Ischemia-reperfusion (IR) injury represents a major clinical challenge, which contributes to morbidity and mortality during surgery. The critical role of natural immunoglobulin M (IgM) and complement in tissue injury has been demonstrated. However, cellular mechanisms that result in the deposition of natural IgM and the activation of complement are still unclear. In this report, using a murine intestinal IR injury model, we demonstrated that the β-actin protein in the small intestine was cleaved and actin filaments in the columnar epithelial cells were aggregated after a transient disruption during 30 min of ischemia. Ischemia also led to deposition of natural IgM and complement 3 (C3). A low dose of cytochalasin D, a depolymerization reagent of the actin cytoskeleton, attenuated this deposition and also attenuated intestinal tissue injury in a dose-dependent manner. In contrast, high doses of cytochalasin D failed to worsen the injury. These data indicate that ischemia-mediated aggregation of the actin cytoskeleton, rather than its disruption, results directly in the deposition of natural IgM and C3. We conclude that ischemia-mediated aggregation of the actin cytoskeleton leads to the deposition of natural IgM and the activation of complement, as well as tissue injury.
- cytochalasin D
ischemia-reperfusion (IR) injury, resulting in damage to local and remote organs after periods of ischemia, is a major contributor to morbidity and mortality during myocardial infarction, transplantation, stroke, surgery, and trauma (6, 8, 9, 25, 47). However, currently there are no effective therapies because the mechanisms that result in tissue injury are not fully understood.
Although IR injury causes a strong inflammatory response (18, 44, 45, 48), accumulated evidence has shown that complement plays a critical role in IR injury. The finding that chemotactic complement (C)3-cleavage products are found in damaged heart tissue indicates that IR injury is complement dependent (17, 42). Additionally, the myocardium is protected from necrosis by cobra venom factor, which depletes C3 activation via the activation of the alternative pathway (12, 17, 25, 27), and by recombinant soluble human complement receptor (CR)1, which promotes inactivation of C3 by factor I (45). Soluble CR1 also reduces cerebral infarct volume (18) and intestinal injury (13). The observation that C5b-9 deposits in human myocardial tissue and C5- or C6-deficient mice and rabbits are protected from IR injury strongly suggests that complement is a major mediator of IR injury (15, 36, 51). Studies in C3- and C4-deficient mice and factor B- or factor D-knockout mice suggest that all three complement pathways mediate IR injury (39, 41, 44, 48). Although the significance of complement has been well established, the cellular mechanisms leading to its activation in tissues are not clear.
In addition to the contribution of complement, natural antibodies are also shown to be involved in the pathogenesis of IR injury (20, 44, 48). Recombination activation gene (RAG)1/2-deficient mice lacking natural antibodies and CR1/2-null mice with defects in T-dependent B-2 B cell responses to foreign antigens undergo less IR injury (14). Furthermore, natural immunoglobulin (Ig) M has been identified as one of culprits of IR injury (3, 48, 50). Recently, nonmuscle myosin type II (NM-II) heavy chain A and C have been identified as self-targets of natural IgM and IR injury in both the small intestine and in the skeletal muscle of mice (49). Thus it has been suggested that neoantigens or modified epitopes presented on the cell surface in ischemic tissues may trigger complement activation via natural IgM deposition (49). However, the mechanisms that lead to the exposure of these neoantigens need further investigation.
The actin cytoskeleton is known to play a crucial role in maintaining the functional and structural integrity of cells (21). The three-dimensional network of the actin cytoskeleton interacts with selected plasmalemmal proteins and ATP depletion or ischemia disrupt the actin cytoskeleton in vascular smooth muscle cells, endothelial cells, as well as epithelial cells (21–23, 32, 40). Multiple cleavage of actin protein has been observed in apoptotic cells (19, 28), and one of these cleaved fragments leads to apoptosis-like morphological changes in cultured cells (29).
We hypothesized that alteration of the actin cytoskeleton mediated by ischemia is one of the major initial events that result in the deposition of natural IgM, activation of complement, and further tissue injury. In this paper we establish a definite link between ischemia-mediated alteration of the actin cytoskeleton and the deposition of natural IgM, the activation of complement, and IR injury. We show in a murine intestinal IR injury model that ischemia induces aggregation of the actin cytoskeleton in columnar epithelial cells in small intestinal villi after a period of transient disruption. We also show that ischemia-mediated aggregation of actin filaments and deposition of IgM and C3/C3d fragment are diminished by an optimal dose of cytochalasin D. In addition, low doses of cytochalasin D attenuate IR injury in a dose-dependent manner whereas high doses do not worsen the injury. Taken together, these data reveal that ischemia-mediated aggregation of the actin cytoskeleton plays a crucial role in mediating the deposition of IgM and C3, as well as IR injury.
