Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2026 May 14;16:21199. doi: 10.1038/s41598-026-47021-8

Encapsulation of colistin using hyaluronic acid and PVA–PSS polymeric nanocarriers for enhanced drug delivery: a molecular dynamics simulation study

Abolfazl Azadi 1, Khadijeh Ahmadi 2,3, Hossein Fazli 4, Hoda Nassira 1,✉
PMCID: PMC13346572  PMID: 42135335

Abstract

Colistin (polymyxin E) is an effective antibiotic against Gram-negative bacteria, but its clinical use is limited due to its high toxicity. Targeted drug delivery, which has made significant progress in reducing drug toxicity in recent years, could be an ideal approach to mitigate the nephrotoxicity and neurotoxicity associated with this drug. However, the complexity of drug design and manufacturing remains a major challenge in the pharmaceutical industry. To address this, we investigated polymeric drug delivery systems for colistin encapsulation using molecular dynamics simulations. In this study, the MARTINI 3 force field within GROMACS 2022 software was employed to model and simulate various systems, including unmodified hyaluronic acid (HA), HA modified with two and four hydrophobic groups, and a combination of polyvinyl alcohol (PVA) and polystyrene sulfonate (PSS). Structural analysis of the systems was performed based on indices such as root-mean-square deviation (RMSD), radius of gyration (Rg), radial distribution function (RDF), and solvent-accessible surface area (SASA). The results indicated that hydrophobic modification of HA increased aggregation, reduced the radius of gyration, and enhanced the system’s stability in the presence of colistin. Among the modified systems, HA with four hydrophobic groups (HA(4)) exhibited the highest structural stability. Additionally, the PVA-PSS polyelectrolyte complex displayed stable behavior, formed a bilayer structure, and demonstrated effective interactions with the drug. Targeted chemical modifications and the use of hybrid polymers can facilitate the design of stable, efficient, and controllable systems for colistin delivery. This study provides a viable pathway for the development of safe nanocarriers with high encapsulation efficiency under realistic conditions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-47021-8.

Subject terms: Biotechnology, Chemistry, Drug discovery, Materials science

Introduction

Hospital-acquired infections and antimicrobial resistance (AMR) are among the most important and costly problems facing health systems worldwide1. Gram-negative bacteria (GNB) are primary contributors to these infections and pose a significant clinical threat in hospital settings due to their high levels of drug resistance2. The complex structure of the bacterial cell envelope plays a key role in their resistance and persistence by providing an innate defense barrier against antimicrobial agents3. The primary stages of Gram-negative sepsis include bacterial entry and colonization, endotoxin release, immune system activation, and systemic inflammatory response syndrome (SIRS)4–7. In this context, polymyxins, especially colistin (polymyxin E), are among the last-resort treatments for severe infections caused by multidrug-resistant Gram-negative bacteria8. These cyclic lipopeptides, consisting of three components—the acyl tail, the outer chain, and the heptapeptide ring (Fig. 1)—displace stabilizing cations by electrostatically binding to lipid A in the LPS of the bacterial outer membrane, damaging the cell membrane and causing leakage of cellular contents, ultimately leading to bacterial death9. However, the clinical use of colistin is limited, particularly in sensitive populations such as pregnant women and children, due to nephrotoxicity and neurotoxicity10–12. Targeted drug delivery is therefore a key strategy to reduce the toxicity of colistin. The goal of this approach is to deliver the drug efficiently and specifically, thereby improving therapeutic efficacy while minimizing side effects13,14. Polymers play an important role in drug delivery systems due to properties such as controlled release, targeted delivery, biocompatibility, and degradability15,16. Polymeric nanoparticles have been recognized as effective drug carriers that can protect drugs from degradation, enable controlled release, and improve bioavailability17. Colistin, as a polycationic antibiotic18, can interact with anionic polymers such as hyaluronic acid (HA) to form targeted19. HA is a linear, biocompatible polysaccharide suitable for targeted drug delivery due to its adhesion to the CD44 receptor, ability to form responsive hydrogels, and potential for chemical modification of functional groups20,21.

Fig. 1.

Fig. 1

Open in a new tab

The structure of colistin consists of an acyl tail, a cyclic heptapeptide ring, and an exocyclic peptide chain.

