Introduction

Globoid cell leukodystrophy (GCL) also known as Krabbe disease is a remarkably rare inherited neuronal disorder, manifests as a consequence of galactocerebrosidase enzyme (encoded by the GALC gene) mutation, leading to galactocerebroside accumulation in various tissues1. Nevertheless, the gravest ramifications manifest in the central nervous system (CNS), predominantly impacting the myelin sheath, thereby instigating a cascade of deleterious processes eliciting inflammatory responses, immune cells increased activity, demyelination, globoid cells formation, progressive neurological deterioration, peripheral neuropathy, and motor function impairment with seizure episodes1,2. The GCL clinical symptoms typically manifest during infancy, causing irritability, feeding disorders, and developmental regression2, with the disease progression significantly impairing both motor and cognitive functions, thereby markedly declining the overall quality of life. Unfortunately, the current scenario for GCL is characterized by a scarcity of therapeutic avenues, thereby markedly constraining the spectrum of viable treatment modalities and presenting formidable obstacles for both healthcare practitioners and afflicted individuals. Despite the potential efficacy of hematopoietic stem cell transplantation, especially in neonates who have not yet developed severe symptoms in the early stages of the condition, the availability of such interventions remains significantly limited3,4.

Differential gene expression (DGE) analyses, through the comparative examination of gene expression patterns between diseased states and healthy or control conditions, facilitate a deeper comprehension of disease progression, thereby aiding in the identification of potential drug targets elucidating the intricate molecular underpinnings of various ailments5,6. This approach may offer invaluable insights into the molecular mechanisms that propel the disease forward, thereby facilitating a comprehensive understanding of pathological processes enabling to identify potential checkpoints in disease intervention. Furthermore, genes demonstrating significant differential expression in disease can be strategically prioritized as promising candidates for therapeutic targeting5,6,7. Importantly, targeting a core node in a disease-associated protein  network  can identify novel therapeutic approach  as this perspective may directly impact over disease system by manipulating the protein interactions. This prioritization would be rational in novel drug discovery efforts or drug repurposing by focusing on specific genes and proteins more likely to be involved in the disease pathogenesis8. Therefore, in this study,  initially we focused to identify the DEGs in GCL, assess  potential pathophysiological-neuronal processes, elucidate pivotal protein node implicated in disease progression, evaluate the neuronal protein circuitry, and finally, screen possible  small molecules that target the central protein node(s) of interest.

Furthermore, it is imperative to acknowledge that protein interactions serve as a cornerstone to fundamental biological processes, thereby intricately shaping  cellular signaling cascades, modulating immune responses, and regulating disease-associated molecular functions, thus significantly propelling forward drug discovery endeavors through the central protein node(s) identification 9. Similarly, molecular docking aids in comprehending the ligand binding affinity with the possible target of interest, thereby discerning the preferred molecule pose, and offering better insights into potential interaction modes and affinities by providing a static snapshot of the interaction10. Due to the inherently dynamic nature of ligand–protein complexes, wherein they incessantly undergo intricate conformational alterations, molecular dynamics simulations emerge as an indispensable approach to facilitate a nuanced comprehension of the atomic-scale physical dynamics across temporal trajectories. By delving into the intricacies of stability, flexibility, and conformational dynamics within a dynamic milieu, molecular dynamics simulations provide a highly detailed depiction, within a  closer physiological conditions11. The amalgamation of these methodologies could discern the core node within the disease-associated neuronal protein network circuit implicated in the pathogenesis, comprehend the energetics governing ligand–protein interactions, and facilitate the transient states and elucidate intermediate conformations to explore pivotal functional mechanisms.

In this study, we also emphasize the interaction between T-5224 (ligand) and the proto-oncogene c-Jun (target) through molecular docking and molecular dynamics simulations, identifying c-Jun as a core node in the GCL-associated neuronal protein network. Furthermore, targeting c-Jun, T-5224 exerts a potent suppressive effect on its activity with the activator protein (AP)-1 transcription factor complex, thereby exerting influence over the intricate transcriptional processes and subsequent gene expression within AP-1 regulated pathways12. Furthermore, c-Jun, functioning as a pivotal transcription factor, intricately orchestrates a multitude of cellular processes, proliferation, differentiation, and apoptosis, exhibiting responsiveness to various stress and injury stimuli13. Elevated c-Jun activity in the neural milieu of GCL is intricately intertwined with the progressive demyelination process14, thereby precipitating apoptosis and fostering inflammatory cascades, which collectively exacerbate the neurodegenerative manifestations severity. This underscores the multifaceted functions of c-Jun, not solely relegated to a mere marker of cellular distress and damage, but actively engaged in pathophysiological mechanisms by orchestrating cellular demise and instigating inflammatory responses. Based on these study bakgrounds, our study aimed to highlight the DEGs-based GCL- neuronal protein network circuit, identify the potential central node, evaluate the binding affinity of T-5224 with c-Jun using molecular docking, and assess complex stability using molecular dynamics simulation.

Materials and methods

Identification of DEGs to map GCL-associated neuronal protein circuit

Initally, we assessed the DEGs from altered developmental pathways in induced pluripotent stem cell (iPSC)-derived neural stem cells from GCL patients, using transcriptome profiling dataset composed of  12 samples (6 each in normal and disease) illustrating how GALC mutation affects signaling pathways of neural development providing a k-iPSCs model to study the molecular basis of GCL development15. We identified the DEGs by adjusting p-values using the Benjamini & Hochberg method (false discovery rate) at a significant level cutoff (p < 0.05, |Fc|> 1).

