Complement Component 3 is Necessary to Preserve Myocardium and Myocardial Function in Chronic Myocardial Infarction

Abstract

Activation of the complement cascade (CC) with myocardial infarction (MI) acutely initiates immune cell infiltration, membrane attack complex formation on injured myocytes, and exacerbates myocardial injury. Recent studies implicate the CC in mobilization of stem/progenitor cells and tissue regeneration. Its role in chronic MI is unknown. Here, we consider complement component C3, in the chronic response to MI. C3 knockout (KO) mice were studied after permanent coronary artery ligation. C3 deficiency exacerbated myocardial dysfunction 28 days after MI compared to WT with further impaired systolic function and LV dilation despite similar infarct size 24 hours post-MI. Morphometric analysis 28 days post-MI showed C3 KO mice had more scar tissue with less viable myocardium within the infarct zone which correlated with decreased c-kit^pos cardiac stem/progenitor cells (CPSC), decreased proliferating Ki67^pos CSPCs and decreased formation of new BrdU^pos/α-sarcomeric actin^pos myocytes and increased apoptosis compared to WT. Decreased CSPCs and increased apoptosis were evident 7 days post-MI in C3 KO hearts. The inflammatory response with MI was attenuated in the C3 KO and was accompanied by attenuated hematopoietic, pluripotent, and cardiac stem/progenitor cell mobilization into the peripheral blood 72 hours post-MI. These results are the first to demonstrate the CC, through C3, contributes to myocardial preservation and regeneration in response to chronic MI. Responses in the C3 KO infer that C3 activation in response to MI expands the resident CSPC population, increases new myocyte formation, increases and preserves myocardium, inflammatory response, and bone marrow stem/progenitor cell mobilization to preserve myocardial function.

Introduction

The complement cascade (CC), integral to first line host defense and innate immunity, plays a significant role in the immune response subsequent to myocardial infarction (MI). After MI, myocardial tissue damage serves as a potent CC activator which then contributes to tissue damage by promoting infiltration of immune cells and formation of the membrane attack complex (MAC) on the host cells within the ischemic region [1]. Necrotic cells expose intracellular antigens leading to robust activation of CC through classical or alternative pathways [1]. Necrotic cells release proteolytic enzymes that can directly cleave and activate complement component C3 and C5 without activation of classical or alternative pathways [2, 3]. C3 and C5 cleavage fragments contribute to initiation of the immune response in ischemia/reperfusion injury by chemoattraction of immune cells and MAC formation on cells within the ischemic region [4–6]. CC becomes activated acutely within hours of myocardial infarction [7]. Hence, it has been thought to be implicated in immune response initiation and to exacerbate injury in infarcted mice. In experimental models of myocardial infarction, complement depletion by cobra venom and blocking complement component C5 activation with a monoclonal antibody masking its cleavage site have been shown to be beneficial reducing injury size [5, 8, 9]. However, these studies focused on the CC in acute MI models without considering its role in chronic scar formation, remodeling, or regeneration. Clinically, trials testing CC blockade at C5 to prevent MAC formation and chemoattraction of neutrophils in patients suffering with myocardial infarction were equivocal as no clear benefit was observed in long term follow up including survival and LV systolic function[10–13]. These results suggest the CC may be more complex than initially perceived.

Recent evidence suggests that components of the CC may serve novel functions that modulate diverse regenerative processes, such as cell survival, growth, differentiation and trafficking of the stem/progenitor cells (mobilization/homing). Complement has recently been implicated as a mediator of lens and limb regeneration in lower vertebrates and brain, liver, bone and bone marrow in rodents [14–21]. Early studies of liver regeneration established classical pathway activation through C3a and C5a and respective receptors, C3aR and C5aR, is necessary. Regeneration in mice with targeted deletion of C3, C5, C3aR, or C5aR was compromised demonstrating cleavage products C3a and C5a were necessary for the response. Complement initiates the priming phase of liver regeneration that requires proliferative and growth signals to replace lost mass. C3a and C5a also increase hepatocyte survival during regeneration [16, 17, 21]. In the brain, complement contributes to neurogenesis in both the basal state and in response to ischemia (middle cerebral artery occlusion) [22]. Neurogenesis in C3 KO mice was impaired in the ischemic area and subventricular zone. C3a promoted neural differentiation of progenitors and accentuated stromal cell-derived factor 1α (SDF-1) stimulated migration and mitogen activated protein kinase (MAPK) signaling[20]. CC activation in bone marrow was necessary for stress induced bone marrow hematopoiesis but not homeostasis. Regeneration of bone marrow hematopoiesis after hematopoietic stem/progenitor cell (HSPC) transplantation was impaired in C3, C5, C3aR, and C5aR KO mice and was also associated with impaired HSPC homing [18, 23–28]. Recent studies suggest that CC might favor bone formation where C3a and C5a influence bone cell migration, osteoblast-osteoclast interaction, and modulation of the inflammatory response by osteoblasts [19].