This polymer is composed of repeating disaccharide units, including D-glucuronic acid and N-acetyl-D-glucosamine22. A key property of HA is its capacity for chemical modification with hydrophobic groups, which leads to the formation of self-assembled nanoparticles with stable hydrophobic cores. Such modifications enhance mechanical stability, reduce biodegradation rates, and improve drug loading capacity23–25. Polyelectrolyte complexes have also attracted attention in pharmaceutical system design due to their unique physicochemical properties and high biocompatibility26. For example, polyanionic poly(styrene sulfonate) complexes can form stronger interactions with colistin27. Given the challenges of developing new drug compounds28–30, there has been growing interest in optimizing existing formulations. In this regard, computational pharmacology, using big data and artificial intelligence, offers a novel and efficient approach for developing drug delivery systems31. Molecular dynamics are among the most accurate computational methods at the nanoscale, modeling temporal behavior of molecular systems32. The accuracy and efficiency of MD simulations primarily depend on the choice of the force field (FF), which defines the potential energy of the system as a function of atomic coordinates33–35. Molecular modeling is typically categorized into four levels: quantum, all-atom, coarse-grained, and mesoscale, each with distinct applications depending on the purpose and size of the system36. Coarse-grained molecular dynamics (CGMD) reduces the number of degrees of freedom by treating groups of atoms as single particles, allowing more efficient simulations of larger molecular systems over longer time scales37–39. One of the most widely used coarse-grained models is the Martini force field developed by Marrink et al.40. In this study, interactions between colistin and polymeric carriers were investigated using molecular dynamics simulations with the Martini 3 force field41,42. Despite the critical role of colistin against Gram-negative bacteria, computational studies on polymeric carriers specifically designed for its delivery are scarce. Site-specific hydrophobic modification of HA and the use of hybrid polyelectrolyte systems such as PVA–PSS have not been previously explored, and detailed molecular-level insights into polymer–colistin interactions and carrier stability remain lacking. This study addresses these gaps by applying advanced MARTINI 3 coarse-grained simulations to systematically evaluate polymer modifications and hybrid complexes, providing a novel framework for designing stable and efficient colistin nanocarriers.

Methods

Software. Molecular dynamics simulations in this study were performed using GROMACS (version 2022.2; https://manual.gromacs.org/2022.2/index.html, RRID: SCR_014565) on an NVIDIA GTX 1080 Ti graphics card. The chemical structures of the polymers were modified in GaussView, and the resulting structures were prepared for molecular simulations in Chimera.

Tools. PyCGTOOL is a software package designed to derive coarse-grained model parameters, such as bonds and angles, by combining atomic simulations with user-defined mapping43. Polyply is a Python-based package that streamlines and automates the molecular dynamics simulations of macromolecules and nanomaterials44. The Automated Topology Builder (ATB) was employed to generate topology files for the molecules of interest45.

All-atom to coarse-grained model conversion

Colistin and Hyaluronic Acid. Bead-based structures of colistin and hyaluronic acid were generated following the MARTINI 3 guidelines (Fig. 2). The initial structure of colistin was obtained from the PubChem database (RRID: SCR_004284), while hyaluronic acid was retrieved from the PDB database (RRID: SCR_012820, ID 2BVK). Corresponding topology files were prepared using the ATB database. For CG parameterization, after mapping the atoms to their respective beads (Fig. S1 and S2), PyCGTOOL was used to calculate bond angles and bead interactions.

Fig. 2.

Fig. 2

Open in a new tab

A, B All-atom structures of colistin and hyaluronic acid, respectively, and C, D their corresponding coarse-grained models.

PVA and PSS Polymers. Using the Polyply tool, vinyl alcohol and polystyrene sulfonate monomers were mapped according to the MARTINI scheme and placed into the simulation box. Following a melting and linking process (with the addition of sodium ions for the anionic sample), they were converted into polymeric chains (Fig. 3). The polymer chain length of PVA consisted of 50 monomer units, while the polymer chain length of PSS was set to 20 monomer units.

Fig. 3.

Fig. 3

Open in a new tab

A, B Monomeric units of PVA and PSS polymers, respectively, and C, D their coarse-grained representations generated using the Polyply tool.

Structure Modification. The extracted HA structure was modified at two and four sites by substituting the hydrogen atoms of the hydroxyl (–OH) groups in the D-glucuronic acid units with C₈H₁₇ alkyl chains using GaussView, resulting in the formation of ether linkages (–O–C₈H₁₇). The carboxylate groups responsible for the negative charge of hyaluronic acid were not affected. After structural optimization, the final structures were prepared for simulation (Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

A, B Modified hyaluronic acid structures with two and four sites, respectively, and C, D their corresponding coarse-grained representations.

Coarse-grained simulation. To simulate the interaction between HA and colistin, the processes of energy minimization, pre-equilibration, equilibration, and production were carried out. Initially, energy minimization was performed using the Steepest Descent algorithm until the system reached an energy tolerance of 100 kJ/mol/nm. Then, the system was stabilized for 100 ns under NVT and NPT conditions (temperature: 300 K, pressure: 1 bar, temperature and pressure coupling using the Berendsen and v-rescale algorithms, respectively) until equilibrium was reached. Subsequently, to improve sampling accuracy, a 500 ns production simulation was performed under NPT conditions using the Parrinello–Rahman pressure coupling. At all stages, electrostatic interactions were modeled using the PME method, and van der Waals interactions with a cut-off radius of 1.1 nm. Periodic boundary conditions (PBC) were applied in all directions, and coarse-grained mapping was performed based on the MARTINI force field. Structural analysis of the systems was performed using indices such as RMSD, Rg, RDF, and SASA.