Further, genes with |Fc|> 1 were subjected to construct a GCL-associated neuronal protein network circuit using the Search Tool for the Retrieval of Interacting Genes/Proteins (https://www.string-db.org/) version 11.5 with a size cutoff of less than 10 for interactors and a high confidence score to evaluate related biological processes. Subsequently, protein network  was mapped using Cytoscape (https://cytoscape.org/) version 3.9.1, and integrated centrality (IC) was calculated for each node i as explained by Li et al.16 with minor modifications, using the following equation.

$$ICi= \frac{1}{4}\left(\sum_{1}^{7}(a1+a2+\dots a7)\right)$$

In the above equation, 'a1 to 7' represents the ratio of the difference in minimum functional score with a given node to the difference between the maximum and minimum functional score for neighborhood connectivity degree (a1), betweenness (a2), closeness (a3), average shortest path length (a4), indegree (a5), outdegree (a6), and stress (a7). The value of IC ranges from 0 to 1. From a topological perspective, the significance of a target within its functional module increases proportionally with its higher IC value, indicating a greater centrality and pivotal role in the disease protein network16. Subsequently, small molecules were screened from molecular atlas and pharma-information of all drugs17 against GCL-associated protein network circuit nodes with at least one interactive path (edge).

Contrasting core node expression in the GCL-associated protein network

The pivotal role in unravelling complex molecular mechanisms underlying cellular functions can be assessed by comparing hub gene(s) expression  which delve the differential regulatory expression, and help to focus on core nodes decoding close disease protein association. To embark on this analysis, we scrutinized the expression of top interacting genes with maximum IC score within GCL-associated neuronal protein circuit comparing GCL samples with normal using an unpaired t-test with Welch’s correction, assuming a Gaussian distribution in GraphPad Prism (https://www.graphpad.com/), and the expression was considered significant if p < 0.05.

Molecular docking and molecular dynamics simulation to evaluate ligand–protein interaction

In the GCL-associated neuronal protein network circuit, we discern c-Jun (JUN node) exhibiting the utmost IC, thereby positing it as a pivotal hub crucial for the GCL developmental processes . Consequently, we speculated that interfering with the c-Jun function in the aforementioned network may have the potential to exert a significant effect on GCL development. Moreover, our investigation pinpointed T-5224 as a candidate for c-Jun activity manipulation embarking upon a series of molecular docking and simulation studies involving T-5224 and Jun/CRE complex, aiming to assess their binding affinity and stability.

The 3D structure of T-5224 was retrieved from the PubChem small molecule database (https://pubchem.ncbi.nlm.nih.gov/compound/23626877) in ‘.sdf’ format. This molecule was further processed to ‘.pdb’ format and underwent minimization using the 'mmff94' forcefield with the ‘POAP_lig.bash’ script of the POAP pipeline. The structure was then converted into the ‘.pdbqt’ format by incorporating Gasteiger charges and polar hydrogens. The 3D structure of the macromolecule Jun/CRE complex (PDB: 1JNM) was obtained from the Research Collaboratory for Structural Bioinformatics structural database (https://www.rcsb.org/structure/1jnm) and information regarding its binding pocket residues was assessed using the computed atlas of surface topography of proteins (CASTp) server (http://sts.bioe.uic.edu/castp/index.html?201l). The ligand was docked within the identified binding pocket setting grid map center (x, y, z = 9.36, − 5.34, 14.10) and size (x, y, z) = 25.0 using AutoDock Vina (https://autodock.scripps.edu/) via the POAP pipeline, resulting in nine docked conformations, and the conformation with the lowest root-mean-square deviation (RMSD) was generated. After docking, interactions between T-5224 and c-Jun/CRE were visualized using BIOVIA Discovery Studio Visualizer (https://discover.3ds.com/discovery-studio-visualizer-download) version 2019.

Further, the Jun/CRE apo form and complex with T-5224 underwent a 200 ns all-atom molecular dynamics simulation to illustrate the conformational changes and intermolecular interaction stabilities. The GROMACS 2021.3 (https://www.gromacs.org/) software package was employed with the Amber ff99SB-ildn force field (https://ambermd.org/index.php). The partial charge of the ligand (T-5224) was calculated through quantum calculations using an antechamber, employing a ‘bcc’ charge model. The topological parameters for both the ligands and the entire complex were generated using the xleap module of AmberTools (https://ambermd.org/AmberTools.php). The system was solvated with a three-site water model in a rectangular box, with boundary conditions set at a distance of 10.0 Å from the protein borders in all directions. Neutralization was achieved by adding the necessary amount of counter ions. Near-global state least energy conformations were obtained using a combination of steepest descent and conjugate gradient energy minimization approaches. The systems were maintained with ‘canonical (NVT) and isobaric (NPT)’ ensembles throughout 1 ns. The NVT equilibration process used a modified Berendsen thermostat algorithm to ensure consistent volume and temperature (300 K). The Parrinello-Rahman barostat was employed during NPT equilibration to maintain a constant pressure of 1 bar. Furthermore, the Particle Mesh Ewald approximation was utilized to calculate coulomb, van der Waals, and long-range electrostatic interactions, using a cut-off value of 1 nm. Similarly, the LINear Constraint Solver method was used to restrict bond lengths. The resulting trajectories were evaluated using built-in GROMACS tools, and other software packages were employed for specific analyses as needed. Additionally, the ‘g_mmpbsa’ program18 was utilized to determine the relative binding energy between T-5224 and Jun/CRE complex.

