Background

The clinical actions of sugammadex have been well studied, but the detailed molecular mechanism of the drug encapsulation process has not been systematically documented. The hypothesis was that sugammadex would attract rocuronium and vecuronium via interaction with the sugammadex side-chain “tentacles,” as previously suggested.

Methods

Computational molecular dynamics simulations were done to investigate docking of sugammadex with rocuronium and vecuronium. To validate these methods, strength of binding was assessed between sugammadex and a heterogeneous group of nine other drugs, the binding affinities of which have been experimentally determined. These observations hinted that high concentrations of unbound sugammadex could bind to propofol, potentially altering its pharmacokinetic profile. This was tested experimentally in in vitro cortical slices.

Results

Sugammadex encapsulation of rocuronium involved a sequential progression down a series of metastable states. After initially binding beside the sugammadex molecule (mean ± SD center-of-mass distance = 1.17 ± 0.13 nm), rocuronium then moved to the opposite side to that hypothesized, where it optimally aligned with the 16 hydroxyl groups (distance, 0.82 ± 0.04 nm) before entering the sugammadex cavity to achieve energetically stable encapsulation by approximately 120 ns (distance, 0.35 ± 0.12 nm). Vecuronium formed fewer hydrogen bonds with sugammadex than did rocuronium; hence, it was less avidly bound. For the other molecules, the computational results showed good agreement with the available experimental data, showing a clear bilogarithmic relation between the relative binding free energy and the association constant (R2 = 0.98). Weaker binding was manifest by periodic unbinding. The brain slice results confirmed the presence of a weak propofol–sugammadex interaction.

Conclusions

Computational simulations demonstrate the dynamics of neuromuscular blocking drug encapsulation by sugammadex occurring from the opposite direction to that hypothesized and also how high concentrations of unbound sugammadex can potentially weakly bind to other drugs given during general anesthesia.

Editor’s Perspective
What We Already Know about This Topic
  • Sugammadex encapsulates rocuronium and vecuronium to provide rapid effective reversal of neuromuscular blockade

  • Encapsulation dynamics might involve the eight negatively charged carboxyl thioether side chains at the edge of the hydrophobic cavity of cyclodextrin (the primary face) attracting the positively charged aminosteroid neuromuscular blocking drugs, followed by encapsulation of the muscle relaxant into the central core of sugammadex

What This Article Tells Us That Is New
  • A molecular dynamics computer simulation of the molecular interaction between rocuronium and sugammadex tested the hypothesis that it would follow this sequence of rocuronium attraction by the sugammadex carboxyl thioether side chains, followed by encapsulation

  • In the process of encapsulation, the rocuronium molecule entered from the direction opposite to the sugammadex carboxyl thioether side chains, where it optimally aligned with the 16 hydroxyl groups (the secondary face) and occupied a series of metastable states before finally occupying the hydrophobic cavity of the sugammadex molecule

  • Vecuronium was less avidly bound to sugammadex than rocuronium because it formed fewer hydrogen bonds with sugammadex

Sugammadex encapsulates neuromuscular blocking drugs (rocuronium and vecuronium) via a supramolecular mechanism of action to provide the rapid effective reversal of neuromuscular blockade seen in clinical practice.1  During the design process, cyclodextrin host molecules were systematically modified by the addition of eight carboxyl thioether groups to achieve the highest affinity with rocuronium, thus creating a stable complex. The size of the hydrophobic binding cavity, electrostatic interactions, and van der Waals forces were found to be important determinants of structure–activity relationships, consistent with known cyclodextrin chemistry (fig. 1, A and B).2  The negatively charged carboxyl thioether side chains at the edge of the hydrophobic cavity of the cyclodextrin molecule bind the positively charged quaternary amine in the structure of rocuronium. This increases the affinity of sugammadex for rocuronium, compared with the unmodified cyclodextrin molecule, which has a neutral charge.2  The formation of the sugammadex–rocuronium host–guest complex has been confirmed by x-ray crystallography.3  Previous publications suggested that the dynamics of the process of binding might involve these negatively charged side-chain tentacles of the sugammadex molecule attracting the positively charged aminosteroidal neuromuscular blocking drugs, followed by encapsulation of the muscle relaxant into the central core of sugammadex, where it is irreversibly fixed.4,5  In this scenario, the rocuronium molecule enters the sugammadex molecule from the direction of the carboxyl thioether tentacles (the primary face), with sugammadex encapsulating the rocuronium much like the classic video game Pac-Man. To test this proposed mechanism, we created a molecular dynamics computer simulation model that would allow visualization of the dynamic interaction between the host and guest molecules. We hypothesized that the simulation of the molecular interaction would follow this sequence of rocuronium attraction by the sugammadex tentacles, followed by encapsulation. For comparison, we also investigated the binding characteristics of vecuronium and pancuronium, which are reversed less effectively than rocuronium by sugammadex.6  In part, this comparison would serve to validate the accuracy of the simulations.

Fig. 1.

(A) Two-dimensional structure of sugammadex. (B) Three-dimensional configuration of sugammadex. (C) Trimer structure of the sugammadex unit that was used for the quantum mechanical calculations (partial charges calculations).

Fig. 1.

(A) Two-dimensional structure of sugammadex. (B) Three-dimensional configuration of sugammadex. (C) Trimer structure of the sugammadex unit that was used for the quantum mechanical calculations (partial charges calculations).

