Accelerated wound healing with topical application of complement C5.

BACKGROUND: Delayed-healing traumatic, surgical, and chronic wounds can be detrimental to patients and the health care system. The authors set out to investigate the effects of complement C5, a naturally occurring chemotactic cytokine, on wounds. METHODS: The authors examined the effects of complement C5 on the rat paired skin incision model. Each rat served as its own control where topical collagen was applied to one incision and 100 nM of C5 in collagen vehicle was applied to the other incision. Rats were killed on days 3 (n = 6), 7 (n = 6), and 28 (n = 5) after wounding. RESULTS: There was a statistically significant, 65 percent increase in maximum wound breaking strength with the topical application of C5 at day 3 (p < 0.01). The increase persisted to 14 percent at 7 days after wounding (p < 0.05). When compared with the sham group, the C5-treated wound strength increased by 83 percent at day 3 and 64 percent at day 7. There was no change in breaking strength at 28 days. Western blot analysis demonstrated a significant increase in collagen and fibronectin content in the C5-treated wounds. CONCLUSIONS: Topical application of C5 to skin wounds significantly increases wound healing maximum breaking strength as early as 3 days and up to 7 days after wounding. C5 accelerated wound healing by at least 4 days in the first week of wounding. This was correlated with an increase in vascular permeability, increased inflammatory cell recruitment, subsequent fibroblast migration, and increased collagen deposition.

(99m)Tc-labeled murine ior C5 monoclonal antibody in colorectal carcinoma patients: pharmacokinetics, biodistribution, absorbed radiation doses to normal organs and tissues and tumor localization.

Monoclonal Antibody (mAb) ior C5 is a murine IgG(1) that recognizes the tumor associated antigen (TAA) ior C2, a cell surface O-linked glycoprotein carbohydrate chain not present in most normal tissues and homogeneously expressed in the cytoplasm of normal colon epithelium and heterogeneously expressed in more than 83% of primary colorectal carcinomas. This study was designed to investigate the pharmacokinetics, biodistribution and the absorbed radiation doses of (99m)Tc-labeled mAb ior C5 antibody in colorectal tumor patients. Ten patients were administered 3 mg of anti-O-linked glycoprotein carbohydrate chain TAA ior C2 murine monoclonal antibody ior C5 radiolabeled with (99m)Tc activity of 1435.0 +/- 123 MBq by intravenous (i.v.) bolus infusion. Blood and urine samples were collected from 4 out of 10 patients at timed intervals from 10 min and up to 24 h after injection of the (99m)Tc-labeled mAb ior C5 for pharmacokinetic studies. Whole body images were taken in 5 out of 10 patients for quantitative normal organ biodistribution and dosimetry studies and planar anterior and posterior and SPECT images were taken in 5 out of 10 patients for tumor localization. Mean absorbed doses were estimated using the methods developed by the Medical Internal Radiation Dose (MIRD) committee. The effective dose equivalent (EDE) and effective dose (ED) were calculated as prescribed in International Commission on Radiological Protection (ICRP) publications 30 and 60. Plasma disappearance curves of (99m)Tc-labeled murine antibody ior C5 were best fit by a two-compartment model in all patients with (t(1/2alpha)) of 4.32 +/- 2.18 h and (t(1/2beta) of 32.6 +/- 3.82 h. Among the main target organs, accumulation of the radiolabeled antibody was found in liver (9.38 +/- 0.80%), heart (8.92 +/- 0.94%) and spleen (1.37 +/- 0.30%) at 5 min post-administration. These values were reduced at 24 h to (5.91 +/- 0.73%) and (0.62 +/- 0.22%), respectively, for the heart and spleen and increased to (9.78 +/- 1.99%) for liver. Estimates of radiation absorbed dose to normal organs in rad/mCi administered were: whole body, 0.0181 +/- 0.0017; heart wall, 0.0768 +/- 0.0090; kidneys, 0.0530 +/- 0.0260; liver, 0.0565 +/- 0.0109 and spleen, 0.0540 +/- 0.0128. The effective dose equivalent and effective dose estimates for adults were 0.0314 +/- 0.0031 and 0.0249 +/- 0.0027 rem/mCi administered. This feasibility study indicates that the O-linked glycoprotein carbohydrate chain TAA ior C2 is expressed in primary and metastatic colorectal carcinomas and shows very limited expression in normal adult tissues. The very good pattern of biodistribution of (99m)Tc-labeled mAb ior C5 in patients will allow imaging of colorectal carcinoma lesions.