In this study, we examine the effect of C3 on the chronic response to MI that includes scar formation, remodeling, and regeneration. Our studies show that LV function is impaired and progression towards heart failure is accelerated in C3 KO mice. Increased scar with decreased viable myocardium and fewer new myocytes in the C3 KO suggests regeneration was impaired. Impaired inflammatory response and mobilization of bone marrow stem/progenitor cells and reduced proliferative response of c-kit^pos resident cardiac stem cells in infarcted hearts of C3 KO mice contribute to attenuated regeneration. These findings provide the first evidence that, in addition to responding to acute injury, the complement system is involved in the chronic response to MI, which includes preservation and regeneration of the myocardium.

Methods

A full Methods section is available in the online-only Data Supplement.

Murine model of myocardial infarction

All procedures were conducted under the approval of the University of Louisville IACUC in accordance with the NIH Guide for the Care and Use of Laboratory Animals (DHHS publication No. [NIH] 85-23, rev. 1996) as previously described [29]. Female C57BL/6 (WT, Sham n=7, MI n=9) and C3 KO (Jackson Laboratories Stock 003641, Sham n=6, MI n=8) mice (mice were bred and used at 10–14 weeks when available) were anesthetized and the coronary artery occlusion or sham surgery performed as previously described [29]. Mortality of WT and C3 KO was 22.7% and 31.6%, respectively. Mice were injected daily with BrdC (80mg/kg, i.p.) beginning the day of operation until sacrifice. Group assignment was randomized at the time of surgery. In vivo cardiac function was assessed immediately prior to sacrifice 28 days after ligation by echocardiography as previously described [29, 30].

Echocardiography

In vivo cardiac function was assessed immediately prior to sacrifice 28 days after ligation by echocardiography as previously described [29, 30]. M-mode, 2D, and Doppler echocardiography (Vevo 770, Visual Sonics, 15 MHz linear array transducer, 120 Hz frame rate) analysis provided LV end-systolic or diastolic short axis or long axis areas (ESA, EDA, LVALS, and LVALD), short-axis end-systolic (ES) and end-diastolic (ED) diameter (D) and wall thickness (WT) and long-axis end-systolic and end-diastolic volume (ESV and EDV). LV systolic function was measured as ejection fraction (EF) and fractional shortening (FS).

Complement C5 activation

Complement C5 activation was measured as the formation of the cleavage product C5b-9 at indicated time points after MI or sham operation by ELISA (Kamiya Biomedical Company, Seattle, Wa).

Histology

After final echocardiography and peripheral blood collection, hearts were harvested in diastole with saturated KCl and CdCl (100 mmol/L) injected through the apex into the LV cavity. An additional group of WT (WT, Sham n=5, MI n=6) and C3 KO (WT, Sham=6, MI=6) mice were infarcted as above and processed for c-kit and TUNEL staining. Hearts were fixed in formalin sectioned and histology and morphometry performed as previously described [29, 31]

Mobilization and blood counts

WT and C3 KO mice without surgery, after sham surgery, and MI (n=6–8 for each group at 72 hours, Figure 7; n=5–9 for each group at each time point for time course, Supplemental Figure 7) were bled from the retro-orbital plexus for complete blood counts (WBC, NE, and LY, Hemavet) before, and flow cytometry 24, 48, 72 hours and 7 days after MI.

Figure 7. Reduced mobilization of bone marrow stem/progenitor cells in C3 KO mice after MI.

Mobilization of bone marrow cells seventy two hours after MI and sham operation was determined in peripheral blood of WT and C3 KO mice. Panel A – flow cytometry gating strategy for determination of Lin⁻Sca-1⁺c-kit⁺ hematopoietic, and Lin⁻CD45⁻Sca-1⁺ or Lin⁻CD45⁻c-kit⁺ non-hematopoietic stem cells. Panel B – number of circulating CFU-GMs, Lin⁻Sca-1⁺c-kit⁺ (SKL), Lin⁻CD45⁻Sca-1⁺ and Lin⁻CD45⁻c-kit⁺ was reduced in C3 KO mice after MI. Reduced mRNA level for cardiac (Nkx2.5 and GATA4) (Panel C) and pluripotent specific markers (Oct4, Nanog, Dppa3, and Rex-1) (Panel D) in peripheral blood cells of C3 KO mice after MI, evaluated by real time PCR and marker specific primers. The number of WBC, NE, and LY in peripheral blood was measured before and 24, 48, 72 hours and 7 days post-MI (Panel E). Values are the mean ± SEM (n=6–8). *P<0.05 vs. WT MI.

Colony forming unit-granulocytes/macrophage (CFU-GM) assay

Red blood cells (RBCs) from WT and C3 KO mice before and 24, 48, 72, 96 hours and 7 days after sham surgery or MI (n=6–8 for each group, Figure 7B and Supplemental Figure 7A) were lysed and nucleated cells used for CFU-GM assays.

Real-time RT-PCR and RT-PCR

Total mRNA was isolated from the WBC fraction from WT and C3 KO mice before and 72 hours after sham surgery or MI (n=6–8) to measure Oct-4, Nanog, Rex-1, Rif1, Dppa1, Gata-4 and NKX2.5, and β2-microglobulin mRNA levels by qRT-PCR and RT-PCR.