Close modal

Interactions of sugammadex with other drugs have also been investigated, because of the concern that displacement of rocuronium from sugammadex might cause reestablishment of neuromuscular blockade—or that the pharmacokinetics of other drugs might be affected. Clinically important interactions seem rare, although a potential change in the effectiveness of sugammadex could not be excluded for toremifene (an estrogen receptor modulator) and fusidic acid (a bacteriostatic antibiotic), and the effectiveness of combined oral contraceptive pills may be disrupted by sugammadex.7  There are also anecdotal reports of inadequate propofol anesthesia after neuromuscular reversal by sugammadex.8  To investigate this further, the binding strengths between various drugs and sugammadex were estimated in our simulations and compared to the published association constants obtained from the experimental work by Zwiers et al.9  Focusing on propofol, we tested in in vitro cortical slices the model prediction that—in the absence of neuromuscular blocking drugs—even the low-affinity relationship between propofol and sugammadex might be sufficient to cause some reduced in vitro propofol effect.

The computational work consisted of a series of in silico simulations of the interaction between sugammadex and various molecules. The computer simulations were to examine the dynamics of the process of encapsulation of rocuronium and vecuronium by sugammadex; to validate the simulation technique by comparing the simulated binding strengths between sugammadex and nine other compounds with previously published, experimentally obtained affinity data; and to characterize the patterns of sugammadex low-affinity binding with propofol. This body of work therefore included three sets of 500-ns molecular dynamics simulations, in which the muscle relaxant was initially located within the simulation box but outside the sugammadex molecule (four replications for rocuronium [simulations 1 to 4 in Supplemental Digital Content table S1, http://links.lww.com/ALN/C984], three replications for vecuronium [simulations 5 to 7 in table S1, http://links.lww.com/ALN/C984], and three replications for pancuronium [simulations 22 to 24 in table S1, http://links.lww.com/ALN/C984]). To validate and compare our model with the experimental work by Zwiers et al.,9  we did 5-ns simulations for each of 11 drug molecules (rocuronium and vecuronium and a heterogenous group of 9 other compounds) predocked into the sugammadex (simulations 8 to 18) to calculate the docking energies. The stability of the propofol–sugammadex interaction was investigated by an additional long (1 µs) simulation (simulation 19). The center-of-mass distance between the sugammadex–drug complex during the simulations was used to quantify the association process, and the strength of molecular association was estimated by calculating the relative binding free energy of the system over the last 3 ns of the 5-ns simulations. The details of the computational methods are described in the following sections.

Computational Simulation Methods

Molecular dynamics simulations initially require a quantum mechanical characterization of the molecular structures, followed by allowing the movement of the molecules of interest to evolve over time, subject to their various intermolecular forces. The structures of all studied drugs were obtained from PubChem,10  and the initial structure of sugammadex11  was taken from the Cambridge Crystallographic Data Center (CCDC 1495256; Cambridge‚ United Kingdom).12 Figure 1 shows the two-dimensional (A) and three-dimensional (B) structures of sugammadex used for the molecular dynamics simulations. Figure 1C shows a trimer of three structural units of sugammadex for the quantum mechanical calculations to get the partial charges of each atom. The partial charges of the middle unit were used for the repeating unit of sugammadex to avoid any terminal effects on the partial charges. The quantum calculations were performed using density functional theory at the B3LYP/6-31G* level of theory in water. Gaussian0913  software (USA) was used to perform density functional theory calculations, and then the restrained electrostatic potential fitting14  was used to derive partial charges of atoms using Antechamber15  embedded in the AmberTools20 (University of California‚ San Francisco‚ USA).16  The other topological parameters for the simulation of sugammadex were retrieved from the GLYCAM force field.17 

Figure 2 shows the two-dimensional structures of all drugs tested. Density functional theory calculations and restrained electrostatic potential fitting were done to get the partial charges of all atoms. AMBER force fields16  were used for the topology parameters using the AmberTools20 package. Docking of drug molecules was performed using AutoDock Vina (Center for Computational Structural Biology, Scripps Research Institute, USA)18  to be used as an initial structure for the molecular dynamics simulations (fig. 3). The GROMACS 202019  package was used to carry out molecular dynamics simulations. Explicit SPC/E water molecules20  were used. The simulated temperature was maintained constant (298 K and 310 K) using a Berendsen thermostat,21  and the pressure was kept fixed at 1.0 bar using the Parrinello–Rahman algorithm.22  Periodic boundary conditions were applied to the cubic simulation box, and the integration step was set to 2 fs. The particle-mesh Ewald method23  was used for long-range electrostatics. Sodium chloride was added to the simulation box to a 150 mM ionic concentration, and then the whole system was neutralized with the addition of a sufficient number of ions.

Fig. 2.

Two-dimensional structures of studied drugs: rocuronium, vecuronium, and pancuronium (steroidal neuromuscular blocking drugs), atracurium (nonsteroidal neuromuscular blocking drug), dexamethasone and betamethasone (corticosteroids), fusidic acid (antibacterial), flucloxacillin (penicillin), toremifene (selective estrogen receptor modulator), and propofol and ketamine (anesthetics).

Fig. 2.

Two-dimensional structures of studied drugs: rocuronium, vecuronium, and pancuronium (steroidal neuromuscular blocking drugs), atracurium (nonsteroidal neuromuscular blocking drug), dexamethasone and betamethasone (corticosteroids), fusidic acid (antibacterial), flucloxacillin (penicillin), toremifene (selective estrogen receptor modulator), and propofol and ketamine (anesthetics).

Close modal
Fig. 3.

Computational molecular docking of 11 studied drug molecules in the sugammadex, along with their docking score calculated by AutoDock Vina software (Center for Computational Structural Biology‚ Scripps Research Institute‚ USA).

Fig. 3.

Computational molecular docking of 11 studied drug molecules in the sugammadex, along with their docking score calculated by AutoDock Vina software (Center for Computational Structural Biology‚ Scripps Research Institute‚ USA).