MATERIALS AND METHODS
Cytochalasin D was purchased from Sigma-Aldrich (St. Louis, MO). Hoechst 33342, Alexa Fluor 546-labeled phalloidin, and Alexa Fluor 488-labeled DNase I were bought from Invitrogen (Carlsbad, CA). Microvascular clips were obtained from Biomedical Research Instruments (Silver Spring MD). FITC-anti-mouse IgM and C3 antibodies were purchased from Immunology Consultant Laboratory (Newberg, OR). Goat anti-mouse C3d was purchased from R & D Systems (Minneapolis, MN). Mouse anti-β-actin NH2-terminus antibody was bought from Abcam (Cambridge, MA), and mouse anti-β-actin COOH-terminus antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
Animal model of murine intestinal IR.
All mice used in this study were maintained in specific pathogen-free conditions in the animal research facility at the Beth Israel Deaconess Medical Center (BIDMC). All experiments were performed in accordance with the guidelines and approval of the Institutional Animal Care and Use Committee of the BIDMC. Male C57BL/6 mice aged 8 to 10 wk were purchased from Jackson Laboratory (Bar Harbor, ME) and acclimated for 1 wk. Mice were anesthetized by intraperitoneal injection of a combination of ketamine-xylazine-acepromazine (100:20:3 mg/kg) (1) and subjected to IR as described previously (14) with some modifications that included ischemia for 20 to 30 min and reperfusion for 10 min to 2 h. At various times after ischemia and reperfusion, mice were euthanized and tissues were harvested. Sham mice were subjected to an identical surgical protocol aside from artery clamping. In inhibition experiments, food, but not water, was withdrawn for 24 h before anesthesia. All procedures were performed while maintaining mouse body temperature at 37°C using a controlled heating pad. Mice subjected to inhibition experiments underwent small intestine intraluminal injection 0.5–1.0 ml of cytochalasin D dissolved in dimethyl sulfoxide (DMSO)-PBS immediately after arterial clamping. Vehicle-treated mice were given an equal volume of DMSO.
About 15 cm of jejunal segment was rinsed and fixed immediately with cold 10% phosphate-buffered formalin. The tissues were then embedded in paraffin, sectioned transversely (5 μm), and stained with hematoxylin and eosin (H & E). Villus damage was scored according to the severity of injury. The complete destruction of the villus was scored “6” and no injury in the villus was scored “0” (14). The scores of injury were calculated by the following equation: injury score = ∑(score × Ni)/N. Ni represents the number of villi with the same injury score. N represents the total number of villi counted (250 villi).
Jejunal segments obtained from mice (n = 3/group) subjected to sham, 30-min ischemia, and 30-min ischemia-2 h reperfusion were harvested and sent on dry ice directly for two-dimensional difference gel electrophoresis (2-DIGE) and mass spectrometry analysis provided by Applied Biomics (Hayward, CA).
Western blot assay.
Jejunal segments from individual mice subjected to various durations of ischemia and reperfusion were homogenized in buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, and 1 mM DTT) with complete protease inhibitor cocktail (Roche, Mannheim, Germany). Concentrations of protein extracts were determined using Micro BCA protein assay kit (Pierce, Rockford, IL). Protein extracts were processed under reducing conditions and subjected to standard Western blot analysis. Blots were developed with mouse anti-β-actin antibodies (NH2 terminus and COOH terminus) and horseradish peroxidase-conjugated secondary antibodies and images were captured using a Fujifilm LAS-40000 luminescent image analyzer (Fujifilm, Valhalla, NY).
Jejunal segments were rinsed with cold PBS and snap frozen in frozen tissue embedding media. The tissues were sectioned transversely (6 μM), fixed with 10% phosphate-buffered formalin for 10 min, and permeabilized with 0.1% Triton X-100-PBS for another 10 min. Sections were blocked with 1% BSA-PBS at room temperature for 1 h and then incubated with fluorochrome-labeled primary or secondary antibodies at room temperature for 1 h. After being washed and mounted on slides, sections were analyzed by confocal microscopy (Nikon Eclipse Ti, Nikon Instruments, Melville, NY).