Results

Colistin–HA delivery system

To investigate the interaction between colistin and HA, a 500 ns simulation was performed at 300 K. The system consisted of 10 HA chains, one colistin molecule, sodium ions, and water, all modeled using the Martini force field. As shown in (Fig. 5A), around 100 ns, HA chains begin to aggregate around colistin. These aggregates persisted for up to 250 ns, after which they decreased slightly as the system’s mobility increased. This behavior could affect the performance of the drug delivery system.

Fig. 5.

Fig. 5

Open in a new tab

Schematic representations of colistin–HA drug delivery after 500 ns. A HA-Colistin, B HA(2)-Colistin, C HA(4)-Colistin. *Colistin is shown in red.

Colistin–HA(2) delivery system

Modification of HA by adding two short hydrophobic chains enhances chain aggregation. To test whether this modification affects chain aggregation as intended, a 500 ns molecular dynamics simulation was performed under the same conditions as before. Preliminary results show that the modified structure, while still similar to natural HA, aggregates more effectively (Fig. 5B). The data indicate that the modified hyaluronic chains HA(2) cluster more clearly around the colistin molecule, potentially improving the performance of the drug delivery system.

Colistin–HA(4) delivery system

To further enhance HA’s structure, four hydrophobic groups were added to it (HA(4)), and a 500 ns simulation was performed under the same conditions. As shown in (Fig. 5C), HA(4) exhibited faster chain aggregation than HA(2). Within the first 100 ns, initial aggregations formed, and by 250 ns, the system reached optimal structural equilibrium for colistin delivery. This stability indicates improved performance of the drug delivery system due to increased hydrophobic modifications.

Analysis of simulation parameters for HA-colistin delivery systems

After simulating the interaction between hyaluronic acid and the drug colistin, RMSD, Rg, SASA, and RDF parameters were analyzed. The results show that the systems’ structural behavior arises from the simultaneous interactions of electrostatic forces, hydrophobic interactions, and the conformational entropy of the polymer chain. Specifically, to investigate the stability of the carrier structure, Rg was calculated during the simulation (Fig. 6A). In HA, Rg fluctuated between ~ 2 and 4 nm, indicating an extended, open, and highly dynamic conformation. This behavior is mainly due to strong electrostatic repulsions between the negative HA beads, which prevent formation of a compact, dense conformation in the chain, even in the presence of electrostatic interactions with the cationic drug colistin. As a result, the polymer maintains high conformational flexibility and shows a limited tendency for stable spatial organization.

Fig. 6.

Fig. 6

Open in a new tab

Schematic plots of A Rg, B RMSD, C RDF, and D SASA for the colistin–HA, HA(2), and HA(4) systems.

In the HA(2), the addition of two hydrophobic groups leads to competing interactions: local chain compaction versus electrostatic repulsions and chain entropy. This creates an unstable conformational state, characterized by repeated opening and closing cycles, sharp Rg and RMSD fluctuations, and broad, low-intensity RDF peaks indicating transient contacts.

In contrast, the HA(4) shows a significant and rapid decrease in Rg from ~ 3.5 to 1.5 nm, reflecting a distinct structural transition from an extended coil to a compact, dense globule. In this system, the higher density of hydrophobic groups leads to their cooperative aggregation, forming a continuous hydrophobic core. This dense core effectively screens electrostatic repulsions and stabilizes the carrier structure by reducing the chain’s freedom and conformational flexibility, as evidenced by reduced Rg oscillations.

The RMSD analysis is a key indicator of structural stability and conformational changes during simulation. The RMSD analysis further corroborates these trends (Fig. 6B). While absolute RMSD values are comparable across systems, the HA(4) system exhibits limited, and uniform fluctuations, (~ 5–5.5 nm). This behavior indicates the suppression of large-scale internal rearrangements following globule formation and reflects the attainment of a stable equilibrium state. In contrast, HA(2) displays broader fluctuations accompanied by a pronounced jump at ~ 300 ns, which can be attributed to repeated structural rearrangements and incomplete stabilization of hydrophobic interactions. Pristine HA shows an intermediate behavior, consistent with a balance between electrostatic repulsions and chain flexibility.

To investigate the structural organization and spatial interactions of molecules in the different systems, the RDF was calculated for HA, HA(2), and HA(4) (Fig. 6C). All three systems, exhibited a multilayer pattern with two main peaks, indicating ordered molecular structures at short distances. In the HA(4) system, a sharp, pronounced peak was observed at 0.5 nm with g(r) ≈ 20, reflecting high structural order and strong interactions in the near-neighbor region. A second peak appeared at 1 nm with g(r) ≈ 10, confirming the ordering of the second layer. Together with a gradual, uniform decrease in g(r) at larger distances, these observations indicate high system stability. Conversely, HA(2) displayed weaker and broader peaks, suggesting greater flexibility and a less coherent structure. The unmodified HA system exhibited intermediate behavior, with relatively regular, moderate-intensity peaks, indicating a balance between structural stability and molecular flexibility. Collectively, the RDF results suggest that HA(4), with its more ordered structure and greater steric stability, a promising candidate for slow, controlled-release drug delivery applications.