Statistical analysis

All data are expressed as the mean ± standard error (SEM). Statistical analysis of normally distributed data (Kolmogorov-Smirnov) was performed by unpaired T-test. Non-normal data was analysed by the Mann-Whitney test. A value of p<0.05 was considered significant.

Results

C3 deficiency exacerbates myocardial dysfunction after coronary artery ligation

Complement system activation with myocardial injury has characteristically been associated with exacerbated injury when studied in acute ischemia reperfusion MI models without further follow up to determine impact on post MI remodeling and contribution to failure [7, 8, 10, 11, 13]. Planimetric analysis of heart sections from WT and C3 KO mice 24 hours post-MI indicated no significant differences in infarct size (32.23±2.13% of LV WT vs 32.16±4.97% of LV C3KO, Supplemental Figure 1). Thus, C3 deficiency had no effect on infarct size acutely. Despite C3 deletion, C5 activation 48 hours after MI remained similar to that in WT mice (Supplemental Figure 2). In chronic studies, LV function was evaluated by echocardiography four weeks after infarction. No differences in LV systolic function measured by ejection fraction (EF) and fractional shortening (FS), LV end-systolic or diastolic short axis or long axis areas (ESA, EDA, LVALS, and LVALD) or LV geometry evaluated by LV inner diameter in systole and diastole (LVIDD, LVISD) and end-systolic and diastolic volumes (EDV, ESV) were observed between sham operated WT and C3 KO mice (Figure 1 A–F, Supplemental Figure 3A–D). However, function in infarcted C3 KO mice was significantly impaired compared to WT with decreased EF and FS (Figure 1A and 1B). This was associated with exacerbated dilation in C3 KO hearts compared to WT with increased ESA, EDA, LVALS, and LVALD (Figure 1 C–F) and LVEDV, LVESV, LVIDD, and LVISD (Supplemental Figure 3A–D). Thus, chronic myocardial remodeling after MI is exacerbated in C3 KO mice despite the absence of an effect on acute MI.

Figure 1. Echocardiographic assessment of LV function, volumes and diameters.

C3 deficiency attenuates cardiac function and accentuates myocardial injury after coronary artery occlusion. C3 KO mice and their wild type controls underwent coronary artery ligation or sham operation. LV function, ejection fraction (EF, Panel A), fractional shortening (FS, Panel B), LV end-systolic or diastolic short axis or long axis areas (ESA, EDA, LVALS, and LVALD) (Panel C–F) were assessed 28 days after surgery in vivo by echocardiography in Sham WT (n=7), Sham C3 KO (n=6), WT MI wild type (n=9), and C3 KO MI (n=8). Values are the mean ± SEM *P<0.05 vs WT MI.

C3 deficiency increases scar size and remodeling

The effect of C3 deficiency on LV dilation, remodeling, and failure after MI was further characterized, in Masson's trichrome stained sections from WT and C3 KO hearts after MI or sham surgery (Figure 2A). Morphometric analysis at sacrifice 4 weeks after MI demonstrated C3 deficiency significantly increased myocardial scar compared to WT (Figure 2B). The increase in scar was accompanied by a decrease in viable myocardium within the risk region (Figure 2C). Morphometry confirmed echocardiography data showing increased expansion index (Figure 2D), thinned infarct wall and exacerbated myocyte hypertrophy in risk and remote regions in the C3 KO compared to WT (Supplementary Figure 4, Figure 2E and F).

Figure 2. Morphometric analysis and assessment of LV hypertrophy.

C3 deficiency increases myocardial injury and accentuates LV hypertrophy after MI. Scar and viable myocardium in the risk region were determined in Masson's trichrome stained LV sections 28 days after surgery. Myocyte cross-sectional area was measured in sections stained with FITC-conjugated wheat germ agglutinin (WGA). Panel A - Representative Masson's trichrome stained heart short axis sections. Scar as percent of LV was increased in C3 KO hearts after infarction (Panel B) as viable myocardium in the risk region was decreased (Panel C). Expansion index in hearts of C3KO mice after infarction was elevated compared with ligated hearts of wild type animals (Panel D). Myocyte cross-sectional area in FITC-labeled WGA stained sections was significantly increased in both risk and remote regions in C3 KO hearts after MI (Panel E). Representative FITC-labeled WGA stained cross-sections (Panel F). Scale bars = 50 µm. Values are the mean ± SEM. Sham WT (n=7), Sham C3 KO (n=6), WT MI wild type (n=9), and C3 KO MI (n=8)*P<0.05 vs. WT MI.

To examine the mechanisms by which C3 deficiency could impair the response to injury leading to LV dilatation and dysfunction, capillary density was assessed as an index of the heart's capacity to respond to increased stress [29]. Myocardial capillary density in risk and remote regions was measured in sections stained with the endothelial specific FITC-isolectin B4. Capillary density was similar in sham C3 KO and WT mice. Capillary density in WT hearts 28 days post-MI decreased progressively in remote and risk areas. This decrease was exacerbated in the infarcted C3 KO hearts with a significant decrease in the risk area a smaller insignificant decline in the remote area (Figure 3A and B). The further decline in capillary density in the infarcted C3 KO risk area suggests neovascularization may be impaired and could contribute to exacerbation of MI.