Close modal

To ensure full equilibration of temperature, pressure, and density, initial simulations of 100-ps duration were performed for both isothermal and isobaric canonical ensembles, before the start of the 5-ns simulation proper. The results were visualized using VMD (University of Illinois Urbana–Champaign, USA)24  and PyMOL.25  The relative binding free energy of these drugs encapsulated in sugammadex was calculated using gmx_MMPBSA software26  at 1-ps intervals over the last 3 ns of the simulations. These simulations were performed on a desktop with an AMD (USA) Ryzen Threadripper 3990X 64-Core processor and two GeForce (Nvidia; USA) RTX 2080 Ti GPUs with the average performance of 500 ns/day for each simulation.

Comparison of Experimental and Computer Molecular Binding Strengths

To validate the molecular dynamics simulations, we calculated the relative binding free energy of 11 drugs encapsulated in sugammadex and compared these to experimental data. Initially, the docking process was done using the AutoDock Vina software.18  The docking scores indicate how strongly that particular drug fits into the sugammadex cavity. However, as the sugammadex structure is considered as a rigid molecule while working with the AutoDock Vina software, these docking scores are just initial indicators of how the drug molecules might associate with sugammadex. To get the relative binding free energy, we need to start with the Gibbs free energy change (ΔG) relation:

ΔG=ΔHTΔS
(1)

where ΔH in the gmx_MMPBSA package is the relative binding free energy, T is the temperature of the system (in a unit of kelvin), and ΔS is the entropy change. The gx_MMPBSA package was used to calculate the relative binding free energy (ΔH) of the system when we have energy of a complex of sugammadex and drugs subtracted from the energy of the sugammadex and the drug molecules individually:

ΔH=HComplex(HSugammadex+HDrug)
(2)

where HComplex is the energy of the system when the sugammadex and the drug molecules are interacting with each other, while HSugammadex and HDrug are the energy of isolated individual sugammadex and drug molecule, respectively. Using gmx_MMPBSA software, energies were calculated in gas and solvation states for each term in equation 2. In the gas state, combinations of van der Waals and electrostatics energy were calculated, while for the solvation state, polar and nonpolar energies were included. The gmx_MMPBSA software calculates those energies for each snapshot taken from the molecular dynamics simulation. The Gibbs free energy change (ΔG ) under standard conditions (1 atm and 298 K) has a logarithmic relationship with the association constant (kass), which can be given by the following:

ΔG=RTlnkass
(3)

where R is the universal gas constant (1.987 cal · K−1 · mol−1), and T is the temperature of the system in degrees of kelvin.27  Having small molecules in the system—as in our simulations—leads to very high uncertainty in entropy estimation. As is suggested in several refences,28–30  to avoid this distortion, we omitted the entropy term (TΔS) and focused on the calculation of the relative binding free energy (ΔH). Combining equations 1 and 3, we can find the relation between the relative binding free energy and the association constant:

ΔH=TΔSRTlnkass
(4)

where ΔH can be calculated using equation 2 from molecular dynamics trajectory over the last 3 ns of simulations at 1-ps intervals. kass (the association constant) comes from experimental data,9  and T is the temperature for the system.

Cortical Slice Methods

Sugammadex was tested for a biologic effect in reducing propofol action in a cortical slice model. The purpose of these experiments was to validate the computational results in a simple biologic model (without the confounding effects of blood flow and protein binding) rather than to closely replicate or mimic a clinical drug-reversal scenario. Coronal mouse brain slices were prepared from adult male and female C57 mice housed in a temperature-controlled room with unlimited access to food and water. The tissue recovery methods for this work were approved by the Animal Ethics Committee at the University of Waikato (Hamilton, New Zealand). The details of the methods are described in the supplemental digital content (http://links.lww.com/ALN/C984). Briefly, 400-µm cortical slices were perfused continuously with artificial cerebrospinal fluid and electrical field potential activity recorded (75 µm silver/silver chloride), amplified, and saved for analysis. The effect of propofol (34 µM) combined with sugammadex (20 µM; n = 6, 6 slices, 2 animals) was compared to equivalent experiments run with propofol (34 µM) only (n = 6, 4 slices, 2 animals).

Statistics

Continuous data are described as means ± SD, and the relationship between relative binding free energy and association constant described using linear regression (least-squares fit). The 95% CI was calculated using 1.96SDn, where n is the number of steps. Changes in cortical slice activity were checked for normality and then compared using an unpaired t test. The gmx_rms package available in Gromacs was used to calculate the root mean square deviation. A P value less than 0.05 was considered statistically significant.