The SASA analysis is an indicator of the extent of contact between the structure and the aqueous environment, playing a crucial role in evaluating the stability, solvent interactions, and hydrophobic/hydrophilic properties of drug delivery systems. According to (Fig. 6D), HA exhibited the lowest SASA (~ 120–145 nm²)) with regular fluctuations averaging 135 nm². This pattern indicates high structural stability and limited contact with the solvent, a feature desirable for slow and controlled drug release. In contrast, HA(2) displayed the largest solvent contact area (~ 150–170 nm²) and more prnounced fluctuations. The HA(4) system initially exhibited a higher SASA (~ 180 nm²), which rapidly decreased and stabilized around 160 nm², indicating a balance between structural stability and flexibility.

Collectively, these results demonstrate that increasing the degree of hydrophobic modification in the hyaluronic acid chain transitions the system from a flexible and predominantly electrostatic polymer (HA), through an unstable and frustrated intermediate state HA(2), to a self-assembled, compact, and structured nanocarrier HA(4). This shift highlights the key role of hydrophobic group density in controlling carrier structural stability, bead-level contact patterns, polymer flexibility, and drug entrapment potential.

Simulation of colistin with the polyelectrolyte complex

The anionic PSS polymer is suitable for loading the cationic drug colistin due to its negative charge and high flexibility. The inclusion of PVA chains further enhances the system’s strength and flexibility46. Accordingly, these two polymers were combined to design a colistin drug delivery system. The system was simulated for 500 ns, during which polymer chains began to forming aggregates after 200 ns and established effective interactions with the drug (Fig. 7). This aggregation continued until 250 ns, at which point the drug delivery system reached complete stability, with no signs of structural collapse.

Fig. 7.

Fig. 7

Open in a new tab

Schematic representation of the PVA–PSS–colistin polyelectrolyte system. Red: colistin, green: PSS, and blue: PVA.

Analysis of simulation parameters for PVA-PSS–colistin delivery systems

Analysis of the Rg and RMSD (Fig. 8A, B) indicates the stability and favorable dynamic properties of the PVA-PSS drug delivery system. The Rg profile (Fig. 8A) shows that the system has an open, unstable structure, with Rg values ranging from 4 to 6 nm during the early stages of the simulation. This reflects the initial rearrangement of the polymer chains as the system searches for a stable equilibrium. As the simulation progresses, Rg gradually decreases to approximately 2 nm, stabilizing at this value after around 200 ns. This gradual compaction indicates a controlled aggregation process, mainly driven by electrostatic attractions between the negative groups of PSS and the positive drug colistin, while the highly flexible PVA chains allow for a smooth structural rearrangement.

Fig. 8.

Fig. 8

Open in a new tab

Schematic plots of A Rg, B RMSD, C RDF, and D SASA for the colistin-PVA-PSS systems.

RMSD profile (Fig. 8B) also supports this interpretation. An initial increase in RMSD is followed by limited fluctuations in the 7–9 nm range, indicating that the system reaches a stable topology while maintaining dynamic flexibility. This suggests that although the system is not rigidly locked, large-scale structural rearrangements do not occur, and the overall carrier framework is preserved, a feature critical for proper function under physiological conditions.

The RDF analysis (Fig. 8C) provides further insight into the structural organization at the bead scale. A sharp peak at short distances (0.4–0.5 nm) indicates strong, close-range interactions, primarily resulting from electrostatic complexation between PSS and colistin. A broader secondary peak at ~ 1 nm corresponds to a less ordered outer layer, likely due to the presence of PVA chains and more loosely packed polymer segments. The gradual decay of g(r) at larger distances reflects the transition to less ordered regions in the system’s periphery. Overall, this pattern is consistent with the formation of a core–shell–like structure, featuring an electrostatically stabilized core surrounded by flexible outer layers.

The SASA results (Fig. 8D) further support this behavior. The high initial SASA value (~ 300 nm²) indicates extensive solvent contact during the early stages of the simulation, consistent with the system’s open structure. Over time, the SASA gradually decreases and stabilizes in the range of 240–250 nm². This gradual decrease indicates surface rearrangement and the partial burial of the drug and charged groups within the carrier structure, while still maintaining sufficient contact with the aqueous environment for controlled exchange. The limited fluctuations in SASA also indicate that the system maintains its dynamic flexibility while remaining structurally stable.

Taken together, these results demonstrate that the PVA–PSS–colistin system forms a stable yet dynamic polyelectrolyte complex in which gradual structural compaction is primarily driven by electrostatic attractions. The presence of PVA, a neutral and flexible polymer, prevents sudden structural collapse and helps maintain a balance between structural stability and molecular mobility. These characteristics make this system a suitable candidate for effective drug entrapment and controlled release in drug delivery applications.

Drug-loaded vs. drug-free systems

HA(4). Analysis of HA(4) (Fig. S3) revealed that the carrier beads exhibit aggregation even without the drug. This observation indicates that intrinsic electrostatic and hydrophobic interactions within the polymer structure are sufficient to drive self-assembly and cluster formation.

