Skip main navigation

Complement Component C3 Mediates Inflammatory Injury Following Focal Cerebral Ischemia

Originally published Research. 2006;99:209–217


The complement cascade has been implicated in ischemia/reperfusion injury, and recent studies have shown that complement inhibition is a promising treatment option for acute stroke. The development of clinically useful therapies has been hindered, however, by insufficient understanding of which complement subcomponents contribute to post-ischemic injury. To address this issue, we subjected mice deficient in selected complement proteins (C1q, C3, C5) to transient focal cerebral ischemia. Of the strains investigated, only C3−/− mice were protected, as demonstrated by 34% reductions in both infarct volume (P<0.01) and neurological deficit score (P<0.05). C3-deficient mice also manifested decreased granulocyte infiltration (P<0.02) and reduced oxidative stress (P<0.05). Finally, administration of a C3a-receptor antagonist resulted in commensurate neurological improvement and stroke volume reduction (P<0.05). Together, these results establish C3 activation as the key constituent in complement-related inflammatory tissue injury following stroke and suggest a C3a anaphylatoxin-mediated mechanism.

Inflammation contributes to progressive neurological injury following experimental stroke.1,2 The complement cascade is integral in initiating and modifying the inflammatory process through its influence on adhesion molecule upregulation, neutrophil chemotaxis, platelet activation, and generation of reactive oxygen species.3,4 It follows that complement might play a deleterious role in cerebral ischemia/reperfusion (I/R) injury. To date, however, the precise contribution of the complement cascade to progressive neurological injury following cerebral I/R remains undefined.

Although early work involving pharmacological complement inhibition using cobra venom factor5,6 and C1-esterase inhibitor7,8 suggested that complement activation contributes to cerebral I/R injury, subsequent studies using the more specific complement inhibitor, soluble complement receptor-1 (sCR1), demonstrated only modest reduction of cerebral infarct volume.9 However, given the relative lack of specificity of these agents, it is unclear which complement components are critical to the pathogenesis of cerebral I/R injury or whether the protective benefits occur through a complement-independent mechanism.

To address this, we subjected mice genetically deficient in selected complement components (C1q, C3, C5) to focal cerebral I/R injury. Semiquantitative immunohistochemistry was used to confirm that complement cascade activation was suppressed as expected. In the case where genetic deletion suggested neuroprotection, knockout animals were reconstituted with deleted protein to demonstrate loss of protection; furthermore, wild-type (WT) mice were treated with a specific antagonist to the receptor for the cleavage product of the protein. Granulocyte infiltration was quantified immunohistochemically, and free radical production and lipid peroxidation were assessed by determining cerebral malondialdehyde (MDA) levels.

Materials and Methods


All experiments were approved the Columbia University Institutional Animal Care and Use Committee. Male C1qa−/−, C3−/−, and C5 sufficient/deficient mice (The Jackson Laboratory, Bar Harbor, Me) were backcrossed into the C57BL/6 background for 10 generations.10,11 All experimental animals were aged 8 to 10 weeks and weighed between 22 to 26 g at operation. They were housed in certified barrier facilities in microisolator cages with free access to food/water on a 12-hour light/12-hour dark cycle. The operator was blinded to the genotype of experimental animals at all times. Genotype was confirmed by polymerase chain reaction following backcrossing.

Murine Model

Studies used the intraluminal filament model described previously, with minor modifications.12 Briefly, mice were anesthetized with inhaled isoflurane. Middle cerebral artery occlusion was performed by advancing a heat-blunted, silicon-coated 7-0 nylon monofilament to the middle cerebral artery origin. Following 60 minutes of ischemia, the occluding filament was withdrawn to establish reperfusion.

Transcranial cerebral blood flow (CBF) was measured using laser-Doppler flowmetry (Periflux System 5000, Perimed Inc),13 with 0.5-mm flexible fiberoptic Doppler probes (Perimed) attached to intact skull over previously published landmarks. Strict criteria were used to prospectively exclude animals that do not experience adequate CBF drop-off.

A separate cohort of animals was subjected to middle cerebral artery occlusion and used to obtain physiological parameters. Mean arterial blood pressure, pH, Pao2, and Paco2 recorded pre-ischemia, intraischemia, and post-reperfusion were compared between groups.


Primary antibodies and reagents used for immunohistochemistry were commercially available. These include anti-Microtubule Associated Protein (MAP-2) (Sigma, St Louis, Mo), monoclonal anti-C1q (Connex GmbH, Martinsried, Germany), monoclonal anti-C3 (Connex GmbH), anti-C5 (Santa Cruz Biotechnology, Santa Cruz, Calif), anti-Ly-6G (BD-Pharmingen, San Diego, Calif), and Neurotrace 640/660 Fluorescent Nissl (Molecular Probes, Eugene, Ore). Secondary antibodies included: Alexa Fluor-488 or -594–conjugated anti-mouse IgG, anti-rat IgG, and anti-goat IgG (Molecular Probes).


Mice underwent focal cerebral ischemia and were euthanized after 24 hours. Following transcardiac PBS perfusion, brains were rapidly harvested, fixed in 4% paraformaldehyde, and cryoprotected. Brains were frozen in Optimal Cutting Temperature Compound, placed at −80°C and cut using a cryostat (20-μm coronal sections).

For immunofluorescent staining, sections were blocked with the secondary antibody-appropriate serum (10% donkey serum) with 0.2% Triton X-100 for 30 minutes at room temperature. Primary antibodies were diluted in PBS containing 0.2% Triton X-100. Sections were washed and incubated with secondary antibodies for 1 hour. Nissl was used for counterstaining. Sections were then mounted on slides in Vectashield (Vector laboratory) and visualized using a Bio-Rad 2000 confocal laser-scanning device with a Nikon E800 microscope.

