Systemic Inflammation Leads to Resistance to Atracurium without Increasing Membrane Expression of Acetylcholine Receptors

After we obtained governmental approval for the study (Regierung von Oberbayern, AZ 211–2531–70/97), 99 male Sprague-Dawley rats (Charles River GmbH, Kisslegg, Germany; weight, 250–300 g) were allowed to accommodate to the standard conditions of our animal facility with free access to chow and water for 14 days. Seventy-two rats received an intravenous injection of 56 mg/kg of a whole cell preparation of heat-killed corynebacterium parvum (Roche, Penzberg, Germany) in a total volume of 0.5 ml saline. The injection of the bacteria induces a granulomatous liver inflammation, with peak systemic inflammatory signs at approximately 5 days and an average length of the inflammation of 8 days. 17To determine the time course of the inflammation on neuromuscular pharmacodynamics, the infected rats were randomly assigned to be examined at days 2, 4, 6, 8, 10, 12, 14, or 16 after injection (11 rats per time point). The in vivo  and in vitro  variables studied at these periods included atracurium pharmacodynamics, AChR concentrations on gastrocnemius, and α¹-AGP, as well as nitrite–nitrate (NO²–NO³) concentrations in plasma. Controls (n = 11) received the same volume of saline with no bacterium. In vivo  experiments of control animals were performed at day 2 after saline injection. Because of rat deaths during induction of anesthesia or during the surgical procedures, the number of rats per group decreased. In addition, rats were excluded if they were hemodynamically unstable at the beginning of the experiments or if their hemodynamic or blood gas status at the designated measuring points were not within our predetermined range. This finally led to in vivo  experiments in 69 rats with a group split-up of 8 animals at day 2, 11 at day 4, 8 at day 6, 7 at day 8, 8 at day 10, 6 at day 12, 8 at day 14, 5 at day 16, and 8 in the control group.

Anesthesia was induced by inhalation of sevoflurane in a glass cylinder. After loss of consciousness, the rats were endotracheally intubated and mechanically ventilated with oxygen in nitrous oxide (ratio 1:2). Anesthesia was maintained with 4–6% sevoflurane. After cannulation of the left external jugular vein, anesthesia was switched to a continuous infusion of propofol (20–40 mg · kg⁻¹· h⁻¹) and fentanyl (4 μg · kg⁻¹· h⁻¹) and maintained according to cardiovascular signs of adequate anesthesia. The left carotid artery was cannulated to measure mean arterial pressure and perform blood gas analyses. Following cannulation and start of intravenous anesthesia, all animals were allowed to stabilize over a period of 60 min to eliminate any major effects of sevoflurane on neuromuscular transmission.

Before the neuromuscular experiments, ventilation was adjusted to maintain an arterial carbon dioxide partial pressure between 36 and 40 mmHg. Whenever necessary, base excess was corrected with 1 mm sodium bicarbonate to values between −2 ± 2 mm. Arterial oxygen partial pressure, heart rate, and mean arterial pressure were continuously monitored to ensure stable hemodynamic conditions throughout the experiment. Rectal temperature was controlled between 36.8 and 37.2°C with a warming blanket and heating lamp. Blood gases and hemodynamic variables were documented at two time points: (1) directly before determination of atracurium neuromuscular pharmacodynamics and (2) during maintenance of steady state neuromuscular paralysis at 50%.