To assess the structural stability of the drug delivery systems, the RMSD and Rg were analyzed. The Rg profiles revealed initial fluctuations in the range of 2.00 to 4.50 nm during the early stages of the simulation for both systems (Fig. 9A). However, the subsequent decrease in these fluctuations indicates structural coherence and overall stability of the systems under simulated conditions. As shown in Fig. 9A, the radius of gyration (Rg) of both systems exhibit significant fluctuations (~ 2.0–4.5 nm) in the early stages of the simulation, reflecting the high flexibility of the polymer chains at the beginning of the process. This behavior is due to the inherent flexibility of the hyaluronic acid backbone, combined with the presence of hydrophobically modified groups, that induce transient intra- and intermolecular contacts. In the drug-free HA(4) system, the gradual decrease and stabilization of the Rg value at around 1.7–1.8 nm indicate spontaneous polymer compaction, driven primarily by hydrophobic interactions between the hydrophobically modified beads. Despite the net negative charge of hyaluronic acid, these hydrophobic interactions alone are sufficient to overcome electrostatic repulsions, leading to self-assembly and polymer cluster formation even in the absence of the drug. Upon the addition of colistin, a further decrease in Rg is observed, reaching approximately 1.6–1.7 nm, which indicates the formation of a more compact and ordered structure. This additional compaction can be attributed to strong electrostatic attractions between the cationic colistin beads and the negatively charged hyaluronic acid beads, which act as a bridging mechanism between polymer chains. At the bead scale, colistin reduces interpolymer repulsion and simultaneously promotes a more compact arrangement through multivalent electrostatic interactions.

Fig. 9.

Fig. 9

Open in a new tab

Schematic plots of A Rg and B RMSD of HA(4) with and without colistin.

To calculate the carrier radius, the average Rg value was used and substituted into the equation Inline graphic, which represents the effective radius of a spherical structure47,48. Based on this calculations, the Rg of the system without drug was 22.7 Å (2.27 nm), while in the presence of the drug it decreased to 21.2 Å (2.12 nm). This slight reduction in Rg indicates an increased density of the carrier structure upon drug loading, likely resulting from stronger interactions between colistin and the polymer chains, thereby contributing to enhanced structural integrity of the drug delivery system. The RMSD profiles (Fig. 9B) also support this interpretation. The HA(4) system shows larger RMSD fluctuations during the early stages of the simulation compared with the drug-loaded system, indicating greater polymer mobility and structural rearrangements in the absence of stabilizing interactions. In contrast, the HA(4)+Colistin system quickly reaches a stable RMSD region, reflecting reduced structural mobility due to persistent drug–polymer interactions.

Taken together, these results demonstrate that hydrophobic interactions dominate the initial stages of self-assembly, whereas electrostatic forces introduced by colistin play a key role in stabilizing the structure and driving its final compaction. Moreover, the reduced RMSD fluctuations observed in the drug-loaded system indicate decreased polymer flexibility and a transition from a relatively loose polymer cluster to a stable drug–carrier complex. This synergistic interplay between hydrophobic aggregation and electrostatic binding underlies the enhanced stability of the hyaluronic acid–based nanocarrier in the presence of colistin.

PVA-PSS. The PVA-PSS system was also examined. As shown in Fig. 10A, the drug-free PVA–PSS system exhibits a relatively stable radius of gyration (approximately 2.0–2.1 nm) throughout the simulation, indicating the formation of a compact, mechanically stable polymer network. This stability arises from the inherent rigidity of the PSS backbone, along with favorable interchain electrostatic interactions that prevent large-scale structural rearrangements and limit polymer’s flexibility. Upon addition of colistin, the system initially exhibits significant fluctuations in the Rg value (approximately 4.0–6.0 nm), reflecting extensive structural rearrangements driven by strong electrostatic attraction between the cationic colistin beads and the negatively charged sulfonate beads of PSS. At the bead scale, colistin acts as a multivalent electrostatic linker, temporarily disrupting the preorganized polymer network and leading to more elongated structures during the early stages of the simulation.

Fig. 10.

Fig. 10

Open in a new tab

Schematic plots of A Rg and B RMSD of PVA-PSS with and without colistin.

As the simulation progresses, these electrostatic rearrangements gradually lead to the formation of more compact and energetically stable configurations. After about 200 ns, the Rg value of the drug-containing system approaches that of the drug-free system, indicating the re-formation of a stable polymer network, now reinforced by stable drug–polymer electrostatic contacts. This convergence suggests that colistin does not disrupt the long-term stability of the PVA–PSS carrier, but instead reorganizes its internal contact pattern. The calculated carrier radius in the absence of colistin is 25.6 Å (2.56 nm), whereas in its presence it increases to 27.9 Å (2.79 nm). This increase in size indicates the presence of stable and effective interactions between the drug and the polymeric carrier structure. The RMSD profiles (Fig. 10B) also confirm this behavior. The drug-free system exhibits small, low-amplitude RMSD fluctuations, indicative of limited molecular mobility and high structural rigidity. In contrast, the colistin-containing system shows larger RMSD fluctuations in the early stages, consistent with polymer rearrangements induced by electrostatic interactions. However, these fluctuations remain limited, and the system reaches a stable RMSD region after about 250 ns.