Dynamics of the Association of Sugammadex with Rocuronium and Vecuronium

Rocuronium and vecuronium were chosen to study the encapsulation process as they have a high affinity with sugammadex. It was also of interest to compare these two neuromuscular blockers, because sugammadex more strongly reverses rocuronium clinically.6  We started with the configuration of separated sugammadex and drug molecules in the simulation box. Four simulations for rocuronium are shown in figure 4. It takes between 75 and 150 ns for the sugammadex molecule to encapsulate the rocuronium. Vecuronium (and pancuronium) follow a similar process but showed more freedom, resulting in slightly less affinity for sugammadex (figs. S1 and S2, http://links.lww.com/ALN/C984; movie S3, http://links.lww.com/ALN/C987, respectively). We found that the process of encapsulation followed a consistent route. There were discrete metastable and stable molecular configurations, shown by the plateaus in figure 4A. Initially due to the thermal fluctuations, rocuronium moved all around the sugammadex molecule and flipped head over tail many times, thus properly randomizing the direction of approach, with center-of-mass distances from approximately 1.5 to approximately 4 nm from the sugammadex molecule. It then settled to an intermediate stage of weak binding (from about 50 to 100 ns for run 1 in fig. 4A) in which the rocuronium molecule initially aligned itself beside (mean ± SD distance = 1.17 ± 0.13 nm) and then bound to the secondary face of the sugammadex molecule, where there are 16 hydroxyl groups available (as seen by the distance plateau of 0.82 ± 0.04 nm). This is the opposite side of the sugammadex molecule to the hypothesized attraction between the negatively charged carboxyl thioether side-chain tentacles (primary face) and the positively charged aminosteroid. However, this first touch led to a more stable connection involving intermolecular hydrogen bonds between the rocuronium and the sugammadex hydroxyl groups, where we can see the dipole–dipole attraction between hydrogen atoms bonded to a strongly electronegative atom (oxygen) and another electronegative atom from the other molecule in its vicinity. Because the rocuronium molecule is now optimally aligned, it moved quickly into the cavity of the sugammadex ring to achieve energetically stable encapsulation by approximately120 ns (center-of-mass distance, 0.35 ± 0.12 nm). Other directions of proximity between rocuronium and sugammadex did not result in encapsulation (movie S1, http://links.lww.com/ALN/C985). Histograms of distances between centers of mass are shown in figure 4 (B to E) for runs 1 to 4. respectively. The shaded area (distance, less than 0.6 nm) in those figures represents the stable encapsulation region. The histogram peaks (arrows) outside the encapsulation region indicate the existence of metastable configurations that precede the stable encapsulation state. In addition, it can be seen that there are two possible alignments of rocuronium entry: from the positively charged nitrogen side of the rocuronium (cationic head) or from the opposite (tail). In the first three runs (simulations 1 to 3), rocuronium entered into the sugammadex cavity tail-first, but in the fourth run, it entered cationic head-first. In this case, the negatively charged carboxyl groups of sugammadex are able to hold it deeper in the sugammadex cavity. We see that tail-first rocuronium alignment is manifest by two peaks in the encapsulation region around 0.3 and 0.5 nm. However, for the fourth (cationic head-first) run in figure 4E, the rocuronium is held more tightly. There is only one peak in the distance histogram: the encapsulation region between 0.1 and 0.2 nm, indicating that the negative charge of carboxyl groups in sugammadex holds the rocuronium, so it is not able to fluctuate within the sugammadex cavity.

Fig. 4.

(A) Encapsulation process of a rocuronium molecule by sugammadex. Distances between centers of mass of rocuronium and sugammadex during a 500-ns simulation for four replication simulations are shown. (Inset) Configuration of the complex after 500 ns. (B to E) Histograms of distances for each replication simulation. The shadow area represents the encapsulation region. The arrows are pointing to the metastable configurations of the rocuronium–sugammadex association. (F) Cationic head-in encapsulation process (run 4). (G) Tail-in encapsulation process (runs 1, 2, and 3).

Fig. 4.

(A) Encapsulation process of a rocuronium molecule by sugammadex. Distances between centers of mass of rocuronium and sugammadex during a 500-ns simulation for four replication simulations are shown. (Inset) Configuration of the complex after 500 ns. (B to E) Histograms of distances for each replication simulation. The shadow area represents the encapsulation region. The arrows are pointing to the metastable configurations of the rocuronium–sugammadex association. (F) Cationic head-in encapsulation process (run 4). (G) Tail-in encapsulation process (runs 1, 2, and 3).

Close modal

Comparison of Rocuronium and Vecuronium Binding

Figure 5 (A and B) shows the distance between the center of mass of the neuromuscular blocking drugs (rocuronium and vecuronium) and sugammadex along with the number of hydrogen bonds (seen in red) during the time course of simulation. It can be seen that the number of hydrogen bonds between rocuronium and sugammadex is higher than for vecuronium and sugammadex, after each of the aminosteroids has entered the sugammadex cavity. Having three oxygen atoms in the rocuronium structure, as highlighted in the inset of figure 5A, leads to a 22.5% hydrogen bond occupancy over the last 200 ns of the four simulations. In contrast, the average hydrogen bond occupancy between vecuronium and sugammadex over the last 200 ns of three simulations is only 3.5%. Vecuronium has only two carbonyl groups (as highlighted in figure 5B) to achieve the hydrogen bond connection with sugammadex. In comparison, rocuronium can achieve hydrogen bonds with one carbonyl group, one hydroxyl group, and a third highlighted oxygen atom. Figure 5C shows a comparison of hydrogen bond occupancy percentage of rocuronium with sugammadex versus vecuronium with sugammadex. Rocuronium (averaged over four simulations) has 6.4 times more hydrogen bonds occupancy than vecuronium (averaged over three simulations) with sugammadex. These findings mirror the clinical scenario in which sugammadex reverses rocuronium more strongly than vecuronium and validates the computational methods used in this study.6 

Fig. 5.

Number of hydrogen bonds (red) between rocuronium (A), vecuronium (B), and sugammadex over the simulation course. (Insets) Two-dimensional structures of rocuronium and vecuronium indicate the oxygen atoms were involved in hydrogen bonds. The black lines represent the distances between the centers of mass of rocuronium (A) and vecuronium (B) and sugammadex. (C) The hydrogen bond occupancy percentage between drug and sugammadex over the last 200 ns. The error bar is the SD over four replications of rocuronium complex simulation and three replications for vecuronium complex simulation.

Fig. 5.

Number of hydrogen bonds (red) between rocuronium (A), vecuronium (B), and sugammadex over the simulation course. (Insets) Two-dimensional structures of rocuronium and vecuronium indicate the oxygen atoms were involved in hydrogen bonds. The black lines represent the distances between the centers of mass of rocuronium (A) and vecuronium (B) and sugammadex. (C) The hydrogen bond occupancy percentage between drug and sugammadex over the last 200 ns. The error bar is the SD over four replications of rocuronium complex simulation and three replications for vecuronium complex simulation.

Close modal

Further Validation of Simulation Methods: Affinities of Different Drugs with Sugammadex

One of the limitations of molecular dynamics simulations is the requirement for the model to approximate atomic and molecular forces and interactions. While this is necessary to reduce the high computational cost of the calculations, it creates the potential for inaccuracies in the simulation outcomes. Validation of the model against experimentally derived data is an important facet of studies utilizing this approach. Comparison was made between the simulated molecular relative binding free energy (ΔH) and experimentally determined association affinities (association constant) for a range of molecules.9  A clear correspondence between computational and experimental data would further confirm the suitability of the computational methods for the dynamical simulation of the sugammadex–drug interactions.