Figure 3. C3 deficiency decreases capillary density and increases apoptosis in infarcted mice.

Capillaries in risk and remote regions of the LV were labeled with FITC-conjugated isolectin B4 and counted. Panel A - Quantitative analysis of capillary density demonstrating decreased risk region capillary density in C3 KO mice after MI. Panel B - representative fluorescent confocal images of FITC-isolectin stained cross sections from the border zone of wild type and C3 KO mice after MI or anterior LV after sham operation. Apoptotic cell death was assessed in WT and C3 KO hearts 7 (Panel C) and 28 (Panel D) days after MI and sham operation by counting the total number of TUNEL⁺ cells in risk and remote regions Panels C and D - TUNEL staining and apoptotic cell death is increased in the LV of C3 KO hearts after 7 and 28 days after MI. Panel E - Representative fluorescent confocal images of TUNEL stained apoptotic nuclei from WT and C3KO mice after MI or sham operation. Arrows denote apoptotic nuclei. Panel F - Representative image of TUNEL⁺ myocyte undergoing apoptosis 28 days post-MI. Scale bars = 50 µm. Values are the mean ± SEM (n=6–8). *P<0.05 vs. WT MI; N.S. – not significant.

Heart failure is associated with an increase in apoptotic cell death that contributes to myocyte loss and decreased functional capacity. C3 deficiency had no effect on apoptosis (assessed by TUNEL staining) in the sham-operated group at the 7 or 28-day time points. Apoptosis was significantly increased in the C3 KO LV after MI compared to WT LV (Figure 3C–E). TUNEL positive cells in C3 KO LV after MI were found only in the risk region with no observed difference in the remote region (Supplemental Figure 5A and B). Significant TUNEL staining was seen in myocytes in border zone and risk regions 7days after MI that then decreased substantially by 28 days and accounted for a small fraction of the total apoptotic cells (yellow arrow highlights an apoptotic myocyte, Figure 3E and Supplemental Figure 5C). Increased apoptosis in the risk region would contribute to decreased viable myocardium in the risk region and impair regenerative processes to further aggravate injury.

Reduced c-kit^pos CSPCs and impaired regeneration in C3 KO mice

The heart has a limited capacity to regenerate new myocardium that may depend on interplay between peripheral c-kit^pos cell populations resident in the bone marrow and c-kit^pos resident cardiac stem/progenitor cells (CSPC) [32–35]. Complement activation is known to affect hematopoietic stem and progenitor cell mobilization from the bone marrow, however, the role of complement in mobilization post-MI is unknown. The decrease in capillary density in the risk region of C3 KO mice post-MI suggests altered angiogenesis/neovasculogenesis and implicates mobilization of peripheral endothelial progenitor cell (EPC) populations. After MI, the number of endogenous c-kit^pos CSPCs increase within the first few days post injury and remain elevated for several weeks [33]. Resident c-kit^pos CSPCs in heart sections from WT and C3 KO mice four weeks after MI or sham operation were quantitated after identification by immunofluorescent confocal microscopy as previously described [29]. These cells were characterized to be negative for the hematopoietic and lymphoid markers CD34 and CD45, supporting the identification of these cells as cardiac resident and excluding mast cells that express CD34. In sham operated mice low levels of c-kit^pos CPCs were identified with no differences observed between WT and C3 KO hearts at 7 and 28 days post-MI. C3 deletion also had no effect on the number of Iba-1^pos macrophages in remote, border or infarct zones. This macrophage population did not contribute to the c-kit^pos cell population with only ~0.002% of Iba-1^pos macrophages staining positively for c-kit (data not shown). In WT hearts, the number of c-kit^pos CSPCs were markedly increased 7 days after MI and these levels declined significantly by 28 days post-MI (Figures 4C – E). In C3 KO hearts, the number of CSPCs was significantly reduced 7 and 28 days post-MI compared to those in WT mice (Figures 4C – E). Notably, levels in C3 KO hearts levels were only slightly elevated and did not reach significance over sham operated C3 KO mice (less than 2 fold) (Figure 4A–E). Thus, the increase in c-kit^pos CSPCs in the LV after MI was significantly attenuated in C3 KO mice 7 and 28 days after MI. Further, the number of c-kit^pos CSPCs in C3 KO LV after MI was reduced in both remote and risk regions (Figure 4E).

Figure 4. Reduced numbers of CSPCs in hearts of C3 KO mice after infarction.

c-kit⁺ CSPCs in risk and remote regions of LV sections after MI or sham surgery were counted after detection by immunofluorescent staining and confocal microscopy. Panel A – confocal microscopy images of c-kit⁺ CSPCs in non-ischemic region of LV section. Panel B – cluster of c-kit⁺ CSPCs in infarcted region of LV. Panel C and D – number of c-kit⁺ CSPCs were reduced in C3 KO mice LV compared to wild type controls 7 and 28 days, respectively, after MI. Panel E – number of c-kit⁺ CSPCs were lower in both risk and remote regions in C3 KO mice after MI. αSA (αSarcomeric Actin). Values are the mean ± SEM (n = 6–8). *P<0.05 vs. WT MI.