Assessment of c-Jun/CRE apo form and complex motions and their interdependencies

Each eigenvector is intrinsically linked to a specific eigenvalue and quantifies the corresponding energy contribution of a particular motion component to the aggregate dynamic behavior of the molecule. Projecting the molecular trajectory onto a specific eigenvector intricately reveals the temporal evolution of the components within a distinct vibrational mode and insights into the atomic vibration effects providing a comprehensive understanding of vibrational phenomena and their impact on the overall dynamical behavior19. In this study, the c-Jun/CRE apo form and the T-5224 complex were analyzed to assess their motions and interrelations, employing Principal Component Analysis (PCA) and Dynamic Cross-Correlation Matrix (DCCM). For this, a covariance matrix was initially derived from a linear transformation of the Cartesian coordinate space, which was subsequently subjected to diagonalization, culminating in the eigenvectors set extraction that delineate the motion directions. The computation of trajectory eigenvectors and eigenvalues was performed using the ‘g_covar’ tool in Gromacs which involved calculation and diagonalization of the covariance matrix. Additionally, the ‘g_anaeig’ tool was utilized to examine and illustrate the eigenvectors. Furthermore, the free energy landscape was constructed employing the first two principal components (PC1 and PC2) as interaction coordinates, thereby facilitating a multidimensional representation of the system's energy states20,21. Similarly, the  DCCM was examined to assess the amplitude of all pairwise cross-correlation coefficients, identifying correlated motion (positive or negative) between atom pairs reflecting the intricate dynamic relationships and interdependencies within molecular structures. A value of Cij = 1 denotes that the time frames and phases are identical, indicating a perfect positive correlation, whereas Cij = 0 signifies an absence of any correlation. Conversely, Cij =  − 1 implies that the fluctuations of atoms i and j are in complete antiphase, thereby demonstrating a perfect negative association22.

Results and discussion

A comprehensive DEGs analysis revealed a total of 19,036 genes exhibiting differential expression, of which a subset of 1913 genes demonstrated statistically significant expression levels (p < 0.05). Within this subset, 922 genes were down-regulated, while 991 were up-regulated. Further screening based on |log Fc|> 1 revealed 87 down-regulated genes and 205 up-regulated genes (Fig. 1). These genes (encompassing both upregulated and downregulated) were  utilized to construct a GCL-associated neuronal protein network circuit identifying 53 interacting genes (p = 5.55e−15) corresponded to diverse neurophysiological processes, ascertained through scrutinizing their expression patterns, homology relationships, database annotations, and textual mining. Furthermore, the integrated score was obtained by correcting the probability associated with randomly observed interactions (Fig. 2), elucidating 5 pivotal genes in 4 interrelations that included c-Fos (FOS)- FosB (FOSB), c-Fos (FOS)- transcription factor c-Jun (JUN), FosB (FOSB)- JUN, and glial cell line-derived neurotrophic factor (GDNF)- glial cell line-derived neurotrophic factor receptor α (GFRA1).

Figure 1
figure 1

Differential gene expression on induced pluripotent stem-cell-derived neural stem cells derived from Globoid cell leukodystrophy (GCL) patients using transcriptome profiling. (a) Volcano plot presenting differential expressed genes. A total of 87 genes were downregulated, 205 were upregulated and 1621 were non-significant. The absolute fold change criterion (|Fc|> 1) corresponds to a fold change greater than 2 or less than 0.5 on a linear scale. In log2 scale, this criterion translates to log2(Fold Change) greater than 1 or less than − 1 . The p-value threshold of 0.05, adjusted using the Benjamini & Hochberg method, ensures that the genes considered as differentially expressed meet the statistical significance required for multiple testing corrections. (b) GCL-associated neuronal protein network circuit of differentially expressed genes in the GCL with at least one interaction. The larger the node size, the more central the protein it represents.

Full size image
Figure 2
figure 2

Identification of physiologic-neurobiological process of differentially expressed genes. (a) Neurological processes involved in the brain development were significantly affected. 1 biological regulation, 2 system development, 3 cell differentiation, 4 nervous system development, 5 cell development, 6 neurogenesis, 7 neuron differentiation, 8 cell–cell signaling, 9 neuron development, 10 axon development, 11 axonogenesis, 12 nerve development, 13 neural crest cell differentiation. (b) Heat map presenting the protein node scores. ASPL average shortest path length, BC BetweennessCentrality, CC ClosenessCentrality, EC EdgeCount, ID Indegree, NC neighborhood connectivity, OD Outdegree, St stress, IC integrated centrality. (c) Heat map presenting the gene expression of differentially expressed genes based on fragments per kilobase of transcript per million mapped reads. Genes traced in the Search Tool for the Retrieval of Interacting Genes/Proteins with at least one interaction were evaluated.