Neuromuscular function was monitored by evoked mechanomyography (Myograph; Biometer, Copenhagen, Denmark). The sciatic nerve of the left leg was exposed at its exit from the lumbosacral plexus and stimulated using the train-of-four pattern (2 Hz for 2 s every 12 s). The knee was pinned and firmly fixed. A force transducer was connected to the Achilles’ tendon, and the contraction of the gastrocnemius muscle was measured. Supramaximal stimulus and control twitch height (T0) were established. Baseline mechanomyographical response was stabilized over a period of 10 min before determining the individual dose–response relation of atracurium in each rat using the cumulative dose–response method described previously. 20Bolus doses of atracurium were given intravenously in increments between 0.2 and 0.8 mg/kg until the first twitch of the train-of-four (T1) was below 5% of baseline values. Each incremental dose was given only when the previous dose had produced maximal effect, as indicated by three equal consecutive T1 twitches. After the last dose of atracurium, twitch response was allowed to recover to baseline values. The recovery interval was calculated as follows: recovery interval [s]= time at (T1/T0 = 75%) − time at (T1/T0 = 25%). Following complete recovery of T1, a continuous infusion of atracurium was started, and the infusion rate was adjusted to achieve a constant T1/T0 of 50%. Following 10 min of stable T1/T0 = 50% at a certain infusion rate, steady state conditions were assumed. The required infusion rate was documented, and 1 ml of heparinized blood was withdrawn to determine total plasma concentrations of atracurium. The blood was immediately transferred to Eppendorff tubes containing 20 μl 1 m H²SO⁴and centrifuged (3,500 rpm, 10 min, 4°C). The supernatant was collected, and 0.2-ml portions were aliquoted into Eppendorff tubes containing 0.8 ml 15 mm H²SO⁴. The samples were immediately frozen at −70°C. Directly after blood sampling, both gastrocnemii muscles were dissected from of the surrounding structures and rapidly frozen in isopentane precooled in liquid nitrogen and stored at −70°C. Following this, animals were killed by exsanguination. Blood collected during exsanguination was separated into plasma by centrifugation (3,500 rpm, 10 min, 4°C) and immediately stored at −70°C for later determination of α¹-AGP and NO²–NO³plasma concentrations.

The muscle was homogenized, the protein extracted, and the amount of membrane AChRs quantified by the ¹²⁵I-α-bungarotoxin binding assay as described previously. 21The protein concentration of the muscle extract was assayed using the Bio-Rad DC protein assay kit (Bio-Rad Laboratories, Hercules, CA), and the content of AChRs calculated and expressed in femtomoles per milligrams of protein.

Production of NO, a product of inducible NO synthase as a consequence of inflammation and release of cytokines, was assessed by measuring the stable metabolite, NO²–NO³, concentrations in plasma. Plasma samples were deproteinized with 0.5 m NaOH and 10% ZnSO⁴. Nitrate was then converted to nitrite using high-performance liquid chromatography (HPLC) on a cadmium column. Nitrite concentrations were determined spectrophotometrically at 540 nm using a method based on the Griess reaction. 22 

To determine the α¹-AGP in the rat serum, we established a competitive chemiluminescence immunoassay. Signaling was performed by applying horseradish peroxidase–conjugated streptavidin and quantifying the enzyme activity using an enhanced chemiluminometric method (Amerlite System; Ortho Clinical Diagnostics, Neckargemünd, Germany). Polyclonal antibodies against rat α¹-AGP were raised in rabbits using pure commercially available rat α¹-AGP (Sigma, Deisenhofen, Germany). The specific α¹-AGP antibodies were coated onto the microwell surface (Dynatech Laboratories, Chantilly, VA) via  goat antirabbit catcher antibodies (Biogenesis, Berlin, Germany). The tracer in the assay is biotinylated rat α¹-AGP. Biotinylation was performed using biotin-N -hydroxysuccinimide ester. The biotinylated protein was purified by size-exclusion chromatography using a Sephadex G-25 column (Amersham Pharmacia, Freiburg, Germany), followed by a reverse-phase HPLC on a 5-μm C¹⁸column (Machery und Nagel, Düren, Germany). The calibration curves were also produced from pure rat α¹-AGP by serial dilutions in phosphate-buffered saline containing 0.5% bovine serum albumin in the range of 0.1–100 mg/ml. Lower detection limit was found to be 0.1 mg/ml in serum.

