intracutaneous injections of saline, mock cells, or CD40L cells as described in Materials and Methods. Skin specimens were harvested after 7 days and were evaluated for angiogenesis by immunohistochemical staining for vWf. Illustrated is representative staining of vWf (rose-red color) in a skin injected with mock cells (C) or CD40L cells (D).E, The mean vessel density was evaluated in skin grafts by standard calibrated grid counting of vWf-positive vessels at ϫ400 magnification. Bars illustrate the vessel counts from skin specimens injected with saline (n ϭ 4), mock cells (n ϭ 8), or CD40L cells (n ϭ 11) Ϯ 1 SD. F and G, Skin specimens were stained for PCNA to identify cells undergoing active proliferation. Shown is enhanced PCNA staining of EC in a skin treated with CD40L cells (G) compared with skin treated with mock cells (F). Magnification: C and D, ϫ400; F and G, ϳϫ800.

4–6 wk, mice were divided into different experimental groups. In each RT-PCR and RNase protection assays group the human skin was injected intracutaneously with 35 ␮l of Matrigel (growth factor-reduced Matrigel; BD Biosciences, Bedford, MA) and ei- RNA was isolated from cultured EC or skin samples using the Ultraspec ther 35 ␮l of saline (as a control) or 2.5 ϫ 106 murine fibroblasts stably RNA isolation system (Biotecx, Houston, TX) and was reverse transcribed, transfected with human CD40L (CD40L cells) or mock transfectants and PCR was performed using standard techniques (33). Sequence-specific (mock cells) each in 35 ␮l of saline. Before injection, mock and CD40L primers for PCR of human KDR (also called VEGF receptor-2) were: Ј Ј Ј cells were irradiated at 7500 rad. sense, 5 -CGAAGCATCAGCATAAGAAAC-3 ; and antisense, 5 - Ј ␤ SCID mice were also treated with a neutralizing anti-human VEGF Ab ACATACACAACCAGAG-AGACC-3 . -Actin (Stratagene, La Jolla, (Genentech) or human IgG at a dose of 5 mg/kg in 100 ␮l of saline i.p. Ab CA) was used as an internal control. The PCR conditions were 94°C for 5 therapy was begun 2 days before and every other day after the injection of min, followed by 35 cycles at 94°C for 1 min, 60°C for 1 min, and 72°C CD40L cells into the skin. Animals were sacrificed, and skin grafts were for 1 min. The last cycle was extended to 7 min at 72°C. The amplified harvested after 7 days. Animal care and anesthesia were performed in products were resolved by electrophoresis in an ethidium bromide-stained compliance with guidelines established by the animal care and use com- 1.5% agarose gel. mittee at Children’s Hospital (Boston, MA). RNase protection assays were performed using the RiboQuant multi- probe template system (BD PharMingen, San Diego, CA), according to the Immunohistochemistry manufacturer’s instructions and as previously described (33). Briefly, equal amounts of RNA (10 ␮g) were hybridized with [32P]UTP-labeled ribo- Frozen specimens were fixed in acetone, and formalin-fixed specimens probes that were synthesized from the template. After hybridization, sam- were deparaffined. All specimens were then incubated in primary and sec- ples were digested with RNase A, RNase T1, and proteinase K, and pro- ondary Abs as previously described (25, 33) using the Vectastain Kit (Vec- tected RNA was resolved on a denaturing polyacrylamide gel. Relative tor Laboratories, Burlingame, CA). Finally, specimens were counterstained signals were detected by autoradiography with Kodak MR film (Eastman in Gill’s hematoxylin and mounted in glycerol gelatin. The PCNA kit was Kodak, Rochester, NY), and expression was quantified by densitometry by used according to the manufacturer’s instructions. means of an AlphaImager 2000 system (Alpha Innotech, San Leandro, CA). For quantification, signals were standardized to the internal house- Quantification of angiogenesis keeping gene GAPDH. Vessels were identified by staining for human vWf, and angiogenesis was Statistical analyses quantified by a standard grid-counting method at ϫ400 magnification as previously described (40). Four to six adjacent nonoverlapping fields of Data were compared by nonparametric analysis using the Mann-Whitney each specimen were analyzed in a blinded manner, and the mean vessel test for data analysis. Differences with p Ͻ 0.05 were considered statisti- counts for each specimen were calculated. cally significant. The Journal of Immunology 1537

FIGURE 3. Expression pattern of CD40-expressing cells within skin grafts: representative immunohistochemical staining of CD40, CD40L, CD20, and CD68 in untreated skins (Untreated) or skins treated with CD40L-transfected cells (Treated). CD40 (upper panels) was expressed on EC and resident macrophages in all skins examined. There was no difference in the pattern or intensity of expression of CD40 in untreated or treated skins. In contrast, CD40L was present only in the skins treated with the CD40L transfectants, where its expression was most apparent on groups of cells in clusters (middle panel, right). There was no staining of CD20 in any of the skin samples (middle panel, left), whereas CD68-expressing monocyte/macrophages (lower panels) were present in all skins examined. The expression of CD68 was similar in mock- and CD40L-treated skins. Data are representative of three skins examined in each group. Magniﬁcation of micrographs: ϫ200 (upper left panel) and ϫ400.

