Sepsis Increases the Plasma Membrane Content of α1 and α2 Isoforms of Na+-K+ Adenosine Triphosphatase in Rat Skeletal Muscle

Hypothesis  Increased Na⁺-K⁺ adenosine triphosphatase (ATPase) activity in skeletal muscle during sepsis is caused by transient increases in enzyme content within the plasma membrane.

Design  Randomized controlled study.

Setting  University laboratory.

Intervention  Eighty-eight adult male Wistar rats were randomly assigned to undergo cecal ligation and puncture (CLP) or sham operation.

Main Outcome Measures  Gastrocnemius muscles were harvested 6, 12, 24, and 48 hours after operation and Na⁺-K⁺ ATPase activities were measured spectrofluorimetrically. Messenger RNA (mRNA) levels for the α1 and α2 isoforms of Na⁺-K⁺ ATPase were determined by Northern blot analysis. Crude membranes, internal membranes, and purified plasma membranes were isolated from gastrocnemius muscles and protein levels of α1 and α2 isoforms were determined by Western blot analysis.

Results  Na⁺-K⁺ ATPase activity in the CLP group was significantly higher compared with the sham group 24 hours after operation (P<.05). However, there were no differences between the sham and CLP groups 6, 12, or 48 hours after operation. No significant differences between the CLP and sham groups were noted in mRNA levels for Na⁺-K⁺ ATPase α1 and α2 isoforms. Western blot analysis revealed that the plasma membrane (but not internal membrane or crude membrane) content of α2 and α1 isoforms from the CLP group was significantly increased compared with the sham group 24 hours after operation (P<.05).

Conclusions  Na⁺-K⁺ ATPase activity increases 24 hours after CLP in gastrocnemius muscle and then declines. This increase is caused by increased Na⁺-K⁺ ATPase protein levels in the plasma membrane.

THE EFFECTS of sepsis on skeletal muscle energetics and membrane function are poorly understood. Previous studies have demonstrated that Na⁺-K⁺ adenosine triphosphatase (ATPase) activity in skeletal muscle increases during sepsis, but the molecular mechanisms by which this increase occurs have not been thoroughly investigated. Changes in Na⁺-K⁺ ATPase activity occur to modulate intracellular sodium concentrations to maintain intracellular cation homeostasis and to regulate cell volume. An increase in Na⁺-K⁺ ATPase activity increases ATP consumption. Energy demands that outstrip production cause tissue energy failure and may ultimately contribute to cell death. A better understanding of the changes in cellular bioenergetics that occur during severe infection may improve our ability to care for sick patients. This study is the first to show that the plasma membrane (PM) levels of catalytic subunits α2 and α1 of Na⁺-K⁺ ATPase are transiently increased by systemic infection.

Despite advances in critical care treatment and antibiotics therapy, sepsis and its sequelae continue to be the leading causes of death in medical and surgical intensive care units, contributing to 30% to 60% of all mortality.^1,2 One of the most important metabolic responses to severe infection is a redistribution of sodium and water between the intracellular and extracellular spaces in skeletal muscle. Na⁺-K⁺ ATPase is a transmembrane protein complex that transports 3 Na⁺ molecules out of and 2 K⁺ molecules into the cell by hydrolysis of ATP. Thereby, it maintains an electrogenic gradient that is crucial for many cellular processes. The effects of sepsis on Na⁺-K⁺ ATPase activity are controversial. Some investigators have described decreased activity, whereas in other studies the activity of the enzyme has been shown to increase.^3-5 Our previous investigations demonstrated that Na⁺-K⁺ ATPase activity is increased in the rat gastrocnemius muscle during sepsis,^6,7 but the exact mechanism of this change was unclear.

An increase in Na⁺-K⁺ ATPase activity could be provoked by an increase in intracellular sodium concentration^8-10 secondary to changes in sodium permeability. Total Na⁺-K⁺ ATPase activity could also increase secondary to changes in the activity rate of each pump or because of an increased number of pump units at the PM. The latter could occur secondary to enhanced synthesis, increased stability of existing surface enzymes, or recruitment of units from other intracellular compartments.