To evaluate the ratio of global to filamentous (G/F)-actin using quantitative fluorescence image analysis (35), tissue sections were stained with fluorochrome-labeled DNase I and phalloidin, which bind to global actin and filamentous actin, respectively (24). Fluorescence from the columnar epithelial cells of small intestinal villi was obtained by confocal microscopy under identical conditions. Intensity of fluorescence was analyzed by Nikon EZ-C1 FreeViewer 3_20_615 Gold (Nikon Instruments, Melville, NY). The ratio of G/F-actin was calculated by dividing DNase I fluorescent intensity by phalloidin fluorescent intensity. Increased ratios of G/F-actin reflect disruption of actin filaments, and decreased ratios of G/F-actin indicate aggregation of actin filaments.
Data are expressed as means ± SD. Differences between groups were evaluated by Student's t-test. P ≤ 0.05 was considered statistically significant. A two-tailed distribution and paired t-test were used to evaluate the differences in the ratios of G/F-actin. A two-tailed distribution and unpaired t-test were employed for the differences in injury scores.
Ischemia induces cleavage of β-actin protein.
On the basis of a previous study showing that a 15-kDa fragment of actin leads to morphological changes of cells undergoing apoptosis (29) and our interest to find potentially cleaved proteins in the small intestine, we conducted experiments to compare the differential expression of proteins in sham-, ischemia-, or IR-treated small intestine. The superior mesenteric artery (SMA) was clamped with a microvascular clip to generate ischemia in the jejunum for 30 min followed by reperfusion 2 h. Jejunal segments harvested without clip clamping were designated as sham samples, before clip removal as ischemia samples, and after reperfusion as IR samples. Jejunal segments were processed and subjected to 2-DIGE. One spot with a molecular weight of 30 kDa identified in the ischemia samples (Fig. 1, A and B) was isolated and sequenced. The abundance of this protein was increased 11.43-fold after ischemia and 0.78-fold after IR compared with sham samples. The MASCOT search engine analysis identified this protein to be β-actin.
Mouse anti-β-actin COOH-terminus antibody identified a band of 40 kDa in the ischemia (Fig. 1C, lanes 2 and 3) and IR samples (Fig. 1C, lanes 4 and 5) but not from sham samples (Fig. 1C, lane 1), indicating that β-actin protein (42 kDa) was cleaved during ischemia and reperfusion. A faint 30-kDa band in 20-min ischemia samples (Fig. 1D, lane 2) and a strong band in 30-min ischemia sample were also visualized in Western blot (Fig. 1D, lane 3). This band was also seen in 10-min reperfusion samples (Fig. 1D, lane 4). However, it disappeared in 2-h reperfusion samples (Fig. 1D, lane 5). These results are consistent with the 2-DIGE results above, indicating that ischemia induces cleavage of β-actin and the cleaved fragment is destroyed by reperfusion.
Mouse anti-β-actin NH2-terminus antibody failed to identify this 30-kDa band (data not shown), indicating that the cleaved fragment represents the COOH terminus of β-actin. In this experiment, 25-kDa bands of the Ig light chain were considered as loading controls since the primary antibody used in the Western blot assay was a mouse immunoglobulin (Fig. 1D).
Ischemia with or without reperfusion results in alteration of the actin cytoskeleton.
Disruption of the actin cytoskeleton induced by ATP depletion or ischemia have been reported in various cell types (21–23, 32). However, it has also been demonstrated that the actin cytoskeleton polymerizes at 2 h and then depolymerizes in a camptothecin-induced apoptotic model in HL-60 cells (35). In addition, β-actin is cleaved to yield 40- and 30-kDa bands during ischemia in our intestinal IR model (Fig. 1). Therefore, it is possible that ischemia may alter the actin cytoskeletal network in tissues. We conducted experiments focusing on the alteration of the actin cytoskeleton to address this hypothesis. The SMA was clamped with a microvascular clip to generate jejunal ischemia for 20 or 30 min and reperfused for 10 min or 2 h. Jejunal segments were processed and then stained with fluorochrome-labeled phalloidin and DNase I, Hoechst 33342 or H & E.
H & E staining revealed that 30-min ischemia caused minor injury in small intestinal villi (injury score: 1.54 ± 0.58, n = 5; P = 0.003) whereas 2-h reperfusion following 30-min ischemia resulted in severe injury (injury score: 4.18 ± 0.44, n = 8; P < 0.0001) compared with sham segments (injury score: 0.18 ± 0.07, n = 4).