To begin to delineate the mechanism by which C3 deletion affected c-kit^pos CPC numbers we first examined whether the rate of c-kit^pos CSPC proliferation was affected. c-kit^pos CSPC proliferation in LV sections was measured by staining with the proliferation/mitosis marker, Ki67 (Figure 5A and B). Quantitative analysis of c-kit^posKi67^pos cells in LV sections from sham operated mice showed no differences between WT and C3 KOs. c-kit^posKi67^pos CSPCs were increased post-MI, however the level to which proliferating CSPCs increased in C3 KO hearts was significantly less than that in WT (Figure 5C). The reduction of c-kit^posKi67^pos cells in C3 KO hearts was consistently found in both the remote and risk regions (Figure 5D). Thus, these findings indicate that c-kit^pos CSPC proliferation was reduced in C3 KO hearts.

Figure 5. Impaired proliferation of CSPCs in hearts of C3 KO mice after infarction.

c-kit⁺ CSPC Proliferation was determined by Ki67 staining of heart sections of C3 KO and WT mice after infarction or sham operation. Panel A – confocal microscope image of proliferating c-kit⁺Ki67⁺ CSPCs in non-ischemic region. Panel B – images of proliferating (Ki67⁺) and quiescent (Ki67⁻) c-kit expressing CSPCs in infarcted region of LV. Panel C – reduced number of proliferating CSPCs in LV sections of C3 KO mice after MI. Panel D – reduced proliferation of c-kit⁺ CSPCs in both risk and remote regions in heart sections of C3 KO mice. Values are the mean ± SEM (n=6–8). *P<0.05 vs. WT MI.

c-kit^pos CSPCs have the ability to proliferate and differentiate to cardiac myocyte precursors and adult myocytes [32–35]. Adult myocytes were considered terminally differentiated and did not divide implying that any new myocytes with an adult phenotype were of a stem/progenitor cell origin. However, recently, adult myocytes have been shown to divide [36]. Thus differentiating α sarcomeric actin (αSA^pos) and dividing (BrdU^pos) CSPCs which were small and of an immature phenotype (Figure 6A and Supplemental Figure 5A and B) and large mature adult myocytes labeled with BrdU (Figure 6A and Supplemental Figure 5C and D) were identified as newly formed BrdU^posαSA^pos myocytes [32, 36]. To further verify the BrdU labeled nuclei associated with αSA^pos cardiomyocytes were of cardiac origin, sections were stained for cardiac-specific transcription factors GATA4 and Nkx2.5. In Figures 6B and 6C, the BrdU labeling of myocyte nuclei are counterstained with both GATA4 and Nkx2.5 further demonstrating BrdU labeling as a consequence of DNA replication occurs in cardiomyocytes. The number of newly formed myocytes was determined by counting BrdU^posαSA^pos cells in WT and C3 KO LV sections after MI and sham surgery (Figure 6A and D, Supplemental Figure 5). We found that in LV sections of infarcted C3 KO mice the number of newly formed myocytes was significantly reduced compared to WT. BrdU^posαSA^pos cells were reduced in both risk and remote regions (Figure 6D). These data indicate that the regenerative response in C3 deficient mice is diminished and correlates with impaired LV function measured by echocardiography and LV remodeling.

Figure 6. Diminished cardiomyogenesis in C3 KO mice after MI.

Newly formed myocytes, BrdU⁺αSA⁺ in LV sections of infarcted WT and C3 KO mice were counted after immunofluorescent staining and confocal microscopy. BrdU⁺αSA⁺ myocytes in the risk region (Panel A) are shown in confocal microscopic images. Sections are counterstained with cardiac myocyte specific transcription factors GATA4 (Panel B) and Nkx2.5 (Panel C) and images acquired from the border zone to verify BrdU-labeled nuclei are of myocyte origin. Scale bars = 20 µm. Panel D – quantitative analysis of BrdU⁺αSA⁺ myoyctes in risk and remote regions of LV sections from WT and C3 KO mice after MI. Arrows denote BrdU⁺αSA⁺ myocytes. Values are the mean ± SEM (n=6–8). *P<0.05 vs. WT MI.

Impaired mobilization of bone marrow stem/progenitor cells and inflammatory response in C3 KO mice

As noted above, complement activation affects mobilization but whether it participates in MI induced mobilization of bone marrow stem/progenitor cells is unknown. Mobilization after infarction in C3 KO mice did not correlate with size of the initial injury. Maximal mobilization of bone marrow hematopoietic stem/progenitor cells occurs 72–96 hours after MI as shown by the number of circulating clonogenic CFU-GM progenitors and early circulating Sca-1^posc-kit^posLin^neg cells (SKL cells) hematopoietic stem cells as shown by flow cytometry (Supplementary Figure 6A and B). Notably, CFU-GM and SKL numbers were significantly reduced in C3 KO mice compared to WT from 48 – 96 hours after MI, whereas levels were significantly increased in WT mice (Supplementary Figure 6A and B; Figure 7A and B). Importantly, the degree of bone marrow mobilization in patients after MI correlates with long-term outcome measured by LV function [37]. Thus reduced SKL mobilization would contribute to exacerbated injury by attenuating regeneration.