Primary antibody specificity was confirmed by visualizing infarcted tissue from C57Bl/6 mice incubated with only secondary antibodies using confocal microscopy. To further confirm complement antibody specificity, infarcted tissue from C57Bl/6 mice stained with primary antibodies was blocked with the following peptides: purified human C1q protein (Quidel Corp, San Diego, Calif), C3 blocking peptide, and C5 blocking peptide (Santa Cruz Biotechnology). In all cases, peptide blocking abolished complement antibody staining.

Necrotic core was further identified by regional loss of MAP-2 immunopositivity, which has been demonstrated as an early immunochemical marker of ischemic neuronal injury.14 Cells were identified as neurons based on morphology/size on light microscopy and presence of remnant MAP-2 immunopositivity surrounding Nissl-stained nuclei in fluorescence microscopy.

Semiquantitative analysis of immunohistochemistry was performed to assess total complement immunopositivity. Multiple (10 to 15 images per slide) nonoverlapping ×40 medium-power fields were imaged to cover the entire infarct. Adobe Photoshop version 5.5 was used to acquire and process the images, which were then analyzed using Image Pro-Plus 4.5 (Media Cybernetics, Silver Spring, Md) software. Percentage area occupied by positive immunostaining was calculated for each image, and mean value for each animal was determined by averaging values from all images taken from that animal.

Infarct Volume and Neurological Function Assessment

Stroke outcomes were assessed at 24 hours post-ischemia using both cerebral infarct volume and a 4-tiered neurological score. A score of 1 reflects normal spontaneous movements; 2 indicates the animal was circling clockwise; 3 indicates the animal was observed to spin clockwise longitudinally; and 4 reflects an animal was unresponsive to noxious stimuli. This scoring system has been described previously in mice.13 For both primary outcome assessments, mice were compared with their respective WT controls. All assessments were conducted by an observer blinded to the identity of individual animals. Following neurological examination, mice were euthanized, and 1-mm sections of brain were stained with TTC. Infarct volumes were calculated by 2 blinded independent observers and expressed as percentage of the ipsilateral hemisphere occupied by infarct.13

Reconstitution of C3−/− Mice

To purify murine C3, 20 parts murine plasma was treated with 1 part inhibitor containing 1 mol/L KH2PO4, 0.2 mol/L Na4EDTA, 0.2 mol/L benzamidine, and 1 mmol/L phenylmethanesulfonyl fluoride (PMSF). C3 was precipitated from plasma first with 4.5% polyethylene glycol (PEG) and then with 12% PEG at 0°C. The pellet was dissolved in buffer A (20 mmol/L Na2HPO4, pH 7.4) and loaded onto a DEAE 40HR column (5×6.3 cm; Waters) preequilibrated with buffer A. The bound proteins were eluted using a 0 to 0.5 mol/L NaCl linear salt gradient. Fractions containing C3 were detected by SDS-PAGE and immunodiffusion, dialyzed against buffer A, loaded onto a Mono Q HR 10/10 column (Amersham Pharmacia Biotech) and eluted with a linear salt gradient to 0.5 mol/L NaCl. Homogeneous C3 fractions were identified by SDS-PAGE and immunodiffusion, pooled, and dialyzed against PBS (10 mmol/L Na2HPO4 and 145 mmol/L NaCl, pH 7.4). Using this method, we obtain a purified C3 containing 80% native C3 and 20% C3(H2O) by analyzing the protein sample on a Mono S column (Amersham Pharmacia Biotech).15 C3-containing fractions are tested for reactivity with a specific anti-mouse C3 antibody using Western blot and hemolytic activity, according to the method of Sahu et al using rabbit erythrocytes and C3-deficient mouse serum reconstituted with the purified C3 preparation.16

C3 reconstitution was performed by blinded administration of a single 0.5-mg IP injection of purified murine C3 to C3−/− mice 30 minutes before ischemic induction. An additional cohort of C3−/− mice received equal volumes of pH-matched saline.

Assessment of Infiltrating Granulocytes

C3−/− mice (n=7) and their WT littermates (n=9) underwent focal cerebral ischemia as described above and were euthanized after 24 hours. Tissue sections were prepared as detailed above using anti-Ly-6G primary antibody. An observer blinded to the identity of the mice counted Ly-6G+ cells in 5 representative medium-power fields (×40 objective) in each of 4 sections from each mouse. The fields were chosen in the ischemic region in an anatomically consistent manner between mice. Results are reported as mean number of cells per field in both ipsilateral and contralateral hemispheres.

Assessment of Brain MDA Levels

Mice used in this portion of the experiment were euthanized after 24 hours. Brains were snap frozen in liquid nitrogen and stored at −80°C until assayed. Using commercially available kits (Oxis Bioxytech, Montreal, Canada), MDA levels were measured in ipsilateral and contralateral hemispheres to quantify oxidative stress damage, and results are expressed as the ratio of MDA in the ipsilateral/contralateral hemispheres.17

C3a-Receptor Antagonist Experiments

C3a-receptor antagonist (C3aRA) (SB290157), purchased from Calbiochem (Darmstadt, Germany),18 has been shown to specifically block C3a-mediated effects in rodent disease models.19 Lipopolysaccharide (LPS) content was analyzed using a limulus assay (Pyrochrome) and was <1.5 ng/mg protein. C3aRA or equal volume of vehicle (10% EtOH in PBS) were administered via intraperitoneal injection (1 mg/kg) to C57Bl/6 mice in a blinded fashion 45 minutes before ischemia.

Statistical Analyses

Between-group analyses of semiquantitative immunostaining data were performed using ANOVA and multiple post hoc comparisons with a Bonferroni correction. Mortality was compared using χ2 analysis. All other between-group differences were made using 2-tailed unpaired Student’s t tests. All values are expressed as means±SE, with P<0.05 considered statistically significant.