Mechanical properties of rat bone marrow and circulating neutrophils and their responses to inflammatory mediators.

Neutrophils are continuously released from the bone marrow (BM), and this release is accelerated during inflammation. This study compared the mechanical properties of mature neutrophils within the BM and the circulating blood, as well as the role of microtubule rearrangement in the release of neutrophils from the BM in rats. Neutrophils isolated from the BM were stiffer than neutrophils in the circulating blood, using magnetic twisting cytometry. BM neutrophils also contained more F-actin within the submembrane region than circulating neutrophils when examined using confocal microscopy, suggesting that mature quiescent neutrophils within the BM are stiffer than circulating neutrophils because of increased formation of F-actin beneath the plasma membrane. Complement protein 5 fragments or formylmethionyl-leucylphenylalanine (fMLP) induced a stiffening response within 2 minutes that was greater in circulating than in BM neutrophils. This stiffening required F-actin formation within the submembrane region but not microtubule rearrangement in both circulating and BM neutrophils. fMLP-induced shape changes were more pronounced in circulating than in BM neutrophils, which showed fewer and smaller pseudopods and fewer membrane irregularities. In vivo, fMLP induced neutropenia, sequestration of neutrophils within the pulmonary capillaries, and release of neutrophils from the BM. Studies using colchicine demonstrated that rearrangement of microtubules was not required for any of these processes but was required for normal trafficking of neutrophils through the pulmonary capillaries.

A novel role of complement: mice deficient in the fifth component of complement (C5) exhibit impaired liver regeneration.

Components of innate immunity have recently been implicated in the regulation of developmental processes. Most strikingly, complement factors appear to be involved in limb regeneration in certain urodele species. Prompted by these observations and anticipating a conserved role of complement in mammalian regeneration, we have now investigated the involvement of complement component C5 in liver regeneration, using a murine model of CCl(4)-induced liver toxicity and mice genetically deficient in C5. C5-deficient mice showed severely defective liver regeneration and persistent parenchymal necrosis after exposure to CCl(4.) In addition, these mice showed a marked delay in the re-entry of hepatocytes into the cell cycle (S phase) and diminished mitotic activity, as demonstrated, respectively, by the absence of 5-bromo-2'-deoxyuridine incorporation in hepatocytes, and the rare occurrence of mitoses in the liver parenchyma. Reconstitution of C5-deficient mice with murine C5 or C5a significantly restored hepatocyte regeneration after toxic injury. Furthermore, blockade of the C5a receptor (C5aR) abrogated the ability of hepatocytes to proliferate in response to liver injury, providing a mechanism by which C5 exerts its function, and establishing a critical role for C5aR signaling in the early events leading to hepatocyte proliferation. These results support a novel role for C5 in liver regeneration and strongly implicate the complement system as an important immunoregulatory component of hepatic homeostasis.

Interleukin 3-dependent mediator release in basophils triggered by C5a.

The anaphylatoxin C5a is a potent trigger for basophil degranulations, but in contrast to IgE-dependent basophil activation, it does not result in the synthesis of sulfidoleukotrienes (leukotriene C4/D4/E4). Thus, degranulation and the generation of lipid mediators are separately regulated cellular responses. Exposure of human blood basophils to the cytokine IL-3 alone does not induce the release of histamine in cells from most donors and never leads to the generation of LTC4, indicating that IL-3 is not a direct agonist for basophil mediator release. However, preincubation of basophils with IL-3 enhances the degranulation response to C5a. Most importantly, IL-3 "primes" basophils to release large amounts of leukotriene C4 after challenge with C5a (mean of 50 gp LTC4 per nanograms cellular histamine), while neither peptide alone is capable of inducing the formation of bioactive lipids. This effect is dose dependent, occurring at IL-3 concentrations considerably lower than are required to stimulate the growth of bone marrow progenitor cells. IL-3 affects the extent but not the time course of basophil degranulation, and leukotriene release of cells sequentially exposed to IL-3 and C5a occurs very rapidly concomitant with degranulation. A preincubation of the basophils with IL-3 is strictly required for C5a-induced LTC4 synthesis, but not for an enhancement of degranulation. Priming for C5a-induced lipid mediator generation occurs rapidly after exposure of the cells to IL-3, starting at 1 min and reaching maximal effects at 5 min, but this altered state of responsiveness is relatively long lasting. Cell fractionation studies indicate that the basophil is the source of lipid mediators and that IL-3 affects the basophil response directly. This study demonstrates that IL-3 is a potent modifier of effector functions of mature basophils; this is possibly of greater in vivo significance than its growth factor properties. The large amounts of LTC4 formed after triggering of IL-3-primed basophils may not only enhance but also qualitatively change the pathophysiological consequences of complement activation, and this might be important in the pathogenesis of immediate type hypersensitivity reactions, shock syndromes, and inflammation.