In addition to being a source of hematopoietic stem/progenitor cells, bone marrow is also a source of pluripotent Very Small Embryonic Like Stem Cells (VSEL cells) and tissue committed stem/progenitor cells including cardiac and endothelial progenitors [38, 39]. These, cells are also mobilized in response to tissue injury including myocardial infarction [38, 40, 41]and have been associated with myocardial regeneration [34, 42]. We used flow cytometry to evaluate the number of circulating Lin^negCD45^negc-kit^pos and Lin^negCD45^negSca-1^pos tissue committed and pluripotent stem cells (Figure 7A). Similar to hematopoietic stem cells, populations of Lin^negCD45^negc-kit^pos and Lin^negCD45^negSca-1^pos cells are robustly mobilized at 72 hours post-MI in WT mice. Notably, mobilization of these populations was also significantly reduced in infarcted C3 deficient mice (Figure 7A and B). These results demonstrate that C3 is an important factor triggering mobilization of hematopoietic, VSEL, and tissue-committed bone marrow stem/progenitor cells after MI.

Stem and progenitor cells mobilized after MI are enriched in subpopulations that express pluripotent and lineage specific factors that are considered important for tissue regeneration [38]. To confirm the enrichment of peripheral blood (PB) with VSELs and cardiac progenitor cells after MI, we measured mRNA expression levels of cardiac specific transcription factors GATA-4 and Nkx2.5 (Figure 7C) and pluripotency factors Oct-4, Nanog, Rex-1, Rif1, Dppa3 (Figure 7D in PB-derived cells harvested 72 hours after coronary ligation by qRT-PCR. We found that PB of WT mice after infarction was indeed enriched in cells containing mRNA for these markers compared to sham operated animals. Notably, mRNA levels for both pluripotency and cardiac transcription factors in PB were significantly reduced in infarcted C3 KO mice compared to WT (Figure 7C and 7D).

Complement is a first line contributor to innate immunity and its activation results in a strong inflammatory response and both complement and the inflammatory are activated with myocardial injury. C3 contribution to the inflammatory response was assessed by measuring white blood cells (WBC), neutrophils (NE), and lymphocytes(LY) in peripheral blood before and 24, 48, 72 hours and 7 days after MI (Figure 7E). WT mice responded to MI with increased WBCs, NEs, and LYs at 24 hours, which remained elevated through 72 hours and returned towards baseline levels by 7days. In contrast, the increase in WBCs, NEs, and LYs with MI was blocked in C3 KO mice demonstrating C3 contributes significantly to the MI induced inflammatory response.

Discussion

Our findings provide evidence of a novel role for complement in the chronic response to myocardial injury in which complement component C3 activation facilitates myocardial preservation and regeneration. Targeted C3 deletion in mice impairs the chronic response to MI leading to exacerbated dysfunction and remodeling, LV dilation, increased scar, and decreased viable myocardium within the scar. Importantly, these effects occurred despite full activation of C5 demonstrating C3 specifically plays an important role in the response to MI. These changes are accompanied by a constellation of contributing effects: increased apoptosis, decreased capillary density, and increased hypertrophy. The increase in resident c-kit^pos CSPCs that has been shown to play a significant role in the myocardial response to injury was significantly attenuated at early and late time points in C3 KO hearts. The attenuated c-kit^pos CSPC response in the C3 KO was due in part to decreased proliferation. Decreased total new and cycling c-kit^pos CSPCs was consequently associated with decreased generation of new myocytes. This attenuated activation and expansion of c-kit^pos CSPCs in the C3 KO was accompanied by reduced mobilization of distinct bone marrow stem/progenitor populations after MI. In addition, C3 deletion significantly attenuated the MI inflammatory response. Thus this is the first report to demonstrate that C3 activation is necessary for mobilization of stem cell populations that activate resident c-kit^pos CSPCs which may then contribute to formation of new myocardium in the chronic response to myocardial injury.

This study in a chronic MI model is particularly relevant as complement activation has typically been shown to exacerbate myocardial injury [3, 43–48]. The absence of a significant effect on infarct size measured 24 hours post-MI in C3 KO hearts compared to WT, although suggesting C3 does not play a role in the initial extent of injury, implicates C5 as contributing significantly to the initial response as its activation with MI was unaffected by C3 deficiency. Interestingly, the similarity of acute injury in WT and C3 KO mouse occurred despite a significant attenuation of the inflammatory response. It may be interesting to postulate that the decrease in inflammatory response counterbalances attenuated stem cell mobilization. This is also significant as it can thus be stated that the effects observed throughout the chronic response to injury were not due to differences in acute injury.