The smallest functional unit of Na⁺-K⁺ ATPase is an α-β subunit heterodimer.^11,12 The α subunit (112 kd) contains the catalytic activity of the enzyme and possesses binding sites for all substrates. However, its enzymatic activity relies on its association with a glycosylated β subunit (35 kd). Two α (α1 and α2) and 2 β (β1 and β2) isoforms are expressed in rat muscle. The α2 isoform is the primary isoform in skeletal muscle and α2 messenger RNA (mRNA) expression in skeletal muscle is 15- to 20-fold higher than α1 mRNA expression.^13,14 O'Brien et al¹⁵ reported that sepsis did not affect Na⁺-K⁺ ATPase protein levels; however, the PM was not actually separated from the crude membrane (CM) fraction. Therefore, the objectives of our experiments were (1) to measure the time course of changes in Na⁺-K⁺ ATPase activity after the onset of sepsis; (2) to determine the effect of sepsis on the amount of α1 and α2 isoforms in purified PM using discontinuous sucrose gradient centrifugation and Western blot analyses; and (3) to determine whether sepsis affects the mRNA expression of α1 and α2 isoforms by Northern blot analysis. We found that Na⁺-K⁺ ATPase activity increased 24 hours after cecal ligation and puncture (CLP) and then declined 24 hours later. Western blot analysis using monoclonal antibodies specific for the α1 and α2 isoforms revealed that Na⁺-K⁺ ATPase protein levels in the PM were significantly increased. Sepsis did not change mRNA levels of the α1 and α2 isoforms of Na⁺-K⁺ ATPase.

Adult male Wistar rats weighing 250 to 300 g were purchased from Harlan Sprague-Dawley Laboratories (Indianapolis, Ind) and housed in groups for at least 3 days at a constant temperature of 22°C with 12-hour periods of light and dark exposure. During acclimatization the animals were allowed rat chow and water ad libitum.

The study was approved by the animal care committee of the Harvard Medical School (Boston, Mass). The rats were weighed and anesthetized with sodium pentobarbital (50 mg/kg intraperitoneally) and randomized to receive either CLP or sham operation. The abdominal skin was shaved and scrubbed with povidone-iodine solution and the peritoneal cavity was entered using aseptic technique. Peritonitis and sepsis were induced by ligating the cecum 2 cm from its distal end to avoid intestinal obstruction and then puncturing the ligated segment once with an 18-gauge needle at the antimesenteric border. Penetration was confirmed visually. In the control rats who received the sham operation, the peritoneal cavity was entered and the cecum and colon were exteriorized, manipulated, and returned to the peritoneal cavity. Six milliliters of isotonic sodium chloride solution per 100 g of body weight was injected into the intraperitoneal cavity of all rats to compensate for volume depletion. The abdomen was then closed in 2 layers. After operation, all animals were housed individually and allowed free access to water but not to food. Mortality rates were recorded 6, 12, 24, and 48 hours after operation and fresh gastrocnemius muscles were harvested to measure Na⁺-K⁺ ATPase activity (n = 6 for each group) and stored at −80°C until use.

In the second set of experiments, rats were randomly assigned to undergo CLP or sham operation as described above. Mortality rates were recorded and gastrocnemius muscles were harvested 12 hours (n = 4 for each group) and 24 hours (n = 6 for each group) after operation for Northern blot analysis and stored at −80°C until use.

In the third set of experiments, gastrocnemius muscles were harvested from both hindlimbs 24 hours after operation (n = 6 for each group) for membrane preparation and Western blot analyses, and stored at −80°C until use.

Na⁺-K⁺ ATPase activity was determined spectrofluorimetrically in muscle homogenates from changes in 3-O-methylfluorescein phosphatase activity before and after the addition of potassium.¹⁶ The K⁺-dependent phosphatase activity correlated directly with Na⁺-K⁺ ATPase activity and the number of titrated ouabain binding sites.

Total RNA was isolated from individual muscles by the acid guanidinium thiocyanate-phenol-chloroform method¹⁷ using TRIZOL reagent (GIBCO/BRL, Rockville, Md). RNA was quantified and total RNA (20 µg/lane) was dissolved in 50% formamide with 2.2-mmol/L formaldehyde and electrophoresed in a 1% agarose, 2.2-mmol/L formaldehyde gel in 20 mmol/L 3-(N-morpholino) propanesulfonic acid. RNA samples were transferred overnight by capillary action to a nylon membrane and fixed on the membrane by UV cross-linking. The membrane was prehybridized for 15 minutes at 68°C in Quick-Hyb (Stratagene, La Jolla, Calif) to shorten hybridization time. The membrane was then hybridized with the radiolabeled α1 isoform complementary DNA (cDNA) probe for 2 hours at the same temperature. The membrane was washed as follows: 2 × SSC (1 × SSC is 0.15-mmol/L sodium chloride and 0.015-mmol/L sodium citrate, pH 7.0), 0.1% sodium dodecyl sulfate for 2 × 20 minutes at room temperature and 0.1 × SSC, 0.1% sodium dodecyl sulfate for 30 minutes at 60°C. It was next subjected to autoradiography at −80°C using intensifying screens for 24 hours. After autoradiography, the membrane was stripped and rehybridized with the α2 isoform cDNA probe, washed, and subjected to autoradiography as described above. The cDNA probes encoding the Na⁺-K⁺ ATPase α1 and α2 isoforms were obtained from American Type Culture Collection (Manassas, Va). Then the membrane was stripped and rehybridized with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA to verify the integrity of the RNA sample and the equality of the amounts of RNA loaded onto all lanes. Signal density was measured using a laser scanning densitometer and National Institutes of Health Image 1.60 software. The GAPDH probe was made using the reverse transcriptase-polymerase chain reaction. The first strand of cDNA was prepared from total RNA extracted from adult rat gastrocnemius muscles. Polymerase chain reaction amplification of the template cDNA was carried out using 5′-CAAGA TGGTG AAGGT CGGTG TCAAC G -3′ and 5′-CACAG TCTTC TGAGT GGCAG TGATG G -3′ as primers for 30 cycles of 1 minute at 95°C, 1 minute at 65°C, and 1 minute at 72°C.

Membrane preparation was performed using a modification of a previously described procedure.¹⁸ Briefly, frozen rat gastrocnemius muscles (2.5-3.3 g) were cleaned of visible fat and minced in ice-cold homogenization buffer 1 (20 mmol/L NaHCO₃, pH 7.0; 0.25-mmol/L sucrose, 5-mmol/L NaN₃, 100-µmol/L phenylmethylsulfonylfluoride, 1-µmol/L leupeptin, 1-µmol/L pepstatin A, and 10-µmol/L E-64, freshly prepared) at a ratio of 1 gram of muscle per 12 milliliters. Samples were homogenized in a glass homogenizer. The homogenate was centrifuged at 1200g for 10 minutes. The pellet was resuspended in the same volume of buffer, rehomogenized, and recentrifuged to remove debris. The supernatants were combined and centrifuged at 9000g for 10 minutes to sediment mitochondria and nuclei, and the resulting supernatant was centrifuged at 190 000g for 60 minutes to obtain the CM pellet. The CM was resuspended in homogenization buffer 1, layered on a discontinuous sucrose gradient (25%, 30%, 35% sucrose, w/w), and centrifuged at 150 000g for 16 hours at 4°C in a SW-41 rotor (Beckman Instruments, Palo Alto, Calif). Plasma membrane and internal membrane (IM) fractions were collected from the top of 25% and 35% sucrose interface, respectively, diluted 10-fold with buffer 2 (20 mmol/L NaHCO₃, pH 7.0, and 5 mmol/L NaN₃), recovered by high-speed centrifugation (200 000g for 90 minutes), and resuspended in homogenization buffer 1. Protein was determined by the bicinchoninic acid method¹⁹ using a standard kit. The PM marker enzyme (5′-nucleotidase) activity was measured in CM, PM, and IM fractions using 5′-nucleotidase reagent (Sigma-Aldrich Corp, St Louis, Mo). The fractions were then stored at −20°C and used for Western blot analysis.

Equal amounts of total protein (3, 1.5, and 1 µg) obtained from CM, IM, and PM, respectively, were separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose membranes in the presence of 20% methanol, 25-mmol/L tris(hydroxymethyl)aminomethane, and 192-mmol/L glycine, at pH 8.3. Nonspecific binding to the membrane was blocked by 10% nonfat dry milk in phosphate-buffered saline at 4°C overnight. The blots were washed in phosphate-buffered saline and then incubated with the monoclonal antibody for the α2 isoform (McB2) at 1:500 dilution for 2 hours at room temperature. For analysis of α1 isoform expression, equal amounts of total protein (13.5, 6.75, and 4.5 µg) obtained from CM, IM, and PM, respectively, were analyzed by immunoblot and probed with monoclonal antibody for the α1 isoform as described above. The monoclonal antibodies specific for the α2 isoform and the α1 isoform were donated by K. Sweadner, MD, Massachusetts General Hospital, Boston.¹² Membranes were subjected to 3 washing procedures and incubated with antimouse horseradish peroxidase-conjugated second antibody at 1:8000 dilution for 1 hour, then washed again. After incubation, membranes were washed 3 times with phosphate-buffered saline. The signal was detected by the chemoluminescence methodology of the ECL Plus System (Amersham, Arlington Heights, Ill) and exposure to x-ray film. In preliminary studies, multiple exposures were analyzed and the optimal amounts for sample loading were determined. It was ascertained that the relative intensities were directly proportional to the amount of sample loaded.