Full size image

The GCL-associated DEGs network identifies potential druggable targets

In our present investigation, we spotlight five potential targets (FOS, FOSB, GDNF, GFRA1, and JUN) for possible drug intervention against GCL by examining the network interactions of DEGs in GCL-associated protein network  at high confidence score (0.7)  with DEGs exhibiting statistical significance (p < 0.05) and a fold change (|Fc|> 1). In this network (Fig. 1b),  our focus was directed towards protein nodes exhibiting at least one edge relation to calculate IC identifying the c-Jun protein (JUN node) which suggests its pivotal role in the GCL disease network. Furthermore, GFRA1, GDNF, FOS, and FOSB stand out as key nodes in the proteins interaction network ranking top five nodes (FOS, FOSB, GDNF, GFRA1, and JUN) due to their notably high interaction, and is presented by their substantial node sizes (Fig. 1b). Consequently, we highlight c-Jun (JUN) as a potential druggable target for GCL and screen potential small molecules for JUN, including other hub targets from the drug information related molecular atlas.

The functional enrichment analysis revealed crucial biological processes to encompass nervous system development, axonogenesis, and signal transmission (Fig. 2a), all of which are directly affected by galactolipid accumulation-induced demyelination, axonal damage, neuronal death, and systemic toxic effects. Additionally, in the GCL-associated neuronal protein network, c-Jun (JUN) emerged as a central node and may exhibit hub role in signal transduction (Fig. 1b) and interestingly, all 5 genes (GFRA1, GDNF, FOS, FOSB, and JUN) were significantly down expressed in the GCL (Fig. 3, S1-S4). The pivotal role of GFRA1 in development and survival of nerve cells23, evidenced by its diminished expression in the GCL (Fig. S1) underscores its significance in neuronal function. So, it can be further assessed whether GFRA1 downregulation favors myelin maintenance, precipitating myelin loss and GCL onset highlighting the intricate interplay between GFRA1 signaling and myelin integrity. In this study, our assessment showed that GFRA1 also served as one of the central protein node in GCL-associated protein network, displaying a higher average shortest path length potentially influencing signal transmission to other proteins. Furthermore, the potential role of GFRA1-related growth factors has been investigated to maintain and promote nerve cell survival in myelin damage23. While preclinical studies on GDNF and associated factors promise for neurological disorder24, it is currently unclear how these factors affect  GCL or how their regulators can be utilized against it. Furthermore, the FOS is a component of the AP-1 transcription factor complex contributing to several cellular functions, such as differentiation, growth, and responsiveness to varied stimuli, as well as gene control25, and was found to be down-expressed in GCL (Fig. S3). Similarly, FOSB is also involved in multiple cellular processes responsible for neuron growth, function, and plasticity26. As a transcription factor, FOSB exerts regulatory effects by selectively binding to DNA thereby modulating the transcriptional activity of target genes intricately linked with the complex process of neuronal development27. In GCL, its expression was observed to be significantly lower (p = 0.0002) (Fig. S4), and this manipulated expression may directly affect multiple cellular processes related to neuron growth and function. Furthermore, the c-Jun is critical for neuronal proliferation in the CNS and is decreased in the GCL. This significant decreased (p = 0.0004) expression of JUN (Fig. 3) could be a compensatory mechanism, as its overexpression may increase pro-inflammatory cytokines and chemokines production and trigger an immune response leading to demyelination28,29. Furthermore, c-Jun (JUN) is intricately integrated into the AP-1 complex under its active participation in GCL-associated neuronal protein network circuit with congaing for neuronal growth as functional enrichment analysis elucidated its involvement across broad  neurofunctional process, encompassing response to lipids, transcriptional modulation, tissue morphogenesis, and oxidative stress mitigation. Noteworthy associations extend to nucleoplasmic dynamics, DNA binding affinities, integrated stress response pathways including regulatory functions in nuclear protein-containing complexes and chromosomal architecture. Furthermore, in the GCL pathogenesis, c-Jun emerges as a pivotal regulatory node, orchestrating neuroinflammatory cascades and neurodegenerative pathways in response to neurotoxic insult posed by psychosine accumulation. Its nuanced modulation of gene expression towards psychosine-induced stress underscores its potential as a key player warranting further exploration for therapeutic interventions in GCL (Fig. 4). Consequently, its aberrant upregulation has the potential to dysregulate critical signaling cascades indispensable for myelination, precipitate apoptosis in oligodendrocytes, and provoke oxidative stress, thereby directly compromising the integrity of myelin maintenance28,29.

Figure 3
figure 3

Evaluation of interacting edges in Globoid cell leukodystrophy (GCL)-associated neuronal protein circuit. (a) Identification of central interacting nodes based on the combined score (CS). The combination score was based on phylogenetic co-occurrence (PCO), homology (HOMO), coexpression (COEXP), experimentally determined interaction (EDI), database annotated (DA), and automated text mining (ATM). (b) Fragments per kilobase of transcript per million mapped reads of JUN expression in induced pluripotent stem-cell-derived neural stem cells derived from GCL patients using transcriptome profiling. Data were analyzed using unpaired t-test with Welch’s correction assuming Gaussian distribution, *p = 0.0004 (c) JUN expression in each sample compared with control.

