Protein kinase C-βII (PKCβII) is an important modulator of cellular stress responses. To test the hypothesis that PKCβII modulates the response to myocardial ischemia-reperfusion (I/R) injury, we subjected mice to occlusion and reperfusion of the left anterior descending coronary artery. Homozygous PKCβ-null (PKCβ−/−) and wild-type mice fed the PKCβ inhibitor ruboxistaurin displayed significantly decreased infarct size and enhanced recovery of left ventricular (LV) function and reduced markers of cellular necrosis and serum creatine phosphokinase and lactate dehydrogenase levels compared with wild-type or vehicle-treated animals after 30 min of ischemia followed by 48 h of reperfusion. Our studies revealed that membrane translocation of PKCβII in LV tissue was sustained after I/R and that gene deletion or pharmacological blockade of PKCβ protected ischemic myocardium. Homozygous deletion of PKCβ significantly diminished phosphorylation of c-Jun NH2-terminal mitogen-activated protein kinase and expression of activated caspase-3 in LV tissue of mice subjected to I/R. These data implicate PKCβ in I/R-mediated myocardial injury, at least in part via phosphorylation of JNK, and suggest that blockade of PKCβ may represent a potent strategy to protect the vulnerable myocardium.
- myocardial ischemia
ischemia-reperfusion (I/R) injury in the heart is caused by blockage of coronary blood flow followed by restoration of blood flow, resulting in cardiac contractile dysfunction and arrhythmias, as well as irreversible myocyte damage (21, 26, 30). Such injury represents a clinically relevant problem associated with thrombolysis, angioplasty, and coronary bypass surgery. Although the essential factors, including oxidative stress, intracellular Ca2+ overload, neutrophil activation, and excessive intracellular osmotic load leading to I/R-induced cellular injury, have been well documented to explain the pathogenesis and functional consequences of the inflammatory injury in the I/R-exposed myocardium (5, 9, 11, 17, 39), the precise intracellular signaling elements responsible for cardiac damage in I/R have yet to be fully elucidated.
Protein kinase C (PKC) family members play integral roles as signal transducers of cellular stress responses. PKC is a family of serine/threonine protein kinases that comprise at least 12 members (23). Of the various PKC isoforms, PKCβII has been reported to be preferentially activated in the membranous fraction of heart, aorta, and retina, as well as in other vascular tissues in diabetic animals (13, 14, 32). The transgenic mouse model of targeted overexpression of PKCβII in the myocardium evoked cardiac hypertrophy with decreased cardiac performance, establishing a role for PKCβII in modulating cardiac function in vivo (41). In addition, in the failing human myocardium, increased expression and enzymatic activity of the PKCβII protein have been observed (3). Our previous studies of hypoxia and I/R of the lung, as well as acute arterial injury, provided insights into a likely mechanism by which PKCβ, particularly PKCβII, was upregulated in injured vasculature and contributed to tissue injury (1, 7, 43).
Myocardial I/R injury-induced apoptotic and necrotic myocyte death is a critical event leading to heart failure and death (19). Therefore, understanding the intracellular signaling pathways that control cardiomyocyte survival and developing a logical strategy to treat this disease through suppression of cell death pathways have significant clinical implications (12, 22). It has been reported that the PKCβ inhibitor ruboxistaurin (LY-333531) could ameliorate pathological phenotypes in the hearts of the above-described PKCβII transgenic mice (36, 41). Furthermore, it has also been shown that a PKCβII peptide inhibitor attenuated polymorphonuclear leukocyte-induced post-I/R cardiac contractile dysfunction in an ex vivo model (27).
In the present study, to interrogate the hypothesis that PKCβ inactivation by gene deletion or pharmacological blockade would attenuate myocardial I/R injury in an in vivo I/R model of ischemia and reperfusion of the left anterior descending (LAD) coronary artery in mice, we employed multiple strategies to dissect the role of PKCβ and downstream target events in vivo.
MATERIALS AND METHODS
Murine model of I/R.