Overall, these dynamics indicate that electrostatic interactions play a dominant role in both the initial destabilization and the subsequent stabilization of the PVA–PSS system. In contrast to the hyaluronic acid-based carrier, in which hydrophobic interactions initiate self-assembly, the PVA–PSS carrier forms a robust electrostatic network with inherently reduced polymer flexibility. Colistin is incorporated into this network through strong electrostatic interactions at the bead scale, ultimately yielding a stable yet dynamic drug–carrier complex suitable for controlled drug-delivery applications.

High-concentration simulations

HA(4)-Colistin. The results of molecular dynamics simulations over 500 ns at high concentrations indicate a significant structural stability of the drug delivery system (Fig. 11A). The decrease in Rg from 4 to 2.5 nm reflects effective structural compaction (Fig. 11B), while the RMSD reached a stable plateau (~5 nm) after approximately 100 ns (Fig. 11C). The RDF profile, characterized by a primary peak at 0.5 nm (𝑔(𝑟) ≈ 11) and a secondary peak at 1 nm, indicates a high degree of molecular ordering and a uniform distribution of the drug within the carrier matrix (Fig. 11D).

Fig. 11.

Fig. 11

Open in a new tab

Schematic representation of A the HA(4)-colistin system at high concentration, along with the corresponding B Rg, C RMSD, and D RDF plots.

PVA-PSS-Colistin. For the PVA-PSS system, the final frame at 500 ns was centered, and four colistin molecules were introduced around the carrier (Fig. 12A). The Rg exhibited controlled fluctuations between 1.9 and 2.4 nm, reflecting the structural flexibility of the system and its rapid reversibility following perturbation (Fig. 12B). The RMSD remained close to 6 nm throughout the simulation, with only minor fluctuations observed between 400 and 450 ns that rapidly stabilized (Fig. 12C). (Fig. 12C). The final visualization confirms the formation of a coherent and well-organized structure. The RDF analysis, showing a pronounced peak at 0.5 nm, indicates strong drug–carrier interactions and favorable structural organization, while the distribution range extending up to 4 nm further confirms the spatial integrity of the system (Fig. 12D).

Fig. 12.

Fig. 12

Open in a new tab

Schematic representation of A colistin encapsulation within the PVA–PSS complex, along with the corresponding B Rg, C RMSD, and D RDF plots.

Limitation

A key limitation of this study is that it is entirely computational, and no in vitro or in vivo validation of toxicity, biocompatibility, or drug release was performed. While the simulations provide detailed atomistic insights into polymer–colistin interactions, structural stability, and polymer flexibility, these predictions require experimental confirmation. In Addition, only selected polymer modifications (HA(2), HA(4), and PVA–PSS) and specific concentrations were investigated, and alternative polymer architectures or environmental conditions may influence system behavior. Despite these limitations, the present study establishes a predictive framework to support experimental prioritization and the rational design of safer and more effective colistin nanocarriers, which will be evaluated in future biological studies.

Conclusion

The present computational study demonstrates that both hyaluronic acid–based carriers and the PVA–PSS polyelectrolyte complex are promising candidates for the targeted delivery of colistin, with the potential to mitigate its well-known nephrotoxicity and neurotoxicity. The findings highlight that polymer engineering strategies combined with site-specific chemical modifications—such as the introduction of four hydrophobic groups into HA or the formation of PVA–PSS complexes—can be effectively employed to design stable and efficient drug delivery systems. Among the investigated systems, hyaluronic acid modified with four hydrophobic groups (HA(4)) exhibited the highest structural stability, suggesting superior suitability for colistin encapsulation.

Moreover, molecular dynamics simulations provided detailed atomistic insights into polymer–drug interactions, aggregation behavior, and structural dynamics, underscoring their value as a predictive tool in nanocarrier design. Overall, these results represent a meaningful step toward the development of safer and more effective colistin formulations for combating resistant bacterial infections, offering a promising strategy to reduce the serious toxicities associated with this essential antibiotic.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1. (2.2MB, docx)

Author contributions

H.N. designed and supervised the research. H.F. provided scientific guidance and contributed to the project supervision. K.A. contributed to the study conceptualization and interpretation of the results. A.Z. performed the molecular dynamics simulations, processed the trajectory data, and conducted the computational analyses. The manuscript was written through contributions of all authors. All authors have reviewed and approved the final version of the manuscript.