The eventual further deterioration of function and remodeling in the C3 KO heart after MI is likely a consequence of the absence of C3 influence at earlier time points. Attenuated mobilization, CSPC recruitment and expansion in the risk region and increased apoptosis at earlier time points clearly would contribute to the overall outcome at 28 days. This timing is relevant as blockade of CC activation is a therapeutic approach to limit MI. This approach primarily considers the immune and inflammatory response, however, our studies clearly demonstrate that the CC server beneficial functions.

However, chronic injury in the infarcted C3 KO hearts measured four weeks post-MI was exacerbated compared to WT hearts with increased LV dysfunction and dilatation implying further progression towards heart failure. Changes in key indices, increased myocyte hypertrophy, decreased capillary density, and increased apoptosis contribute to this dysfunction and dilatation. EPC mobilization after MI [49] has been associated with increased neovascularization [50], increased EPC recruitment, and improved function [51, 52]. Whereas the role of complement in HSPC retention and mobilization in bone marrow is increasingly understood, its impact on these actions and in particular for those of EPCs in the context of MI is unknown. However, as SKL mobilization was attenuated in the C3 KO after MI, EPC mobilization may also be impaired. Impaired EPC mobilization may contribute to decreased capillary density seen in the C3 KO post MI. Morphometric analysis of heart sections from C3 KOs indicated less viable myocardium in the risk region. This was correlated with impaired regeneration measured by formation of new myocytes indicated by BrdU^posαSA^pos immunofluorescent staining. The reduced number of new myocytes was associated with decreased c-kit^pos CSPCs in the LV of C3 KO mice 7 days and 4 weeks after MI. This relationship between the number of activated c-kit^pos CSPCs, the regenerative potential, and improved functional outcome, has been reported by others [29, 31–33]. The decreased number of CSPCs was likely a consequence of several factors, increased apoptosis, reduced proliferation of CSPCs and attenuated mobilization of stem/progenitors from the periphery and engraftment into infarcted C3 KO hearts. The impaired mobilization in the C3 KO is an important factor that contributes to several consequences of regeneration. Recently, c-kit^pos bone marrow cells were found to be important for increases in new myocytes and EPCs also home and engraft to injured myocardium [34, 53]. Thus, attenuated mobilization would limit c-kit^pos CSPC expansion, activation and formation of new myocardium. Taken together these findings demonstrate C3 is a key factor that promotes myocardial preservation and regeneration after MI.

Initial studies in acute MI models demonstrated that blocking CC activation was beneficial and reduced the extent of injury [5, 9]. Studies focused on the C5 component, which upon cleavage, generates anaphylatoxin C5a and C5b initiating formation of the MAC on the surface of cells within the infarcted region. The strategy of blocking C5 activation with monoclonal antibodies that mask the cleavage site was successful in experimental models of myocardial ischemia/reperfusion in mice, rats and pigs [5, 8, 54]. Systemic C5 inhibition in clinical studies has resulted in mixed results. Pexelizumab (C5 blocking antibody), tested in COMPLY, COMMA, PRIMO-CABG, and APEX-AMI trials did not show significant decreases in infarct size [10–13]. These findings raised concerns whether complement blockade at C5 activation is sufficient to reduce injury and further therapies should focus on blocking directly C5a and MAC formation or upstream components of the CC. Experimental models of myocardial infarction in mice using a C3aR antagonist have shown no beneficial effect on the size of the injury measured acutely [8]. C5 activation after MI in the C3 KO demonstrates the importance of alternate C5 convertase-independent pathways, such as thrombin activation with MI [55]. Our studies in C3 KO mice with permanent coronary ligation confirmed that C3 also has no effect on the size of the injury measured acutely 24 hours after MI.

In the past decade several studies have shown the myocardium has the capacity to regenerate with the creation of new myocytes, disproving the paradigm that the myocardium is post mitotic and unable to regenerate. It is now clear the heart has the ability to regenerate with some discordance regarding the rate with which this occurs [34, 56]. Regeneration depends on the presence of c-kit^pos resident CSPCs that, in response to injury, have the ability to proliferate and differentiate into cells expressing adult myocyte markers (αMHC, cTnI, αSA) [31, 32]. C3 KO mice four weeks post-MI have more scar tissue in the LV and less viable myocardium in the risk region. The role of C3 in this remodeling process can be a consequence of increased myocyte survival in the infarct and increased regenerative response with formation of new myocardium including vasculature and myocytes. Based on our findings C3 would contribute to both functions to preserve myocardium. We found increased apoptosis in heart sections of C3 KO mice that demonstrates C3 plays a protective rather than detrimental role in chronic MI. In support of this observation it has been shown that both C3a and C5a increase survival of mesenchymal stem cells in response to hydrogen peroxide treatment [57] and we also find that the C3aR is present and functional in CSPCs. We also evaluated the role of C3 in formation of new myocytes. We found that C3 KO hearts after MI have fewer newly formed BrdU^posαSA^pos myocytes at four weeks post MI in the ischemic zone than in WT. This provides support for C3 involvement in regeneration of heart muscle in addition to increased survival. A role for C3 in tissue regeneration has also been reported in liver and brain models of injury [15, 16, 21, 22, 58]. Importantly, C3a receptors, are present in myocytes, fibroblasts and CSPCs and functional in CSPCs providing the opportunity for C3a to act directly on myocardium. These findings constitute a novel and beneficial role of C3 in the response to myocardial injury.