There were no significant differences in animal weight between any cohorts. Analysis of physiological parameters performed on a separate cohort of mice demonstrated no significant differences in mean arterial blood pressure, pH, Pao2, Paco2, and rectal temperature between any groups at any time point in the experiment. Continuous transcranial Doppler recordings of CBF demonstrated equivalent blood flow between groups at all time points. Additionally, there were no statistical differences in the number of animals in each group that were excluded for failure to meet CBF criteria.

Immunohistochemical Staining for Complement Components in WT Mice

Examination of immunostained uninjured C57Bl/6 mouse brain sections revealed normal nuclei surrounded by abundant MAP-2–staining processes, representing dendrites and cell bodies. In all cases, by 24 hours post-insult, cerebral ischemia caused profound focal reduction of MAP-2 immunostaining in the ipsilateral, but not contralateral, hemisphere (Figure 1). Nissl staining demonstrated widespread nuclear pyknosis and karyorrhexis limited to the infarct.

Figure 1. Stroke architecture. Representative images of Nissl-stained (blue) and MAP-2–stained (green) tissue are depicted from penumbral regions of the ipsilateral hemisphere, as well as an anatomically equivalent region of the contralateral hemisphere of C57Bl/6 mice 24 hours following onset of cerebral ischemia. Infarcted area was identified by profound loss of MAP-2 immunostaining (a) and high-power image displaying typical histological features including nuclear pyknosis (b). Scale bar=20 μm.

Immunofluorescent staining for C1q, C3, and C5 in brains of C57Bl/6 mice at 24 hours post-ischemia revealed positive staining for these components in the infarct (Figure 2c, 2g, 2k). No immunopositivity was visible on cells in the contralateral cortex (Figure 2a, 2e, 2i). C1q, C3, and C5 were frequently found to colocalize with MAP-2 in the infarct (Figure 2d, 2h, and 2l). In addition, complement immunopositivity was identified on diffuse material that likely represents necrotic cellular debris. The lack of immunostaining seen in the antibody blocking experiments suggest that the complement staining in WT brains resulted from in vivo complement deposition, as opposed to artifactual deposition of complement in necrotic brain tissue from the blocking serum (data not shown).

Figure 2. Cerebral complement immunopositivity is apparent at 24 hours post-ischemia. Representative images from ipsilateral and contralateral hemispheres of C57Bl/6 mice 24 hours following onset of cerebral ischemia stained for C1q (a through d), C3 (e through h), and C5 (i through l). Images were taken from the ischemic regions of ipsilateral hemisphere, as well as from anatomically equivalent regions of contralateral hemisphere. Nissl staining is represented in blue, MAP-2 staining in green, and C′ components in red. The left panels (a, e, and i) demonstrate staining from anatomically equivalent regions of the contralateral hemisphere. The center panels (b, c, f, g, j, k) are split to demonstrate Nissl and MAP-2 staining without concomitant C′ component staining (b, f, j) or Nissl and C′ component staining without concomitant MAP-2 staining (c, g, k) from the ischemic region of the ipsilateral hemisphere. d, h, and l, Merged images of Nissl, MAP-2, and C′ component staining from ischemic region. Note the strong neuronal immunopositivity for C1q, C3, and C5 at 24 hours post-ischemia. Scale bar=20 μm.

Effects of Genetic C1q Deletion on Ischemic Brain Injury

Semiquantitative immunohistochemistry of brains taken from C1qa−/− mice at 24 hours following ischemia demonstrated negligible immunoreactivity for C1q in C1qa−/− mice (WT: 0.74±0.14%, n=4; C1qa−/−: 0.019±0.002%, n=3; P<0.01) (Figure 3a). C1q deletion had little effect on C3 staining in these mice, as amount of C3 in these brains was comparable to that seen in WT mice (WT: 2.36±0.37%, n=4; C1qa−/−: 2.04±0.37%, n=3; P=NS) (Figure 3b). Total amount of C5 immunostaining was moderately decreased in these mice (WT: n=4, 0.77±0.12%; C1qa−/−: n=3, 0.39±0.058%; P<0.05) (Figure 3c).

Figure 3. Semiquantitative immunohistochemical analysis demonstrates expected complement deposition. Mean percentage area occupied by C1q (a), C3 (b), and C5 (c) immunostaining in the brains of WT and gene-deficient mice. C1q immunopositivity was decreased only in C1q-deleted mice. Likewise, C3 immunostaining was depressed only in C3−/−. By contrast, significant reductions of C5 immunostaining were observed in C1q-deficient animals, with an even greater level of C3 inhibition in C3- and C5-deficient animals. Data presented represent grouped per animal averages.

C1qa−/− mice did not experience improvement in cerebral infarct volumes (WT: 38±2%, n=9, versus C1qa−/−: 38±2%, n=12; P=NS), neurological deficit scores (WT: 2.4±0.2, n=9; C1q−/−: 2.4±0.2, n=12; P=NS), or brain MDA levels compared with WT controls (WT: 1.86±0.24, n=4; C1qa−/−: 1.93±0.35, n=3; P=NS), implying that C1q expression does not directly mediate neuronal injury following stroke (Figures 4a, 5a, and 6a). There were no significant differences in overnight mortality between the C1qa−/− (n=1) and WT control (n=1) cohorts (P=NS).

Figure 4. Infarct volumes are improved in C3-null mice. C1q−/− (a), C3−/− (b), reconstituted C3−/− (c), C5-deficient (d) mice, and their respective WT or placebo cohort were subjected to cerebral I/R and euthanized at 24 hours. C3−/− animals demonstrated significantly smaller infarct volumes than their WT littermates. Reconstitution of these knockouts with purified murine C3 resulted in significant increase in infarct size. Representative TTC-stained brains obtained at 24 hours following MCAO in a WT (e) and a C3−/− (f) mouse.