Statistical comparisons were made using analysis of variance techniques. Post hoc intergroup comparisons were performed using Statistica (StatSoft, Tulsa, Okla). P values less than .05 were considered significant.

The mortality rate after CLP was 14.3% (3/21) and 40.0% (4/10) 24 and 48 hours after operation, respectively. No rats that received sham operations died. Six animals chosen at random were subsequently analyzed in each group.

Na⁺-K⁺ ATPase activity in the CLP group increased after operation during the first 24 hours, then declined. There were no statistically significant differences in Na⁺-K⁺ ATPase activity 6, 12, or 48 hours after operation (mean ± SEM, 29.8 ± 1.9 vs 29.4 ± 2.5, 33.4 ± 1.7 vs 37.7 ± 4.2 and 22.6 ± 2.1 vs 27.9 ± 2.1 nmol/min per gram, respectively) between the sham and CLP groups. However, Na⁺-K⁺ ATPase activity was significantly higher in the CLP group compared with controls after 24 hours (mean ± SEM, 39.4 ± 5.8 vs 28.9 ± 2.6 nmol/min per gram; P<.05) (Figure 1).

To determine whether mRNA levels of α1 and α2 isoforms of Na⁺-K⁺ ATPase changed during sepsis, Northern blot analysis was performed. Because there is a time lag between changes in mRNA and protein expression, gastrocnemius muscles were harvested 12 and 24 hours after operation. Figure 2A shows representative Northern blots of α1 isoform of Na⁺-K⁺ ATPase. The α1 isoform of Na⁺-K⁺ ATPase migrated as a single mRNA band of approximately 3.7 kb. The signal intensities were normalized to the GAPDH bands (Figure 2 C). There were no statistically significant differences in mRNA levels of α1 isoform between animals in the sham group and those in the CLP group 12 and 24 hours after operation (Table 1). Figure 2 B shows representative Northern blots of the α2 isoform of Na⁺-K⁺ ATPase. As reported previously,¹³ the α2 isoform migrated as 2 mRNA species, 5.3 kb and 3.4 kb. No statistically significant differences in the α2 isoform of Na⁺-K⁺ ATPase after normalization to the GAPDH bands were noted between the 2 study groups 12 and 24 hours after operation (Table 1).

Similar amount of gastrocnemius muscle tissue from sham and CLP groups were isolated from both legs and separation of the muscle membrane fraction was performed using sucrose density gradient centrifugation. The amount of protein yielded in the CM, PM, and IM isolated by the fractionation procedure was similar between the sham and CLP groups (Table 2).

The 5′-nucleotidase activity, a PM marker, was much higher in the 25% sucrose fraction compared with the 35% sucrose fraction or CM preparation from both groups of animals. The 5′-nucleotidase activity in the 25% sucrose membrane fraction from the CLP group was significantly decreased compared with the sham group, but no changes in the 35% sucrose fraction or CM portion were detected (Table 3).

Figure 3 shows representative immunoblots of experiments investigating the effect of sepsis on the distribution of α1 and α2 isoforms in CM, PM, and IM isolated from gastrocnemius muscle. Sepsis did not alter the content of α1 and α2 isoforms in the CM fraction (mean ± SEM, 1.00 ± 0.12 vs 1.13 ± 0.16, 1.00 ± 0.03 vs 0.90 ± 0.06, respectively). The content of the α1 and α2 isoforms in the PM was significantly increased in the CLP group compared with the sham group 24 hours after operation (mean ± SEM, 1.63 ± 0.21 vs 1.00 ± 0.18 and 1.49 ± 0.09 vs 1.00 ± 0.17, respectively [ P<.05]). This increase was not associated with a reduction in the content of these isoforms in the IM fraction (mean ± SEM, 1.04 ± 0.17 vs 1.00 ± 0.15 and 1.06 ± 0.10 vs 1.00 ± 0.12, respectively) (Figure 4).

In this study, Na⁺-K⁺ ATPase activity increased by approximately 36% within 24 hours of CLP and then declined. O'Brien et al¹⁵ reported that Na⁺-K⁺ ATPase activity increased during sepsis, but that mRNA and enzyme protein levels, as determined by Northern and Western blot analyses, respectively, were unaffected. Our findings of increased Na⁺-K⁺ ATPase activity and unchanged mRNA levels of α2 and α1 isoforms are similar to those of O'Brien and colleagues. However, we discovered that α2 and α1 isoform protein levels in the PM were significantly increased. The reason for this discrepancy may be the method used to purify the PM fraction. We used sucrose density gradient centrifugation, which allowed us to study the PM and IM fractions separately. Measurements of changes in 5′-nucleotidase activity showed that the 25% sucrose fraction is enriched with PM, whereas the 35% fraction is depleted of this enzyme in both sham and CLP groups. Interestingly, sepsis decreased 5′-nucleotidase activity in the 25% sucrose membrane fraction, but had no effect on the 35% sucrose fraction. The biological significance of this finding remains to be determined.