After MI, the number of c-kit^pos CSPCs in the heart rapidly increases. However, the molecular mechanism responsible for expansion of the pool of resident stem cells in response to injury is poorly understood. Studies have implicated cells mobilized from the bone marrow or transplanted to play a role in activating resident c-kit^pos CSPCs however the mechanism responsible for this is unknown [34]. Our finding that the number of c-kit^pos CSPCs four weeks after MI were significantly attenuated in C3 KO mice compared to WT to an extent where they were not significantly increased compared to the Sham KO mice, suggests C3 is a significant factor in regulating c-kit^pos CSPC activation. This difference in c-kit^pos CSPCs in C3 KO versus WT hearts is partly attributed to decreased proliferation defined by the lower number of c-kit^posKi67^pos CSPCs. Thus the action of C3 on CSPC proliferation after MI is an important link between activation of complement and expansion of c-kit^pos CSPCs.

The ability of bone marrow stem cells to facilitate regeneration of the heart after MI has been demonstrated in animal models and in several clinical studies [59]. Bone marrow is a source of not only hematopoietic but also pluripotent (VSELs), cardiac, endothelial, and mesenchymal stem/progenitor cells that have therapeutic potential for regenerating infarcted myocardium [38, 39, 53, 60]. These cells are mobilized from the bone marrow into peripheral blood in response to MI as demonstrated in murine models and in patients [38, 40, 41, 61]. Importantly, the degree of bone marrow stem cell mobilization post-MI is positively correlated with LV function in patients [37]. We find that optimal mobilization of bone marrow hematopoietic stem/progenitor cells after permanent coronary artery ligation in WT mice occurs 72 hours after surgery. Within this time frame mobilization of hematopoietic, pluripotent, and cardiac stem cells is severely attenuated in C3 KO mice compared to WT controls demonstrating that C3 is necessary for mobilization of bone marrow cells after MI. We have previously shown that G-CSF and AMD3100 activate the complement cascade where it then promotes mobilization [23, 24, 62–64]. Thus a role for C3 in myocardial regeneration can be derived from its effect on stem/progenitor cell mobilization. The subsequent involvement in myocardial regeneration is, however, controversial. Both direct and indirect actions of these different stem cell populations on myocardial regeneration and CSPC activation have been reported. Direct effects, cardiomyogenesis, have been demonstrated for Sca1 or cardiac side population [12, 65, 66] VSELS and SKL cells [60]. Importantly, we show here for the first time evidence that C3 promotes mobilization of several types of bone marrow derived stem/progenitor cells in response to MI.

The attenuated inflammatory response is concordant with attenuated stem cell mobilization highlighting the importance of C3 activation for both. The relative contribution of C3 and C5 to injury may be illustrated here where acutely attenuated C3 has no effect on C5 activation yet results in similar extent of injury. This suggests C5 may play the predominant role and is justifiably a therapeutic target. The relative contribution of inflammatory response to injury versus remodeling and regeneration is complex as several subpopulations of inflammatory cells including Gr1^low monocytes will contribute to the healing phase with myofibroblast recruitment, collagen production and angiogenesis [67]. Thus attenuated contribution of this Gr1^low population to late phase responses associated with healing may contribute to the attenuated response to MI in the C3 KO.

In conclusion, our findings demonstrate that C3 deficiency has no effect on acute injury after permanent LAD ligation, however, C3 is necessary to preserve myocardial function and myocardium and prevent adverse remodeling in a chronic MI model. C3 contributes to myocardial survival and to CSPC activation as C3 KO hearts have reduced numbers of c-kit^pos CSPCs due to increased apoptosis and reduced proliferation after MI. Mobilization of hematopoietic, VSEL, and tissue-committed bone marrow stem/progenitor cells after MI was also found to require C3 providing an alternate mechanism that can contribute to the formation of new BrdU^posαSA^pos myocytes and myocardial regeneration. Dysfunction and adverse remodeling in the C3 KO despite an attenuated inflammatory response suggests important components of the healing phase may play a role in the chronic phase of C3 action. All together, these data indicate that C3 is an important factor in regeneration and preservation of myocardium and myocardial function in response to MI. These findings provide the first evidence that the complement system is important for the mobilization and activation of cardiac stem/progenitor cells that promote myocardial regeneration after MI.

Non-standard Abbreviations

BM

Bone Marrow

C3

Complement protein 3

C3aR

C3a Receptor

CC

Complement Cascade

CSPC

Cardiac Stem Cell

EF

Ejection Fraction

EDV

End Diastolic Volume

ESV

End Systolic Volume

FS

Fraction Shortening

HSPC

Hematopoietic Stem/Progenitor Cell

LV

Left Ventricle

LVID

Left Ventricular Inner Diameter

MAC

Membrane Attack Complex

MAPK

Mitogen activated protein kinase

MI

Myocardial Infarction

PB

Peripheral Blood

SDF1

Stromal cell-derived factor 1α