Figure 5. Neurological function scores are improved in C3-null mice. Neurological scores for C1qa−/− (a), C3−/− (b), reconstituted C3−/− (c), C5-deficient (d) mice, and their respective WT or placebo cohort assessed at 24 hours post-ischemia. C3−/− mice demonstrated significantly improved neurological function, whereas reconstituted C3 nulls showed strong trends toward worsened neurological function.

Figure 6. Malondialdehyde levels are decreased in genetic C3 deficiency and C3a-receptor antagonism. Oxidative stress damage, quantified by relative MDA levels for C1qa−/− (a), C3−/− (b), C5-deficient (c) mice, and their respective WT controls assessed at 24 hours post-ischemia. Note significant reductions in oxidative stress in both the C3−/− and C3aRA-treated cohorts.

Effect of Genetic C3 Deletion on Ischemic Brain Injury

To explore the possibility that depletion of C3 might be required to mediate neuroprotection, we subjected C3−/− mice to focal ischemic brain injury. Immunoreactivity for C1q in C3−/− mice was similar to that seen in C57Bl/6 mice (WT: 0.74±0.14%, n=4; C3−/−: 0.56±0.13%, n=2; P=NS) (Figure 3a). Amount of C3 in these brains was negligible (WT: 2.36±0.37%, n=4; C3−/−: 0.018±0.003%, n=2; P<0.01) (Figure 3b). The degree of staining seen in the C3−/− infarcted hemispheres was consistent with the background staining seen in the contralateral hemisphere of C3−/− mice and noninfarcted brain (C3−/− contralateral hemisphere: 0.036±0.01%, n=2; noninfarcted WT: 0.023±0.006%, n=2). As well, C5 immunostaining was nearly absent in these mice (WT: 0.77±0.12%, n=4; C3−/−: 0.015±0.003%, n=2; P<0.001) (Figure 3c).

C3−/− mice experienced significant reductions in infarct volume (WT: 41±3%, n=8; C3−/−: 27±3%, n=10; P<0.01), neurological deficit scores (WT: 2.1±0.3, n=8; C3−/−: 1.4±0.2, n=10, P<0.05), and brain MDA levels (WT: 1.92±0.25, n=9, C3−/−: 1.26±0.12, n=7; P<0.05) compared with WT mice (Figures 4b, 5b, and 6b). As well, C3-deficient mice exhibited a reduction in granulocyte infiltration (WT: 2.33±0.33(ipsilateral), 0.09±0.01(contralateral), n=9; C3−/−: 1.19±0.22(ipsilateral), 0.10±0.01(contralateral), n=7; P=0.02 for ipsilateral comparisons) relative to WT mice (Figure 7). Despite the robust protection demonstrated, there were no significant differences in mortality between the C3−/− (n=2) and WT (n=1) cohorts (P=NS).

Figure 7. Infiltrating granulocytes are decreased in C3-deficient mice. WT (5 Ly-6G+ cells) (a) and C3−/− (2 Ly-6G+ cells) (b) mice euthanized at 24 hours and stained for Ly-6G (red). Nuclei (Nissl) are stained in blue. c, Graphical representation of granulocyte cell count per high-powered field (×40 objective) for WT and C3−/− mice. †Signifies comparison of ipsilateral hemisphere counts. Scale bar=20 μm.

This neuroprotective benefit was ablated by reconstitution of C3−/− mice with exogenous C3 protein. C3−/− mice receiving intraperitoneal purified murine C3 exhibited larger infarct volumes and worse neurological scores than C3−/− mice given placebo (C3−/−: 27±1%, n=4, reconstituted C3−/−: 38±4%, n=6; P<0.05 and C3−/−: 1.8±0.2, n=4, reconstituted C3−/−: 2.7±0.3, n=6; P=0.08, respectively) (Figures 4c and 5c), as well as similar infarct volumes and neurological scores as WT animals (WT: 41±3%, n=8, reconstituted C3−/−: 38±4%, n=6; P=NS). There were no differences in mortality between the groups receiving C3 (n=2) or placebo (n=1) in this portion of the study (P=NS).

Effect of Genetic C5 Deficiency on Ischemic Brain Injury

Brains from C5-deficient mice were stained for C1q, C3 and C5. Neuronal localization of C1q and C3 antigen were observed in the infarct, however no C5 immunopositivity was observed. Our analysis revealed that total amount of C1q (WT: 0.74±0.14%, n=4, C5-deficient: 0.59±0.12%, n=3; P=NS) (Figure 3a) and C3 (WT: 2.36±0.37%, n=4; C5 deficient: 2.06±0.18%, n=3; P=NS) immunostaining (Figure 3b) was similar to that seen in WT mice, whereas C5 staining was negligible (WT: 0.77±0.12%, n=4; C5 deficient: 0.015±0.003%, n=3; P<0.001) (Figure 3c).

Neither infarct volume (WT: 37±3%, n=12; C5 deficient: 40±5%, n=9; P=NS), neurological score (WT: 2.3±0.3, n=9; C5 deficient: 2.6±0.2, n=9; P=NS), nor MDA levels (WT: 1.67±0.19, n=4; C5 deficient: 1.98±0.47, n=3; P=NS) differed between WT and C5-deficient cohorts (Figures 4d, 5d, and 6c). There were no significant differences in mortality between C5-deficient (n=6) and WT (n=2) cohorts (P=NS).

Effect of C3aRA Administration on Ischemic Brain Injury

C3aRA-treated mice experienced significant reduction in infarct volumes (vehicle: 38±3%, n=14; C3aRA: 27±3%, n=11; P<0.05), neurological deficit scores (vehicle: 2.9±0.2, n=14; C3aRA: 2.3±0.2, n=11; P<0.05), and cerebral MDA levels (vehicle: 1.72±0.20, n=6; C3aRA: 1.01±0.18, n=6; P=0.03) compared with vehicle-treated WT controls (Figure 8a through 8c). There were no differences in mortality between the C3aRA (n=4) and vehicle-treated (n=2) cohorts (P=NS).

Figure 8. Cerebral infarct volumes, neurological scores, and oxidative stress are improved in C3aRA-treated animals. Animals treated with C3aRA vs vehicle were assessed at 24 hours post-ischemia. Antagonist-treated cohorts demonstrated significant improvements in infarct volumes (a), neurological deficit scores (b), and MDA levels (c).


This investigation used complement-deficient mice to evaluate the relative contributions of specific complement components to cerebral I/R injury. The study demonstrated that C3−/− mice experienced significant neuroprotection, an effect that was annulled by reconstitution with C3 protein. However, mice genetically deficient in C1q and C5 did not experience improved outcomes when compared with WT mice subjected to cerebral I/R. This lack of neuroprotection occurred despite qualitative and quantitative immunohistochemical evidence demonstrating the expected reduction in cerebral complement deposition.

Cerebral I/R injury leads to deposition of complement components in ischemic brain tissue,9 and many believe this accumulation plays a role in stroke pathogenesis.6,20 Studies of I/R in other organs have demonstrated that complement contributes to ischemic injury and that complement inhibition improves outcome.6,8,9,20,21 Cells in the central nervous system (CNS) are capable of synthesizing a complete set of complement components and upregulating this production in response to injury.22–25 Additionally, it has been demonstrated that complement activation has deleterious proinflammatory effects in other CNS processes, such as Alzheimer disease.26–30 Finally, neurons in vitro are highly susceptible to complement-mediated lysis, leading to the hypothesis that complement can cause neuronal death through membrane attack complex (MAC) deposition.31–33

C1qa−/− mice were used to assess whether C1q deletion provides neuroprotection following cerebral I/R. A role for C1q in the pathogenesis of inflammatory tissue injury has been supported by the recent discovery of C1q receptors.34 C1q deposition has been shown to promote phagocytosis, chemotaxis, cytotoxicity, and adhesion-receptor expression.34 Furthermore, we have recently demonstrated a C1q-mediated amplification of hypoxic-ischemic cerebral injury in immature mice.35 In the current study, whereas our immunohistochemistry suggests that deletion of C1q reduces C5 immunopositivity, both C3 and C5 are still detected in the infarcts of C1q-deficient mice. Our results with adult knockout mice suggest that direct C1q-mediated mechanisms do not contribute considerably to neurological injury following cerebral I/R. This is consistent with the recent study demonstrating that C1qa−/− mice were not protected from stroke.21 Previous studies have demonstrated robust localization of C1q in the CNS and increased C1q functional activity in CSF following cerebral ischemia.9,36 Furthermore, early results with sCR1 were interpreted to suggest that C1q mediated post-I/R cerebral injury.9,37 CR1, however, binds not only C1q but also C3b and C4b, facilitating factor I-mediated degradation of these proteins.38 Therefore, our findings suggest that the modest cerebral protection afforded by sCR1 in adult mice occurs via non-C1q-mediated mechanisms. Similarly, the neuroprotection from stroke observed using C1-esterase inhibitor, long attributed to classical pathway inhibition, has now been shown to be C1q independent.21

In contrast, C3−/− mice demonstrate not only a complete lack of post-ischemic cerebral C3 immunopositivity but also a marked resistance to cerebral injury as evidenced by improved stroke outcomes. Additionally, the neuroprotection observed in mice treated with the C3aRA suggests C3a-mediated events as the mechanism by which C3 causes injury in WT mice. A role for C3a receptor in post-ischemic cerebral injury is supported by its upregulation following murine stroke.39,40 Studies have demonstrated constitutive neuronal expression of C3aR, with profound post-ischemic upregulation on infiltrating inflammatory cells and ischemic endothelium.39,40 Considered together with our observation that C3−/− mice demonstrate significantly reduced infiltrating granulocytes and decreased oxidant stress levels, and that C3aRA-administration likewise suppresses oxidant stress and improves outcome, these data suggest that C3a mediates post-ischemic cerebral inflammation. Furthermore, although recent evidence has led to speculation regarding receptor density-dependent effects of the C3aRA, its efficacy in ameliorating stroke outcome in our model supports our findings in C3-deficient animals and suggests a C3a-dependent neuroprotective mechanism.41

It remains possible that the protection conferred by genetic C3 deletion was attributable to unknown effects of this manipulation; however, because reconstitution of C3−/− animals with C3 protein results in outcome similar to WT mice, and C3aRA-treated animals demonstrate significant neuroprotection, this is unlikely. An alternative interpretation may be that C3 deletion confers protection by preventing downstream C5 cleavage, forming the potent C5a anaphylatoxin the MAC. However, our C5-deficient mouse cohort did not exhibit neuroprotection. Although this result was surprising given the potent proinflammatory effects of C5a, genetic C5 deficiency has previously been shown to negatively impact the susceptibility of these mice to cerebral injury. For example, C5-deficient mice are more susceptible to intraventricular kainate toxicity through the overproduction of proinflammatory cytokines and alterations of Ca2+ influx.42 Furthermore, work in a murine intracerebral hemorrhage model indicates that C5-deficient mice exhibit increased brain water content.43 This evidence points to a potential difference in the function of genetic C5 deficiency in the brain than in visceral organs, where anti-C5 strategies have proven beneficial.44–47