Atracurium plasma concentrations were determined by HPLC. Briefly, 20 μg verapamil as an internal standard was added to the pretreated samples (200 μl serum + 800 μl 15 mm H²SO⁴). Samples were deproteinized using 1 ml acetonitrile and centrifuged at 5,000 g . A total of 50 μl of the supernatant was injected into the HPLC column (RP18, LiChrospher, 5 μm, 100 mm, 4.6 ID; Merck, Darmstadt, Germany). Atracurium was separated using a linear gradient elution system: from A to 100% B in 8 min (A [23.5 ml acetonitrile, 5 ml methanol, and 0.03 m K²HPO⁴, 57.5 ml, adjusted to pH 5], B [35.5 ml acetonitrile, 15 ml methanol, and 0.1 m K²HPO⁴, 47.5 ml, adjusted to pH 5], flow rate 1.7 ml/min, fluorescence detection at 240 nm excitation and 320 nm emission). A calibration solution was prepared from authentic substances with 5.0, 15.0, and 24.0 μg/ml atracurium and 1.0, 2.0, and 3.0 μg/ml laudanosine in 0.005 m H²SO⁴, pH 2.5. The plasma clearance of atracurium during steady state conditions was calculated by the equation: clearance = infusion rate/plasma concentration.

The impact of altered α¹-AGP concentrations on the plasma binding (or free fraction) of atracurium was evaluated post hoc  using the plasma samples that were drawn during exsanguination. The total plasma concentrations of atracurium in these (not acidified) samples are expected to be lower than the values collected and measured during steady state 50% depression. This lower concentration is due to the time effect of later sampling (during exsanguination) and the spontaneous degradation of atracurium at pH 7.4 even though in the freezer. The binding conditions, however, are maintained. We therefore divided each plasma sample into two parts: one part was processed as it was, and the other was warmed to 37°C followed by the addition of atracurium to increase plasma concentrations to an arbitrary concentration. In both (original as well as atracurium-spiked) samples, we assessed the total (bound and unbound) plasma concentrations of atracurium and, in addition, the plasma concentrations of bound atracurium. For the assessment of the protein-bound component, atracurium was separated from the unbound fraction using a microdialyzer system (Pierce, Rockford, IL) with an appropriate dialysis membrane (molecular weight cutoff of 8 kd). A 0.08 m phosphate buffer, pH 7.2, was used as dialysis buffer. All plasma samples were dialyzed for 40 min at room temperature. The plasma concentrations of bound and total atracurium were measured using the same HPLC set-up as for total atracurium plasma concentrations. The free atracurium fraction of each sample was then calculated by the equation: 1 − bound plasma concentration/total plasma concentration, and the mean value of the two sample parts was calculated. The plasma concentrations of unbound atracurium that represent the steady state condition during the 50% neuromuscular paralysis were calculated using the respective total plasma concentration multiplied by the free fraction of atracurium.

Data are presented as mean and SD. The effective doses to achieve a 50% and 95% neuromuscular paralysis (ED⁵⁰and ED⁹⁵) were calculated from the cumulative dose–response curve for each rat. The individual ED⁵⁰and ED⁹⁵values were calculated by interpolation from the linear regression of the degree of blockade in logit scale and the respective cumulative dose of atracurium in log scale.

Statistical analyses were conducted using a factorial analysis of variance. To evaluate the effect of inflammation, we tested post hoc  with the Dunnett-T procedure comparing the values of groups day 2, 4, …, and 16 with the values of control rats at a 5% level of significance. To address differences at the day of inflammation, the analysis of body weight included the body weight at the day of inflammation as a covariable.

After injection of the corynebacterium parvum suspension, body weights of the rats were significantly decreased at days 4 and 6. At days 8, 10, and 12, the weights were stable and were not different from controls. At days 14 and 16, the weights were significantly increased compared with controls. NO²–NO³plasma concentrations were significantly elevated by day 4 and peaked at day 6 (relative to control rats). After this period, the elevated NO²–NO³concentrations did not reach statistical significance and gradually decreased over time (table 1). The α¹-AGP concentrations were significantly increased at day 2, with a peak elevation at days 4–6. Thereafter, these values gradually decreased and were normalized by days 12–16 (fig. 1).