The factors that regulate the expression of Na⁺-K⁺ ATPase isoforms in muscle during sepsis are largely unknown but may include the effects of hormones and metabolic demand. For example, aldosterone increases Na⁺-K⁺ ATPase mRNA levels in rodent smooth muscle.²⁰ Azuma et al²¹ reported that α2 and β2 mRNA and protein levels were modulated by changes in thyroid hormone. Tsakiridis et al²² showed that exercise increased the amount of α1 mRNA in soleus muscle fibers and the amount of β2 mRNA in type IIb muscle fibers, while the levels of α2 or β1 transcripts in the soleus and α1 transcripts in type IIb muscles were unaltered. In the latter instance, the exact cellular signals responsible for increasing PM subunits in response to exercise are unknown. Hundal et al,²⁵ using sucrose density centrifugation, reported that insulin induced translocation of the α2 and β1 subunits of the Na⁺-K⁺ ATPase from intracellular compartments to the PM. Further investigation is required to determine how changes in Na⁺-K⁺ ATPase isoforms and their activities are modulated in health and disease. In principle, the number of enzyme units at the cell surface could increase because synthesis increases, because existing units become more stable, or by recruitment of pump units from other intracellular compartments. Because mRNA levels of the α1 and α2 isoforms and the amounts of the Na⁺-K⁺ ATPase isoforms in the CM preparation were not changed in our study, it seems unlikely that new subunit synthesis contributed to the observed increase in PM content. Instead, our findings suggest that translocation from intracellular compartments to the PM occurs during sepsis. An increase in the content of α1 and α2 isoforms in the PM could at least partially explain the observed increase of pump activity because the α isoforms possess the catalytic activity of the enzyme. It is possible that α isoforms are recruited to the PM where they associate with β isoforms. This phenomenon is known to occur in insect sf-9 cells.²³

Both α and β subunits are required for normal activity. The β subunit is important for functional assembly of the full enzyme complex and its insertion into the PM.²⁴ Our data do not support the existence of a donor pool of preexisting pumps because the content of α1 and α2 isoforms in the intracellular fractions did not change. However, it is possible that small membrane pools were not detected by the fractionation procedure. Tsakiridis et al²² reported that exercise recruited pump subunits to the membrane from an intracellular compartment that was not isolated by fractionation. In the case of insulin, the donor pools that supply pump subunits to the PM have been identified.²⁵ Translocation of the α2 isoform in response to insulin has also been demonstrated using immunogold electron microscopy.²⁶

It is not known whether changes in hormone elaboration or changes in hormone activity affected Na⁺-K⁺ ATPase activity in our experiments. However, it is unlikely that insulin was responsible for the observed changes because sepsis is known to decrease insulin levels.²⁷ The observed changes in Na⁺-K⁺ ATPase activity may have been in response to an increase in intracellular sodium concentration. Our previous studies have shown that the intracellular to extracellular sodium content ratio, determined directly from the ²³Na magnetic resonance spectroscopy, was 95% higher in septic animals than in controls, although no differences in intracellular pH were detected.⁸ An alteration in membrane permeability in response to sepsis has been documented.²⁸

To our knowledge, our study is the first to show that the content of catalytic subunits α2 and α1 of the Na⁺-K⁺ ATPase in the PM is increased by systemic infection. De novo protein synthesis of pump subunits was not detected 24 hours after CLP. Although an intracellular membrane pool was not specifically identified, the results are compatible with either the recruitment of new subunits or with increased retention of pumps at the cell surface. Our results suggest that increased Na⁺-K⁺ ATPase protein levels in the PM are at least in part responsible for the changes in pump activity during severe systemic infection. These alterations in Na⁺-K⁺ ATPase activity may be mediated by changes in intracellular sodium content.⁸

This work was supported by grant 1-P50 GM52585 from the National Institutes of Health, Bethesda, Md.

We would like to thank Kathleen Sweadner, MD for the supply of the monoclonal anti-α1 and α2 antibodies.

Presented at the 12th annual meeting of the Surgical Infection Society–Europe, Oslo, Norway, June 5, 1999.

Corresponding author: Danny O. Jacobs, MD, MPH, Department of Surgery, Creighton University, 601 North 30th St, Suite 3740, Omaha, NE 68131 (e-mail: djacobs@creighton.edu).