Although these data clearly implicate C3 as a critical mediator of cerebral injury following stroke, the source of initial C3 activation in ischemic brain remains unclear. A recent study has identified the endothelial neoantigen that serves as a target for natural antibody that triggers complement-mediated injury in intestinal and muscle I/R.48 Additional work has demonstrated that intestinal I/R is lectin pathway dependent and C1q independent.49 In the present study, C1q-deleted mice demonstrate preserved C3 activation and a lack of neuroprotection, indicating that the classical pathway is not critical, at least in these deficient animals. This suggests that complement activation in stroke in all likelihood occurs either through the alternative pathway or by MBL recognition of exposed structures on ischemic cells. Although future experiments using MBL/MASP and Factor B knockouts may help address the importance of these additional pathways of C3 activation, phenotypic alterations and compensatory upregulations are likely to complicate interpretation of these findings to some degree. Furthermore, although natural antibody normally is excluded from the brain, breakdown of the blood-brain barrier occurs in many strokes. Additional studies of the role of the natural antibody in cerebral ischemia, in part through administration of a novel peptide that binds pathogenic IgM, should therefore be considered.48

In conclusion, our results establish that C3 mediates inflammatory neuronal injury following reperfused stroke in adult mice. Given the emerging data that C3a may be more critical in the acute period following human stroke than C5a,50 and the recent development of effective anti-C3 strategies in primates,51 additional preclinical experimental efforts to elucidate the therapeutic window for these anticomplement therapeutics are of critical importance.

*Both authors contributed equally to this study.

Original received January 24, 2006; revision received April 20, 2006; accepted June 7, 2006.

We thank Dr John Lambris for helpful suggestions and donation of murine C3 and C3a-receptor antagonist, Dr Marina Botto for donation of the C1qa−/− mice, and Dr Vadim Ten for suggestions.

Sources of Funding

This work was funded by NIH grant RO1 NS40409 (to E.S.C.). J.M. was supported by an American Heart Association, Heritage Affiliate, Postdoctoral Research Fellowship. W.J.M. was supported by the New York Academy of Medicine Glorney-Raisbeck Fellowship in Cardiovascular Medicine. D.J.P. was supported by NIH grants HL59488 and NS41460.




Correspondence to E. Sander Connolly Jr, MD, Department of Neurosurgery, Columbia University, 710 W 168th St, New York, NY 10032. E-mail