PKCβ−/− mice (20) have been backcrossed more than 10 generations into C57BL/6 mice in our laboratory. Wild-type C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, ME) and used as PKCβ+/+ controls. Male PKCβ−/− and PKCβ+/+ mice (8–12 wk old, 20–25 g body wt) were anesthetized and subjected to I/R according to protocols approved by the Institutional Animal Care and Use Committee at Columbia University. The coronary artery occlusion-and-reperfusion procedure was performed as described previously (10). Briefly, mice were placed in a supine position, and a length of endotracheal polyethylene (PE-90) tubing was used to provide ventilation (450 μl, 150 cycles/min) via a rodent ventilator (model 683, Harvard, South Natick, MA). The chest was opened by a lateral incision along the upper margin of the fourth rib. Ligation was performed using an 8-0 silk suture, and a tapered needle was passed under the LAD coronary artery branch; a 1-mm-section of PE-10 tubing was placed on top of the vessel, and a knot was tied in the tubing vessel to occlude the coronary artery. To ensure that no veins were occluded with this maneuver, we observed the operative area under a microscope and blanched the tissue distal to the occlusion. If veins were ligated, the ligation-affected myocardium would not change color (index of ischemia) because of blood retention. After 30 min of coronary artery ligation, we cut the knot in the PE-10 tubing to establish reperfusion. The chest wall was closed, and the animal was removed from the ventilator and kept warm by a homeothermic blanket system (Harvard). In other studies, C57BL/6 mice were fed chow containing the PKCβ inhibitor ruboxistaurin (LY-333531; 10 mg/kg daily) or vehicle chow without inhibitor from 3 wk before they were killed (1, 7). Ruboxistaurin and vehicle chow were generously supplied by Dr. Louis Vignati (Eli Lilly, Indianapolis, IN), who provided specific instructions regarding the appropriate dose of ruboxistaurin.
After 48 h of reperfusion, the area at risk and infarct size were determined as described previously (24). Briefly, the chest was reopened, and the LAD coronary artery was reoccluded through the previous ligation site. The aorta was cannulated using a section of PE-10 tubing, and 1% Evans blue dye was perfused retrogradely into the aorta and coronary artery system to allow distribution throughout the ventricular wall proximal to the coronary artery ligature to demarcate the ischemic area at risk. The nonischemic area was stained blue. The LV was then excised and sliced into five ∼1-mm cross sections below the ligature. A custom-made cassette into which four stainless steel razor blades were inserted equidistant to one another was used to produce slices of equal thickness. Sections of the ventricle were then incubated in 1.5% triphenyltetrazolium chloride at 37°C for 15 min. After the procedures, viable myocardium stained red, and the infarct appeared pale. Each slice was imaged using a microscope and a digital camera (DSL-F717 Cyber Shot, Sony). The infarct area (pale), the area at risk (not blue), and the total LV from both sides of each section were measured using an image analyzer (Axiovision Area Measurement software, Zeiss). The ratio of area at risk to LV and the ratio of infarct area to area at risk were calculated and expressed as percentages.
For antigen retrieval, paraffin-embedded sections (4 μm thick) were deparaffinized, rehydrated, quenched for endogenous peroxidase activity with 3% H2O2, and boiled in 0.01 M citrate buffer (pH 6) for 15 min in a microwave oven. After the sections were incubated for 20 min in 10% goat serum to reduce nonspecific immunoreactivity of the secondary antibody, anti-cleaved (Asp175) caspase-3 (1:200 dilution; Cell Signaling Technology, Beverly, MA) was applied for 2 h at room temperature. Binding of the primary antibody to the target antigen was detected with the Vectastain anti-rabbit kit (Vector Laboratories, Burlingame, CA) using peroxidase as the reporter and 3,3′-diaminobenzidine (brown reaction product) as its substrate. Sections incubated with rabbit nonimmune serum, instead of the primary antibody, were used as negative control.
The number of caspase-3-positive cells were counted against the number of caspase-3-negative cells under a light microscope at a magnification of ×400, the risk areas were searched, and five visual fields were chosen through an orderly shifting (epicardial, middle, and endocardial) on each slide. Four animals from each group were examined.
A spectrophotometric method (Sigma) was used for the assessment of creatine phosphokinase (CPK) and lactate dehydrogenase (LDH) in serum 48 h after reperfusion as an indicator for loss of cell membrane integrity and as a sign of cell death.
Two-dimensional echocardiography was performed in conscious mice using techniques described previously (Sonos 5500 System, Philips Medical Systems, Andover, MA) (37). Two-dimensional echocardiographic images were obtained and recorded in a digital format. Images were then analyzed offline by a single observer unaware of the murine genotypes or conditions (42).