Hemodynamic and metabolic variables were kept stable throughout the experiment in each rat. The values were not different between groups, and therefore the group means are reported. At the beginning of the experiments, the mean arterial pressure was 108 ± 14 mmHg, heart rate was 344 ± 35 beats/min, arterial blood pH ranged between 7.352 and 7.440, base excess was −2.6 ± 1.2 mm, and body temperature was 37.1 ± 0.2°C. During steady state atracurium infusion, the mean arterial pressure was 98 ± 13 mmHg, heart rate was 355 ± 32 beats/min, arterial blood pH ranged between 7.344 and 7.432, base excess was −2.4 ± 1.2 mm, and body temperature was 37.1 ± 0.2°C.

The ED⁵⁰and ED⁹⁵were increased in the experimental group compared with control rats starting at day 4. The increased ED values persisted until day 8 and returned to normal at later time points. The recovery from T1 = 25% to 75% did not differ between groups (table 2).

The infusion rate necessary to maintain a steady state 50% neuromuscular paralysis was significantly increased in the experimental groups at days 4, 6, 8, and 10 compared with control rats; the α¹-AGP concentrations were also the highest at the same periods. The total plasma concentrations of atracurium to maintain a steady state 50% paralysis were significantly increased at days 4, 6, and 8 compared with control rats. The plasma concentrations of unbound (free) atracurium to achieve 50% neuromuscular paralysis, however, did not differ significantly between groups. The higher atracurium requirement paralleled the higher concentrations of α¹-AGP at the same periods (fig. 1). The plasma clearance of atracurium did not differ between groups (table 2). Expression of nicotinic AChRs in the gastrocnemic muscle did not differ between groups (fig. 1).

In this study, we have demonstrated that systemic inflammation results in resistance to the neuromuscular effect of the nondepolarizing neuromuscular relaxant, atracurium. This resistance manifested itself as early as 4 days after corynebacterium parvum injection, persisted until day 8, and returned to normal at and beyond day 10. The resistance to atracurium was evidenced as increase in ED, an increase in continuous infusion rate requirement, and an increase in plasma concentration requirement for 50% paralysis. As indicated by figures 1A and 1C, the changes in plasma concentrations of atracurium to maintain a steady state 50% paralysis paralleled the changes in α¹-AGP. The similar plasma concentrations of unbound atracurium to achieve 50% neuromuscular paralysis (fig. 1B) suggest that target organ sensitivity is not changed. That target organ sensitivity is unchanged is consistent with the finding of unaltered AChR number on the muscle membrane at periods of atracurium resistance. Together, these findings suggest that increased binding of atracurium by α¹-AGP caused the increased atracurium demand. The systemic inflammatory response to corynebacterium parvum, as demonstrated previously, 7is self-limiting. Our study confirms that with decrease of inflammatory response, evidenced as normalization or near-normalization of α¹-AGP and NO²–NO³concentrations, the resistance to atracurium dissipates.

In a previous study, we demonstrated resistance to vecuronium during systemic inflammation, evidenced as increased plasma concentration requirement of vecuronium during a steady state neuromuscular paralysis. In addition, a prolonged duration of action of vecuronium was observed. 7Decreased metabolism and clearance of vecuronium by the liver explained the prolonged recovery time. Atracurium, however, is a quaternary ammonium compound degraded primarily via  Hoffmann elimination. 23Uniform distribution and degradation of atracurium were attempted by maintaining hemodynamic variables, acid-base balance, and temperature constant, resulting in comparable plasma clearance among groups. We are therefore confident that the observed differences in the effective doses and plasma concentrations of atracurium are unrelated to altered metabolism of atracurium.