Results with human CD40 ligand (CD40L cells; Fig. 2, A and B) were Ligation of CD40 induces FGF-2, Flt-4, and Flt-1 mRNA injected intracutaneously into the healed human skins on the SCID expression in human EC in vitro mice. Injection of saline served as a negative control. The skins were harvested after 7 days and were evaluated for the develop- We first assessed the effect of CD40L-CD40 interactions on the ment of angiogenesis. By quantitative grid counting, we found that expression of a panel of angiogenesis factors and receptors in cul- the number of vessels in skins injected with saline or mock cells tured EC in vitro using RNase protection assays. Human EC were was comparable to that in untreated skins, as determined by stan- starved overnight in 0.5% FCS and were subsequently treated with stimulating anti-CD40. Anti-CD40 induced the expression of dard immunostaining with vWF. In contrast, the mean vessel den- FGF-2, Flt-1 and Flt-4 in a dose- and a time-dependent manner sity was significantly increased in the skins injected with CD40L (Fig. 1, A and B). FGF-2, Flt-1, and Flt-4 increased in expression cells vs controls (2.3-fold vs saline and 2.1-fold vs mock cells, Ͻ as early as 3 h following activation with anti-CD40, with peak both p 0.01; Fig. 2, C–E). Similar to untreated human skins (39), expression occurring after 6 h (Fig. 1B). Similar results were ob- the skins injected with mock-transfectant cells had an approxi- tained following treatment with soluble CD40L, and under all con- mately equal number of human and mouse EC, as determined by ditions resulted in the expression of VEGF (data not shown) (33). immunohistochemical analysis of human vs mouse CD31. In con- In contrast, the mRNA expression of angiopoietin, endoglin, trast, we found a greater percentage of cells expressing human CD31, TIE, TIE2, thrombin receptor, and KDR was unaltered after CD31 in skins treated with CD40L cells (not shown). Furthermore, treatment with anti-CD40 (Fig. 1, C and D). These findings suggest in the CD40L-injected skins there were numerous EC expressing that CD40L-CD40 interactions have the potential to be potent for PCNA in their nucleus, a marker that identifies cells that have angiogenesis in vivo. entered the cell cycle (Fig. 2, F and G). This suggests active angiogenesis in skins that received CD40L cells. Infiltration of tissues with CD40L-expressing cells induces We also wished to establish which intragraft cells may respond angiogenesis in vivo to the injected CD40L cells. By immunohistochemistry, CD40 was We next assessed whether cell surface CD40L on infiltrating cells expressed by multiple cells in untreated skins, including EC, res- might function to mediate immune-mediated angiogenesis in vivo. ident macrophages, and keratinocytes with a similar pattern of ex- It is well established that activated platelets and lymphocytes expression in all groups of skins (Fig. 3). CD68-expressing mono- pressing CD40L are present in tissues in cell-mediated immune cyte/macrophages were present in both untreated and treated skins inflammation. We used an established model of angiogenesis in in variable numbers, but overall there appeared to be slightly more which human skin is transplanted onto SCID mice (25). Skin is cells in the mock- and CD40L transfectant-injected skins than in allowed to heal for 4–6 wk to enable stabilization of the healing uninjected skins (Fig. 3). In contrast, we did not find any CD20- angiogenesis reaction in the skin graft (39). Untransfected murine expressing B cells in any of the