Full size image
Figure 4
figure 4

Assessment of Jun-associated proteins circuit. (a) Heat map presenting the edge (interaction) score for node(protein) pair. The edge score for each protein pair was based on phylogenetic co-occurrence (PCO), homology (HOMO), coexpression (COEXP), experimentally determined interaction (EDI), database annotated (DA), automated text mining (ATM), and combined score (CS). (b) Functional enrichment analysis of JUN-associated biological process (integrated stress response signaling, response to oxidative stress, tissue development, response to lipid), molecular function (transcription regulator activity, DNA binding, transcription coregulator binding), and cellular components (chromosome, nuclear protein-containing complex, nucleoplasm). (c) JUN was associated with 10 different genes. The red node is the query node (JUN) and the rest represent the interacting nodes. The surface dimer structure of JUN is presented.

Full size image

Insight into the JUN inhibitory activity by T-5224 as a therapeutic approach against GCL

Assessment of fragments per kilobase of transcript per million mapped reads reveals suppressed JUN expression (p = 0.0004) in iPSC from GCL patients compared to normal (Fig. 3b) which could be one of the compensatory mechanisms against GCL. The aberrant activation of the c-Jun transcription factor has significant importance in GCL as it may parallel to disease progression with toxic lipids accumulation, thereby instigating cellular stress. This can initiate a complex series of interconnected processes including demyelination, neuroinflammation, and the demise of oligodendrocytes, thereby exacerbating the pathological conditions in the neural environment. Furthermore, it is noteworthy that the c-Jun transcription factor not only governs the gene expression linked to oxidative stress29, inflammation30, and apoptosis31 but also orchestrates these pivotal pathogenic mechanisms in GCL, thereby underscoring its multifaceted role in disease. Nevertheless, achieving a comprehensive understanding of the intricate mechanisms through which the c-Jun transcription factor contributes to GCL and identifying precise therapeutic interventions necessitate the undertaking additional investigative endeavors. Significantly, within the scope of this investigation, the JUN node emerged as the entity possessing the highest integrated central score (Fig. 1b) within the neuronal protein circuit associated with GCL, thereby indicating the potential efficacy of targeting the entirety of the pathogenic network inherent. Henceforth, our inquiry proceeded to put additional emphasis on the c-Jun (JUN) activity inhibition, although we listed  distinct small molecules targeting GCL-related proteins (Table 1, Fig. 5), directing our interest towards a central protein node in the GCL-associated proteins , which profoundly influences the signals interplay among the network constituent proteins. Consequently, our scrutiny pinpointed c-Jun (JUN) as one of the pivotal hub nodes for neuronal demyelination synchronization with other proteins (Fig. 2b), thereby prompting subsequent comprehensive evaluations. Moreover, T-5224 exerts its inhibitory action on the JUN/AP-1 complex by targeting the c-Fos DNA-binding domain, thereby impeding the association of the AP-1 complex with DNA and disrupting the transcriptional activity of the JUN-FOS dimer without affecting other transcription factors, such as NF-Κβ32. However, to confirm T-5224 mediated c-Jun inhibition for potential therapeutic effects on GCL, experimental studies are needed to assess its efficacy in modulating c-Jun activity and subsequent downstream effects, with an expectation of preventing neuronal demyelination and reducing neuronal apoptosis, which may improve neuronal function and needs further investigation.

Table 1 Small molecules identified from molecular atlas of drugs as potential therapeutic moieties towards GCL.
Full size table
Figure 5
figure 5

Screening of small molecules for differentially expressed genes. (a) Targets were initially screened for a different class if successful, under trial, discontinued, or unclear activity in different conditions like cardiovascular, diabetes, cancer, etc. Targets were also screened for their drug availability. (b) 2D sketch of representative drugs availability from DrugMAP: molecular atlas and pharma-information of all drugs (http://drugmap.idrblab.net/) for different GCL-related targets.

Full size image

Structural insight into the T-5224 binding with Jun/CRE complex

In this study, we describe the T-5224 binding affinity with the Jun/CRE complex using molecular docking and assessing the complex for its stability using molecular dynamics simulation. Molecular docking is employed to assess the interaction between ligand(s) and protein(s) by predicting their binding modes which primarily emphasize on the binding energy estimation providing  possible  hydrophilic and hydrophobic interactions33. The interactions of T-5224 with the Jun/CRE complex involve DNA bases and amino acid residues that highlight the molecular intricacies of its binding (Fig. 6, Supplementary Movie 1). The DNA bases involved in these interactions include DG C:211, DA C:209, DC C:210, DC D:309, DC D:310, and DG D:308. Meanwhile, the Jun/CRE complex amino acid residues participating include Arg A:270, Lys B:271, and Leu B:274. T-5224 forming five hydrophilic interactions, specifically with Arg A:270 (4.0 Å), DG C:211 (5.46 Å), DA C:209 (6.43 Å), DC C:210 (6.08 Å), and DC D:309 (5.01 Å). These hydrophilic interactions suggest that T-5224 can establish bonds with both the DNA and protein elements of the complex in an aqueous environment, stabilizing its position. Additionally, nine hydrophobic interactions were also formed, which involve Arg A:270 (4.99 Å), Lys B:271 (6.92 Å), Leu B:274 (6.72 Å and 5.48 Å), Arg B:270 (5.11 Å, 5.59 Å, and 6.49 Å), DC D:310 (7.53 Å), and DG D:308 (4.42 Å). These hydrophobic interactions indicate regions where T-5224 interfaces with nonpolar parts of the Jun/CRE complex, further contributing to its binding affinity and specificity. Importantly T-5224-DNA interaction may also influence its structure and function by binding with DNA base pairs showing groove binding fitting into the grooves of the helix forming both hydrophilic and hydrophobic interactions which can affect replication and transcription affecting the Jun/CRE complex function. Together, these interactions underline the dual nature of T-5224's binding mechanism, engaging both hydrophilic and hydrophobic elements to stabilize its association with the Jun/CRE complex, potentially affecting its biological activity and regulatory functions. Importantly, we also highlight all the binding energy (Fig. 6c) for all T-5224 docked poses which ranged from − 10.2 kcal/mol (pose 2) to − 9.4 kcal/mol (pose 9) including their hydrophilic and hydrophobic interactions (Fig. S6).