Cessation of neutrophil influx in C5a-induced acute experimental arthritis is associated with loss of chemoattractant activity from the joint space.

Neutrophil emigration is a critical component of the inflammatory process and is generally thought to play a role in host defense as well as in the tissue injury that often accompanies inflammation. Most inflammatory reactions exhibit a sequence of emigrating cell types, thus clearly demonstrating that the neutrophil influx eventually ceases and that the neutrophils are then removed from the lesion. It has been our premise that in order to understand the processes that lead to the progressive inflammatory reactions that underly so many disease processes, it is important to determine the mechanism by which the "normal" inflammatory response resolves. The purpose of this study was to identify the time of cessation of neutrophil influx in experimental arthritis induced by the injection of C5 fragments (C5f) and to investigate mechanisms underlying the cessation process. The migration of i.v. delivery pulses to inflamed joints was assessed by lavage of the joint space and by external scintigraphy. We found no evidence for the development of inhibitory systems against chemotactic factors or "desensitization" of the inflamed site, because a second injection of C5f into joints which had been injected previously with C5f resulted in enhancement rather than inhibition of migration. Neither was evidence found for altered tissue barriers to migration or for desensitization of neutrophils as possible explanations for cessation of influx. The major mechanism appeared to be a loss of chemoattractant activity in the joint space between 2 h and 6 h after C5f injection which was detected by transfer into a fresh joint. Radiolabeled C5a des-Arg had a t1/2 of disappearance from the joint of less than 1 h, which suggested that the transferred chemoattractant must, in part, have been due to the generation of new chemotaxins by C5f injection. These observations suggest that continued generation of chemoattractants or failure of their subsequent removal may be mechanisms leading to persistent neutrophil influx in chronic inflammation.

T cell immunity or tolerance as a consequence of self antigen presentation.

In this study we investigated the basis for immunity or tolerance to a mouse serum protein, the fifth component of complement (C5). In C5-deficient mice this protein is absent from serum and therefore they are not tolerized. Immunization of C5-deficient mice with C5-sufficient serum generates CD4+ T cells, which recognize C5 presented in the context of class II. No C5-specific responses were observed in T cells from C5-sufficient mice. We show that this self protein is processed and presented with class II by cells from C5-sufficient tolerant mice and can be recognized by C5-specific T cell clones and hybrids in the absence of exogenously added antigen. The stimulation of C5-specific T cells by C5-sufficient antigen-presenting cells is not a consequence of C5 secretion and subsequent processing in vitro but rather employs C5 peptide/class II complexes generated in vivo. We conclude that this self antigen is presented in normal mice in a form recognizable by T cells to induce and maintain immunological tolerance.

A direct in vivo comparison of the inflammatory properties of human C5a and C5a des Arg in human skin.

C5a is an 11,000-Da complement-derived inflammatory glycoprotein that has been shown to mediate inflammatory reactions in vitro as well as in vivo in human skin. The C5a degradation product, C5a des Arg, is rapidly formed after exposure of C5a to serum carboxypeptidase N and may represent the relevant C5-derived inflammatory peptide in vivo. To examine the biologic activity of human C5a des Arg in vivo and to compare it with that seen with human C5a, we purified and characterized homogeneous preparations of human C5a and C5a des Arg and injected them intradermally into seven normal volunteers. C5a des Arg exhibited biochemical and biologic properties in vitro that were different from those of C5a. When injected into human skin, C5a des Arg was less potent than C5a, in respect to both minimal dose eliciting wheal and flare reactions and maximal wheal and flare elicited at a given dose, but C5a des Arg still elicited cutaneous wheal and flare reactions at physiologically relevant concentrations. Histologically, C5a des Arg skin test sites showed dense polymorphonuclear neutrophil-rich infiltrates associated with leukocytoclasis, dermal mast cell degranulation, and endothelial cell swelling. These were virtually indistinguishable from reactions elicited by C5a and occurred with concentrations attainable in vivo. Cutaneous wheal and flare reactions elicited by either C5a or C5a des Arg were partially inhibited by H1 antihistamines but were unaffected by selected nonsteroidal anti-inflammatory agents.