untreated, mock-injected, or fibroblasts (mock cells) or murine fibroblasts stably transfected CD40L transfectant-injected skins. CD40L was minimal or absent 1538 CD40-DEPENDENT ANGIOGENESIS in untreated skins and in skins injected with mock transfectant the mice were treated with either anti-human VEGF or IgG as a cells. In contrast, staining for CD40L was found on discrete clus- control by i.p. injection. There was no change in baseline vessel ters of cells in the skins injected with CD40L transfectants (Fig. 3). density in anti-VEGF or IgG-treated mice when skins were treated By double immunohistochemical staining (not shown), there was with mock cells (not shown). In contrast, skins treated with CD40L no distinct pattern of contact between the CD40L-expressing clus- cells in IgG-treated mice had significantly increased vessel num- ters and EC or macrophages. We interpret this to suggest that the bers over baseline saline-treated skins (Fig. 6; p Ͻ 0.01). However, injected CD40L-expressing cells may interact with several resident injection of CD40L cells into skins on mice treated with anti- CD40-expressing skin cells spatially associated with the CD40L VEGF failed to result in an angiogenesis reaction (Fig. 6; p Ͻ cell clusters. Alternatively, it is possible that CD40L shed from 0.03). Thus, the effect of CD40L for the stimulation of an angio- these cells may bind the several CD40-expressing cells within the genesis reaction involves the critical expression of VEGF. skin to mediate the angiogenesis reaction. Discussion CD40L induces the expression of multiple angiogenesis factors In this study we show that local infiltration of skin with cells ex- in vivo pressing human CD40L on their cell surface mediates the induc- We next examined the mRNA expression of a panel of angiogen- tion of VEGF and FGF in vivo and results in a marked angiogenesis factors and receptors (including VEGF, FGF, and angiopoe- esis reaction. These findings have important implications for the itin-1) in skins harvested 7 days after intracutaneous injection of function of CD40L-CD40 interactions in immunity and chronic saline, mock cells, or CD40L cells. There was a low and variable inflammation. First, it is well established that CD40L-CD40 inter- expression of all factors in skins injected with saline or mock cells actions are of central importance in the development of an effective (Fig. 4, A–C). In contrast, the angiogenesis factors VEGF, FGF, immune response involving T cell activation, T cell-B cell inter- and angiopoeitin-1 were significantly up-regulated after treatment actions, and Ab switching as well as endothelial activation re- with CD40L cells. Also, the angiogenesis receptors TIE-2, Flt-4, sponses (30–32). Our findings here extend the functions of and Flt-1 as well as the EC markers endoglin and CD31 were CD40L-CD40 to include regulation of VEGF and FGF as well as significantly increased in skins injected with CD40L cells (Fig. 4, the angiogenesis receptors Flt-1 and Flt-4 in vitro and in vivo. A–C). Second, inflammatory cells, including lymphocytes and mono- By immunohistochemistry, little FGF-2 and VEGF was found in cytes, can elicit an angiogenesis reaction, a process called leuko- saline- and mock cell-treated skins (Fig. 5, A–D). In contrast, a cyte-induced angiogenesis (24, 25). This process is well charac- marked induction of both FGF-2 and VEGF expression was found terized in the literature and is reported to be of pathological in skins treated with CD40L cells (Fig. 5, E and F). FGF-2 and significance in several chronic inflammatory