Figure 6
figure 6

Illustration depicting the T-5224 docking with the Jun/CRE binding site. (a) Crystal Structure of the Jun/CRE Complex (PDB: IJNM). X-ray diffraction experimental data snapshot with 2.20 Å resolution, 0.286 free R-value, 0.228 work R-value, and 0.228 observed R-value. The surface region represents the Jun, and the cartoon with mesh represents the CRE. (b) Docked position of T-5224 with Jun/CRE Complex. The surface region represents the JUN and the mesh represents the CRE. (c) The binding energy of 9 different ligands poses with Jun/CRE Complex. Pose 1 was identified to have the lowest binding energy (− 11.6 kcal/mol) which was later subjected to molecular dynamics simulation. (d) Interaction of T-5224 pose 1 with Jun/CRE complex. Five hydrophilic interactions were observed between T-5224 and Jun/CRE Complex.

Full size image

RMSD is crucial in dynamics simulations as it measures the average distance between the atomic coordinates of a simulated structure and a reference structure, providing a quantitative assessment of structural accuracy and it serves as a key metric for evaluating the reliability of molecular dynamics simulations by indicating how well the simulated structure aligns with the expected conformation34. The all-atom explicit molecular dynamics simulation trajectory revealed stable dynamics during the 200 ns period (Fig. 7). The average RMSD difference of backbone atoms for Jun/CRE and in complex with T-5224 was observed to be ∼ 0.9 Å (Fig. 7a). Similarly, the average RMSD difference for Jun/CRE, both alone and in complex with T-5224 (Fig. 7b, Supplementary Movie 2), was observed to be ∼ 0.52 Å. The considerable disparity observed in the T-5224 bound complex stemmed from the alterations in Jun/CRE structural conformations occurring over the equilibration period, specifically characterized by the transformation of a helical structure into a loop within the Lys268 to Glu275 region and this may be attributed to the stable hydrogen bond formation with Arg270. Similarly, RMSF complements RMSD by highlighting the variability of atomic positions over time, offering insights into the flexibility and dynamic behavior of individual residues in a biomolecular system35. The structural conformational changes, therefore, were confirmed by observing the residual fluctuations, with the Jun/CRE and T-5224 complex exhibiting a larger fluctuation compared to the Jun/CRE apo form (Fig. 7c). The residues spanning from Lys268 to Glu275 displayed diminished fluctuations within the Jun/CRE apo form, yet manifested increased fluctuations, in the T-5224 bound complex, a consequence attributed to the helix-to-loop conversion. Furthermore, the solvent-accessible surface area (SASA) was analyzed to distinguish the protein solvent accessibility behavior. The Jun/CRE apo form area was stable throughout the simulation (~ 150 nm2) and the T-5224 bound complex showed a steady decrease in the surface area from ~ 270 nm2 to 220 nm2 during the 200 ns simulation (Fig. 7d). The radius of gyration in ligand–protein simulation is crucial as it assesses the compactness and structural stability of the complex, offering insights into the distribution of mass around the center and the overall conformational dynamics and flexibility of the ligand–protein interaction36. In the present study, analysis reveals that the JUN apo form exhibited consistent stability in its radius of gyration across 200 ns whereas the complexation of Jun/CRE with T-5224 demonstrated a gradual reduction in radius of gyration (Fig. 7e) indicating a sustained folding behavior throughout the simulation period. Moreover, T-5224 established a total of 8 hydrophilic bonds, 3 of which exhibited consistency across all simulation iterations (Fig. 7f), leading to the inference that the sustained formation of the stable complex primarily stemmed from enduring interactions of hydrophilic bonds.

Figure 7
figure 7

Parameters describing Jun/CRE-T-5224 complex structural stabilities. (a) Backbone atoms RMSD for Jun/CRE apo (green), in complex with T-5224 (red), and ligand T-5224 (blue), (b) complex atoms RMSD for Jun/CRE apo (green) and in complex with T-5224 (red), (c) RMSF Jun/CRE C-alpha RMSF apo (green) and in complex with T-5224 (red), (d) Protein SASA for Jun/CRE apo (green) and in complex with T-5224 (red), (e) Protein radius of gyration for Jun/CRE apo (green) and in complex with T-5224 (red), (f) Number of hydrogen bond interactions formed between Jun/CRE  and T-5224, (b) The first 50 eigenvectors were plotted versus eigenvalue for Jun/CRE apo (green) and T-5224 bound complex (red), (h) The collective motion of Jun/CRE apo (green) and in complex with T-5224 (red) using projections of molecular dynamics trajectories on two eigenvectors corresponding to the first two principal components.