Western blot analysis.
For detection of membrane translocation of PKCα, PKCβ isoforms I and II, PKCδ, and PKCε, cytosolic and membrane protein fractions were prepared from ventricle tissue as described elsewhere (44), and each blot was incubated with one of the following primary antibodies: anti-PKCα IgG, anti-PKCβ isoform I and II IgG, anti-PKCδ IgG, anti-PKCε IgG (Santa Cruz Biotechnology, Santa Cruz, CA), and anti-phosphorylated PKCβII IgG (Cell Signaling), as well as anti-GAPDH IgG (Abcam, Cambridge, MA). For detection of phosphorylated JNK, ERK1/2, and signal transducer and activator of transcription type 3 (STAT3) and total JNK, ERK1/2, and STAT3, total protein extracts were prepared from the LV of murine hearts using cell lysis buffer (Cell Signaling Technology). Particulate material was removed by centrifugation, and protein concentration was determined using the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA). Equal amounts of total protein (50 μg/sample) were subjected to SDS-PAGE (10%) and then electrophoretically transferred to a nitrocellulose membrane. Nonspecific binding was blocked by incubation of membranes with nonfat dry milk (5%) for 1 h at room temperature. The blots were incubated with the following antibodies: anti-phosphorylated JNK, ERK1/2, and STAT3 IgG and anti-total JNK, ERK1/2, and STAT3 IgG (Cell Signaling Technology). Each primary antibody was used at a dilution of 1:1,000, except 1:10,000 of GAPDH IgG, for 1–3 h or overnight according to the manufacturer's instructions. Horseradish peroxidase-conjugated donkey anti-rabbit IgG secondary antibody (1:2,000 dilution; Amersham Biosciences) was used to identify primary antibody binding sites.
Values are means ± SE. All analyses were performed using the Statview Statistical package (version 5.0.1). P < 0.05 was considered statistically significant.
Coronary artery occlusion and reperfusion in vivo: protective effect of PKCβ deletion on I/R injury.
To study the effect of the deletion or blockade of PKCβ on I/R injury in vivo, PKCβ+/+, PKCβ−/−, and PKCβ+/+ mice fed vehicle chow or the PKCβ inhibitor ruboxistaurin 3 wk before instrumentation were subjected to 30 min of LAD coronary artery ligation followed by 48 h of reperfusion. The infarct size, determined from triphenyltetrazolium chloride-stained sections and reported as percentage of area at risk, was ∼2.6 fold lower in PKCβ−/− than PKCβ+/+ mice (Fig. 1, A–C; P = 0.0004). We next tested the effect of the PKCβ inhibitor ruboxistaurin in PKCβ+/+ mice. Ruboxistaurin treatment was begun 3 wk before the infarction to ensure adequate levels of the agent at the time of I/R. Consistent with the significantly decreased infarct size in the PKCβ−/− mice, an ∼2.4-fold reduction in infarct size was observed in PKCβ+/+ mice fed ruboxistaurin subjected to heart I/R compared with PKCβ+/+ mice fed vehicle (Fig. 1, D–F; P = 0.0027). No differences in percentage of LV at risk were observed between PKCβ+/+ and PKCβ−/− mouse hearts or between hearts from PKCβ+/+ mice fed vehicle chow and those fed ruboxistaurin (data not shown).
To formally assess indexes of injury, we measured CPK activity and LDH release in the serum in PKCβ+/+, PKCβ−/−, and PKCβ+/+ mice fed vehicle chow or ruboxistaurin after 30 min of ischemia followed by 48 h of reperfusion. Total CPK activity was ∼5.5-fold lower in PKCβ−/− than PKCβ+/+ mice (Table 1; P = 0.0002). Similarly, total release of LDH in the serum was ∼3.9-fold lower in PKCβ−/− than PKCβ+/+ mice (Table 1; P < 0.0001). Consistent with these data, PKCβ+/+ mice fed ruboxistaurin displayed significantly reduced (∼4.7-fold) total CPK activity (Table 1; P < 0.0001) and LDH release (∼3.8-fold) compared with those fed vehicle chow (Table 1; P < 0.0001). Total CPK activity and LDH release in the serum were uniformly very low in all the sham-treated groups. Our results indicate that the inhibition of PKCβ limits I/R injury in vivo and does not differ between PKCβ−/− and PKCβ+/+ mice in the absence of I/R.