diseases, including VEGF were expressed mainly by vascular EC. Thus, it is likely arthritis, atherosclerosis, and diabetes (1, 3, 42–44). Our finding that CD40L-induced angiogenesis in vivo involves the expression here that CD40L-expressing cells within a tissue can functionally and function of multiple angiogenesis factors. induce VEGF-dependent angiogenesis is consistent with these re- ports and further suggests that CD40L-CD40 interactions may rep- Function of VEGF in CD40L-induced angiogenesis resent an intermediary between cell-mediated immune inflamma- It is well established that angiogenesis factors mediate angiogen- tion and angiogenesis. esis synergistically with VEGF (41). Thus, it is possible that Activated T cells and platelets express CD40L, and CD40 is VEGF is a critical factor in CD40L-induced angiogenesis. To test expressed on multiple cell types, including EC, fibroblasts, mono- this possibility, human skins on SCID mice were injected with cytes, and epithelial cells (45, 46). Our findings demonstrate that either saline or CD40L cells, as described above. Simultaneously, CD40L promotes the functional expression of several angiogenesis

FIGURE 4. Ligation of CD40 induces mRNA expression of multiple angiogenic factors in vivo. A and B, Analysis of angiogenesis factor expression by RNase protection assay from skins treated with saline, mock cells, or CD40L cells.C, mRNA expression was quantiﬁed by densitometry as the relative expression of the angiogenesis factor compared with the expression of the housekeeping GAPDH signal. Bar graphs represent the mean expression (Ϯ1 SD) of three skin specimens injected with saline (control), mock cells, .p Ͻ 0.05 ,ء .or CD40L cells The Journal of Immunology 1539

FIGURE 5. Ligation of CD40 induces the expression of FGF-2 and VEGF in vivo. A, Immuno- histochemical staining of FGF-2 and VEGF in vivo in representative human skin grafts on SCID mice following intracutaneous injection of saline (A and B), mock cells (C and D), or CD40L cells (E and F). There is enhanced expression of FGF-2 and VEGF protein in skins treated with CD40L cells compared with the control skins. Data are representative of 10 similar experiments. Magniﬁcation of A–F, ϫ400. B, A representative skin from each group was evaluated for VEGF mRNA expression by RT-PCR. The expression of VEGF164 and VEGF121 (upper and lower bands) is increased in skins treated with CD40L cells. The expression of ␤-actin is illustrated as an internal control. Data are representative of two experiments with similar results.

factors in vivo. These findings are consistent with in vitro obser- as allografts undergoing rejection (49, 50). Interestingly, these vations by several groups that CD40L-CD40 interactions may re- same disease processes have been found to be associated with sult in angiogenesis via enhanced VEGF, tissue factor, or matrix intense VEGF expression (43, 44). Moreover, FGF has been metalloproteinase expression and via signaling through the phos- shown to act directly on smooth muscle cells, which might have phoinositol 3-kinase/Akt pathway (33–37, 47). However, despite important implications for inflammatory disorders, such as athero- these observations, a recent report by Urbich et al. (48) suggested sclerosis and graft arteriosclerosis (51). Furthermore, blockade of that CD40L may have antiangiogenesis effects. In the report the CD40L-CD40 interactions has been found to prevent the develop- authors cited that a possible reason for their different findings was ment of acute and chronic inflammation, including allograft rejec- the cell culture conditions used in their analysis. Nevertheless, our tion and atherosclerosis (52–56). It is thus possible that these ef- studies here demonstrate that cell surface CD40L is sufficient to fects may in part be related to the CD40L-dependent angiogenesis facilitate angiogenesis factor expression and angiogenesis in vivo. response. Therefore, our findings here that local infiltration of Thus, we suggest that it is most likely that the physiological effect CD40L-expressing cells at a peripheral inflammatory site mediate of CD40L-CD40 interactions is proangiogenic. However, in light both angiogenesis factor production and angiogenesis are probably of the report by Urbich (48), it will be important to determine whether there is a factor or physiologic condition that can limit or of pathophysiological importance. inhibit this proangiogenic effect of CD40L. In conclusion, in this report we define a role for CD40L-CD40 The expression of CD40 has been reported to be prominent in interactions in angiogenesis in vitro and in vivo, and we suggest processes known to be associated with angiogenesis and inflam- that a major component of immune-mediated angiogenesis is mation (32, 45, 46). These include atherosclerosis, arthritis, as well VEGF dependent. Our findings imply that the infiltration of cells 1540 CD40-DEPENDENT ANGIOGENESIS

gration, through phosphatidylinositol 3-kinase-dependent pathways. J. Biol. Chem. 276:26969. 14. Soker, S., S. Takashima, H. Q. Miao, G. Neufeld, and M. Klagsbrun. 1998. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for vascular endothelial growth factor. Cell 92:735. 15. Carmeliet, P., Y. S. Ng, D. Nuyens, G. Theilmeier, K. Brusselmans, I. Cornelissen, E. Ehler, V. V. Kakkar, I. Stalmans, V. Mattot, et al. 1999. Im- paired myocardial angiogenesis and ischemic cardiomyopathy in mice lacking the vascular endothelial growth factor isoforms VEGF164 and VEGF188. Nat. Med. 5:495. 16. Stalmans, I., Y. S. Ng, R. Rohan, M. Fruttiger, A. Bouche, A. Yuce, H. Fujisawa, B. Hermans, M. Shani, S. Jansen, et al. 2002. Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms. J. Clin. Invest. 109:327. 17. Stalmans, I., D. Lambrechts, F. De Smet, S. Jansen, J. Wang, S. Maity, P. Kneer, M. Von Der Ohe, A. Swillen, C. Maes, et al. 2003. VEGF: a modifier of the del22q11 (DiGeorge) syndrome? Nat. Med. 9:173. 18. Pepper, M. S., N. Ferrara, L. Orci, and R. Montesano. 1992. Potent synergism FIGURE 6. VEGF is a critical angiogenesis factor in CD40L-induced between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitro. Biochim. Biophys. Acta 189:824. angiogenesis. SCID mice bearing human skin grafts were treated with a 19. Couper, L. L., S. R. Bryant, J. Eldrup-Jorgensen, C. E. Bredenberg, and neutralizing anti-VEGF Ab (n ϭ 9) or control IgG (n ϭ 4) by i.p. injection V. Lindner. 1997. Vascular endothelial growth factor increases the mitogenic (5 mg/kg on days Ϫ2, 0, 2, 4, and 6). Skins were treated with either saline response to fibroblast growth factor-2 in vascular smooth muscle cells in vivo via as a control (n ϭ 4) or CD40L-transfected cells (n ϭ 9). The mean vessel expression of fms-like tyrosine kinase-1. Circ. Res. 81:932. 20. Seghezzi, G., S. Patel, C. J. Ren, A. Gualandris, G. Pintucci, E. S. Robbins, density was evaluated in skin grafts by standard calibrated grid counting of R. L. Shapiro, A. C. Galloway, D. B. Rifkin, and P. Mignatti. 1998. Fibroblast vWf-positive vessels at ϫ400 magnification. Bars illustrate the mean ves- growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) ex- sel counts Ϯ 1 SEM. The CD40L cells failed to elicit an angiogenesis pression in the endothelial cells of forming capillaries: an autocrine mechanism reaction in animals treated with anti-VEGF Ab. contributing to angiogenesis. J. Cell Biol. 141:1659. 21. Alon, T., I. Hemo, A. Itin, J. Pe’er, J. Stone, and E. Keshet. 1995. Vascular endothelial growth factor acts as a survival factor for newly formed retinal vessels and has implications for retinopathy of prematurity. Nat. Med. 1:1024. 22. Sidky, Y. A., and R. Auerbach. 1975. Lymphocyte-induced angiogenesis: a quantitative and sensitive assay of the graft-vs.-host reaction. J. Exp. Med. 141:1084. expressing CD40L at a peripheral site of inflammation is mecha- 23. Polverini, P. J., P. S. Cotran, M. A. Gimbrone, and E. R. Unanue. 1977. Activated nistic for an interplay between the immune response and macrophages induce vascular proliferation. Nature 269:804. angiogenesis. 24. Auerbach, R., and Y. A. Sidky. 1979. Nature of the stimulus leading to lymphocyte-induced angiogenesis. J. Immunol. 123:751. 25. Moulton, K. S., R. J. Melder, V. R. Dharnidharka, J. Hardin-Young, R. K. Jain, Acknowledgments and D. M. Briscoe. 1999. Angiogenesis in the huPBL-SCID model of human We thank William Wong for technical assistance, Ingrid Vos, Ph.D., for transplant rejection. Transplantation 67:1626. 26. Leibovich, S. J., P. J. Polverini, H. M. Shepard, D. M. Wiseman, V. Shively, and help with animal studies, and Cees van Kooten for the gift of CD40L- N. Nuseir. 1987. Macrophage-induced angiogenesis is mediated by tumour ne- transfected fibroblasts. crosis factor-␣. Nature 329:630. 27. Koch, A. E., P. J. Polverini, and S. J. Leibovich. 1986. Induction of neovascu- larization by activated human monocytes. J. Leukocyte Biol. 39:233. References 28. Wiseman, D. M., P. J. Polverini, D. W. Kamp, and S. J. Leibovich. 1988. Trans- 1. Cotran, R. S. 1994. Inflammation and repair. In Pathologic Basis of Disease, forming growth factor-␤ (TGF␤) is chemotactic for human monocytes and in- Chapt. 3. R. S. Cotran, V. Kumar, and S. L. Robbins, eds. Saunders, Philidelphia, duces their expression of angiogenic activity. Biochim. Biophys. Acta 157:793. pp. 51–92. 29. Leibovich, S. J., P. J. Polverini, T. W. Fong, L. A. Harlow, and A. E. Koch. 1994. 2. Brown, L. F., M. Detmar, K. Claffey, J. A. Nagy, D. Feng, A. M. Dvorak, and Production of angiogenic activity by human monocytes requires an L-arginine/ H. F. Dvorak. 1997. Vascular permeability factor/vascular endothelial growth nitric oxide-synthase-dependent effector mechanism. Proc. Natl. Acad. Sci. USA factor: a multifunctional angiogenic cytokine. EXS 79:233. 91:4190. 3. Folkman, J. 1995. Seminars in Medicine of the Beth Israel Hospital, Boston: 30. Yellin, M. J., J. Brett, D. Baum, A. Matsushima, M. Szabolcs, D. Steern, and clinical applications of research on angiogenesis. N. Engl. J. Med. 333:1757. L. Chess. 1995. Functional interactions of T cells with endothelial cells: the role 4. Carmeliet, P., and R. K. Jain. 2000. Angiogenesis in cancer and other diseases. of CD40L-CD40-mediated signals. J. Exp. Med. 182:1857. Nature 407:249. 31. Alderson, M. R., R. J. Armitage, T. W. Tough, L. Strockbine, W. C. Fanslow, and 5. Leung, D. W., G. Cachianes, W. J. Kuang, D. V. Goeddel, and N. Ferrara. 1989. M. K. Spriggs. 1993. CD40 expression by human monocytes: regulation by cy- Vascular endothelial growth factor is a secreted angiogenic mitogen. Science tokines and activation of monocytes by the ligand for CD40. J. Exp. Med. 178: 246:1306. 669. 6. Tischer, E., R. Mitchell, T. Hartman, M. Silva, D. Gospodarowicz, J. C. Fiddes, 32. Grewal, I. S., and R. A. Flavell. 1996. A central role of CD40 ligand in the and J. A. Abraham. 1991. The human gene for vascular endothelial growth factor: ϩ regulation of CD4 T-cell responses. Immunol. Today 17:410. multiple protein forms are encoded through alternative exon splicing. J. Biol. Chem. 266:11947. 33. Melter, M., M. E. Reinders, M. Sho, S. Pal, C. Geehan, M. D. Denton, 7. Polverini, P. J. 1997. Role of the macrophage in angiogenesis-dependent dis- D. Mukhopadhyay, and D. M. Briscoe. 2000. Ligation of CD40 induces the eases. EXS 79:11. expression of vascular endothelial growth factor by endothelial cells and mono- 8. Bottomley, M. J., N. J. Webb, C. J. Watson, P. J. Holt, A. J. Freemont, and cytes and promotes angiogenesis in vivo. Blood 96:3801. P. E. Brenchley. 1999. Peripheral blood mononuclear cells from patients with 34. Cho, C. S., M. L. Cho, S. Y. Min, W. U. Kim, D. J. Min, S. S. Lee, S. H. Park, rheumatoid arthritis spontaneously secrete vascular endothelial growth factor J. Choe, and H. Y. Kim. 2000. CD40 engagement on synovial fibroblast up- (VEGF): specific up-regulation by tumour necrosis factor-alpha (TNF-␣)insy- regulates production of vascular endothelial growth factor. J. Immunol. 164:5055. novial fluid. Clin. Exp. Immunol. 117:171. 35. Tai, Y. T., K. Podar, D. Gupta, B. Lin, G. Young, M. Akiyama, 9. Freeman, M. R., F. X. Schneck, M. L. Gagnon, C. Corless, S. Soker, K. Niknejad, K. C. and Anderson. 2002. CD40 activation induces p53-dependent vascular G. E. Peoples, and M. Klagsbrun. 1995. Peripheral blood T lymphocytes and endothelial growth factor secretion in human multiple myeloma cells. Blood 99: lymphocytes infiltrating human cancers express vascular endothelial growth fac- 1419. tor: a potential role for T cells in angiogenesis. Cancer Res. 55:4140. 36. Mach, F., U. Schonbeck, R. P. Fabunmi, C. Murphy, E. Atkinson, 10. Peoples, G. E., S. Blotnick, K. Takahashi, M. R. Freeman, M. Klagsbrun, and J. Y. Bonnefoy, P. Graber, and P. Libby. 1999. T lymphocytes induce endothelial T. J. Eberlein. 1995. T lymphocytes that infiltrate tumors and atherosclerotic cell matrix metalloproteinase expression by a CD40L-dependent mechanism: im- plaques produce heparin-binding epidermal growth factor-like growth factor and plications for tubule formation. Am. J. Pathol. 154:229. basic fibroblast growth factor: a potential pathologic role. Proc. Natl. Acad. Sci. 37. Deregibus, M. C., S. Buttiglieri, S. Russo, B. Bussolati, and G. Camussi. 2002. USA 92:6547. CD40-dependent activation of phosphatidylinositol 3-kinase/Akt pathway medi- 11. Ferrara, N. 2000. VEGF: an update on biological and therapeutic aspects. Curr. ates endothelial cell survival and in vitro angiogenesis. J. Biol. Chem. 277:25195. Opin. Biotechnol. 11:617. 38. van Kooten, C., X. van der Linde, A. M. Woltman, L. A. van Es, and M. R. Daha. 12. Ferrara, N., K. A. Houck, L. B. Jakeman, J. Winer, and D. W. Leung. 1991. The 1999. Synergistic effect of interleukin-1 and CD40L on the activation of human vascular endothelial growth factor family of polypeptides. J. Cell Biochem. 47: renal tubular epithelial cells. Kidney Int. 56:41. 211. 39. Briscoe, D. M., V. R. Dharnidharka, C. Isaacs, G. Downing, S. Prosky, P. Shaw, 13. Zeng, H., H. F. Dvorak, and D. Mukhopadhyay. 2001. Vascular permeability N. L. Parenteau, and J. Hardin-Young. 1999. The allogeneic response to cultured factor (VPF)/vascular endothelial growth factor (VEGF) receptor-1 down-mod- human skin equivalent in the hu-PBL-SCID mouse model of skin rejecton. Trans- ulates VPF/VEGF receptor-2-mediated endothelial cell proliferation, but not mi- plantation 67:1590. The Journal of Immunology 1541