Full size image

Likewise, PCA helps to identify the most significant motions and conformational changes in the ligand–protein complex and assess the dynamic interactions20,21. In the present study, the maximum collective motion of the complex was investigated using the first two principal components (PC1 and PC2) and captured by the first 50 eigenvectors. It was observed that the T-5224 – Jun/CRE complex exhibited higher conformational flexibility, i.e., − 6 nm to 6 nm (PC1) and − 5 nm to 5 nm (PC2), with a larger diversity of conformations (eigenvalue: 5.5 nm2) during the simulations (Fig. 7). However, the Jun/CRE apo form showed lower conformational flexibility, ranging from − 5 nm to 5 nm2 (PC1) and − 3 nm to 4 nm (PC2), with a smaller diversity of conformations (eigenvalue: 4 nm2) during simulation (Fig. 7). This evidence into complex evinces a superior degree of flexibility, manifesting through a spectrum of structural permutations that spans a broader landscape, juxtaposed with the Jun/CRE apo form comparatively subdued flexibility, marked by a discernibly diminished repertoire of sampled conformations. Furthermore, an analysis of the collective motion of the ligand-binding domains utilizing DCCM discerned that the amplitude of anti-correlation exhibited a notable elevation in the Jun/CRE apo form, whereas in the complex T-5224 – Jun/CRE, the degree of anti-correlation was considerably diminished (Fig. S5). Consequently, we posit that the  T-5224 binding would incline towards facilitating the conformational transition, primarily spanning from Lys268 to Glu275, thereby fostering the stable complex formation through augmented non-bonded interactions.

Furthermore, the molecular mechanics poisson-boltzmann surface area analysis identified the relative binding energy (− 448 ± 41.73 kJ/mol) of the T-5224 – Jun/CRE complex. Similarly, the estimated van der Waals, electrostatic, polar solvation, and SASA energies were -238.96 ± 22.81, − 444.39 ± 37.85, 258.28 ± 50.99, and − 23.69 ± 2.05 kJ/mol, respectively. These results demonstrate the stable T-5224-Jun/CRE-complex formation, as shown on different ligand-protein complexes in our previous study37. Importantly, the most energetically favorable structure (corresponding to global minima) of Jun/CRE apo and in complex with T-5224 (Fig. 6) was extracted from the free energy landscape. It was observed that the dynamics of the T-5224 binding was highly governed by the flexible dynamics of ligand-binding residues (Fig. 8). In the Jun/CRE apo form, the low energy conformation was discerned around 143 ns (coordinates near − 2.96 and 4.54) with a conspicuous absence of significant conformational alterations. Conversely, in the T-5224 bound form, the low energy conformation manifested around 112.11 ns (coordinates near − 25.63 and − 4.97) precipitating a substantial structural metamorphosis at the ligand-binding locale, concomitant with perceptible adjustments in the secondary structure, notably transitioning from helix to loop configuration.

Figure 8
figure 8

Free energy landscape (a) Jun/CRE apo form and (b) Jun/CRE-T-5224 complex. The dark blue region represents the coordinate with the lowest energy state (visualized using Mathematica (https://www.wolfram.com/mathematica/) 13.2.1). The snapshot for the Jun/CRE apo form was presented around 143 ns and, in the complex, around 112.11 ns.

Full size image

Prospective of T-5224 mediated JUN inhibition against GCL using experimental evidence

Although our current investigation delineates c-Jun as a promising druggable target and proposes T-5224 as a plausible candidate against GCL, it is imperative to underscore the necessity of corroborating this conjecture through an experimental approach. This imperative arises from the fact that the present findings stem exclusively from computational models, which, despite their intricate nature, are predicated upon approximations and presumptions that may not comprehensively encapsulate the intricacies of biological systems. Experimental validation in the wet lab milieu furnishes empirical substantiation to either corroborate or rebut these models, thereby ensuring congruence between predictions and empirical reality. Consequently, it is imperative to ascertain the feasibility of T-5224-induced c-Jun inhibition in GCL treatment. Given the constraints of our study owing to limited resources, our delineation has been confined to spotlighting the potential of T-5224-mediated c-Jun inhibition in GCL treatment through a computational lens. Nonetheless, we endeavor to underscore the viability of this hypothesis by advocating for the integration of both in vitro and in vivo studies.

The focal point of the cellular studies would entail an investigation into the effects of T-5224 on c-Jun inhibition within oligodendrocytes, Schwann cells, and assorted neuronal cell types (e.g. Neuro-2a (N2a) Cells, PC12, and SH-SY5Y), among others, with the overarching objective of elucidating the potential effect on cell viability, proliferation, myelination, and lysosomal function across these diverse cellular cohorts. It is imperative to ascertain the impact of T-5224 on Jun expression necessitating the assessment of T-5224's influence on mRNA levels through varying concentrations over distinct temporal intervals, juxtaposed with untreated cells serving as experimental controls. If notable variation exists in mRNA levels, it would be prudent to validate Jun protein expression after T-5224 treatment. Given the specific focus of prior research on the attenuation of Jun activity by T-522432 rather than its impact on expression per se, it is plausible that there may be minimal discernible alteration in mRNA levels or expression patterns. Notably, an advantageous strategy could entail augmenting Jun expression and activity via transfection, subsequently mitigating the resultant effects through T-5224 treatment, with implications spanning cell viability, proliferation, apoptosis, as well as myelination, encompassing lysosomal function and psychosine quantification. Examination of cell lines could yield pivotal insights into the molecular and cellular ramifications of T-5224-induced Jun inhibition in the GCL, elucidating potential mechanistic pathways underlying the therapeutic efficacy of T-5224 in mitigating disease pathology. The attainment of favorable outcomes from these investigations (Fig. 9a), should they materialize, could warrant further exploration in animal models as presented.