The excessive production of cytokines is a characteristic feature of inflammation. A key effect of these cytokines is the increased expression of inducible NO synthase with release of NO. 24Recent findings have shown that endogenous NO decreases submaximal force by modulating excitation–contraction coupling 25and there-fore promotes relaxation of the skeletal muscle through the cyclic guanosine monophosphate pathway. 26NO also takes part in regulating acetylcholinesterase expression in the neuromuscular junction and ion channel properties by activating cyclic guanosine monophosphate. 27NO donors have been shown to inhibit the evoked release of acetylcholine, leading to a reduction of the end-plate potential. 28Taken together, one can assume from these findings that the elevated NO concentrations in our model would have made the muscle more susceptible to paralysis rather than enhancing resistance.

Previous studies have documented a relation between up-regulation of AChRs and resistance to nondepolarizing neuromuscular relaxants. The resistance is related to both quantitative increases and qualitative (isoform) changes in the receptor, as well as electrophysiologic properties and differences in ligand-specific sensitivity or affinity between the two receptor subtypes. 11Reports of hyperkalemic responses following administration of succinylcholine in patients with prolonged severe inflammation or sepsis support the idea that sepsis, just as burns, leads to an up-regulation of AChRs. In our studies, a single insult produced by injection of corynebacterium parvum producing an inflammatory response for a few days does not up-regulate AChR. One might pose the question of whether changes in isoform (from mature to immature) may explain the resistance. This possibility is highly unlikely. In all pathologic or acquired disease states where isoform changes occur, there is concomitant up-regulation of AChR. That is, isoform changes occur paripassu  with up-regulation. AChR numbers were unchanged throughout the experiment, and therefore the thesis that changes in AChR from mature to immature isoform caused the resistance is untenable. There is also no evidence in the literature that either isoform of the AChR can exist in high- or low-affinity states when binding to muscle relaxants. In addition, since atracurium shows no differences in potency of inhibition between the two subtypes of receptor, 29an influence of sensitivity differences or binding affinity for our results (resistance to atracurium) can be excluded.

In our model, which served as a paradigm for a systemic inflammatory response, the inflammation was self-limiting, evidenced by lack of loss in body weight and decreasing NO²–NO³and α¹-AGP at or after 8 days after infection. Other factors that induce up-regulation of AChRs include immobilization and concomitant prolonged muscle relaxant therapy. These variables, however, did not confound this study. Changes in AChRs may have occurred if the inflammation had persisted for a longer period. This could be achieved by injecting a second dose of corynebacterium parvum at 4–6 days after the initial dose or by producing a more severe form of prolonged systemic inflammatory response.

Two important drug-binding proteins have been identified: albumin and α¹-AGP. Albumin binds predominantly to anionic drugs (e.g. , barbiturates, benzodiazepines). α¹-AGP is a component of globulin and binds cationic drugs, including nondepolarizing relaxants. 30During many diseases, such as burns, malignancy, chronic inflammation, and sepsis, plasma concentrations of acute-phase reactant proteins, such as α¹-AGP, are increased. 13,30In this model of a systemic inflammation, we could demonstrate that α¹-AGP concentrations were increased starting at day 2, reached the highest concentrations at days 4 and 8, and returned to normal concentrations by day 12. During a target controlled infusion of atracurium, the plasma concentrations of total atracurium to achieve 50% neuromuscular paralysis paralleled the changes in α¹-AGP, while the plasma concentrations of unbound atracurium were not altered. As indicated previously, these findings confirm the lack of change in target organ sensitivity and the role and importance of binding of atracurium to α¹-AGP in serum.

In summary, in our model of systemic inflammation, the resistance to neuromuscular blocking drugs is not due to altered metabolism or AChR expression, but represents an increase in drug binding to α¹-AGP. The resistance is self-limiting and dissipates parallel with the inflammatory response. The triggering mechanisms, mediators, and duration of infection or inflammation necessary to induce an up-regulation of AChR needs further study.