  • 1 Kuroda S, Siesjo BK. Reperfusion damage following focal ischemia: pathophysiology and therapeutic windows. Clin Neurosci. 1997; 4: 199–212.MedlineGoogle Scholar
  • 2 Jean WC, Spellman SR, Nussbaum ES, Low WC. Reperfusion injury after focal cerebral ischemia: the role of inflammation and the therapeutic horizon. Neurosurgery. 1998; 43: 1382–1396.MedlineGoogle Scholar
  • 3 Peerschke EI, Reid KB, Ghebrehiwet B. Platelet activation by c1q results in the induction of alpha IIb/beta 3 integrins (gpIIb-IIIa) and the expression of P-selectin and procoagulant activity. J Exp Med. 1993; 178: 579–587.CrossrefMedlineGoogle Scholar
  • 4 Murphy HS, Shayman JA, Till GO, Mahrougui M, Owens CB, Ryan US, Ward PA. Superoxide responses of endothelial cells to C5a and TNF-alpha: divergent signal transduction pathways. Am J Physiol. 1992; 263: L51–L59.MedlineGoogle Scholar
  • 5 Vasthare US, Barone FC, Sarau HM, Rosenwasser RH, DiMartino M, Young WF, Tuma RF. Complement depletion improves neurological function in cerebral ischemia. Brain Res Bull. 1998; 45: 413–419.CrossrefMedlineGoogle Scholar
  • 6 Cowell RM, Plane JM, Silverstein FS. Complement activation contributes to hypoxic-ischemic brain injury in neonatal rats. J Neurosci. 2003; 23: 9459–9468.CrossrefMedlineGoogle Scholar
  • 7 Akita N, Nakase H, Kanemoto Y, Kaido T, Nishioka T, Sakaki T. [The effect of C1 esterase inhibitor on ischemia: reperfusion injury in the rat brain]. No to Shinkei. 2001; 53: 641–644.MedlineGoogle Scholar
  • 8 De Simoni MG, Storini C, Barba M, Catapano L, Arabia AM, Rossi E, Bergamaschini L. Neuroprotection by complement (C1) inhibitor in mouse transient brain ischemia. J Cereb Blood Flow Metab. 2003; 23: 232–239.CrossrefMedlineGoogle Scholar
  • 9 Huang J, Kim LJ, Mealey R, Marsh HC Jr, Zhang Y, Tenner AJ, Connolly ES Jr, Pinsky DJ. Neuronal protection in stroke by an sLex-glycosylated complement inhibitory protein. Science. 1999; 285: 595–599.CrossrefMedlineGoogle Scholar
  • 10 Botto M, Dell’Agnola C, Bygrave AE, Thompson EM, Cook HT, Petry F, Loos M, Pandolfi PP, Walport MJ. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nat Genet. 1998; 19: 56–59.CrossrefMedlineGoogle Scholar
  • 11 Circolo A, Garnier G, Fukuda W, Wang X, Hidvegi T, Szalai AJ, Briles DE, Volanakis JE, Wetsel RA, Colten HR. Genetic disruption of the murine complement C3 promoter region generates deficient mice with extrahepatic expression of C3 mRNA. Immunopharmacology. 1999; 42: 135–149.CrossrefMedlineGoogle Scholar
  • 12 Connolly ES Jr, Winfree CJ, Stern DM, Solomon RA, Pinsky DJ. Procedural and strain-related variables significantly affect outcome in a murine model of focal cerebral ischemia. Neurosurgery. 1996; 38: 523–531.MedlineGoogle Scholar
  • 13 Connolly ES Jr, Winfree CJ, Springer TA, Naka Y, Liao H, Yan SD, Stern DM, Solomon RA, Gutierrez-Ramos JC, Pinsky DJ. Cerebral protection in homozygous null ICAM-1 mice after middle cerebral artery occlusion. Role of neutrophil adhesion in the pathogenesis of stroke. J Clin Invest. 1996; 97: 209–216.CrossrefMedlineGoogle Scholar
  • 14 Li Y, Jiang N, Powers C, Chopp M. Neuronal damage and plasticity identified by microtubule-associated protein 2, growth-associated protein 43, and cyclin D1 immunoreactivity after focal cerebral ischemia in rats. Stroke. 1998; 29: 1972–1980.CrossrefMedlineGoogle Scholar
  • 15 Lubinski J, Wang L, Mastellos D, Sahu A, Lambris JD, Friedman HM. In vivo role of complement-interacting domains of herpes simplex virus type 1 glycoprotein gC. J Exp Med. 1999; 190: 1637–1646.CrossrefMedlineGoogle Scholar
  • 16 Sahu A, Pangburn MK. Investigation of mechanism-based inhibitors of complement targeting the activated thioester of human C3. Biochem Pharmacol. 1996; 51: 797–804.CrossrefMedlineGoogle Scholar
  • 17 Gerard-Monnier D, Erdelmeier I, Regnard K, Moze-Henry N, Yadan JC, Chaudiere J. Reactions of 1-methyl-2-phenylindole with malondialdehyde and 4-hydroxyalkenals. Analytical applications to a colorimetric assay of lipid peroxidation. Chem Res Toxicol. 1998; 11: 1176–1183.CrossrefMedlineGoogle Scholar
  • 18 Ames RS, Lee D, Foley JJ, Jurewicz AJ, Tornetta MA, Bautsch W, Settmacher B, Klos A, Erhard KF, Cousins RD, Sulpizio AC, Hieble JP, McCafferty G, Ward KW, Adams JL, Bondinell WE, Underwood DC, Osborn RR, Badger AM, Sarau HM. Identification of a selective nonpeptide antagonist of the anaphylatoxin C3a receptor that demonstrates antiinflammatory activity in animal models. J Immunol. 2001; 166: 6341–6348.CrossrefMedlineGoogle Scholar
  • 19 Morikis D, Roy M, Sahu A, Troganis A, Jennings PA, Tsokos GC, Lambris JD. The structural basis of compstatin activity examined by structure-function-based design of peptide analogs and NMR. J Biol Chem. 2002; 277: 14942–14953.CrossrefMedlineGoogle Scholar
  • 20 D’Ambrosio AL, Pinsky DJ, Connolly ES. The role of the complement cascade in ischemia/reperfusion injury: implications for neuroprotection. Mol Med. 2001; 7: 367–382.CrossrefMedlineGoogle Scholar
  • 21 De Simoni MG, Rossi E, Storini C, Pizzimenti S, Echart C, Bergamaschini L. The powerful neuroprotective action of C1-inhibitor on brain ischemia-reperfusion injury does not require C1q. Am J Pathol. 2004; 164: 1857–1863.CrossrefMedlineGoogle Scholar
  • 22 Fischer B, Schmoll H, Riederer P, Bauer J, Platt D, Popa-Wagner A. Complement C1q and C3 mRNA expression in the frontal cortex of Alzheimer’s patients. J Mol Med. 1995; 73: 465–471.MedlineGoogle Scholar
  • 23 Gasque P, Ischenko A, Legoedec J, Mauger C, Schouft MT, Fontaine M. Expression of the complement classical pathway by human glioma in culture. A model for complement expression by nerve cells. J Biol Chem. 1993; 268: 25068–25074.CrossrefMedlineGoogle Scholar
  • 24 Levi-Strauss M, Mallat M. Primary cultures of murine astrocytes produce C3 and factor B, two components of the alternative pathway of complement activation. J Immunol. 1987; 139: 2361–2366.MedlineGoogle Scholar
  • 25 Johnson SA, Lampert-Etchells M, Pasinetti GM, Rozovsky I, Finch CE. Complement mRNA in the mammalian brain: responses to Alzheimer’s disease and experimental brain lesioning. Neurobiol Aging. 1992; 13: 641–648.CrossrefMedlineGoogle Scholar