Figure 9
figure 9

Exploring T-5224 mediated c-Jun inhibition against Globoid cell leukodystrophy using experimental evidence. (a) The cellular studies can aim to investigate T-5224 effects on c-Jun activity and/or expression inhibition in various neuronal cells, focusing on cell viability, proliferation, myelination, and lysosomal function. The goal is to provide novel insights into T-5224 potential therapeutic mechanisms and justify further animal model studies. (b) The in vivo study may aim to evaluate T-5224 therapeutic potential in Twitcher mouse model with key goals to assess survival, and improvements in biochemical and histological markers. The study can use multiple doses of T-5224, assess motor functions, and analyze myelination, neuroinflammation, lysosomal function, and psychosine levels to determine treatment efficacy.

Full size image

If the aforementioned experiments prove successful, they could provide compelling grounds for progressing toward involving animal models (Fig. 9b). The principal objective of the in vivo investigation would be to evaluate the therapeutic efficacy of T-5224 in a murine model of GCL, aiming to ascertain its capacity to ameliorate motor function, prolong survival, and enhance various biochemical and histological indicators associated with the disease. The Twitcher mouse stands as a GCL model, distinguished by compromised GALC activity resulting psychosine accumulation. It would be crucial to select mice from early postnatal period (for example P10–P14) so that therapy can begin early in the course of the illness by dividing animals into different groups, including (a) Wild-type control group—non-Twitcher mice treated with the vehicle solution, (b) Vehicle group—Twitcher mice treated with the vehicle solution without T-5224, (c) Treatment group 1—Twitcher mice treated low dose of T-5224 and (d) Treatment group 2—Twitcher mice treated with high dose of T-5224. T-5224 can be administered p.o. as its oral administration has been shown to produce anti-arthritic activity in vivo38 (confirming the T-5224 blood brain barrier permeability is equally important) and assess the behavioral and motor functions periodically. In this study framework, use of rotarod and grip strength assessments emerges as a judicious strategy, given their capacity to comprehensively gauge motor coordination and balance, while concurrently quantifying forelimb and hindlimb strength. The administration of treatment protocols may persist until reaching either a predetermined experimental terminus or until the natural disease progression manifests. Decisions can be made from  nuanced observations of behavioral adaptations, alongside vigilant monitoring to ascertain the threshold of mortality. Such observations not only serve to elucidate the progression of the ailment but also furnish invaluable data to construct Kaplan–Meier survival curves, facilitating comparative analyses of lifespan across distinct experimental cohorts. At the culmination of the investigative process, brain, spinal cord, and peripheral nerves dissection is imperative to conduct a thorough evaluation encompassing myelination using myelin basic protein staining, and neuroinflammation, discernible through markers like glial fibrillary acidic protein for astrocytes and ionized calcium-binding adaptor molecule 1 for microglia. Furthermore, an indispensable aspect pertains to assessing the lysosomal function that necessitates the scrutiny of β-galactosidase activity within the harvested tissue homogenates. Of equal significance is the meticulous quantification of psychosine and other pertinent metabolites present in tissue homogenates through  suitable spectrometric methodologies, thereby augmenting comprehension regarding the impact of T-5224 intervention on GCL pathophysiology. The employment of suitable statistical analyses, including but not limited to Analysis of Variance and t-tests, is imperative to discern discernible trends and ascertain statistical significance in diverse experimental groups, while duly accounting for potential confounders through multiple comparison corrections. The culmination of this exhaustive animal model investigation (Fig. 9b) may hold promise in furnishing pivotal insights into both the therapeutic efficacy and mechanistic underpinnings of T-5224 as a prospective pharmacological intervention for GCL.

Conclusion

In conclusion, this comprehensive study delves into the intricate dynamics of the GCL-associated neuronal protein network  circuit, shedding on potential transformative shifts therein. Notably, c-Jun emerges as a pivotal figure in this network, orchestrating key signaling cascades crucial to the GCL progression. The elucidation of c-Jun centrality prompts an exploration into novel therapeutic avenues, with a keen focus on identifying small molecules capable of modulating its activity. Among these, T-5224 emerges as a promising candidate, displaying a propensity to form a robust and enduring bond with Jun/CRE complex, thereby disrupting its function. However, the complexity of the GCL protein network extends beyond Jun, encompassing different  interconnections among multiple proteins. While their roles remain largely uncharted in this study, their significance in disease warrants further investigation, particularly concerning their interactions with potential inhibitors. It is imperative to acknowledge the limitations of this study, primarily its reliance on computational methodologies, necessitating robust experimental validation to corroborate the findings and pave the way for translational applications in combating GCL-related pathologies.