Confocal microscopic images showed that phalloidin staining of the actin cytoskeleton in the columnar epithelial cells of the villi is lighter and less dense after 20-min ischemia (Fig. 2B) whereas staining becomes brighter and more dense in 30-min ischemia tissues (Fig. 2C) compared with sham-treated tissues (Fig. 2A). This indicates that ischemia results in an initial disruption of the actin cytoskeleton, followed by aggregation. This was confirmed by a quantitative analysis of G/F-actin; the ratio of G/F-actin was 2.08 ± 1.34 (n = 6, Fig. 3) in the columnar epithelial cells from sham mice, which significantly increased to 4.93 ± 3.62 (n = 6, P = 0.038, Fig. 3) after 20-min ischemia. However, after 30-min ischemia the ratio of G/F-actin was significantly decreased to 2.54 ± 2.52 (n = 6, P = 0.007, Fig. 3) consistent with the observation above that the actin cytoskeleton aggregates after 30-min ischemia (Fig. 2C). These results indicate that ischemia results in aggregation of actin filaments after an initial transient disruption. However, the actin cytoskeleton was disrupted again when the small intestine was reperfused for 10 min (Fig. 2D) and was destroyed completely after reperfusion for 2 h (Fig. 2E). The disruption induced by reperfusion was also demonstrated by an increase in the ratio of G/F-actin to 4.79 ± 4.58 (n = 6) or 4.87 ± 3.69 (n = 6) after 10 min or 2 h reperfusion, respectively. These results indicate that the restoration of blood supply damages the actin cytoskeleton.
Ischemia with or without reperfusion induces natural IgM and C3 deposition.
To investigate whether ischemia with or without reperfusion results in natural IgM and C3 deposition, we evaluated this deposition in small intestinal villi from mice subjected to various durations of ischemia with or without reperfusion by confocal microscopy. Jejunal segments were processed as described above and stained with FITC-anti-IgM or anti-C3 antibodies, or anti-C3d antibody/FITC-secondary antibody, Alexa Fluor 546-phalloidin, and Hoechst 33342. Confocal microscopy showed IgM deposits on the membranes of villus epithelial cells, which are in close proximity to blood vessels and also on the microvilli of the columnar epithelial cells from mice subjected to 10 min (Fig. 4D) or 2 h reperfusion (Fig. 4E). IgM deposits were clearly visualized (Fig. 4C) in 30-min ischemia samples although minor IgM deposits were observed on cores of the villi in 20-min ischemia samples (Fig. 4B), indicating that ischemia efficiently induces deposition of natural IgM. The IgM deposits on the villi from mice subjected to 30- or 10-min IR confirmed that the deposition of IgM in ischemic tissues is due to specific binding of IgM to the tissues. In contrast, IgM deposits were only detected in blood vessels of sham samples (Fig. 4A).
Both C3 and C3d deposits were observed on the microvilli of the columnar epithelial cells from mice subjected to 30-min or 2-h IR (Fig. 4, J and O). Both C3 and C3d deposits were also visualized on the villi from mice subjected to 30- or 10-min IR (Fig. 4, I and N). Deposits of C3 and C3d were clearly detected in tissues from mice subjected to 30-min ischemia (Fig. 4, H and M) but barely from those subjected to 20-min ischemia (Fig. 4, G and L). This indicates that ischemia results in deposition of C3 and more specifically C3d. These results reveal that not only reperfusion but also ischemia without reperfusion activates complement in the small intestine.
Ischemia-mediated aggregation of the actin cytoskeleton results in IR-mediated injury.