40. Weidner, N., J. P. Semple, W. R. Welch, and J. Folkman. 1991. Tumor angio- 49. Mach, F., U. Schonbeck, and P. Libby. 1998. CD40 signaling in vascular cells: genesis and metastasis: correlation in invasive breast carcinoma. N. Engl. J. Med. a key role in atherosclerosis? Atherosclerosis 137(Suppl.):S89. 324:1. 50. Reul, R. M., J. C. Fang, M. D. Denton, C. Geehan, C. Long, R. N. Mitchell, 41. Ng, Y. S., and P. A. D’Amore. 2001. Therapeutic angiogenesis for cardiovascular P. Ganz, and D. M. Briscoe. 1997. CD40 and CD40 ligand (CD154) are coex- disease. Curr. Control Trials Cardiovasc. Med. 2:278. pressed on microvessels in vivo in human cardiac allograft rejection. Transplan- 42. Folkman, J. 1995. Angiogenesis in cancer, vascular, rheumatoid and other dis- tation 64:1765. ease. Nat. Med. 1:27. 51. Hansson, G. K., P. Libby, U. Schonbeck, and Z. Q. Yan. 2002. Innate and adap- 43. Celletti, F. L., J. M. Waugh, P. G. Amabile, A. Brendolan, P. R. Hilﬁker, and tive immunity in the pathogenesis of atherosclerosis. Circ. Res. 91:281. M. D. Dake. 2001. Vascular endothelial growth factor enhances atherosclerotic 52. Larsen, C. P., D. Z. Alexander, D. Hollenbaugh, E. T. Elwood, S. C. Ritchie, plaque progression. Nat. Med. 7:425. A. Aruffo, R. Hendrix, and T. C. Pearson. 1996. CD40-gp39 interactions play a 44. Lu, J., T. Kasama, K. Kobayashi, Y. Yoda, F. Shiozawa, M. Hanyuda, critical role during allograft rejection. Transplantation 61:4. M. Negishi, H. Ide, and M. Adachi. 2000. Vascular endothelial growth factor 53. Larsen, C. P., E. T. Elwood, D. Z. Alexander, S. C. Ritchie, R. Hendrix, expression and regulation of murine collagen-induced arthritis. J. Immunol. 164: C. Tucker-Burden, H. R. Cho, A. Aruffo, D. Hollenbaugh, P. S. Linsley, et al. 5922. 1996. Long-term acceptance of skin and cardiac allografts after blocking CD40 45. Denton, M. D., R. M. Reul, V. R. Dharnidharka, J. C. Fang, P. Ganz, and and CD28 pathways. Nature 381:434. D. M. Briscoe. 1998. Central role for CD40/CD40 ligand (CD154) interactions in 54. Hancock, W. W., M. H. Sayegh, X. G. Zheng, R. Peach, P. S. Linsley, and transplant rejection. Pediatr. Transplant. 2:6. L. A. Turka. 1996. Costimulatory function and expresson of CD40 ligand, CD80, 46. Van Kooten, C., and J. Banchereau. 1996. CD40-CD40-ligand: a multifunctional CD86 in vascularized murine cardiac allograft rejection. Proc. Natl. Acad. Sci. receptor-ligand pair. Adv. Immunol. 61:1. USA 93:13967. 47. Mach, F., U. Schonbeck, J. Y. Bonnefoy, J. S. Pober, and P. Libby. 1997. Ac- 55. Kirk, A. D., D. M. Harlan, N. N. Armstrong, T. A. Davis, Y. Dong, G. S. Gray, tivation of monocyte/macrophage functions related to acute atheroma complica- X. Hong, D. Thomas, J. H. Fechner, Jr., and S. J. Knechtle. 1997. CTLA4-Ig and tion by ligation of CD40: induction of collagenase, stromelysin, and tissue factor. anti CD40 ligand prevent renal allograft rejection in primates. Proc. Natl. Acad. Circulation 96:396. Sci. USA 94:8789. 48. Urbich, C., E. Dernbach, A. Aicher, A. M. Zeiher, and S. Dimmeler. 2002. CD40 56. Schonbeck, U., G. K. Sukhova, K. Shimizu, F. Mach, and P. Libby. 2000. Inhi- ligand inhibits endothelial cell migration by increasing production of endothelial bition of CD40 signaling limits evolution of established atherosclerosis in mice. reactive oxygen species. Circulation 106:9816. Proc. Natl. Acad. Sci. USA 97:7458.