Human dendritic cell maturation and cytokine secretion upon stimulation with Bordetella pertussis filamentous haemagglutinin.

In addition to antibodies, Th1-type T cell responses are also important for long-lasting protection against pertussis. However, upon immunization with the current acellular vaccines, many children fail to induce Th1-type responses, potentially due to immunomodulatory effects of some vaccine antigens, such as filamentous haemagglutinin (FHA). We therefore analysed the ability of FHA to modulate immune functions of human monocyte-derived dendritic cells (MDDC). FHA was purified from pertussis toxin (PTX)-deficient or from PTX- and adenylate cyclase-deficient Bordetella pertussis strains, and residual endotoxin was neutralized with polymyxin B. FHA from both strains induced phenotypic maturation of human MDDC and cytokine secretion (IL-10, IL-12p40, IL-12p70, IL-23 and IL-6). To identify the FHA domains responsible for MDDC immunomodulation, MDDC were stimulated with FHA containing a Gly→Ala substitution at its RGD site (FHA-RAD) or with an 80-kDa N-terminal moiety of FHA (Fha44), containing its heparin-binding site. Whereas FHA-RAD induced maturation and cytokine production comparable to those of FHA, Fha44 did not induce IL-10 production, but maturated MDDC at least partially. Nevertheless, Fha44 induced the secretion of IL-12p40, IL-12p70, IL-23 and IL-6 by MDDC, albeit at lower levels than FHA. Thus, FHA can modulate MDDC responses in multiple ways, and IL-10 induction can be dissociated from the induction of other cytokines.

[Immunobiological properties of Bordetella pertussis lipopolysaccharide in the acellular pertussis vaccine].

AIM: Study of Bordetella pertussis lipopolysaccharide (LPS) immunobiological properties in the acellular pertussis vaccine. MATERIALS AND METHODS: Experimental series of acellular pertussis vaccines (APV), lyophilized LPS were used. Antibody titers against LPS in mice sera were evaluated by using EIA with peroxidase conjugate of anti-species antibodies against mice IgG. LPS activity in B. pertussis antigen complex preparations was determined in quantitative chromogenic LAL-test by end point. APV protective activity was determined in mice test during intracerebral infection by B. pertussis strain No. 18323 virulent culture. APV safety was determined in the mice body weight change test. RESULTS: The presence of LPS in APV was shown in immune electrophoresis with purified B. pertussis LPS preparation as a control. Formalin treatment changes immunochemical properties of APV LPS that lead to the shift of precipitation bands with pertussis agglutinating sera from the start zone into cathode. The quantity of LPS in pertussis culture supernatants was on average 49050 +/- 6774 endotoxin units per ml (EU/ml). In APV preparations the quantity of LPS was on average 906 +/- 90 EU/ml, i.e. decreased by more than 50 times. An increase of antibody titers against B. pertussis LPS in mice sera after the APV immunization was shown in EIA, which gives evidence of its presence in immunogenic form in the complex preparations. The preclinical studies carried out show protective activity and specific safety of the experimental APV series. CONCLUSION: Formalin-neutralized APV preparation is a complex of protein antigens in association with LPS. Formalin treatment results in modification of LPS molecule that retains antigenic properties but is significantly less toxic.

Modulation of the host interferon response and ISGylation pathway by B. pertussis filamentous hemagglutinin.

Bordetella pertussis filamentous hemagglutinin (FHA) is a surface-associated and secreted protein that serves as a crucial adherence factor, and displays immunomodulatory activity in human peripheral blood mononuclear cells (PBMCs). In order to appreciate more fully the role of secreted FHA in pathogenesis, we analyzed FHA-induced changes in genome-wide transcript abundance in human PBMCs. Among the 683 known unique genes with greater than 3-fold change in transcript abundance following FHA treatment, 125 (18.3%) were identified as interferon (IFN)-regulated. Among the latter group were genes encoding several members of the IFN type I response, as well as 3 key components of the ISGylation pathway. Using real-time RT-PCR, we confirmed FHA-associated increases in transcript abundance for the genes encoding ubiquitin-like protein, ISG15, and its specific protease USP18. Western-blot analysis demonstrated the presence of both, free ISG15 and several ISGylated conjugates in FHA-stimulated PBMC lysates, but not in unstimulated cells. Intracellular FACS analysis provided evidence that monocytes and a natural killer-enriched cell population were the primary producers of ISG15 in PBMCs after FHA stimulation. Our data reveal previously-unrecognized effects of B. pertussis FHA on host IFN and ISGylation responses, and suggest previously-unsuspected mechanisms by which FHA may alter the outcome of the host-pathogen interaction.