We next hypothesized that ischemia-mediated alteration of the actin cytoskeleton eventually leads to IR tissue injury. Therefore, we used cytochalasin D, a reagent that disrupts and blocks polymerization of actin filaments (37), to alter the actin cytoskeletal network in small intestinal villi to test this hypothesis. Mice were treated as described above. Various doses of cytochalasin D in DMSO-PBS (<0.1% in vol/vol) were injected into the small intestinal lumen immediately after the clip was applied to the SMA. After 30-min or 2-h IR, jejunal segments were processed and stained with H & E. Results of the cytochalasin D dose-response study showed that 1 μM of cytochalasin D effectively inhibited small intestinal injury (injury score: 1.38 ± 0.55, n = 12, P < 0.0001, Fig. 5), compared with the injury score in tissues from vehicle-treated mice (3.83 ± 0.97, n = 16, Fig. 5). This inhibition was dose dependent when doses of cytochalasin D used were less than 1 μM. This indicates that cytochalasin D dose-dependent alteration of the actin cytoskeleton protects mice from the small intestinal injury. Although cytochalasin D lost its inhibitory effect with gradually increased doses (Fig. 5), higher doses (30 and 100 μM) of cytochalasin D did not induce additional injury than that from vehicle-treated mice (P = 0.75 and P = 0.86, respectively). This indicates that cytochalasin D-induced disruption of the actin cytoskeleton does not contribute to the injury. All experimental mice died after 200 μM of cytochalasin D was injected whereas sham-treated mice died within several minutes after 1 μM of cytochalasin D was injected so that we were unable to obtain results from cytochalasin D-treated sham samples. These findings suggest that aggregation of actin filaments ultimately leads to IR tissue injury.
Optimal concentration of cytochalasin D attenuates the alteration of the actin cytoskeleton.
To further investigate how cytochalasin D alters the actin cytoskeleton, we observed the effects of cytochalasin D on IR-induced morphological changes of the actin cytoskeleton. Mice were treated with 1 μM of cytochalasin D or vehicle and subjected to various durations of ischemia and reperfusion. Frozen sections of jejunal segments were stained with fluorochrome-labeled phalloidin and DNase I, as well as Hoechst 33342. Confocal microscopic results showed that the brightness, thickness, and integrity of the actin cytoskeleton in the villus columnar epithelial cells from mice subjected to 20-min ischemia (Fig. 2F) were comparable to those from sham-treated mice without cytochalasin D treatment (Fig. 2A) and were brighter, more dense than those in samples from 20-min ischemia-treated mice treated with vehicle (Fig. 2B). These data indicated that the optimal dose of cytochalasin D blocks the disruption of the actin cytoskeleton. In samples from 30-min ischemia-treated mice treated with cytochalasin D (Fig. 2G) or vehicle (Fig. 2C) the integrity of the actin cytoskeleton was comparable, indicating that the actin cytoskeleton does not undergo aggregation. The actin cytoskeleton network is highly organized in the columnar epithelial cells from mice subjected to 10-min or 2-h reperfusion (Fig. 2, H and I), compared with those from vehicle-treated mice subjected to 10 min or 2 h of reperfusion (Fig. 2D and E). These images demonstrate that cytochalasin D blocks reperfusion-mediated damage of the actin cytoskeleton.
The findings above are supported by changes in the ratio of G/F-actin. The ratio of G/F-actin in the columnar epithelial cells of small intestinal villi from mice subjected to 20-min ischemia (2.19 ± 1.24, n = 6, P = 0.89) was comparable to that from sham mice without cytochalasin D treatment (2.08 ± 1.34, n = 6, Fig. 3) and was significantly lower than that from vehicle-treated mice subjected to 20-min ischemia (4.93 ± 3.62, n = 6, P = 0.04, Fig. 3), consistent with the observation above (Fig. 2F) indicating that 1 μM of cytochalasin D attenuates the disruption of the actin cytoskeleton. The ratio of G/F-actin from mice subjected to 30-min ischemia increased, rather than decreased, slightly to 2.60 ± 2.40 (n = 6, P = 0.54, Fig. 3), compared with those from mice subjected to 20-min ischemia. This result also agrees with the observation above in which cytochalasin D blocks the aggregation of actin filaments (Fig. 2G). This may also mean that the aggregation of actin filaments may be a result of the initial disruption and actin filaments may be unable to aggregate owing to the blockade of the disruption by cytochalasin D as in tissues from vehicle-treated mice. Thus it can be concluded that the disruption of the actin cytoskeleton initiates the aggregation. The ratios of G/F-actin increased significantly (n = 6, P = 0.02, Fig. 3) and went down to initial levels (n = 6, P = 0.64, Fig. 3) after 10 min or 2 h of reperfusion, respectively, compared with that from mice subjected to 20-min ischemia (Fig. 3). The ratios were much lower than those from vehicle-treated mice subjected to 10-min or 2-h reperfusion, respectively (Fig. 3), supporting the conclusion above that cytochalasin D protects the actin cytoskeleton from reperfusion-mediated disruption (Fig. 2, H and I). These data strongly support the hypothesis that cytochalasin D alters ischemia and reperfusion-mediated disruption and aggregation of the actin cytoskeleton.
Ischemia-mediated aggregation of the actin cytoskeleton leads to deposition of IgM and C3.
Ischemia-mediated deposition of natural IgM and C3 in small intestinal villi has been shown here and by others (48), and ischemia-mediated aggregation of the actin cytoskeleton is demonstrated above. On this basis, we hypothesized that ischemia-mediated aggregation of the actin cytoskeleton directly results in the deposition of natural IgM and C3. To test this hypothesis, we determined whether cytochalasin D can attenuate ischemia-mediated deposition of IgM and C3/C3d. Mice treated with 1 μM of cytochalasin D or vehicle were subjected to various durations of ischemia and reperfusion. Frozen sections of jejunal segments were stained for IgM and C3/C3d plus phalloidin and Hoechst 33342. Confocal images demonstrate decreased deposition of IgM and C3/C3d on the villi from cytochalasin D-treated mice subjected to various durations of ischemia and reperfusion (Fig. 6) compared with the villi from vehicle-treated mice (Fig. 4). However, cytochalasin D does not interfere with the natural deposition of IgM or C3/C3d in blood vessels (Fig. 6). These results clearly indicate that the optimal concentration of cytochalasin D attenuates the deposition of IgM and C3, and this strongly supports the hypothesis that ischemia-mediated aggregation of the actin cytoskeleton directly leads to the deposition of natural IgM and C3.
In the present study, we investigated the alteration of the actin cytoskeleton during ischemia with or without reperfusion. We demonstrated that ischemia results in multiple cleavages of β-actin (Fig. 1) and aggregation of the actin cytoskeleton after a transient disruption (Figs. 2 and 3). The aggregation of the actin cytoskeleton mediated by ischemia induces deposition of natural IgM and C3 (Fig. 4). Furthermore, cytochalasin D attenuates the aggregation of the actin cytoskeleton (Fig. 2), the subsequent deposition of natural IgM and C3 (Fig. 6), as well as IR injury (Fig. 5).
It has been reported that the actin protein is cleaved by interleukin-1β converting enzyme (ICE) family proteases, which are active players of programmed cell death (apoptosis) (30), to produce three major peptides (40–41, 30–31, and 14–15 kDa) in vitro and in cultured apoptotic cells (19, 28). We demonstrated, in vivo, employing 2-DIGE and Western blot analysis in this report, that ischemia-mediated cleavage of β-actin in mouse small intestine yields bands with molecular weights of ∼30 and 40 kDa (Fig. 1). Our results reveal that ischemia-mediated cleavage of β-actin occurs, in vivo, in tissues.
Ischemia-mediated disruption of vascular smooth muscle cells and epithelial cells in the kidney has been reported (21, 23). However, ATP depletion induced by antimycin A in the proximal tubule-derived LLC-PK1 cell line results in disruption of the cortical cytoskeleton, and at same time actin monomers significantly convert into filamentous actin to form large cytoplasmic aggregates (32). Furthermore, energy depletion leads to disintegration of filamentous actin and formation of numerous small clumps of filamentous actin in the cytoplasm of cultured aortic endothelial cells (22). Using an intestinal IR injury model in which small intestinal villi are vulnerable to IR insult, we demonstrated in vivo by phalloidin staining that the actin cytoskeleton in the columnar epithelial cells of the small intestinal villi is disrupted initially and then aggregates during a short term of ischemia (Fig. 2). We also demonstrated this result quantitatively by calculating ratios of G/F-actin (Fig. 3). We were unable to measure concentrations of global actin and filamentous actin using the assay of DNase I inhibition described previously (5) because of very high endogenous DNase I activity in intestinal samples, which may come from the pancreas. Our finding of the initial disruption of the actin cytoskeleton is supported by observations that high cytosolic Ca2+ concentrations, which increase rapidly after 2-min exposure to a metabolic inhibitor, result in depolymerization of the actin cytoskeleton (22). Thus it is possible that during early ischemia in our study a rapid increase in free cytosolic Ca2+ results in the disruption of filamentous actin. Free Ca2+ also helps villin, one of the major actin-associated proteins (11), to break or sever filamentous actin (16). The cleavage of actin fragments by ICE family proteases demonstrate impaired capability to block DNase I activity and to polymerize normally (19). Ectopic expression of a 15-kDa fragment of actin, which is cleaved from actin filaments by ICE family proteases during apoptosis, induces morphological changes, very similar to those of apoptotic cells (29). Although we were unable to visualize the 15-kDa fragment of actin by Western blot analysis, the accumulation of these fragments may contribute to the later aggregation of actin filaments. These data support our observation of actin cytoskeletal aggregation in the small intestine. Ischemia-induced consumption of ATP in the epithelial cells may lead to the aggregation too (22, 32, 35).
Several studies have shown that IR results in the deposition of natural IgM and IgG, the activation of complement (3, 9, 12, 17, 18, 27, 36, 44, 45, 48, 51), and the subsequent formation of the membrane attack complex that results in tissue necrosis (2, 36, 51). It was suggested that natural IgM against NM-II directly causes tissue injury after IR and a peptide mimicking an epitope on NM-II blocks complement activation and prevents tissue injury (49, 50). These studies clearly indicate that the deposition of natural IgM and the activation of complement during IR is due to a turnover of antigens on cell membranes, which are not exposed to the innate immune system under physiological conditions (49). Our observations in this report support these findings. We also observed that actin filaments aggregate after a transient disruption. So we hypothesized that the alteration of the actin cytoskeleton results in the deposition of natural IgM and complement. We used cytochalasin D, a fungal metabolite that binds to the barbed end of actin filaments (37), to test this hypothesis. The fact that cytochalasin D does not inhibit glucose transport (34) makes it acceptable for our ischemia-mediated actin cytoskeletal alteration study. It has been reported that cytochalasin D inhibits the rapid polymerization of actin (10, 46). We developed a novel model that makes it possible to study the effects of cytochalasin D on cell function in vivo. In this model, various doses of cytochalasin D were injected immediately into the small intestinal lumen after the SMA was clamped. We believe that SMA clamping blocks the absorption of cytochalasin D into the blood circulation since all animals died several minutes after its injection without SMA clamping. A 1-μM dose of cytochalasin D is optimal to block both disruption and aggregation of actin filaments (Figs. 2 and 3). This finding is consistent with the observation that cytochalasin D treatment of cells induces actin aggregation while simultaneously depolymerizing preexisting actin cytoskeletal components (33). However, higher doses of cytochalasin D, resulting in disruption of actin filaments (43) and detachment of the epithelial cells from the intestinal villi (data not shown), failed to worsen tissue injury (Fig. 5). This indicates that the disruption does not contribute to IgM- or complement-mediated tissue injury (Fig. 4). We conclude that ischemia-mediated aggregation after an initial disruption induces deposition of natural IgM and C3 (Figs. 5 and 6). It is also possible that the subsequent aggregation results from the initial disruption. Under physiological conditions, polymerization and depolymerization may be well balanced whereas this balance is biased to either disruption or aggregation during ischemia.
It has been suggested that NM-II heavy chains, which were identified as targets of natural IgM, may be exposed to the innate immunity during ischemia (49). NM-II binds to actin filaments, playing a critical role in regulating cell motility and polarity (7). It has also been reported that the injection of anti-β2-glycoprotein I (β2-GPI) antibodies and anti-phospholipid antibodies into Rag1- and CR2-deficient mice restore intestinal injury (14). β2-GPI, a plasma protein, was reported to bind to phosphatidylserine (PS) on the outside of the membrane of apoptotic cells (4, 26). These observations indicate that anti-β2-GPI and anti-PS antibodies mediate tissue injury due to the formation of complexes of PS/β2-GPI/antibodies or PS/antibodies on the cell membrane, which lead to the activation of complement. It is possible that ischemia-mediated aggregation of actin filaments results in the turnover or exposure of NM-II, PS, and/or other antigens to the outer surface of cell membrane so that they can be recognized by natural IgM and IgG. Although it is challenging technically for us to test the turnover hypothesis directly using anti-NM-II or anti-β2-GPI antibodies, the blockade of ischemia-mediated deposition of natural IgM and C3 by cytochalasin D clearly supports this notion that needs further investigations. Future studies should focus on how ischemia-mediated aggregation leads to the turnover of antigens in the members.
In conclusion, we revealed that the deposition of IgM and complement during IR is a direct result of ischemia-mediated aggregation of the actin cytoskeleton.
The work was supported by Grant no. W81XWH-07-1-0286 from Medical Research and Materiel Command.
The authors thank Dr. K. Frank Austen for critical reading of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society