Funding

The authors received no funding for this work.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.Hormozi, S. F., Vasei, N., Aminianfar, M., Darvishi, M. & Saeedi, A. A. Antibiotic resistance in patients suffering from nosocomial infections in Besat Hospital. Eur. J. Transl Myol. 28, 304–308 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Feinsod, F. M. Infection management for geriatrics in long-term care facilities. J. Am. Med. Dir. Assoc.4, 180 (2003). [Google Scholar]
  • 3.Zgurskaya, H. I., López, C. A. & Gnanakaran, S. Permeability barrier of gram-negative cell envelopes and approaches to bypass it. ACS Infect. Dis.1, 512–522 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Vincent, J. L. et al. International Study of the Prevalence and Outcomes of Infection in Intensive Care Units. Jama302, 2323–2329 (2014). [DOI] [PubMed] [Google Scholar]
  • 5.Gorbet, M. B. & Sefton, M. V. Biomaterial-Associated Thrombosis: Roles of Coagulation Factors, Complement, Platelets and Leukocytes. The Biomaterials: Silver Jubilee Compendium (Woodhead Publishing Limited, 2006). 10.1016/B978-008045154-1.50025-3 [DOI] [PubMed]
  • 6.Ernst Th. Rietschel & Bruce Beutler. Innate immune sensing and its roots the story of endotoxin. Nat. Rev. Immunol.3, 169–176 (2003). [DOI] [PubMed] [Google Scholar]
  • 7.Bone, R. C. et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest101, 1644–1655 (1992). [DOI] [PubMed] [Google Scholar]
  • 8.Bialvaei, A. Z. & Samadi Kafil, H. Colistin, mechanisms and prevalence of resistance. Curr. Med. Res. Opin.31, 707–721 (2015). [DOI] [PubMed] [Google Scholar]
  • 9.Rhouma, M., Beaudry, F., Thériault, W. & Letellier, A. Colistin in pig production: Chemistry, mechanism of antibacterial action, microbial resistance emergence, and one health perspectives. Front. Microbiol.7, 1–22 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Chen, Z. et al. Meta-analysis of colistin for the treatment of Acinetobacter baumannii infection. Sci. Rep.5, 17091 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Lim, L. M. et al. Resurgence of colistin: a review of resistance, toxicity, pharmacodynamics, and dosing. Pharmacother J. Hum. Pharmacol. Drug Ther.30, 1279–1291 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Spapen, H., Jacobs, R., Van Gorp, V., Troubleyn, J. & Honoré, P. M. Renal and neurological side effects of colistin in critically ill patients. Ann. Intensive Care. 1, 1–7 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Nayak, A. K., Ahmad, S. A., Beg, S., Ara, T. J. & Hasnain, M. S. Drug Delivery: Present, Past, and Future of Medicine. Applications of Nanocomposite Materials in Drug Delivery (Elsevier Inc., 2018). 10.1016/B978-0-12-813741-3.00012-1
  • 14.Liu, Y., Shah, S. & Tan, J. Computational modeling of nanoparticle targeted drug delivery. Rev. Nanosci. Nanatechnol.1, 66–83. 10.1166/rnn.2012.1014 (2012). [Google Scholar]
  • 15.Bae, Y. H. Smart polymers in drug delivery. Pharm. News. 9, 417–424 (2002). [Google Scholar]
  • 16.Gandhi, K. J., Deshmane, S. V. & Biyani, K. R. Polymers in pharmaceutical drug delivery system: A review. Int. J. Pharm. Sci. Rev. Res.14, 57–66 (2012). [Google Scholar]
  • 17.Zielińska, A. et al. Polymeric nanoparticles: production, characterization, toxicology and ecotoxicology. Molecules25, 3731 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Vairo, C., Villar Vidal, M., Hernandez, M., Igartua, R. & Villullas, S. M. Colistin- and amikacin-loaded lipid-based drug delivery systems for resistant gram-negative lung and wound bacterial infections. Int. J. Pharm.10.1016/j.ijpharm.2023.122739 (2023). [DOI] [PubMed] [Google Scholar]
  • 19.Dubashynskaya, N. V. et al. Hyaluronan/colistin polyelectrolyte complexes: Promising antiinfective drug delivery systems. Int. J. Biol. Macromol.187, 157–165 (2021). [DOI] [PubMed] [Google Scholar]
  • 20.Buckley, C., Murphy, E. J., Montgomery, T. R. & Major, I. Hyaluronic acid: A review of the drug delivery capabilities of this naturally occurring polysaccharide. Polym. (Basel). 14, 3442 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Huang, G. & Huang, H. Application of hyaluronic acid as carriers in drug delivery. Drug Deliv. 25, 766–772 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Fraser, J. R. E., Laurent, T. C. & Laurent, U. B. G. The nature of hyaluronan. J. Intern. Med.242, 27–33 (1997). [DOI] [PubMed] [Google Scholar]
  • 23.Hill, T. K. et al. Indocyanine green-loaded nanoparticles for image-guided tumor surgery. Bioconjug. Chem.26, 294–303 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ondreas, F. et al. Self-assembly of hydrophobically modified hyaluronic acid. Appl. Surf. Sci.546, 149161 (2021). [Google Scholar]