C5a-induced bronchoconstriction: absence of a role of circulating white blood cells and platelets.

C5a causes an intense bronchoconstriction when injected intravenously into the guinea pig. The present study was designed to determine if circulating white blood cells or platelets were important target cells for the induction of bronchoconstriction on intravenous injection of C5a. To accomplish this, we examined the effect of C5a on numbers of circulating cells as well as the effect of depletion of various circulating cell populations on C5a-induced bronchoconstriction. Intravenous injection of C5a caused a precipitous drop in circulating granulocytes which preceded the maximum bronchoconstrictor response. The levels of circulating platelets and mononuclear cells were unchanged by intravenous injection of C5a. Depletion of either white blood cells or platelets did not significantly alter either the maximum bronchoconstrictor response to C5a or the time course of the response. Thus, these studies suggest that C5a-induced bronchoconstriction in the guinea pig is not dependent on an interaction of C5a with circulating white blood cells or platelets.

Stimulus-specific effects of endotoxin on superoxide production by rabbit polymorphonuclear leukocytes.

The release of superoxide (O2-) by polymorphonuclear leukocytes (PMN) is an important function that contributes to microbial death. Controversy exists as to the effect of bacterial endotoxin (lipopolysaccharide, or LPS) on the production of O2-. We have injected rabbits with 25 micrograms Escherichia coli LPS intravenously and studied PMN function 18 to 24 hours later. Relative to PMN from saline-injected controls, PMN from LPS-treated rabbits released markedly greater amounts of O2- in response to 10 ng/ml phorbol myristate acetate (PMA) as measured by nmol cytochrome C reduced in 20 minutes (40.8 +/- 7.8 for LPS-treated PMN versus 10.1 +/- 1.6 for control, p less than 0.01). LPS injection, however, significantly reduced O2- release in response to C (complement) 5a (1.4 +/- 0.6 nmole/20 minutes for LPS-treated PMN versus 5.6 +/- 1.3 nmole/20 minutes for control, p less than 0.01). O2- release in response to a third stimulus, n-formyl-methionyl-leucyl-phenylalanine (10(-7) to 10(-9) M), was not affected by LPS. O2- release in response to PMA was enhanced over a wide range of PMA concentrations (10 to 300 ng/ml). Kinetic studies over 30 minutes indicated that, after a brief initial latency in measurable response, LPS enhanced responsiveness to PMA at all time points observed. The reduced responsiveness to C5a corresponds to a previously reported down regulation of receptors for this ligand after intravenous LPS. The observations indicate that intravenous LPS can alter a critical function of PMN for at least 24 hours in a stimulus-specific manner.

Effect of low dose methotrexate on neutrophil chemotaxis induced by leukotriene B4 and complement C5a.

When mediators of inflammation such as complement component C5a or leukotriene B4 are introduced into an air pouch created in mice, these mediators induce the migration of neutrophils into the air pouch. Pretreatment of mice with low doses of methotrexate inhibits leukotriene B4 or C5a induced neutrophil migration into the air pouch. Inhibition of neutrophil chemotaxis by methotrexate may, at least in part, account for the rapid onset of antiinflammatory activity that was observed in clinical trials with methotrexate in rheumatoid arthritis.

Airways hyperreactivity and inflammation produced by aerosolization of human C5A des arg.

The premise of this study was that complement fragments would lead to granulocyte infiltration and to associated changes in airways reactivity. Inflammation was induced in large airways (greater than 1.0 mm) by aerosolization of the complement chemotactic factor, C5a des arg, which was isolated from activated human serum. Pulmonary function (resistance, compliance, and lung volume) and histamine reactivity were assessed at 4 and at 48 h after C5a des arg or a saline sham aerosol. Four hours after exposure to C5a des arg, the animals exhibited evidence of significant bronchoconstriction and hyperinflation. In addition, an increased responsiveness to histamine was demonstrated, which was not correlated with the degree of bronchospasm. By 48 h, these alterations were partially or completely resolved. Histologic examination and quantitation demonstrated significant accumulation of neutrophils, which was limited to airways 0.5 mm in or larger. Saline-treated animals also demonstrated a neutrophil infiltration, but significantly less than those treated with C5a des arg. However, these animals did not demonstrate hyperreactivity to histamine, and there was little change in baseline mechanical function. The mild inflammation associated in the saline control animals was demonstrated to be a result of intubation of the airways. Granulocyte dependence of the effects of C5a des arg was partially established by abrogation of these effects when the animals were rendered granulocytopenic with nitrogen mustard. We conclude that the physiologic responses of the airways to C5a des arg were largely granulocyte dependent. However, since granulocyte accumulation could occur without airways dysfunction, it is suggested that cell activation is an important step in producing neutrophil-associated airways hyperresponsiveness.

Indomethacin inhibits the increased airway responsiveness to histamine following inhalation of C5a des Arg in rabbits.

It has been shown that inhalation of C5a des Arg increases rabbit airway responsiveness to histamine and that this is associated with an influx of neutrophils into the airway walls. This study was undertaken to see if the augmented response to histamine can be blocked by the cyclo-oxygenase inhibitor indomethacin. Spontaneously breathing, anesthetised rabbits were studied in a volume displacement plethysmograph and pulmonary resistance (R1) was measured using the electrical subtraction technique. Histamine does response curves (HDR) were generated by measuring R1 after serial nubulisation of saline and histamine (1, 3, 10, 30 and 100 mg/ml). Aerosols of either saline or C5a des Arg (1.5 ug/ml) were then inhaled by the animals over a time period of 2 min. An HDR was then repeated 4 hours later. In 9 rabbits the inhalation of C5a des Arg resulted in an upward shift of the repeat HDR: the area under the HDR was significantly greater than under the first HDR (p less than 0.05). In 6 rabbits the repeat HDR 4 hours after saline was shifted downwards (N.S.) indicating some degree of tachyphylaxis. When rabbits were pretreated with indomethacin (5 mg/kg i.v.) the repeat HDR following either C5a des Arg (n = 7) or saline (n = 6) were also shifted downwards i.e., the increased airway responsiveness noted after C5a des Arg was abolished. There was no significant difference in baseline saline R1 during the first or second HDR in any group. These results suggest that the increased airway responsiveness following nebulisation of C5a des Arg may be due to release from neutrophils of products of the cyclo-oxygenase pathway.

Generation of biologically active, complement-(C5) derived peptides by cathepsin H.

The thiol proteinase cathepsin H, isolated and purified from rat liver lysosomes, provokes acute inflammation characterized by the accumulation of polymorphonuclear leukocytes (PMN) when injected intracutaneously into newborn rats. We have examined the possibility that the accumulation of PMN at skin sites injected with cathepsin H is due, in part, to generation locally of C-derived chemotactic factors. We have found that cathepsin H acts in a concentration- and time-dependent fashion in whole human (and rat) EDTA-plasma to generate C5-derived peptides with chemotactic activity for PMN. Chemotactic activity was not generated in EDTA-plasma by either heat-inactivated cathepsin H or by a combination of active enzyme and a thiol proteinase inhibitor isolated from rat epidermis. Cathepsin H also acted in a concentration- and time-dependent fashion on isolated (functionally pure) human C5 to yield chemotactic activity for PMN as well as PMN lysosomal enzyme-releasing activity. Whereas 10 ng/ml cathepsin H generated significant chemotactic activity from isolated C5 (1000 CH50 U/ml), 7 to 10 micrograms/ml were required to generate chemotactic activity in whole EDTA-plasma. Cathepsin H not only was capable of generating biologically active, C5-derived peptides, but also was capable of degrading these peptides. Incubation of either whole EDTA-plasma or isolated C5 with high concentrations of cathepsin H (e.g., 25 micrograms/ml and 100 ng/ml, respectively) caused the rapid appearance of chemotactic activity followed by an equally rapid disappearance. PMN accumulated more rapidly in the skin of newborn rats injected with cathepsin H-treated C5 than in the skin of animals injected with cathepsin H alone. These data suggest that generation by cathepsin H of C-derived chemotactic activity contributes to the ability of this enzyme to induce dermal inflammation.

Acute lung inflammation induced in the rabbit by local instillation of 1-0-octadecyl-2-acetyl-sn-glyceryl-3-phosphorylcholine or of native platelet-activating factor.

The intratracheal instillation into rabbits of 1-0-octadecyl-2-acetyl-sn-glyceryl-3-phosphorylcholine (AGEPC) or native platelet-activating factor (PAF) was shown to induce a dose-dependent acute pulmonary inflammation characterized by accumulation of macrophages in the alveolar space, degenerative and necrotic changes of alveolar epithelium, and accumulation of polymorphonuclear leukocytes (PMNs) and platelets in the alveolar capillary lumens with degenerative changes of endothelial cells. Infiltration of alveolar septa by inflammatory cells and, in a later stage, pulmonary fibrosis were also observed. Intrabronchial instillation of lysoglyceryl ether phosphorylcholine (lyso-GEPC) produced no inflammatory changes or only mild ones. In comparison with acute inflammation induced by intratracheal instillation of C5a des Arg, which is mainly characterized by the presence of neutrophils, red blood cells, and fibrin in the alveolar space, AGEPC and native PAF seem to induce a more severe accumulation of macrophages in the alveolar space and septa and of platelet and PMNs in the lumens of alveolar capillaries. These results are compatible with the concept that during inflammatory reaction an intraalveolar release of PAF contributes to the development of pulmonary injury.

Pulmonary injury induced by C3a and C5a anaphylatoxins.

Homogeneous anaphylatoxins C3a (human or porcine), C5a (porcine), and the porcine classic anaphylatoxin, a mixture of C5a and C5a des Arg, isolated from complement-activated serum, were shown to induce acute pulmonary injury in the guinea pig following intrabonchial instillation. The gross physiologic response to these factors is characterized by respiratory distress with rapid, shallow breathing. Administration of 8--17 micrograms/kg of porcine classic anaphylatoxin proved lethal in 50% of the animals treated. The acute response (less than 20 minutes after instillation) of pulmonary tissue to insult by the anaphylatoxins is characterized by constriction of the smooth muscle walls in both bronchioles and pulmonary arteries and by focal atelectasis. Aggregates of platelets and leukocytes in pulmonary vessels and in other organs such as the chambers of the heart were commonly observed after intrabronchial administration of the anaphylatoxins. Although C3a was never lethal in guinea pigs even when doses as high as 500 micrograms/kg were administered by the intrabronchial route, this anaphylatoxin did induce the same pattern of acute pulmonary injury as C5a. In vitro experiments employing guinea pig platelets indicated that these cells aggregate in the presence of 10(-10) M porcine C5a but are not affected by C3a (human or porcine) even at levels up to 10(-6) M. Hence, platelet aggregation as observed in vivo may be directly affected by C5a, but in the case of C3a, secondary mediators must be involved. Anaphylatoxin preparations were also shown to induce contraction of guinea pig lung strips in vitro: this effect was not inhibited by antihistamines at concentrations that blocked contraction to exogenous histamine. The in vivo response to anaphylatoxin could be blocked with high doses of the antihistamine chlorpheniramine but not by corresponding doses of diphenhydramine.

A differential effect of C5a and C5a des Arg in the induction of pulmonary inflammation.

Earlier studies have shown that C5 fragments induce an inflammatory reaction when instilled into the rabbit lung. Because C5a is rapidly converted to C5a des Arg in vivo, experiments were performed to determine which fragment was most effective in producing pulmonary inflammation in this animal model. C5a des Arg consistently produced marked inflammation. This was characterized by neutrophil accumulation, edema, hemorrhage, fibrin formation, and damage to alveolar epithelium. The time course of the inflammatory reaction initiated by C5a des Arg showed pulmonary vascular sequestration of neutrophils with no intra-alveolar migration at 30 minutes after injection. By 2 hours, interstitial and alveolar neutrophils were numerous, with the accumulation of neutrophils in the alveoli increasing to a maximum at 6 hours. At 24 and 48 hours, the predominant cells were mononuclear (macrophages). By 120 hours, the lesions were resolving. In contrast, at all doses examined, a similar instillation of C5a induced either no inflammation or a milder, more focal response than C5a des Arg. This inability of C5a to initiate inflammation was not apparently due to the generation of inhibitors, since mixtures of C5a and C5a des Arg were phlogistic. A prolonged, intrapulmonary infusion of C5a (20 minutes), in contrast to a bolus instillation (1 minute), did initiate an inflammatory response, which may reflect the conversion of the C5a to C5a des Arg in the lung. This study points out the inflammatory potential of products of complement activation, particularly of the C5 fragment C5a des Arg, when applied to the airway side of the lungs. This inflammatory response raises the possibility that cleavage of intrapulmonary C5 may play an important role in the initiation of pulmonary inflammation.