Pleiotropic role of Rac in mast cell activation revealed by a cell permeable Bordetella dermonecrotic fusion toxin.

To activate the GTPase Rac in rat basophilic leukemia (RBL) cells and mouse bone marrow-derived mast cells (BMMC) a TAT fusion toxin of Bordetella dermonecrotic toxin (DNT-TAT) was constructed. The fusion toxin activated Rac1 and RhoA in vitro but only Rac in RBL cells and BMMC. DNT-TAT caused an increase in inositol phosphate formation, calcium mobilization, ERK activation and degranulation of mast cells. All these effects were inhibited by the Rho GTPase-inactivating Clostridium difficile toxin B and Clostridium sordellii lethal toxin. Also the calcium ionophore A23187 caused mast cell activation, including ERK phosphorylation, by processes involving an activation of Rac. The data indicate pleiotropic functions of Rac in mast cell activation.

Monocyte-derived interleukin-10 depresses the Bordetella pertussis- specific gamma interferon response in vaccinated infants.

Antigen-specific gamma interferon (IFN-gamma) has been demonstrated to participate in protection against Bordetella pertussis infection. Circulating mononuclear cells from B. pertussis-infected and from pertussis-vaccinated infants secrete high amounts of IFN-gamma after in vitro stimulation by B. pertussis antigens, but with a large variation in the secreted IFN-gamma levels between individuals. We show here that the inhibition of the specific IFN-gamma response can be at least partially attributed to IL-10 secretion by monocytes. This IL-10 secretion was not associated with polymorphisms at positions -1082, -819, and -592 of the IL-10 gene promoter, suggesting that other genetic or environmental factors affect IL-10 expression and secretion.

Modulation of the NF-kappaB pathway by Bordetella pertussis filamentous hemagglutinin.

BACKGROUND: Filamentous hemagglutinin (FHA) is a cell-associated and secreted adhesin produced by Bordetella pertussis with pro-apoptotic and pro-inflammatory activity in host cells. Given the importance of the NF-kappaB transcription factor family in these host cell responses, we examined the effect of FHA on NF-kappaB activation in macrophages and bronchial epithelial cells, both of which are relevant cell types during natural infection. METHODOLOGY/PRINCIPAL FINDINGS: Exposure to FHA of primary human monocytes and transformed U-937 macrophages, but not BEAS-2B epithelial cells, resulted in early activation of the NF-kappaB pathway, as manifested by the degradation of cytosolic IkappaB alpha, by NF-kappaB DNA binding, and by the subsequent secretion of NF-kappaB-regulated inflammatory cytokines. However, exposure of macrophages and human monocytes to FHA for two hours or more resulted in the accumulation of cytosolic IkappaB alpha, and the failure of TNF-alpha to activate NF-kappaB. Proteasome activity was attenuated following exposure of cells to FHA for 2 hours, as was the nuclear translocation of RelA in BEAS-2B cells. CONCLUSIONS: These results reveal a complex temporal dynamic, and suggest that despite short term effects to the contrary, longer exposures of host cells to this secreted adhesin may block NF-kappaB activation, and perhaps lead to a compromised immune response to this bacterial pathogen.

Lipooligosaccharide from Bordetella pertussis induces mature human monocyte-derived dendritic cells and drives a Th2 biased response.

Bordetella pertussis has a distinctive cell wall lipooligosaccharide (LOS) that is released from the bacterium during bacterial division and killing. LOS directly participates in host-bacterial interactions, in particular influencing the dendritic cells' (DC) immune regulatory ability. We analyze LOS mediated toll-like receptor (TLR) activation and dissect the role played by LOS on human monocyte-derived (MD)DC functions and polarization of the host T cell response. LOS activates TLR4-dependent signaling and induces mature MDDC able to secrete IL-10. LOS-matured MDDC enhance allogeneic presentation and skew T helper (Th) cell polarization towards a Th2 phenotype. LOS protects MDDC from undergoing apoptosis, prolonging their longevity and their functions. Compared to Escherichia coli lipopolysaccharide (LPS), the classical DC maturation stimulus, LOS was a less efficient inducer of TLR4 signaling, MDDC maturation, IL-10 secretion and allogeneic T cell proliferation and it was not able to induce IL-12p70 production in MDDC. However, the MDDC apoptosis protection exerted by LOS and LPS were comparable. In conclusion, LOS treated MDDC are able to perform antigen presentation in a context that promotes licensing of Th2 effectors. Considering these properties, the use of LOS in the formulation of acellular pertussis vaccines to potentiate protective and adjuvant capacity should be taken into consideration.

Receptor-mediated activation of nitric oxide synthesis by arginine in endothelial cells.

Arginine contains the guanidinium group and thus has structural similarity to ligands of imidazoline and alpha-2 adrenoceptors (alpha-2 AR). Therefore, we investigated the possibility that exogenous arginine may act as a ligand for these receptors in human umbilical vein endothelial cells and activate intracellular nitric oxide (NO) synthesis. Idazoxan, a mixed antagonist of imidazoline and alpha-2 adrenoceptors, partly inhibited L-arginine-initiated NO formation as measured by a Griess reaction. Rauwolscine, a highly specific antagonist of alpha-2 AR, at very low concentrations completely inhibited NO formation. Like L-arginine, agmatine (decarboxylated arginine) also activated NO synthesis, however, at much lower concentrations. We found that dexmedetomidine, a specific agonist of alpha-2 AR was very potent in activating cellular NO, thus indicating a possible role for alpha-2 AR in L-arginine-mediated NO synthesis. D-arginine also activated NO production and could be inhibited by imidazoline and alpha-2 AR antagonists, thus indicating nonsubstrate actions of arginine. Pertussis toxin, an inhibitor of G proteins, attenuated L-arginine-mediated NO synthesis, thus indicating mediation via G proteins. L-type Ca(2+) channel blocker nifedipine and phospholipase C inhibitor U73122 inhibited NO formation and thus implicated participation of a second messenger pathway. Finally, in isolated rat gracilis vessels, rauwolscine completely inhibited the L-arginine-initiated vessel relaxation. Taken together, these data provide evidence for binding of arginine to membrane receptor(s), leading to the activation of endothelial NO synthase (eNOS) NO production through a second messenger pathway. These findings provide a previously unrecognized mechanistic explanation for the beneficial effects of L-arginine in the cardiovascular system and thus provide new potential avenues for therapeutic development.

The Rho GTPase activators CNF1 and DNT bacterial toxins have mucosal adjuvant properties.

Cytotoxic necrotizing factor 1 (CNF1) from uropathogenic Escherichia coli belongs to a family of factors activating Rho GTPases. We report the in vivo effects of CNF1 in mice co-fed toxin and the soluble protein antigen ovalbumin (OVA). Similar to cholera toxin, CNF1 elicits adjuvanticity anti-OVA responses, both systemic and mucosal. In contrast, the catalytic inactive mutant CNF1-C866S demonstrated no effects. Using dermonecrotic toxin (DNT), a closely related Rho activating toxin from Bordetella, we discovered that the adjuvant property is within the DNT catalytic domain. Manipulation of Rho proteins thus provides a possible new approach for the development of effective mucosal immunoadjuvants.

Pertussis toxin-sensitive Galphai protein and ERK-dependent pathways mediate ultrasound promotion of osteogenic transcription in human osteoblasts.

Bone cells respond to mechanical stimulation via mechanoreceptors and convert biophysical stimulation into biochemical signals that alter gene expression and cellular adaptation. Pulsed acoustic energy treatment raises membrane potential and induces osteogenic activity. How membrane-bound osteoblast mechanoreceptors convert physical ultrasound (US) stimuli into osteogenic responses is not fully understood. We demonstrated that low-intensity pulsed US treatment (200-micros pulse, 1 kHz, 30 mW/cm2) elevated Cbfa1/Runx2 mRNA expression and progressively promoted osteocalcin mRNA expression in human osteoblasts. Pretreatment with pertussis toxin (PTX), but not with cholera toxin, suppressed US-augmented osteogenic transcription. This indicated that Gi proteins, but not Gs proteins, were involved in US promotion of osteogenic transcription. Further studies demonstrated US treatment could rapidly increase PTX-sensitive Galphai protein levels and subsequently enhanced phosphorylation of extracellular signal-regulated kinase (ERK). PTX pretreatment significantly reduced US promotion of ERK activation. Moreover, inhibition of ERK activity by PD98059 suppressed US augmentation of Cbfa1/Runx2 and osteocalcin mRNA expression. Membranous Galphai proteins and cytosolic ERK pathways acted as potent mechanosensitive signals in the response of osteoblasts to pulsed US stimulation.

Adjuvant activity of free Bordetella pertussis filamentous haemagglutinin delivered by mucosal routes.