Effect of PKCβ on functional recovery after I/R injury.
To compare the functional recovery of the hearts in PKCβ+/+, PKCβ−/−, and PKCβ+/+ mice fed vehicle chow or ruboxistaurin after 30 min of ischemia followed by 48 h of reperfusion, LV function was assessed by echocardiography. Echocardiography revealed no differences in the dimensions and function, determined as the fractional area change, in PKCβ−/− vs. PKCβ+/+ control mice at baseline (Fig. 2A). In addition, heart rate was not different at baseline in PKCβ−/− and PKCβ+/+ mice or PKCβ+/+ mice fed ruboxistaurin or vehicle chow (data not shown). We assessed heart function after injury and found, after 30 min of ischemia followed by 48 h of reperfusion, a significant increase in recovery of function (64.5 ± 6.7 vs. 49.35 ± 6.31%, P = 0.008) in PKCβ−/− compared with PKCβ+/+ mice (Fig. 2A). In parallel, PKCβ+/+ mice fed ruboxistaurin also demonstrated significantly better recovery of function than PKCβ+/+ mice fed vehicle chow after 48 h of reperfusion (63.76 ± 2.66% vs. 42.5 ± 4.3%, P = 0.0007; Fig. 2B). Thus deletion of PKCβ or administration of ruboxistaurin resulted in significant cardioprotective properties in I/R.
Membrane translocation and activation of PKCβII in response to heart I/R.
When PKCβ+/+ mice were subjected to I/R, rapid translocation of PKCβII to the cell membrane was observed. A significant increase of PKCβII associated with the membrane fraction was observed after 30 min of ischemia followed by 30 min of reperfusion (P = 0.0001) and remained significantly activated until 60 min of reperfusion (P = 0.002; Fig. 3A). In parallel, a reciprocal decrease in PKCβII was associated with the cytosol fraction (Fig. 3A). These studies were performed using a PKCβII-specific antibody. In contrast, immunoblotting with an antibody specific for the PKCβI isoform in membrane fractions showed no change during ischemia or reperfusion vs. baseline (Fig. 3B). Next, we examined the patterns of three other PKC isoforms, PKCα, PKCδ, and PKCε. Compared with baseline, PKCα, PKCδ, and PKCε did not reveal any differences in membrane-translocated forms after 30 min of ischemia followed by 30 or 60 min of reperfusion (Fig. 3, C–E). These data are consistent with a previous report that these isoforms do not remain activated throughout I/R (35).
In addition, ruboxistaurin blocked the I/R-mediated increase in PKCβII, especially activated PKCβII, in the membrane fraction, as shown by Western blot with use of phosphorylated PKCβII antibody (Fig. 3F; P < 0.05). Loading controls using anti-GAPDH IgG indicated identical protein loading (Fig. 3F).
Deletion of PKCβ modulates I/R-mediated activation of signaling pathways.
We next sought to examine the signaling pathways modulated by PKCβ deletion in the mouse heart. We first considered the impact of deletion of PKCβ in I/R on activation of JNK, an important regulator of apoptosis. Immunoblots of extracts from I/R-induced hearts retrieved from PKCβ+/+ mice displayed a time-dependent increased intensity of phosphorylated (Thr183/Tyr185)JNK/total JNK. A peak increase (∼6.9-fold) in phosphorylation of JNK was observed after 30 min of ischemia followed by 30 min of reperfusion in PKCβ+/+ hearts compared with uninstrumented PKCβ+/+ hearts (Fig. 4, A and B; P < 0.001). In comparison, PKCβ−/− mice displayed an increase of only ∼2.3-fold in phosphorylation of JNK in I/R-induced hearts compared with uninstrumented PKCβ−/− hearts, representing an ∼3-fold lower level of phosphorylated JNK in PKCβ−/− than PKCβ+/+ hearts in I/R (Fig. 4B; P < 0.01).
Next, we evaluated the effect of ablation of PKCβ in I/R-induced hearts on distinct signaling molecules implicated in cell stress, such as ERK1/2 and STAT3. Immunoblots of extracts from PKCβ+/+ hearts subjected to 30 min of ischemia followed by 30 min of reperfusion displayed ∼3.2- and ∼7.2-fold increases in intensity of phosphorylated ERK1/2 and phosphorylated STAT3, respectively, compared with uninstrumented PKCβ+/+ hearts (Fig. 4, C and D; P < 0.001). In parallel, the level of phosphorylation of ERK1/2 or STAT3 in PKCβ−/− hearts subjected to 30 min of ischemia followed by 30 min of reperfusion was enhanced compared with that in uninstrumented PKCβ−/− hearts to the same degree as in the PKCβ+/+ hearts (Fig. 4, C and D). These data demonstrated that PKCβ modulated phosphorylation of JNK, but not ERK1/2 or STAT3, in the I/R heart.
Deletion of PKCβ modulates activation of caspase-3.
In view of the modulation of proapoptotic signaling pathways in PKCβ−/− hearts, we tested caspase-3 activity in hearts retrieved from PKCβ+/+ and PKCβ−/− mice after 30 min of ischemia followed by 48 h of reperfusion. Positive caspase-3 staining in the I/R heart was reduced ∼9.3-fold in the LV of PKCβ−/− hearts compared with PKCβ+/+ hearts (Fig. 5; P < 0.0001).
The present study presents the following novel findings. 1) In the heart, membrane translocation of PKCβII is sustained during 60 min of reperfusion following 30 min of ischemia. Translocation of PKCβII was selective, and activation (phosphorylation) of PKCβII was blocked by the PKCβ inhibitor ruboxistaurin, inasmuch as membrane translocation of other isoforms of PKC, such as PKCα, PKCβI, PKCδ, and PKCε, was not observed during this time course. Importantly, these studies were performed in the absence of a preconditioning stimulus and in a fully in vivo setting in which flowing blood and cells innate to the heart contributed to I/R injury. 2) With the use of PKCβ−/− mice and PKCβ+/+ mice fed ruboxistaurin, we demonstrated that deletion and pharmacological blockade of PKCβ, respectively, resulted in a significantly decreased area of infarction and attenuation of cardiac dysfunction and reduced markers of cellular necrosis. 3) Genetic deletion of PKCβ significantly decreased phosphorylated JNK, but not ERK1/2 or STAT3, in the heart after I/R and diminished activated caspase-3 expression in the LV, in parallel with significantly improved cardiac function.
A number of studies have focused attention on the PKCβ isoform because of its potential role in cardiovascular dysfunction. For example, aortic banding-induced pressure-overload hypertrophy in the rat heart has been shown to preferentially translocate PKCβII and PKCε (8), whereas angiotensin II-induced hypertrophy in the rat selectively translocates PKCβ alone (29). In contrast with pressure-overload hypertrophy in the rat heart, aortic banding in the guinea pig induces translocation of PKCα, PKCε, and PKCγ (16). In the human myocardium, limited investigation suggests that the development of cardiac hypertrophy and failure may involve PKCβII (3). Support for roles of PKCβII was suggested by experiments in which transgenic overexpression of PKCβII in mice induced cardiac hypertrophy and failure (41). In addition, it appears that very high levels of PKCβ can result in myocyte death, as seen in the mouse overexpressing intact PKCβII from birth (41). Our study focused on PKCβ because of our observations of translocation and activation of PKCβII in PKCβ+/+ mouse hearts on myocardial I/R injury in an in vivo model, and not one characterized by ischemic preconditioning. In addition to studies employing the PKCβII−/− mouse, our experiments elucidated the pathogenic role of PKCβ by demonstrating the cardioprotective effect of oral administration of a PKCβ inhibitor, ruboxistaurin. In the in vivo model employed in our studies, contributions of PKCβ may ensue from multiple cellular sources, such as cardiomyocytes and infiltrating inflammatory cells.
Interestingly, Arikawa et al. (2) showed that the PKCβ inhibitor ruboxistaurin improved glucose utilization and diastolic function in diabetic animals. Since interventions that modulate substrate metabolism favorably have been shown to improve energy metabolism and protect ischemic myocardium (28, 34), it is conceivable that the metabolic benefits shown by Arikawa et al. may have contributed to the protection of ischemic myocardium in PKCβ−/− mice.
Translocation and activation of PKC are controversial in myocardial ischemia. Boyle et al. (4) reported an increase in PKCβI and PKCβII at 4 wk after induction of chronic myocardial infarction in rats, although only PKCβI was reduced with ruboxistaurin treatment. PKCα expression was unaffected by myocardial infarction or ruboxistaurin, and LV fibrosis and dysfunction following myocardial infarction were partially reduced by ruboxistaurin-induced inhibition of PKCβ (4). Boyle et al. began administration of oral ruboxistaurin 1 wk after myocardial infarction, rather than before or immediately after the LAD ligation. In the present studies, we began ruboxistaurin treatment 3 wk before the LAD ligation. In this context, a potential limitation of the present study is that ruboxistaurin may affect PKCβI and PKCβII (14), whereas other isoforms of PKC (i.e., α, γ, δ, ε, and ζ) are not affected (14). However, our data reveal no changes in PKCβI translocation over the same time period in which PKCβII was greatly impacted. Taken together, it is likely that the primary target of ruboxistaurin was the PKCβI.
In the present study, we hypothesized that intracellular signaling stimulated via PKCβ contributed integrally to I/R in the setting of in vivo occlusion and reperfusion of the LAD coronary artery. Our findings suggest key roles for PKCβ in regulation of JNK MAP kinase signaling in the I/R heart, inasmuch as PKCβ−/− mice displayed significantly reduced phosphorylated JNK in the heart after 30 min of ischemia followed by 30 min reperfusion. Consistent with the concept that such reduction in phosphorylated JNK was linked mechanistically to cardioprotection, the work of others illustrated the beneficial effects of blockade of JNK signaling in myocardial I/R injury. Milano et al. (25) showed that administration of a cell-penetrating, protease-resistant peptide inhibitor of JNK before ischemia significantly reduced cardiac dysfunction and enhanced posthypoxic recovery of systolic and diastolic function in the isolated perfused heart in rats. Furthermore, they illustrated the beneficial effects of JNK inhibition on reduction of infarct size in coronary artery occlusion and reperfusion (25). Ferrandi et al. (6) showed that infarct size was significantly reduced by administration of AS-601245, a nonpeptide ATP-competitive JNK inhibitor, in rats undergoing 3 h of coronary artery occlusion followed by 3 h reperfusion. In parallel with significantly reduced infarct size, reduced JNK phosphorylation and apoptotic cells were evident in the heart (6).
In contrast, JNK signaling has been linked to activation of proapoptotic mechanisms and increased activity of cleaved caspase-3 (15, 33). One explanation for these disparate results may be reflected in the experimental conditions. Specifically, distinct effects of JNK inhibition on ischemic preconditioning have been observed. The situation is more complex, however. In a murine model of heart I/R injury, Kaiser and colleagues (18) found that mice deficient in Jnk1/Jnk2 or expressing dominant-negative JNK were protected from injury. Therefore, it was surprising that, in reciprocal gain-of-function studies, mice overexpressing the upstream kinase MKK7 were also significantly protected from I/R (18). It is possible that, in these animals, subtle changes in key compensatory factors within these complex signaling networks may have masked otherwise damaging effects of JNK overactivity. In any case, the data from our present study, performed in the absence of ischemic preconditioning, strongly suggest that downregulation of JNK phosphorylation was linked to decreased caspase-3 activity and, in parallel, decreased infarction and improved functional recovery.
Taken together, deletion of PKCβ and its pharmacological antagonism in a murine model of occlusion-reperfusion of the LAD coronary artery exerted a highly protective effect, as evidenced by reduction of infarct size and preservation of cardiac function. We propose that PKCβ is a potent activator of JNK and proapoptotic signaling in the myocardium subjected to I/R in vivo. Ruboxistaurin is now the focus of clinical trial testing in human subjects (31, 38, 40); hence, administration of this agent to antagonize PKCβ may be a novel strategy to effect cardioprotection in myocardial infarction.
This work was supported by the LeDucq Foundation, the Surgical Research Fund of Columbia University, and National Heart, Lung, and Blood Institute Grant HL-60901.
We are grateful to Latoya Woods for expert assistance preparation of the manuscript.
↵* L. Kong and M. Andrassy contributed equally to this work.
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 © 2008 by the American Physiological Society