Initially, the docking output for all studied compounds along with their docking scores was calculated using AutoDock Vina (fig. 3). During this process, the structure of sugammadex was considered rigid, so these numbers can be considered as a rough estimate for the association of these drugs with sugammadex as an initial step for molecular dynamics simulation. To include the dynamics of the system and study the association of drugs under thermal fluctuations, 5-ns molecular dynamics simulations were performed (which is long enough for calculation of the relative binding free energy (ΔH); table S2; fig. S3L, http://links.lww.com/ALN/C984). The first 2 ns of the simulations were considered to be equilibration time, so the last 3 ns of the simulations were used for calculation of the relative binding free energy (ΔH) using gmx_MMPBSA package (more details in fig. S3, http://links.lww.com/ALN/C984).

We found a clean bilogarithmic correlation between relative binding free energy and association constant (fig. 6; kass). This logarithmic relation is expected from equation 4. The fitted line in figure 6 is ΔH=A+Blnkass. The value for B is −0.58 ± 0.02 kcal · mol−1, and that for A is 5.6 ± 0.3 kcal · mol−1 with R2=0.98. From the theoretical calculations, the coefficient behind the ln function in equation 3 is 0.616 kcal · mol−1 (considering R = 1.9871 × 10−3 kcal · K−1 · mol−1) and T=310K, which has a close agreement with the B value of the fitted line. The error bars indicate the 95% CI.

Fig. 6.

Relative binding free energy of study drugs-sugammadex complex compared to experimental data.9 Error bars indicate 95% CI. The dashed line is the fitted line with the bilogarithmic relation: ∆H = A + B ln(kass). Where the value for the B is –0.58 ± 0.02 (kcal · mol–1) and for the A is 5.6 ± 0.3 (kcal · mol–1).

Fig. 6.

Relative binding free energy of study drugs-sugammadex complex compared to experimental data.9 Error bars indicate 95% CI. The dashed line is the fitted line with the bilogarithmic relation: ∆H = A + B ln(kass). Where the value for the B is –0.58 ± 0.02 (kcal · mol–1) and for the A is 5.6 ± 0.3 (kcal · mol–1).

Close modal

Pattern of Low-affinity Sugammadex Binding of Propofol

As seen in figure 6, experimentally and in computer simulation, propofol shows a relatively low affinity with sugammadex. Because propofol is a commonly used drug clinically, we probed its interaction with sugammadex in more detail. The pattern of association is quite different from the high-affinity binding of rocuronium. A long time-scale simulation of 1 µs is required to get more details of the dynamics of the propofol and sugammadex interaction. The simulation started with encapsulated propofol. As can be seen in figure 7, after about 70 ns, propofol exited the sugammadex cavity, and thereafter, it moved back and forth between encapsulation and unbound states over time scales of hundreds of nanoseconds—a hallmark of low affinity binding. From the simulation (fig. 7), propofol was encapsulated 59% of the simulation time. This suggested that excess free sugammadex might have some effect on propofol availability. We tested this prediction by quantifying the effect of propofol in a series of in vitro cortical slice experiments.

Fig. 7.

Distance between centers of mass of propofol and sugammadex over the simulation course. On the right-hand side is the snapshot of encapsulated configuration and free propofol. Note the time scale is in microseconds‚ not nanoseconds.

Fig. 7.

Distance between centers of mass of propofol and sugammadex over the simulation course. On the right-hand side is the snapshot of encapsulated configuration and free propofol. Note the time scale is in microseconds‚ not nanoseconds.

Close modal

Cortical Slice Experiments

Biological evidence of some affinity of sugammadex for propofol was tested experimentally in in vitro cortical slices (n = 6). We were able to observe a significantly smaller reduction in seizure-like event frequency (fig. 8) when propofol and sugammadex were combined (20% reduction) compared to propofol alone (62% reduction, P = 0.01‚ unpaired t test). This effect was most evident during the drug washout period, which is consistent with the very slow diffusion of propofol into brain slice tissue.9,31  Seizure-like event amplitude and length were not significantly different. Sugammadex on its own had no effect on seizure-like event activity. These results confirm the molecular simulation data.

Fig. 8.

Influence of sugammadex on the inhibitory effect of propofol on cortical slice seizure-like event activity. (A) Seizure-like event frequency. (B) Seizure-like event amplitude. (C) Seizure-like event length. The solid lines show the mean values of the individual experiments (dotted lines).

Fig. 8.

Influence of sugammadex on the inhibitory effect of propofol on cortical slice seizure-like event activity. (A) Seizure-like event frequency. (B) Seizure-like event amplitude. (C) Seizure-like event length. The solid lines show the mean values of the individual experiments (dotted lines).

Close modal

In this study, we describe and validate the molecular dynamics of drug binding with sugammadex. In the process of encapsulation by sugammadex, the rocuronium molecule entered from the direction opposite to the sugammadex carboxyl thioether side chains (the primary face) and occupied a series of metastable states before finally occupying the hydrophobic cavity of the sugammadex. The rocuronium could be adequately encapsulated with either a cationic head-first (where the positively charged quaternary amine side enters first; fig. 4F) or tail-first orientation, but cationic head-first encapsulation was marginally more stable. Vecuronium formed fewer hydrogen bonds with sugammadex than did rocuronium; hence, it was less avidly bound to sugammadex. A variety of other molecules showed weaker binding, which was manifest by periodic unbinding dynamics. Even though propofol is only weakly bound to sugammadex, there appears to be a theoretical possibility for very high concentrations of unoccupied sugammadex to reduce its actions. This effect was confirmed in brain slice experiments.