The development of safe and potent mucosal adjuvants remains a major objective in vaccinology. The potential usefulness of filamentous haemagglutinin (FHA) of Bordetella pertussis as an adjuvant was assessed in a mouse model. The glutathione-S-transferase of Schistosoma mansoni (Sm28GST) was used for intranasal administration, while the gut-resistant keyhole limpet haemocyanin (KLH) was administrated by the oral route. For both antigens, coadministration with FHA increased antigen-specific immunoglobulin titres. This adjuvant effect did not require chemical cross-linking or direct interaction between FHA and the antigen tested. FHA also behaved as an adjuvant by the subcutaneous route, indicating that its adjuvanticity is not restricted to binding to mucosal surfaces. The FHA-induced adjuvanticity was also observed in mice with high anti-FHA antibody titres as a result of antipertussis vaccination, indicating that pre-existing anti-FHA antibodies do not impair FHA adjuvanticity. No mRNA coding for proinflammatory cytokines was induced in the lungs after intranasal FHA administration. However, an increase in the levels of mRNAs coding for B7-1, transforming growth factor (TGF)-beta and major histocompatibility complex (MHC)-II was detected in the lungs after FHA administration. Although the molecular mechanisms of the FHA-induced adjuvanticity remain to be elucidated, the data presented here indicate that this adhesin, already assessed for human use as a pertussis vaccine constituent, represents a promising adjuvant to improve the humoral immune response when given by mucosal routes.

Bordetella pertussis infection in 2-month-old infants promotes type 1 T cell responses.

Neonatal immaturity of the immune system is currently believed to generally limit the induction of immune responses to vaccine Ags and to skew them toward type 2 responses. We demonstrated here that Bordetella pertussis infection in very young infants (median, 2 mo old) as well as the first administration of whole-cell pertussis vaccine induces B. pertussis Ag-specific IFN-gamma secretion by the PBMC of these infants. IFN-gamma was secreted by both CD4(+) and CD8(+) T lymphocytes, and the levels of Ag-induced IFN-gamma secretion did not correlate with the age of the infants. Appearance of the specific Th-1 cell-mediated immunity was accompanied by a general shift of the cytokine secretion profile of these infants toward a stronger Th1 profile, as evidenced by the response to a polyclonal stimulation. We conclude that the immune system of 2-mo-old infants is developmentally mature enough to develop Th1 responses in vivo upon infection by B. pertussis or vaccination with whole-cell pertussis vaccines.

Pulsatile flow-induced angiogenesis: role of G(i) subunits.

OBJECTIVE: Angiogenesis plays a key role in the growth and function of normal and pathological tissues. We investigated the effect of pulsatile flow on endothelial cell (EC) in vitro angiogenic activity. METHODS AND RESULTS: Bovine aortic ECs were exposed to "static" or "flow" (1.2 to 67.0 mL/min, shear stress 1.4 to 19.2 dyne/cm2) conditions for 2 to 24 hours. After exposure, angiogenesis was measured as tubule formation on Matrigel, and EC migration was assessed by filter migration assay. Pulsatile flow increased angiogenesis and EC migration in a temporal and force-dependent manner, with a maximal effect at 16 hours (13.2 dyne/cm2). Pertussis toxin completely inhibited the effect of pulsatile flow on angiogenesis and migration. Transfection of ECs with inhibitory mutants of the alpha subunit of G(i)1 or G(i)3, but not G(i)2, inhibited the flow-induced angiogenic response by 61+/-2% and 32+/-6%, respectively, whereas transfection with constitutively activated mutants of the alpha subunit of G(i)1 or G(i)3, but not G(i)2, increased the flow-induced response by 202+/-23% and 70+/-4%, respectively. In contrast, inhibition of Gbetagamma by the carboxy terminal fragment of beta-adrenergic receptor kinase overexpression increased the flow-induced response by 82+/-8%. CONCLUSIONS: These results suggest that pulsatile flow stimulates angiogenesis and that this effect is mediated by activation of G(ialpha)1 or G(ialpha)3, but not Gbetagamma, subunits.

GABA is toxic for mouse striatal neurones through a transporter-mediated process.

In the present study, GABA was shown to induce a necrotic neuronal death in cultured striatal neurones from mouse embryos. This effect did not depend on the activation of GABA(A), GABA(B) or GABA(C) receptors as it was neither antagonized by bicuculline, saclofen or picrotoxin, respectively, nor reproduced by the GABA receptor agonists, muscimol and baclofen. Excluding the participation of glutamate, GABA neurotoxicity persisted in the presence of either the antagonists of ionotropic and metabotropic glutamate receptors or glutamate pyruvate transaminase, which induces an immediate catabolism of glutamate. A GABA transport-associated process is involved in GABA neurotoxicity as nipecotic acid and NO 711, two inhibitors of the high-affinity neuronal GABA transporters (GAT-1, in particular), completely prevented the neurotoxic effect of GABA. The activation of a subset of G proteins is also implicated in the GABA transport-mediated neuronal death as GABA neurotoxicity was completely suppressed when striatal neurones were pre-treated with pertussis toxin. Further demonstrating the specificity of this neurotoxic process, GABA-induced neurotoxicity was not observed in cortical neurones which, in contrast to striatal neurones, are largely represented by glutamatergic neurones. In conclusion, our study suggests that glutamate is not the sole neurotransmitter that can be responsible for brain damage and that GABA neurotoxicity involves both GABA transport and G protein transduction pathways.

Matrix metalloproteinase-1 activates a pertussis toxin-sensitive signaling pathway that stimulates the release of matrix metalloproteinase-9.

The matrix metalloproteinases (MMPs) are a family of structurally related metalloendopeptidases so named due to their propensity to target extracellular matrix (ECM) proteins. Accumulating evidence, however, suggests that these proteases cleave numerous non-ECM substrates including enzymes and cell surface receptors. MMPs may also bind to cell surface receptors, though such binding has typically been thought to mediate internalization and degradation of the bound protease. More recently, it has been shown that MMP-1 coimmunoprecipitates with the alpha2beta1 integrin, a receptor for collagen. This association may serve to localize the enzymatic activity of MMP-1 so that collagen is cleaved and cell migration is facilitated. In other studies, however, it has been shown that integrin engagement may be linked to the activation of signaling cascades including those mediated by Gialpha containing heterotrimers. As an example, alpha2beta1 can form a complex with CD47 that may associate with Gialpha. In the present study we have therefore investigated the possibility that MMP-1 may affect intracellular changes that are linked to the activation of a Gi protein-coupled receptor. We show that treatment of neural cells with MMP-1 is followed by a rapid reduction in cytosolic levels of cAMP. Moreover, MMP-1 potentiates proteinase activated receptor-1 (PAR-1) agonist-linked increases in intracellular calcium, an effect which is often observed when an agonist of a Gi protein-coupled receptor is administered in association with an agonist of a Gq coupled receptor. In addition, MMP-1 stimulates pertussis toxin sensitive release ofMMP-9 both from cultured neural cells and monocyte/macrophages. Together, these results suggest that MMP-1 signals through a pertussis toxin-sensitive G protein-coupled receptor.

Platelet factor 4 induces human natural killer cells to synthesize and release interleukin-8.

We provide evidence that platelet factor 4 (PF4), but not the related chemokine neutrophil-activating polypeptide-2, induced highly purified human natural killer (NK) cells to produce interleukin (IL)-8 in a time- and dosage-dependent manner. This ability was retained even while PF4 was bound to heparin. PF4 increased the steady state level of IL-8 mRNA, likely implying a transcriptional effect of PF4. Stimulation of NK cells through the Fc receptor for immunoglobulin G-IIIA was found to synergistically increase the effect of PF4 on IL-8 production but did not affect IL-2-related activities such as cytotoxic activity and proliferation. Pertussis toxin did not block the PF4-derived IL-8 production in NK cells, but this response was sensitive to wortmannin, implicating a role of phosphatidylinositol 3-kinase in the intracellular signaling pathway triggered by PF4. Our results characterize a new capacity for PF4 and provide further evidence for the pivotal role of NK cells in the environment of inflammation.

Human dendritic cells express functional formyl peptide receptor-like-2 (FPRL2) throughout maturation.

Immature and mature dendritic cells (iDC and mDC, respectively) migrate to different anatomical sites, e.g., sites of antigen (Ag) deposition and secondary lymphoid organs, respectively, to fulfill their roles in the induction of primary, Ag-specific immune responses. The trafficking pattern of iDC and mDC is based on their expression of functional chemotactic receptors and the in vivo sites expressing the corresponding ligands including chemokines and/or classical chemoattractants. In this study, we have evaluated the expression of the formyl peptide receptor like-2 (FPRL2) by human iDC and mDC. We show that iDC respond chemotactically and by Ca(2+) mobilization to N-formyl-Met-Leu-Phe and a recently identified synthetic peptide Trp-Lys-Tyr-Met-Val-D-Met (WKYMVm), whereas mDC derived from the same donor only respond to WKYMVm. Furthermore, iDC and mDC express FPRL2 mRNA and protein. As mDC do not express any other members of the human FPR subfamily, FPRL2 expressed by DC must be functional and mediate the effect of WKYMVm on DC. Indeed, treatment of iDC and mDC with WKYMVm induces the internalization of FPRL2. Thus, human myeloid DC express functional FPRL2 and maintain its expression even after maturation, suggesting that the interaction of FPRL2 and its endogenous ligand(s) may be involved in regulating DC trafficking during Ag uptake and processing in the periphery as well as the T cell-stimulating phase of the immune responses.

Histidyl-tRNA synthetase and asparaginyl-tRNA synthetase, autoantigens in myositis, activate chemokine receptors on T lymphocytes and immature dendritic cells.

Autoantibodies to histidyl-tRNA synthetase (HisRS) or to alanyl-, asparaginyl-, glycyl-, isoleucyl-, or threonyl-tRNA synthetase occur in approximately 25% of patients with polymyositis or dermatomyositis. We tested the ability of several aminoacyl-tRNA synthetases to induce leukocyte migration. HisRS induced CD4(+) and CD8(+) lymphocytes, interleukin (IL)-2-activated monocytes, and immature dendritic cells (iDCs) to migrate, but not neutrophils, mature DCs, or unstimulated monocytes. An NH(2)-terminal domain, 1-48 HisRS, was chemotactic for lymphocytes and activated monocytes, whereas a deletion mutant, HisRS-M, was inactive. HisRS selectively activated CC chemokine receptor (CCR)5-transfected HEK-293 cells, inducing migration by interacting with extracellular domain three. Furthermore, monoclonal anti-CCR5 blocked HisRS-induced chemotaxis and conversely, HisRS blocked anti-CCR5 binding. Asparaginyl-tRNA synthetase induced migration of lymphocytes, activated monocytes, iDCs, and CCR3-transfected HEK-293 cells. Seryl-tRNA synthetase induced migration of CCR3-transfected cells but not iDCs. Nonautoantigenic aspartyl-tRNA and lysyl-tRNA synthetases were not chemotactic. Thus, autoantigenic aminoacyl-tRNA synthetases, perhaps liberated from damaged muscle cells, may perpetuate the development of myositis by recruiting mononuclear cells that induce innate and adaptive immune responses. Therefore, the selection of a self-molecule as a target for an autoantibody response may be a consequence of the proinflammatory properties of the molecule itself.

Serum factors involved in human microvascular endothelial cell morphogenesis.

Our previous studies have demonstrated that lipid and protein angiogenic factors operate in tandem to induce optimal angiogenic responses in vivo. This study was undertaken to clarify the nature of the substances in human serum that are responsible for its remarkable ability to promote capillary morphogenesis in vitro. The ability of dilute (2%) human serum to promote the morphogenic differentiation of human dermal microvascular endothelial cells on Matrigel supports was depleted by more than 50% by treatment of the serum with activated charcoal, a procedure that effectively removes biologically active lipid growth factors. The remainder of the activity within serum was lost on heating to 60 degrees C for 60 minutes, indicating the involvement of a protein in the response. The ability of charcoal-treated serum to promote capillary morphogenesis was completely restored by the addition of sphingosine 1-phosphate (SPP, 500 nmol/L), but other lipids thought to be released into serum during clotting were ineffective. In addition, basic fibroblast growth factor (bFGF) effectively restored the ability of heat-treated serum to promote endothelial cell morphogenesis, but other protein growth factors, including vascular endothelial growth factor and platelet-derived growth factor, were ineffective. Together, SPP and bFGF were as effective as whole serum in promoting capillary morphogenesis. Responses to purified SPP were entirely sensitive to the effects of preexposure of the cells to pertussis toxin, whereas responses to bFGF were entirely pertussis toxin-resistant. Consistent with our hypothesis that two distinct factors in serum play a role in promoting capillary morphogenesis, responses induced by serum were inhibited approximately 50% by preexposure of endothelial cells to pertussis toxin. We conclude that platelet-released SPP acts in conjunction with circulating bFGF to promote capillary formation by microvascular endothelial cells. Lipid and protein growth factors apparently exert complementary roles in the angiogenic response, as demonstrated by their ability to promote chemotaxis, angiogenic differentiation, and angiogenesis in vivo.