  • 25.Payne, W. M., Svechkarev, D., Kyrychenko, A. & Mohs, A. M. The role of hydrophobic modification on hyaluronic acid dynamics and self-assembly. Carbohydr. Polym.182, 132–141 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Lankalapalli, S. & Kolapalli, V. R. M. Polyelectrolyte complexes: A review of their applicability in drug delivery technology. Indian J. Pharm. Sci.71, 481–487 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Xiao, X., Ji, J., Zhao, W., Nangia, S. & Libera, M. Salt destabilization of cationic colistin complexation within polyanionic microgels. Macromolecules.10.1021/acs.macromol.1c02157 (2022). [Google Scholar]
  • 28.Scannell, J. W., Blanckley, A., Boldon, H. & Warrington, B. Diagnosing the decline in pharmaceutical R&D efficiency. Nat. Rev. Drug Discov. 11, 191–200 (2012). [DOI] [PubMed] [Google Scholar]
  • 29.Chong, C. R. & Sullivan, D. J. Jr New uses for old drugs. Nature448, 645–646 (2007). [DOI] [PubMed] [Google Scholar]
  • 30.Beg, S., Swain, S., Rizwan, M. & Irfanuddin, M. Shobha Malini, D. Bioavailability enhancement strategies: basics, formulation approaches and regulatory considerations. Curr. Drug Deliv. 8, 691–702 (2011). [DOI] [PubMed] [Google Scholar]
  • 31.Wang, W., Ye, Z., Gao, H. & Ouyang, D. Computational pharmaceutics-A new paradigm of drug delivery. J. Control Release. 338, 119–136 (2021). [DOI] [PubMed] [Google Scholar]
  • 32.Saurabh, S. et al. Molecular dynamics simulations in drug discovery and drug delivery. Eng. Mater.10.1007/978-3-030-36260-7_10 (2020). [Google Scholar]
  • 33.Sauceda, H. E., Chmiela, S., Poltavsky, I., Müller, K. R. & Tkatchenko, A. Molecular force fields with gradient-domain machine learning: Construction and application to dynamics of small molecules with coupled cluster forces. J. Chem. Phys.150, 68–70 (2019). [DOI] [PubMed] [Google Scholar]
  • 34.Waidyasooriya, H. M., Hariyama, M. & Kasahara, K. An FPGA accelerator for molecular dynamics simulation using OpenCL. Int. J. Networked Distrib. Comput.5, 52–61 (2017). [Google Scholar]
  • 35.Chang, C. E. A., Huang, Y. M. M., Mueller, L. J. & You, W. Investigation of structural dynamics of enzymes and protonation states of substrates using computational tools. Catalysts6, 82 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lalitha, A. Modelling of MOF/Graphene oxide composites and their performances for CO2 capture. (2020).
  • 37.Noid, W. G. Perspective coarse-grained models for biomolecular systems. J. Chem. Phys.139, 090901 (2013). [DOI] [PubMed] [Google Scholar]
  • 38.Saunders, M. G. & Voth, G. A. Coarse-graining methods for computational biology. Annu. Rev. Biophys.42, 73–93 (2013). [DOI] [PubMed] [Google Scholar]
  • 39.Kmiecik, S. et al. Coarse-grained protein models and their applications. Chem. Rev.116, 7898–7936 (2016). [DOI] [PubMed] [Google Scholar]
  • 40.Marrink, S. J., Risselada, H. J., Yefimov, S., Tieleman, D. P. & De Vries, A. H. The MARTINI force field: Coarse grained model for biomolecular simulations. J. Phys. Chem. B. 111, 7812–7824 (2007). [DOI] [PubMed] [Google Scholar]
  • 41.Souza, P. C. T. et al. Martini 3: a general purpose force field for coarse-grained molecular dynamics. Nat. Methods. 18, 382–388 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Grünewald, F. et al. Martini 3 coarse-grained force field for carbohydrates. J. Chem. Theory Comput.18, 7555–7569 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Graham, J. A., Essex, J. W., & Khalid, S. PyCGTOOL: Automated generation of coarse-grained molecular dynamics models from atomistic trajectories. J. Chem. Inf. Model.57, 650–656 (2017). [DOI] [PubMed] [Google Scholar]
  • 44.Grünewald, F. et al. Polyply; a python suite for facilitating simulations of macromolecules and nanomaterials. Nat. Commun.13, 1–12 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Malde, A. K. et al. An automated force field topology builder (ATB) and repository: version 1.0. J. Chem. Theory Comput.7, 4026–4037 (2011). [DOI] [PubMed] [Google Scholar]
  • 46.Rivera-Hernández, G., Antunes-Ricardo, M., Martínez-Morales, P. & Sánchez, M. L. Polyvinyl alcohol based-drug delivery systems for cancer treatment. Int. J. Pharm.600, 120478 (2021). [DOI] [PubMed] [Google Scholar]
  • 47.Chen, S. H. Small angle neutron scattering studies of the structure and interaction in micellar and microemulsion systems. Annu. Rev. Phys. Chem.37, 351–399 (1986). [Google Scholar]
  • 48.Mobasheri, M., Attar, H., Rezayat Sorkhabadi, S. M., Khamesipour, A. & Jaafari, M. R. Solubilization behavior of polyene antibiotics in nanomicellar system: insights from molecular dynamics simulation of the amphotericin B and nystatin interactions with polysorbate 80. Molecules21, 6 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1. (2.2MB, docx)

Data Availability Statement

Data is provided within the manuscript or supplementary information files.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES