Soluble guanylate cyclase activation during ischemic injury in mice protects against postischemic inflammation at the mitochondrial level

Abstract

The aim was to determine whether treatment with BAY 60-2770, a selective activator of oxidized soluble guanylate cyclase (sGC), near the end of an ischemic event would prevent postischemic inflammation and mitochondrial dysfunction in wild-type (WT) and heme oxygenase-1 KO (HO-1^−/−) mice. This protocol prevented increases in leukocyte rolling (LR) and adhesion (LA) to intestinal venules along with elevated TNFα and circulating neutrophil levels that accompany ischemia-reperfusion (I/R) in both animal models. We further hypothesized that a component of BAY 60-2770 treatment involves maintenance of mitochondrial membrane integrity during I/R. Measurements on isolated enterocytes of calcein fluorescence (mitochondrial permeability) and JC-1 fluorescence ratio (mitochondrial membrane potential) were reduced by I/R, indicating formation of mitochondrial permeability transition pores (mPTP). These effects were abrogated by BAY 60-2770 as well as cyclosporin A and SB-216763, which prevented mPTP opening and inhibited glycogen synthase kinase-3β (GSK-3β), respectively. Western blots of WT and HO-1^−/− enterocytes indicated that GSK-3β phosphorylation on Ser⁹ (inhibitory site) was reduced by half following I/R alone (increased GSK-3β activity) and increased by one-third (reduced GSK-3β activity) following BAY 60-2770. Other investigators have associated phosphorylation of the GSK-3β substrate cyclophilin D (pCyPD) with mPTP formation. We observed a 60% increase in pCyPD after I/R, whereas BAY 60-2770 treatment of sham and I/R groups reduced pCyPD by about 20%. In conclusion, selective activation of oxidized sGC of WT and HO-1^−/− during ischemia protects against I/R-induced inflammation and preserves mucosal integrity in part by reducing pCyPD production and mPTP formation.

previously we reported that the activator of oxidized soluble guanylate cyclase (sGC), BAY 60-2770, conveyed protection from postischemic inflammatory processes when given before the ischemia-reperfusion (I/R) event (38). Preconditioning protection to the small intestine was conveyed in wild-type (WT) as well as heme oxygenase knockout (HO-1^−/−) mice, which exhibit a reduced ability to withstand oxidative stress and develop nitrate tolerance (22, 38, 39). Under these conditions, the oxidized sGC activator BAY 60-2770 was more effective in relaxing mesenteric arteries and reducing the inflammatory response of HO-1^−/− than the stimulator BAY 41-2772, which required reduced heme Fe²⁺ in sGC (14, 38). For this reason, it was suggested that pharmacologically targeting the heme-oxidized form of sGC would be a preferred approach to organ protection during ischemic disorders (30). Since ischemic events present clinically most often before a pretreatment strategy can be invoked, it is important to establish that the pharmacological treatment is effective when applied during the ischemic episode, before reperfusion is initiated. For this reason, we have now extended our previous studies to examine whether treatment with BAY 60-2770 near the end of an ischemic period would prevent postischemic inflammation l.

Mitochondrial dysfunction following I/R is a major contributor to the development of permanent damage to tissues, including the intestine and heart. We have previously reported that I/R of the small intestine induced depolarization of mitochondria membranes, reduction of respiratory activity, and cytochrome c release (29). Likewise, mitochondrial dysfunction was shown to play a major role in the heart, leading to the development of infarcts following I/R (11, 34). An important event in the transition from functional to dysfunctional mitochondria is the formation of permeability transition pores (mPTP). The mPTP allow equilibration of solutes (<1,500 Da), leading to swelling and loss of mitochondrial components and related function (12). Other work indicates that agents that increase protein kinase G (PKG) activity have cardioprotective effects mediated via maintenance of mitochondrial membrane integrity (11, 12, 34, 40). Thus, a second aim of our study was to determine whether treatment with BAY 60-2770 would preserve mitochondrial integrity and limit opening of mPTP pores in enterocytes following I/R.

Recently, it was proposed that dimers of the mitochondrial ATP synthase can form mPTP, which can explain the long-known effect of cyclophilin D (CyPD) as a cofactor in mPTP, and that CyPD binding to the inner membrane of mitochondria was blocked by cyclosporin A (CsA) (5, 12, 17). Other studies have shown that CyPD binds to the outer stalk of ATP synthase by a process that is also antagonized by CsA (4, 16). Also, mice genetically deficient in CyPD are more resistant to mPTP opening than WT mice (2, 3). Both WT mice treated with CsA and CyPD-deficient mice demonstrate reduced injury in postischemic myocardium, observations that strongly support a role for the opening of mPTP in the pathogenesis of I/R (24). However, the efficacy of CsA treatment in the prevention of postischemic inflammation has not been examined in previous work, and thus it constituted an important aim for the present study.

Since CyPD is constitutively present in mitochondria, a question arises as to what factors are involved in promoting and/or inhibiting its interaction with mPTP. It appears that phosphorylation of CyPD via glycogen synthase kinase-3β (GSK-3β), and possibly other kinases, during oxidative stress makes an important contribution to mitochondrial dysfunction (5). Importantly, the formation of mPTP in response to oxidative stress occurred more readily when GSK-3β was not phosphorylated (active form) (33, 42). Since inhibition of GSK-3β is cardioprotective in the setting of I/R (18, 28), and GSK-3β is negatively regulated by phosphorylation on Ser⁹, which is the target for several kinases, including Erk1/2 (25), we hypothesized that the phosphorylation status of these kinases would be altered by I/R and modulated by BAY 60-2770. This is an important question because this kinase relationship has been identified only in transformed cells maintained in cell culture (10, 33) and as yet has not been observed in cells isolated from WT and HO-1^−/− animals subjected to I/R. Likewise, the interplay between oxidized sGC activation by BAY 60-2770 with subsequent GSK-3β phosphorylation and mitochondrial protection has not been addressed in I/R, and so it became an important aim of our study.

MATERIALS AND METHODS

Animals

WT H129 breeder mice were purchased from Harlan Laboratories (Indianapolis, IN), whereas HO-1^−/− breeders that had been derived from H129 mice were received as a gift from Dr. William Fay (University of Missouri-Columbia, Columbia, MO). All of the female HO-1^−/− used in this study were bred in the University of Missouri facilities and used at 8–12 wk of age. Genotyping was done on lysates from tail samples according to previously used methods (22). Experimental procedures were performed in compliance with institutional guidelines for humane animal care and use and approved by the Institutional Animal Care and Use Committee.

Surgical Procedures and Induction of I/R

The procedures and fluorescent labeling of leukocytes were similar to those used previously in our laboratory (15, 43, 44). Briefly, the mice were anesthetized with a mixture of ketamine (150 mg/kg body wt) and xylazine (7.5 mg/kg) followed by a midline abdominal incision. The superior mesenteric artery (SMA) was identified and later occluded for 45 min (0 min for shams). Only for experiments in which leukocytes were to be observed was the left jugular vein cannulated to administrate the carboxyfluorescein diacetate succinimidyl ester (Molecular Probes, Eugene, OR). A 5-min infusion of the fluorescent probe allowed leukocyte-endothelial interactions to be observed via intravital fluorescence microscopy over minutes 30–40 and 60–70 of reperfusion.

Intravital Fluorescence Microscopy

As in previous studies, the mice were placed on a Plexiglas board, and a section of the small intestine was exteriorized over a glass coverslip and superfused with a bicarbonate-buffered saline (37°C, pH 7.4). Body temperature was maintained between 36.5 and 37.5°C. The Plexiglas board was mounted on the stage of an inverted microscope (Eclips TE2000; Nikon), and the submucosal venules in the intestinal microcirculation were observed. Fluorescent images (excitation, 420–490 nm, emission 520 nm) were detected with a charge-coupled device (CCD) camera (Photometrics CoolSnap). Images were projected onto a television monitor (PVM-1953MD; Sony) and recorded on a DVD recorder (DMR-E50; Panasonic). A time-date generator (WJ810; Panasonic) displayed this function on the monitor.

The intestinal segment was scanned, and 10 single unbranched submucosal venules (20–50 μm in diameter and 100 μm long) were observed for ≥30 s. Leukocyte-endothelial interactions (no. of rolling and no. of firmly attached) were quantified in each venule, followed by calculation of the mean for the 10 venules. Leukocytes were considered to be adherent if they did not move for ≥30 s. Rolling cells were defined to be those passing a cross line at a velocity significantly slower than the center line velocity and are expressed as rolling cells per minute. The numbers of adherent cells were normalized in terms of mm² surface area.

In Situ Experimental Protocols

The experimental design for each group appears in Fig. 1 and is described below. Identical protocols were used for WT and HO-1^−/− mice, with five to eight mice in each group.

Fig. 1.

Protocols used for the in situ experiments. Time 0 reference is the time of reperfusion. ▲, Time of drug injection; cross-hatched bars are time of video recording. Tissue samples were taken after the second recording. Bay 60, BAY 60-2770; I/R, ischemia-reperfusion; CsA, cyclosporin A.

Group 1: sham.

The surgical procedures were identical to the other groups. Treatment was with vehicle only (no drugs), and the exposed SMA was not occluded.

Group 2: BAY 60-2770.

The surgical procedures were identical to the other groups. Treatment was with BAY 60-2770 (30 μg/kg iv), and the exposed SMA was not occluded.

Group 3: I/R.

This is similar to group 1, but with added occlusion of the SMA for 45 min.

Group 4: treatment with BAY 60-2770 + I/R.

Mice received BAY 60-2770 (30 μg/kg iv) after 35 min of ischemia (10 min before reperfusion).

Experiments were also conducted in which CsA (CyPD antagonist) or SB-216763 (11, GSK-3β antagonist) was injected (CsA, 10 mg/kg iv; SB-216763, 0.6 mg/kg iv) in place of BAY 60-2770.

Circulating Neutrophil Counts

Whole blood was obtained via cardiac puncture following the reperfusion recordings. Samples were diluted 1:20 with 1% gentian violet solution, and total leukocytes were counted by means of a hemocytometer. To obtain the values for neutrophils, a differential count was done on samples stained with Wright Giesma stain, and the product of the total leukocyte count and the percent neutrophils yielded the number of cells per micoliter whole blood.

TNFα Assay

As in previous studies, segments of jejunum from sham mice or those subjected to the BAY 60-2770 and I/R protocols were ground in liquid N₂ and then homogenized in 1 ml of lysate buffer (10 mM Tris·HCl, pH 7.4, 250 mM sucrose, 20 mM HEPES, 1 mM EDTA, 1 mM PMSF, and 10 μl/ml protease inhibitor cocktail) and sonicated for 20 s. The homogenate was centrifuged at 12,000 g for 20 min at 4°C, and aliquots of the supernatant were stored at −70°C. TNFα was measured in duplicate by ELISA (KMC3012; Invitrogen). The minimum detectable level of TNFα was ≤4 pg/ml, and levels were expressed as picograms per milligram of protein.

Calcein AM

Cells were gently scraped from the inner lining of the jejunum as described and termed “enterocytes” by Henninger et al. (19) and LeGrand and Aw (27). The enterocytes were diluted about 30-fold in HEPES-buffered salt solution and used similarly to our previous study (29). Since openings of the mitochondrial transition pore can be monitored directly in intact cells by changes in mitochondrial calcein fluorescence (32), we incubated the enterocytes with 4 μM calcein AM (M-34153; Molecular Probes) for 45 min in the dark at 35°C. Excess calcein AM was removed by centrifugation at 500 rpm for 5 min at 4°C, and the fluorescence of the resuspended enterocytes was measured in the presence of CoCl₂ (1 mM) at excitation/emission 494/517 nm. Ionomycin (10 μM), which induces mPTP, was incubated with calcein AM for the baseline readings.

JC-1

Enterocytes were isolated as in the calcein AM method but diluted 60-fold into a HEPES-HCO₃-buffered salt solution, as in our previous study (29). The suspension was incubated with 4 μg/ml JC-1 (ab141387; Abcam) for 15 min in the dark at room temperature. The fluorescence was measured in duplicate at excitation/emission 550/600 nm (red) and at excitation/emission 485/535 (green). Mitochondrial depolarization is indicated by a decrease in the red/green fluorescence intensity.

MitoTracker Green FM

Enterocytes were isolated as in the calcein AM method and then diluted about 30-fold in HEPES-buffered salt solution. Cells were incubated for 15 min at 35°C in the dark without (blank) and with 50 nM MitoTracker Green FM (M-7514; Life Technologies), a compound that selectively labels mitochondrial protein (8). A blank with no cells and 50 nM MitoTracker Green was also read. Fluorescence was measured at excitation/emission 485/528 nm, and cell samples were corrected for the blank readings and normalized to the protein content of the sample. Each experimental batch (4 in all) consisted of WT and HO-1^−/− mice, and the HO-1^−/−/WT ratio was determined for each batch. The average and SE for the four batches were calculated and compared with the null hypothesis: ratio = 1.0.

Immunoprecipitation

Because a selective antibody against CyPD phosphorylation (pCyPD) was not commercially available, an immunoprecipitation method was adopted. Enterocytes were isolated as in the calcein AM method before the protein was extracted by the RIPA lysis buffer. Fifty microliters of Protein G UltraLink Resin beads (no. 53125; Thermo Scientific) was incubated with anti-phospho-(Ser/Thr) (no. 9631; Cell Signaling Technology) primary antibody (1:100) before being coimmunoprecipitated with the cell lysate. The beads were separated from the supernatant by centrifugation. After the beads of the nonbound components were washed, the bound protein was eluted with an SDS buffer. The pCyPD that had been eluted was then probed with anti-CyPD (AP1035; Millipore; 1:1,000) using Western blot analysis. The supernatant CyPD was also measured and total CyPD (CyPD + pCyPD) in the starting cell lysate calculated, and the ratio of pCyPD/total CyPD was computed.

Western Blots

The enterocyte proteins were separated by 10% SDS-polyacrylamide gel electrophoresis. After transfer to a nitrocellulose membrane, the blots were blocked with phosphate-buffered saline and nonfat milk (5%) and then incubated with antibodies directed against CyPD, GSK-3β (no. 12456; Cell Signaling Technology), (pSer⁹) GSK-3β (no. 5558; Cell Signaling Technology), ERK1/2 (no. 4695; Cell Signaling Technology), and phospho-ERK1/2(no. 4370; Cell Signaling Technology), along with anti GAPDH (MAB374; Chemicon) as a loading reference. The membranes were washed in phosphate-buffered saline, incubated with horseradish peroxidase-conjugated anti-rabbit IgG (no. 7074S; Cell Signaling Technology) or anti-mouse IgG antibodies (no. 7076S; Cell Signaling Technology), and developed with commercial chemiluminescence reagents (GE Healthcare). The expression of proteins was quantified by standard scanning densitometry.

Drugs

CsA and SB-216763 were from Sigma-Aldrich. BAY 60-2770 was a gift from Bayer Schering Pharma (Wuppertal, Germany). DMSO was used to prepare stock solutions of CsA, SB-216763, and BAY 60-2770 and was used as the control vehicle.

Data Analyses

Data are presented as means ± SE. Student's t-test was used to test for differences between group means where one treatment was being evaluated. ANOVA and the Newman-Keuls method applied post hoc were employed for evaluating multiple comparisons of treatments within a group, e.g., WT. The nonparametric Mann-Whitney U-test was used to evaluate differences between the WT and HO-1 groups. P < 0.05 was deemed to be significant.

RESULTS

Inflammatory Responses

The means for the numbers of rolling and adherent leukocytes for the experimental groups 1–4 appear in Fig. 2, A and B. After 30-min reperfusion, significant increases occurred for both measurements in WT as well as HO-1^−/− mice, which had not received BAY 60-2770. Similar results were also observed at 60-min reperfusion (data not shown). ANOVA indicated that the rolling leukocyte response to I/R was greater in the HO-1^−/− (P < 0.02). No significant increases in leukocyte rolling or adhesion occurred after 30-min reperfusion in WT and HO-1^−/− mice that were treated with the BAY 60-2770 during ischemia. These results were similar to our earlier observations with BAY 60-2770 preteatment at 10 min or 24 h prior to induction of I/R (38). It is also worthwhile to note that CsA and SB-216763 were protective in WT, yielding a response similar to BAY 60-2770 treatment during I/R. Insufficient HO-1^−/− were available to conduct the CsA and SB-216763 measurements.

Fig. 2.

Leukocyte rolling (A) and adhesion (B) in wild-type (WT) and heme-oxygenase-1 knockout (HO-1^−/−) mice after 30-min reperfusion. Sham represents controls. BAY60DIR, BAY 60-2770 injected during ischemia; CsADIR, CsA injected during ischemia; SBDIR, SB-216763 injected during ischemia. The vertical bars represent means + SE. *Significant differences (P < 0.05) for each group (n = 5–7) compared with sham controls.

Measurements of jejunum TNFα, shown in Fig. 3, were done to determine whether treatment with BAY 60-2770 produced an anti-inflammatory effect upstream to adhesive events in postcapillary venules. A significant, three- to fourfold increase in TNFα occurred after I/R in WT and HO-1^−/− mice. Treatment with BAY 60-2770 prevented the increase in WT and HO-1^−/−mice, similar to its effect on the leukocyte-endothelium interactions. Treatment of WT with CsA yielded results similar to BAY 60-2770. Also, there was a tendency for HO-1^−/− to exhibit higher content of TNFα than WT under equivalent conditions (P < 0.05).

Fig. 3.

Effects of treatments on TNFα content of the jejunum from WT and HO-1^−/− mice. The notations along the x-axis are similar to Fig. 2, with the mean shown as a bar with SE. Only the WT BAY60IR group yielded some samples that were below the detection limit. *Significant differences (P < 0.05) for each group (n = 5–7) compared with sham controls.

I/R was also associated with an increased number of circulating neutrophils in WT and HO-1^−/− mice, as shown in Fig. 4. The levels were greater in the HO-1^−/− mice than in WT mice (P < 0.05). BAY 60-2770 prevented the increase in both WT and HO-1^−/− mice. The pattern of circulating neutrophil changes in response to BAY 60-2770 was similar to that for the leukocyte-venule interactions and levels of TNFα. Overall, treatment with BAY 60-2770 during ischemia was effective in preventing the inflammatory response, as monitored by three different measurements. Also, CsA and SB-216763 treatment of WT mice yielded protection similar to BAY 60-2770.

Fig. 4.

Effects of treatments on blood neutrophils from WT and HO-1^−/− mice. The notations along the x-axis are similar to Fig. 2, with the mean shown as a bar with SE. *Significant differences (P < 0.05) for each group (n = 6 for all groups, except WT Sham; n = 9) compared with sham controls.

Mitochondrial Integrity

Two approaches were taken to measure mitochondrial integrity. One employed a fluorescent compound, calcein AM, which can penetrate the membranes of the cell and is subsequently trapped when de-esterified. A quenching agent, Co²⁺, was used to remove the cytosolic signal, with the residual signal taken to be of mitochondrial origin (35). This fluorescence quenching has been exploited for the detection of mPTP openings (32). Therefore, decreases in the calcein signal would indicate the loss of mitochondrial calcein and would be indicative of mPTP formation. Enterocytes that were taken from intestine of both WT and HO-1^−/− showed a significant decrease in calcein fluorescence after I/R, as shown in Fig. 5A. Treatment with BAY 60-2770 not only prevented the decrease in calcein but caused a significant increase compared with sham controls. CsA and SB-216763 had similar effects in WT. The calcein levels in the sham and BAY 60-2770 controls tended to be higher in the HO-1^−/− than in WT mice (P < 0.05). This could result from an increased mitochondrial mass associated with the reduced antioxidant capacity of HO-1^−/− (38, 39). Measurements of a mitochondrial protein marker, MitoTracker Green, were used as an index of mass (8). The uptake was 1.24 ± 0.07 times greater in sham HO-1^−/− than in WT mice (n = 4, P < 0.05), consistent with an increased mitochondrial mass in the HO-1^−/− mice.

Fig. 5.

Effects of treatments (30 min prior to reperfusion) on mitochondrial integrity in enterocytes from WT and HO-1^−/−mice. A: data for calcein [n = 6 for all groups, except WT BAY60Sham (control group administered BAY60 without I/R) and HO-1 KO Sham, where n = 7]. B: data for JC-1 (n = 6 for all groups, except WT Sham, I/R, BAY60Sham, and BAY60IR, where n = 11). Notations along the x-axis are similar to Fig. 2, with the mean shown as a bar with SE. + and *Significant differences (P < 0.001) for each group compared with sham controls.

A second measurement of mitochondrial integrity, JC-1, is less dependent on differences in mitochondrial mass in that it is a ratio metric. The dye that accumulates in the mitochondria under the influence of the membrane potential fluoresces at a wavelength that differs from the less concentrated levels in the cytosol (29, 31). The formation of mPTP causes the mitochondrial membrane potential to collapse with subsequent reduction in the ratio as the dye concentration equalizes. A 50% reduction was observed in both WT and HO-1^−/− following I/R, whereas no such reduction occurred with the group treated with BAY 60-2770 (Fig. 5B). CsA and SB-216763 had similar effects to BAY 60-2770 in WT mice. The two measurements of mitochondrial integrity yielded similar results: a reduced signal (40–67%) associated with I/R that was prevented by treatment with BAY 60-2770, CSA, and SB-216763.

Mitochondrial Signaling

Because increases in pCyPD were shown to facilitate mPTP formation in tumor cells (10), we measured pCyPD/CyPD to determine whether this regulator was also increased in mucosal cells that had undergone mPTP formation during I/R. Representative Western blots presented in Fig. 6A show increased immunoprecipitable pCyPD in cell lysates from I/R and reduced (pSer⁹) GSK-3β and pERk1/2 in the total cell lysate from I/R (Fig. 6A). The analyses of grouped data (Fig. 6, B–D) indicate that a significant increase occurred in pCyPD after I/R, whereas decreases occurred in (pSer⁹) GSK-3β and pErk1/2. Treatment with BAY 60-2770 during I/R prevented the changes in pCyPD and (pSer⁹) GSK-3β associated with untreated I/R. These effects occurred similarly in WT and HO-1^−/−. Moreover, BAY 60-2770 reduced pCyPD by about 35% below sham levels (P < 0.005) and raised (pSer⁹) GSK-3β by about 36% (P < 0.01). The changes observed in enterocytes from I/R are consistent with signaling patterns that were shown to facilitate mPTP formation (33). Furthermore, BAY 60-2770 treatment was efficacious in preventing these changes in both WT and HO-1^−/− mice.

Fig. 6.

Effects of treatments on signaling for permeability transition pore (mPTP) formation. A: p-Ser/Thr was immunoprecipitated from total cell lysates of enterocytes prepared from different groups of mice (see Fig. 1 for protocol); the coimmunoprecipitation of cyclophilin D (CyPD) is shown as pCyPD. Only trace amounts of (p-Ser⁹) GSK-3β and p-ERK1/2 were associated with the beads. IgG is reported as a loading control. CyPD, (p-Ser⁹) GSK-3β, GSK-3β, p-ERK1/2, and ERK1/2 were measured on the supernatants with selective antibodies (materials and methods). The notations along the x-axis are similar to Fig. 2, with the mean shown as a bar with SE. B: averaged data for CyPD. C: averaged data for GSK-3β. D: averaged data for ERK1/2. + and *Significant differences (P < 0.001; n = 6 for each group) from the other groups.

DISCUSSION

To summarize from the introduction, the objectives of this study were to determine whether BAY 60-2770 would limit I/R-induced increases in leukocyte rolling and adhesion, TNF levels, and circulating neutrophils in WT and HO-1^−/− when applied during the ischemic period. Inhibition of mPTP opening by treatment with CsA would also abrogate these postischemic inflammatory responses in WT and HO-1^−/−. This BAY 60-2770 treatment protocol would maintain mitochondrial integrity in intact mucosal epithelial cells after I/R. Changes in GSK-3β Ser⁹ phosphorylation occurred in response to I/R and would be modulated by treatment with BAY 60-2770. Levels of pCyPD are increased after I/R but are reduced by BAY 60-2770 treatment.

Our study provides positive results for each of these objectives, which link sGC activation in a mouse model to tissue and cellular protection at the mitochondrial level. BAY 60-2770 conveyed protection to I/R inflammatory injury in WT and HO-1^−/− mice when applied during the ischemic period. Mitochondrial integrity was compromised during I/R but maintained when treated with BAY 60-2770. CsA (a CyPD antagonist) and SB-216763 (a GSK-3β antagonist) had actions similar to BAY 60-2770 in WT mice. The phosphorylation of GSK-3β on Ser⁹ was reduced during I/R and increased with BAY 60-2770 treatment. CyPD, a substrate for unphosphorylated (active) GSK-3β, showed increased phosphorylation after I/R, which was reversed by BAY 60-2770. An important target of sGC activation appears to involve the inhibition of the GSK-3β-CyPD interaction and subsequent mPTP formation. Selective inhibition of GSK-3β (SB-216763) and CyPD (CSA) produced BAY 60-2770-like effects and confirm the operation of this pathway during I/R.

Inflammatory Responses

The current findings extend our previous study of preconditioning to a more clinically relevant protocol wherein the administration of BAY 60-2770 was initiated during ischemia (38). The I/R-induced changes in leukocyte rolling (LR) and adhesion (LA) were equivalent in both studies, as was the protection conveyed by BAY 60-2770 to both WT and HO-1^−/−. Likewise, the changes in TNFα and circulating neutrophils as well as the protection conveyed by BAY 60-2770 were equivalent to our preconditioning studies. Therefore, treatment with BAY 60-2770 can be instituted just 10 min before reperfusion and still abolish the proinflammatory effects of I/R. This period appears to be sufficient for the drug to circulate to the reperfused tissue during the critical period of reoxygenation when generation of reactive oxygen species (ROS) and oxidation of sGC is initiated (26). This observation is of clinical relevance, since treatment of acute arterial blockage would necessitate the application of protective measures during ischemia or subsequent to reperfusion.

Treatment with the CyPD antagonist CsA during ischemia yielded protection from the inflammatory responses that were similar to that conveyed by BAY 60-2770. This short-term action would appear to represent the acute effect of CsA on the maintenance of mitochondrial integrity rather than its long-term effect as an immunosuppressant. From this it follows that the signals that underlie the endothelial-leukocyte interactions, TNFα formation, and elevated blood neutrophils during I/R likely have a mitochondrial component. This concept is consistent with previous studies that showed that changes in mitochondrial respiration and mPTP opening govern endothelial adhesion molecule expression in response to TNFα (23). In addition, CsA has been shown to limit adhesion molecule expression in models of myocardial I/R and stroke (36, 41). These effects appear to extrapolate to human ischemic disease in that a clinical trial base has shown positive effects of CsA as a cardioprotectant during lengthy bypass procedures (2, 20).

Mitochondrial Integrity

In view of the effects of BAY 60-2770 to limit I/R-induced inflammatory responses reported herein and our previous study (38), we hypothesized that enterocyte mitochondrial function would also be protected by treatment with this BAY compound. It is important to note that an increase in leukocyte influx could increase immunocyte numbers in the epithelial isolates we obtained by mucosal scraping and could contribute to the assessment of mitochondrial dysfunction that has previously been ascribed solely to epithelial cells (19), but this is likely to be a small effect. Two methods, calcein retention and JC-1 distribution, were used to evaluate different aspects of mitochondrial integrity. The calcein AM ester readily crosses membranes but is trapped by esterase activity. After quenching the cytosolic calcein with Co²⁺, the cells exhibited a punctate appearance consistent with mitochondria that excludes Co²⁺ (data not shown and Ref. 35). Cells obtained after I/R retained about one-third of the signal noted in sham controls (Fig. 5A), which indicated mPTP opening (32, 35). The loss of the fluorescent calcein signal was blocked by BAY 60-2770, suggesting that this sGC activator prevented the opening of mPTP. This conclusion is further supported by a similar observation in enterocytes from CsA- and SB-216763-treated WT. Moreover, calcein retention was higher than sham levels in BAY 60-2770-, CsA-, and SB-216763-treated I/R. This phenomenon is less easily explained. One possibility is that the inner membrane of the mitochondria formed fewer mPTP than under sham conditions. This would assume that, under sham conditions, mPTP form spontaneously and equilibrate some of the calcein. However, treatment of the sham groups (no I/R exposure) with BAY 60-2770 did not cause a significant elevation in calcein retention. Elucidating the mechanisms underlying the phenomenon of increased calcein retention will require further study.

The second method employed to assess mitochondrial integrity employed a cationic dye, JC-1, that accumulates in the mitochondria as result of the negative charge (31). As shown in Fig. 5B, the ratio of mitochondrial to cytosolic dye fell significantly during untreated I/R, whereas BAY 60-2770, CsA, and SB-216763 treatment maintained the sham ratios. The ratio metric analysis minimizes potential problems of mitochondrial mass that can complicate the interpretation of calcein data (which may be the case for HO-1^−/−). However, nuances in timing of JC-1 loading and mitochondrial aggregate formation limit its use to qualitative evaluation of mitochondrial membrane potential (31). Despite this qualifier, the changes we observed in I/R were sufficiently large to support the conclusion that the mitochondrial membrane potential collapsed, and that this was prevented by treatment with BAY 60-2770, a conclusion again supported by similar effects of CsA and SB-216763 in maintaining the postischemic JC-1 ratio (Fig. 5B). This observation, along with the calcein data, yield a consistent finding that these treatments maintained mitochondrial integrity via inhibition of signaling events that led to formation of mPTP during I/R. Thus the evaluation of signaling events that might be involved in linking sGC activation to tissue and cellular protection at the mitochondrial level became the next objective of this study.

Mitochondrial Signaling

After a review of the literature, it became apparent that no publications directly link PKG to the regulation of mitochondrial integrity. Several of the reviews noted effects on PKC, cell pH, or Ca entry into mitochondria following I/R (20, 21). However, the preconditioning literature has proposed linkage between NO, sGC stimulation, and downstream effects on ERK and GSK-3β that resulted in cardioprotection (13, 24). BAY 60-2770 has been shown to be an effective activator of oxidized sGC, which in turn produces cGMP and activation of PKG (7, 14). However, PKG has been regarded to be primarily a cytosolic enzyme with various targets that regulate mainly Ca²⁺ transport, release, and sensitivity (6, 9, 38). The net result would be a reduced cytosolic [Ca ²⁺]_cyto. We postulated that a PKG target was also present, which provided an alternate signaling path to the regulation of mPTP formation. Based on our CsA data indicating that mPTP inhibition abrogated postischemic inflammation, we reasoned that CyPD was a downstream target of BAY 60-2770. However, the CsA effects alone do not allow one to differentiate CyPD or a phosphorylated form as the important link to mPTP formation. Experiments on tumor cells by Rasola et al. (33) indicate that pCyPD is the active form involved in mPTP formation. In that study, the pCyPD levels were dependent on the status of ERK1/2 (inactive) and GSK-3β (active). Furthermore, the GSK-3β could be precipitated with a CyPD antibody, which is consistent with CyPD binding to the pocket in the catalytic region of GSK-3β. A possible translocation of GSK-3β and subsequent interaction with CyPD would provide an important link between I/R [which is associated with the generation of reactive oxygen species (ROS)] and mPTP formation, a hypothesis supported by a recent report that showed that ROS (generated with H₂O₂) caused a translocation of GSK-3β from the cytosol to the mitochondria (37). Since the aforementioned studies were conducted on transformed cells, we also considered it necessary to conduct measurements of key signaling components on differentiated cells that were isolated acutely from the preparations used to study the protective effects of BAY 60-2770 in I/R.

Important steps associated with I/R and mPTP formation are diagrammed in Fig. 7. This is meant to provide a summary of recent studies quoted in the discussion in the context of our data as well as a presentation of working hypotheses meant to stimulate future work. We emphasized the steps (indicated by +; Fig. 7) in the hypothetical scheme that were addressed directly by the current work. Of particular importance are the effects of I/R on the progression of events from ERK1/2 (decreased phosphorylation), GSK-3β (decreased phosphorylation), and CyPD (increased phosphorylation) (Fig. 6, B–D). This progression is similar to that reported for tumor cells exposed to agents that produce ROS with resultant formation of mPTP (33). As summarized in Fig. 7, we also observed a loss of mitochondrial integrity after I/R, consistent with mPTP formation. Although we did not measure specific mitochondrial derived signals for necrosis and apoptosis, we observed three aspects of the resultant inflammatory process: increased LR and LA to venular endothelium, intestinal TNFα, and blood neutrophils (Figs. 2, A and B, 3, and 4). To our knowledge, this is the first report that links I/R to phosphorylation of CyPD.

Fig. 7.

Summary diagram based on published reports, current experiments (+) and working hypotheses concerning pathways involved in I/R-induced mPTP formation and BAY 60-2770 conditioning. Arrows represent activation, T, inhibition of GSK-3β with SB-216763 and CypD with CsA and —, the mitochondrial membranes. The perturbations that we imposed are shown inside boxes. At the top appears the BAY 60-2770 activation pathway for PKG, which in turn inactivates cytosolic GSK-3β by phosphorylation to (p-Ser⁹) GSK-3β. I/R is shown to have 3 sites of action: formation of oxidized soluble guanylate cyclase (sGC), activation of cytosolic GSK-3β via dephosphorylation, and translocation of GSK-3β across the mitochondrial membrane. Mitochondrial GSK-3β is shown to phosphorylate CypD, which Ca²⁺ augments the transition of dimeric ATP synthase to mPTP. This transition is associated with the development of highly permeable mitochondria and subsequent inflammatory responses.

We have previously proposed that the protective actions of BAY 60-2770 resulted from its ability to activate sGC/PKG-dependent signaling, which targets various sites to reduce cell [Ca²⁺] (38). The current study suggests that, in addition to this mechanism, BAY 60-2770 also confers protection secondary to maintenance of mitochondrial integrity in postischemic intestine. Of particular importance was the determination of BAY 60-2770's effects on the progression discussed above, which can lead to mPTP formation. The observation that phosphorylation of GSK-3β on Ser⁹ was increased under all conditions of BAY 60-2770 treatment is particularly relevant, as should be clear from Fig. 7. Such phosphorylation has been associated with decreased GSK-3β activity, which in turn should reduce mPTP formation. Although an extensive literature is available on multiple kinases, e.g., p-ERK1/2, that inactivate GSK-3β by this mechanism (20, 24, 33), we provide some of the first evidence establishing this link to sGC/PKG signaling. It is also possible that PKG may act on another unidentified upstream kinase that could be coupled to GSK-3β phosphorylation. We suggest that mitochondrial protection by PKG is initiated in the cytosol via inactivation of GSK-3β and subsequent translocation to the mitochondria of enterocytes. The latter speculation related to translocation will need further study to be established. Treatment with the selective inhibitor (SB-216763) of GSK-3β also provided mitochondrial protection, which is also consistent with the proposed linkages (Fig. 7). Although not a part of this study, it should be noted that Ca²⁺ is a key factor in forming mPTP (Fig. 7). For instance, the electrical activity of isolated ATP synthase and its component c-subunit were shown to exhibit Ca dependence (1, 5, 17). Despite a need to establish details of the mechanisms, regulation of ATP synthase structure is fundamental to the formation of mPTP (5). Reduced availability of Ca²⁺ to the mitochondria and reduced phosphorylation of CyPD during treatment with BAY 60-2770 could stabilize ATP synthase, making BAY 60-2770 a particularly effective conditioning agent.

Effectiveness of BAY 60-2770 in Another Model of Oxidative Stress: HO-1^−/− Mice

In addition to studying the effectiveness of BAY 60-2770 in WT mice, we also sought to determine whether this activator of oxidized sGC would be effective in HO-1^−/− mice. This is an important question because these mice display enhanced oxidative stress and do not exhibit delayed (late-phase) protection from I/R-induced inflammation afforded by preconditioning stimuli such as H₂S (43). Moreover, we showed previously that these mice demonstrate nitrate tolerance as a consequence of reduced levels of soluble guanylate cyclase (22). Our results indicate that the BAY compound confers protection in in mice lacking HO-1, even though these mice are deficient in total sGC (38). HO-1^−/− mice exhibited an exaggerated inflammatory response to I/R similar to that reported in our previous study that focused on the preconditioning effects (38). This enhanced response has been attributed to a reduced capacity of HO-1^−/− to neutralize ROS and produce CO (39) and may have been a factor in the increased mitochondrial mass we observed in the enterocytes. BAY 60-2770, which activates oxidized sGC, was as effective in HO-1^−/− mice as it was WT mice in preventing the inflammatory responses and mPTP formation associated with I/R. Our previous studies provided pharmacological evidence for increased oxidized (heme-free) sGC in the HO-1^−/− mice despite reduced total sGC levels (22, 38). The present study supports the use of a selective oxidized sGC activator under conditions of ROS exposure in models with a compromised redox capability, such as HO-1-deficient mice.

Conclusions

The results of this study indicate that treatment with BAY 60-2770 near the end of the ischemic period, a clinically relevant treatment protocol, completely prevents postischemic inflammatory responses in WT mice and in a murine model that displays nitrate tolerance, enhanced oxidative stress, and reduced total sGC, the HO-1 knockout mouse. In addition, we showed that treatment with CsA or SB-216763, inhibitors of mPTP and GSK-3β, respectively, also confers protection against the deleterious effects of I/R when treatment is initiated over the same time frame. I/R in the absence of treatment induced the formation of mPTP, whereas treatment with BAY 60-2770, CsA, or SB-216763 maintained mitochondrial integrity after I/R. I/R was also associated with reduced phosphorylation of ERK1/2 (inactivated) and GSK-3β (activated) and increased phosphorylation of CyPD. This progression has been proposed by others to be an important factor for the transition of ATP synthase to mPTP during exposure to ROS, and our findings are consistent with this proposal. Indeed, BAY 60-2770 induced an increase in phosphorylated (pSer⁹) GSK-3β (inactivated) and subsequent reduction in pCyPD levels. This sequence was confirmed with the use of the inhibitors CSA and SB-216763. These observations were made in nontransformed enterocytes that were acutely isolated from experimental animals, thus showing the relevance of the above progression to the maintenance of tissue integrity in vivo for the first time.

GRANTS

This work was supported by National Institutes of Health Grants AA-022108 and HL-095486.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.Z.W., W.Z.W., and M.W. performed experiments; D.Z.W., A.W.J., W.Z.W., and M.W. analyzed data; D.Z.W., A.W.J., W.Z.W., and M.W. prepared figures; D.Z.W., A.W.J., W.Z.W., M.W., and R.J.K. approved final version of manuscript; A.W.J., W.Z.W., and R.J.K. conception and design of research; A.W.J. and R.J.K. interpreted results of experiments; A.W.J. drafted manuscript; A.W.J., M.W., and R.J.K. edited and revised manuscript.