Basis and Utility of Molecular Dynamics Modeling

To appreciate the utility of the molecular modeling techniques presented in this article, it is helpful to have a basic understanding of some of the technical aspects. We would point the reader to a recent, technically accessible review for a thorough discussion.32  At the most fundamental level, the aim of molecular dynamics modeling is to predict the movement of every atom in a biomolecular construct over time. This is done by calculating the forces exerted on each atom by all other atoms, using Newton’s laws of motion, integrated over time. The significance of this for biologic research is that it provides a three-dimensional representation of the movement of all atoms in the modeled system, which is unattainable experimentally. This molecular “flexibility” governs the way molecules interact with each other, in contrast to the traditional rigid lock-and-key description of protein–ligand binding.33  While an extremely powerful technique (particularly with modern computing power), it must be noted that calculation efficiency demands several approximations to be made in the force field predictions. Result interpretation must therefore be viewed through the lens of uncertainty and potential errors.34 

Mechanisms of Encapsulation

Previous work demonstrating the encapsulation of rocuronium in sugammadex has shown only the end result of the binding.35  Herein, we describe the details of the dynamics involved in the whole encapsulation process. It was thought that the binding would involve the negatively charged carboxyl thioether side-chain tentacles of the sugammadex molecule (primary face) attracting the positively charged aminosteroidal neuromuscular blocking drugs.36  Contrary to this, our simulations indicate that the rocuronium molecule approaches the sugammadex molecule from the opposite side, settling into a discrete series of metastable states that align the rocuronium molecule, preparing it before entering the central ring of the sugammadex. The rocuronium molecule is also able to be bound in both cationic head-first and tail-first configurations. After the tail-first encapsulation, rocuronium jumps between two positions (fig. 4, B to D), while for the cationic head-first, there is only one tight position (fig. 4E). This shows that the cationic head-first encapsulation is more energetically favorable. However, due to the dynamics of the system and increased probability of establishing hydrogen bonds between the tail of rocuronium (where there is one hydroxyl group) and secondary face of sugammadex (with 16 hydroxyl groups), the tail-first encapsulation appears to be the preferred pathway for encapsulation—even though the subsequent final position of the encapsulated rocuronium is energetically a bit less stable than that for the cationic head-first rocuronium orientation.

The process of sugammadex encapsulation of vecuronium (movie S2, http://links.lww.com/ALN/C986) was broadly similar to that of rocuronium, except that, after the final entry into the sugammadex cavity, the hydrogen bond occupancy was an order of magnitude smaller with vecuronium. In contrast, weakly bound compounds showed a qualitatively different pattern of interaction with the sugammadex. While these observations may not have obvious direct clinical implications, the detailed recognition and understanding of drug interactions such as these is an essential step in guiding future cyclodextrin drug design. This would add a rapid feedback loop to improve structure–activity relationship data used to design cyclodextrin modifications.

Sugammadex and Residual Hypnotic Effects

Signs of light anesthesia and unexpected awakening after sugammadex was given during propofol anesthesia have been noted since 2007.37,38  This phenomenon increased with increasing doses of sugammadex.8  The mechanism of the awakening was hypothesized to be due to increased input into the reticular activating system due to increased muscular activity. An alternative contributing pharmacokinetic mechanism could be a decrease in free propofol concentration in the central compartment due to binding between sugammadex and propofol when there is a high concentration of unbound of sugammadex. This has been dismissed as unlikely, due to low affinity between the molecules. Our computer modeling agrees with this low affinity, but our ex vivo brain-slice data suggest that the possibility of a modest reduction in the action of propofol in the presence of very high concentrations of sugammadex cannot be completely ruled out. However, there are many more layers to the complexities of in vivo pharmacokinetics and pharmacodynamics, and we are unable to make any claims as to whether this is a clinically prevalent issue. For example, the reverse situation can pertain. Cyclodextrins have been studied as solubilizing drugs in pharmaceutical preparations for the delivery and release of many drugs, including propofol.39  In these formulations, the drug is released from its bond with the cyclodextrin host on injection due to dilution and even more so for propofol in human serum because propofol is strongly protein bound. Therefore, although we found a weak affinity between sugammadex and propofol in silico and in brain slices, we would not expect a clinically observed effect commonly.

Conclusions

Molecular dynamics simulations of rocuronium– sugammadex interactions showed a series of metastable states preceding encapsulation of rocuronium from the secondary face of the sugammadex molecule. Based on our computational simulations and experimental results, we cannot rule out the possibility that sugammadex could potentially cause a minor reduction in propofol concentrations in some rare circumstances; however, this is unlikely to be a common cause of clinical observations of lightening of anesthesia.

Research Support

Supported by departmental sources and by Australian and New Zealand College of Anaesthetists (Melbourne‚ Australia) project grant No. 22/007.

Competing Interests

The authors declare no competing interests.

Table S1: Details of molecular dynamics simulations, http://links.lww.com/ALN/C984

Table S2: relative binding free energy from short and long simulations, http://links.lww.com/ALN/C984

Figure S1: Encapsulation process of vecuronium, http://links.lww.com/ALN/C984

Figure S2: Encapsulation process of pancuronium, http://links.lww.com/ALN/C984

Figure S3: Energetic components of relative binding free energy, http://links.lww.com/ALN/C984