  • 26 Head E, Azizeh BY, Lott IT, Tenner AJ, Cotman CW, Cribbs DH. Complement association with neurons and beta-amyloid deposition in the brains of aged individuals with Down syndrome. Neurobiol Dis. 2001; 8: 252–265.CrossrefMedlineGoogle Scholar
  • 27 Matsuoka Y, Picciano M, Malester B, LaFrancois J, Zehr C, Daeschner JM, Olschowka JA, Fonseca MI, O’Banion MK, Tenner AJ, Lemere CA, Duff K. Inflammatory responses to amyloidosis in a transgenic mouse model of Alzheimer’s disease. Am J Pathol. 2001; 158: 1345–1354.CrossrefMedlineGoogle Scholar
  • 28 Tenner AJ. Complement in Alzheimer’s disease: opportunities for modulating protective and pathogenic events. Neurobiol Aging. 2001; 22: 849–861.CrossrefMedlineGoogle Scholar
  • 29 Eikelenboom P, Hack CE, Rozemuller JM, Stam FC. Complement activation in amyloid plaques in Alzheimer’s dementia. Virchows Archiv B Cell Pathol. 1989; 56: 259–262.Google Scholar
  • 30 McGeer PL, Akiyama H, Itagaki S, McGeer EG. Activation of the classical complement pathway in brain tissue of Alzheimer patients. Neurosci Lett. 1989; 107: 341–346.CrossrefMedlineGoogle Scholar
  • 31 Singhrao SK, Neal JW, Rushmere NK, Morgan BP, Gasque P. Spontaneous classical pathway activation and deficiency of membrane regulators render human neurons susceptible to complement lysis. Am J Pathol. 2000; 157: 905–918.CrossrefMedlineGoogle Scholar
  • 32 Gasque P, Thomas A, Fontaine M, Morgan BP. Complement activation on human neuroblastoma cell lines in vitro: route of activation and expression of functional complement regulatory proteins. J Neuroimmunol. 1996; 66: 29–40.CrossrefMedlineGoogle Scholar
  • 33 Singhrao SK, Neal JW, Morgan BP, Gasque P. Increased complement biosynthesis by microglia and complement activation on neurons in Huntington’s disease. Exp Neurol. 1999; 159: 362–376.CrossrefMedlineGoogle Scholar
  • 34 Kishore U, Reid KB. C1q: structure, function, and receptors. Immunopharmacology. 2000; 49: 159–170.CrossrefMedlineGoogle Scholar
  • 35 Ten VS, Sosunov SA, Mazer SP, Stark RI, Caspersen C, Sughrue ME, Botto M, Connolly ES Jr, Pinsky DJ. C1q-deficiency is neuroprotective against hypoxic-ischemic brain injury in neonatal mice. Stroke. 2005; 36: 2244–2250.LinkGoogle Scholar
  • 36 Schafer MK, Schwaeble WJ, Post C, Salvati P, Calabresi M, Sim RB, Petry F, Loos M, Weihe E. Complement C1q is dramatically up-regulated in brain microglia in response to transient global cerebral ischemia. J Immunol. 2000; 164: 5446–5452.CrossrefMedlineGoogle Scholar
  • 37 Klickstein LB, Barbashov SF, Liu T, Jack RM, Nicholson-Weller A. Complement receptor type 1 (CR1, CD35) is a receptor for C1q. Immunity. 1997; 7: 345–355.CrossrefMedlineGoogle Scholar
  • 38 Krych-Goldberg M, Atkinson JP. Structure-function relationships of complement receptor type 1. Immunol Rev. 2001; 180: 112–122.CrossrefMedlineGoogle Scholar
  • 39 Barnum SR, Ames RS, Maycox PR, Hadingham SJ, Meakin J, Harrison D, Parsons AA. Expression of the complement C3a and C5a receptors after permanent focal ischemia: an alternative interpretation. Glia. 2002; 38: 169–173.CrossrefMedlineGoogle Scholar
  • 40 Van Beek J, Bernaudin M, Petit E, Gasque P, Nouvelot A, MacKenzie ET, Fontaine M. Expression of receptors for complement anaphylatoxins C3a and C5a following permanent focal cerebral ischemia in the mouse. Exp Neurol. 2000; 161: 373–382.CrossrefMedlineGoogle Scholar
  • 41 Therien AG. Agonist activity of the small molecule C3aR ligand SB 290157. J Immunol. 2005; 174: 7479.CrossrefMedlineGoogle Scholar
  • 42 Pasinetti GM, Tocco G, Sakhi S, Musleh WD, DeSimoni MG, Mascarucci P, Schreiber S, Baudry M, Finch CE. Hereditary deficiencies in complement C5 are associated with intensified neurodegenerative responses that implicate new roles for the C-system in neuronal and astrocytic functions. Neurobiol Dis. 1996; 3: 197–204.CrossrefMedlineGoogle Scholar
  • 43 Nakamura T, Xi G, Hua Y, Schallert T, Hoff JT, Keep RF. Intracerebral hemorrhage in mice: model characterization and application for genetically modified mice. J Cereb Blood Flow Metab. 2004; 24: 487–494.CrossrefMedlineGoogle Scholar
  • 44 Allegretti M, Moriconi A, Beccari AR, Di Bitondo R, Bizzarri C, Bertini R, Colotta F. Targeting C5a: recent advances in drug discovery. Curr Med Chem. 2005; 12: 217–236.CrossrefMedlineGoogle Scholar
  • 45 Verrier ED, Shernan SK, Taylor KM, Van de Werf F, Newman MF, Chen JC, Carrier M, Haverich A, Malloy KJ, Adams PX, Todaro TG, Mojcik CF, Rollins SA, Levy JH. Terminal complement blockade with pexelizumab during coronary artery bypass graft surgery requiring cardiopulmonary bypass: a randomized trial. JAMA. 2004; 291: 2319–2327.CrossrefMedlineGoogle Scholar
  • 46 Wada K, Montalto MC, Stahl GL. Inhibition of complement C5 reduces local and remote organ injury after intestinal ischemia/reperfusion in the rat. Gastroenterology. 2001; 120: 126–133.CrossrefMedlineGoogle Scholar
  • 47 Zhou W, Farrar CA, Abe K, Pratt JR, Marsh JE, Wang Y, Stahl GL, Sacks SH. Predominant role for C5b-9 in renal ischemia/reperfusion injury. J Clin Invest. 2000; 105: 1363–1371.CrossrefMedlineGoogle Scholar
  • 48 Zhang M, Alicot EM, Chiu I, Li J, Verna N, Vorup-Jensen T, Kessler B, Shimaoka M, Chan R, Friend D, Mahmood U, Weissleder R, Moore FD, Carroll MC. Identification of the target self-antigens in reperfusion injury. J Exp Med. 2006; 203: 141–152.CrossrefMedlineGoogle Scholar
  • 49 Hart ML, Ceonzo KA, Shaffer LA, Takahashi K, Rother RP, Reenstra WR, Buras JA, Stahl GL. Gastrointestinal ischemia-reperfusion injury is lectin complement pathway dependent without involving C1q. J Immunol. 2005; 174: 6373–6380.CrossrefMedlineGoogle Scholar
  • 50 Mocco J, Wilson DA, Komotar RJ, Sughrue ME, Coates K, Sacco RL, Elkind MSV, Connolly ES. Alterations in plasma complement following human ischemic stroke. Neurosurgery. In press.Google Scholar
  • 51 Sahu A, Morikis D, Lambris JD. Compstatin, a peptide inhibitor of complement, exhibits species-specific binding to complement component C3. Mol Immunol. 2003; 39: 557–566.CrossrefMedlineGoogle Scholar