Cortical slice experimental methods, http://links.lww.com/ALN/C984

Movie S1: Encapsulation dynamics of rocuronium, http://links.lww.com/ALN/C985

Movie S2: Encapsulation dynamics of vecuronium, http://links.lww.com/ALN/C986

Movie S3: Encapsulation dynamics of pancuronium, http://links.lww.com/ALN/C987

1.
Pongrácz
A
,
Szatmári
S
,
Nemes
R
,
Fülesdi
B
,
Tassonyi
E
:
Reversal of neuromuscular blockade with sugammadex at the reappearance of four twitches to train-of-four stimulation.
Anesthesiology
2013
;
119
:
36
42
2.
Adam
JM
,
Bennett
DJ
,
Bom
A
,
Clark
JK
,
Feilden
H
,
Hutchinson
EJ
,
Palin
R
,
Prosser
A
,
Rees
DC
,
Rosair
GM
,
Stevenson
D
,
Tarver
GJ
,
Zhang
M-Q
:
Cyclodextrin-derived host molecules as reversal agents for the neuromuscular blocker rocuronium bromide: Synthesis and structure—Activity relationships.
J Med Chem
2002
;
45
:
1806
16
3.
Gijsenbergh
F
,
Ramael
S
,
Houwing
N
,
van Iersel
T
:
First human exposure of Org 25969, a novel agent to reverse the action of rocuronium bromide.
Anesthesiology
2005
;
103
:
695
703
4.
Naguib
M
:
Sugammadex: Another milestone in clinical neuromuscular pharmacology.
Anesth Analg
2007
;
104
:
575
81
5.
Khirwadkar
R
,
Hunter
JM
:
Neuromuscular physiology and pharmacology: An update.
Cont Educ Anaesth Crit Care Pain
2012
;
12
:
237
44
6.
Herring
WJ
,
Woo
T
,
Assaid
CA
,
Lupinacci
RJ
,
Lemmens
HJ
,
Blobner
M
,
Khuenl-Brady
KS
:
Sugammadex efficacy for reversal of rocuronium- and vecuronium-induced neuromuscular blockade: A pooled analysis of 26 studies.
J Clin Anesth
2017
;
41
:
84
91
7.
Schaller
SJ
,
Fink
H
:
Sugammadex as a reversal agent for neuromuscular block: An evidence-based review.
Core Evid
2013
;
8
:
57
67
8.
Sparr
HJ
,
Vermeyen
KM
,
Beaufort
AM
,
Rietbergen
H
,
Proost
JH
,
Saldien
V
,
Velik-Salchner
C
,
Wierda
JMKH
:
Early reversal of profound rocuronium-induced neuromuscular blockade by sugammadex in a randomized multicenter study: Efficacy, safety, and pharmacokinetics.
Anesthesiology
2007
;
106
:
935
43
9.
Zwiers
A
,
Heuve
M van den
,
Smeets
J
,
Rutherford
S
:
Assessment of the potential for displacement interactions with sugammadex.
Clin Drug Investig
2011
;
31
:
101
11
10.
PubChem 2022.
Available at: https://pubchem.ncbi.nlm.nih.gov/. Accessed October 9, 2022.
11.
Xu
H
,
Rodríguez-Hermida
S
,
Pérez-Carvajal
J
,
Juanhuix
J
,
Imaz
I
,
Maspoch
D
:
A first cyclodextrin-transition metal coordination polymer.
Cryst Growth Des
2016
;
16
:
5598
602
12.
Cambridge Crystallographic Data Centre.
13.
Frisch
MJ
,
Trucks
GW
,
Schlegel
HB
,
Scuseria
GE
,
Robb
MA
,
Cheeseman
JR
,
Scalmani
G
,
Barone
V
,
Petersson
GA
,
Nakatsuji
H
,
Li
X
,
Caricato
M
,
Marenich
A
,
Bloino
J
,
Janesko
BG
,
Gomperts
R
,
Mennucci
B
,
Hratchian
HP
,
Ortiz
JV
,
Izmaylov
AF
,
Sonnenberg
JL
,
Williams-Young
D
,
Ding
F
,
Lipparini
F
,
Egidi
F
,
Goings
J
,
Peng
B
,
Petrone
A
,
Henderson
T
,
Ranasinghe
D
,
Zakrzewski
VG
,
Gao
J
,
Rega
N
,
Zheng
G
,
Liang
W
,
Hada
M
,
Ehara
M
,
Toyota
K
,
Fukuda
R
,
Hasegawa
J
,
Ishida
M
,
Nakajima
T
,
Honda
Y
,
Kitao
O
,
Nakai
H
,
Vreven
T
,
Throssell
K
,
Montgomery
JA
, Jr
,
Peralta
JE
,
Ogliaro
F
,
Bearpark
M
,
Heyd
JJ
,
Brothers
E
,
Kudin
KN
,
Staroverov
VN
,
Keith
T
,
Kobayashi
R
,
Normand
J
,
Raghavachari
K
,
Rendell
A
,
Burant
JC
,
Iyengar
SS
,
Tomasi
J
,
Cossi
M
,
Millam
JM
,
Klene
M
,
Adamo
C
,
Cammi
R
,
Ochterski
JW
,
Martin
RL
,
Morokuma
K
,
Farkas
O
,
Foresman
JB
,
Fox
DJ
:
Gaussian 09, revision A.02 2016
14.
Bayly
CI
,
Cieplak
P
,
Cornell
W
,
Kollman
PA
:
A well-behaved electrostatic potential based method using charge restraints for deriving atomic charges: The RESP model.
J Phys Chem
1993
;
97
:
10269
80
15.
Wang
J
,
Wang
W
,
Kollman
PA
,
Case
DA
:
Automatic atom type and bond type perception in molecular mechanical calculations.
J Mol Graph Model
2006
;
25
:
247
60
16.
Case
DA
,
Aktulga
HM
,
Belfon
K
,
Ben-Shalom
I
,
Brozell
SR
,
Cerutti
DS
,
Cheatham
TE
, III
,
Cruzeiro
VWD
,
Darden
TA
,
Duke
RE
:
Amber 2021
.
University of California, San Francisco
,
2021
17.
Woods
RJ
,
Dwek
RA
,
Edge
CJ
,
Fraser-Reid
B
:
Molecular mechanical and molecular dynamic simulations of glycoproteins and oligosaccharides. 1. GLYCAM_93 parameter development.
J Phys Chem
1995
;
99
:
3832
46
18.
Eberhardt
J
,
Santos-Martins
D
,
Tillack
AF
,
Forli
S
:
AutoDock Vina 1.2.0: New docking methods, expanded force field, and Python bindings.
J Chem Inf Model
2021
;
61
:
3891
8
19.
Abraham
MJ
,
Murtola
T
,
Schulz
R
,
Páll
S
,
Smith
JC
,
Hess
B
,
Lindahl
E
:
GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers.
SoftwareX
2015
;
1–2
:
19
25
20.
Mark
P
,
Nilsson
L
:
Structure and dynamics of the TIP3P, SPC, and SPC/E water models at 298 K.
J Phys Chem A
2001
;
105
:
9954
60
21.
Berendsen
HJC
,
Postma
JPM
,
van Gunsteren
WF
,
DiNola
A
,
Haak
JR
:
Molecular dynamics with coupling to an external bath.
J Chem Phys
1984
;
81
:
3684
90
22.
Parrinello
M
,
Rahman
A
:
Polymorphic transitions in single crystals: A new molecular dynamics method.
J Appl Phys
1981
;
52
:
7182
90
23.
Darden
T
,
York
D
,
Pedersen
L
:
Particle mesh Ewald: An log(N) method for Ewald sums in large systems.
J Chem Phys
1993
;
98
:
10089
92
24.
Humphrey
W
,
Dalke
A
,
Schulten
K
:
VMD: Visual molecular dynamics.
J Mol Graph
1996
;
14
:
33
8
25.
Schrodinger
L
:
PyMOL molecular graphics system, version 1.8.
2015
26.
Valdés-Tresanco
MS
,
Valdés-Tresanco
ME
,
Valiente
PA
,
Moreno
E
:
gmx_MMPBSA: A new tool to perform end-state free energy calculations with GROMACS.
J Chem Theory Comput
2021
;
17
:
6281
91
27.
Du
X
,
Li
Y
,
Xia
Y-L
,
Ai
S-M
,
Liang
J
,
Sang
P
,
Ji
X-L
,
Liu
S-Q
:
Insights into protein–ligand interactions: Mechanisms, models, and methods.
Int J Mol Sci
2016
;
17
:
144
28.
Genheden
S
,
Ryde
U
:
The MM/PBSA and MM/GBSA methods to estimate ligand-binding affinities.
Expert Opin Drug Discov
2015
;
10
:
449
61
29.
Špačková
N
,
Cheatham
TE
,
Ryjáček
F
,
Lankaš
F
,
van Meervelt
L
,
Hobza
P
,
Šponer
J
:
Molecular dynamics simulations and thermodynamics analysis of DNA–drug complexes: Minor groove binding between 4′,6-diamidino-2-phenylindole and DNA duplexes in solution.
J Am Chem Soc
2003
;
125
:
1759
69
30.
Yang
T
,
Wu
JC
,
Yan
C
,
Wang
Y
,
Luo
R
,
Gonzales
MB
,
Dalby
KN
,
Ren
P
:
Virtual screening using molecular simulations.
Proteins Struct Funct Bioinf
2011
;
79
:
1940
51
31.
Gredell
JA
,
Turnquist
PA
,
MacIver
MB
,
Pearce
RA
:
Determination of diffusion and partition coefficients of propofol in rat brain tissue: Implications for studies of drug action in vitro.
Br J Anaesth
2004
;
93
:
810
7
32.
Hollingsworth
SA
,
Dror
RO
:
Molecular dynamics simulation for all.
Neuron
2018
;
99
:
1129
43
33.
de Vivo
M
,
Masetti
M
,
Bottegoni
G
,
Cavalli
A
:
Role of molecular dynamics and related methods in drug discovery.
J Med Chem
2016
;
59
:
4035
61
34.
Gunsteren
WF
,
Daura
X
,
Fuchs
PFJ
,
Hansen
N
,
Horta
BAC
,
Hünenberger
PH
,
Mark
AE
,
Pechlaner
M
,
Riniker
S
,
Oostenbrink
C
:
On the effect of the various assumptions and approximations used in molecular simulations on the properties of bio-molecular systems: Overview and perspective on issues.
ChemPhysChem
2021
;
22
:
264
82
35.
Li
L
,
Zhou
Y
,
Wang
Z
,
Wu
C
,
Li
Z
,
Sun
C
,
Sun
T
:
Theoretical studies on the mechanism of sugammadex for the reversal of aminosteroid-induced neuromuscular blockade.
J Mol Liq
2018
;
265
:
450
6
36.
Hunter
JM
,
Naguib
M
:
Sugammadex-induced bradycardia and asystole: How great is the risk?
Br J Anaesth
2018
;
121
:
8
12
37.
Giuffrida
M
,
Ledingham
NS
,
Machi
P
,
Czarnetzki
CA
:
Rapid arousal from anaesthesia after reversal of deep rocuronium-induced neuromuscular block with sugammadex in a neuroradiological procedure.
BMJ Case Rep
2021
;
14
:
e242820
38.
Guen
M
,
Roussel
C
,
Chazot
T
,
Dumont
GA
,
Liu
N
,
Fischler
M
:
Reversal of neuromuscular blockade with sugammadex during continuous administration of anaesthetic agents: A double-blind randomised crossover study using the bispectral index.
Anaesthesia
2020
;
75
:
583
90
39.
Loftsson
T
,
Moya-Ortega
MD
,
Alvarez-Lorenzo
C
,
Concheiro
A
:
Pharmacokinetics of cyclodextrins and drugs after oral and parenteral administration of drug/cyclodextrin complexes.
J Pharm Pharmacol
2016
;
68
:
544
55