Peptide–protein interactions are involved in various fundamental cellular functions, and their identification is crucial for designing efficacious peptide therapeutics. Drug–target interactions can be inferred by in silico prediction using bioinformatics and computational tools. We patented the Tn P family of synthetic cyclic peptides, which is in the preclinical stage of developmental studies for chronic inflammatory diseases such as multiple sclerosis. In an experimental autoimmune enceph-alomyelitis model, we found that Tn P controls neuroinflammation and prevents demyelination due to its capacity to cross the blood–brain barrier and to act in the central nervous system blocking the migration of inflammatory cells responsible for neuronal degeneration. Therefore, the identification of potential targets for Tn P is the objective of this research. In this study, we used bioinformatics and computational approaches, as well as bioactivity databases, to evaluate Tn P–target prediction for proteins that were not experimentally tested, specifically predicting the 3D structure of Tn P and its biochemical characteristics, Tn P–target protein binding and docking properties, and dynamics of Tn P competition for the protein/receptor complex interaction, construction of a network of con-nectivity and interactions between molecules as a result of Tn P blockade, and analysis of similarities with bioactive molecules. Based on our results, integrins were identified as important key proteins and considered responsible to regulate Tn P-governed pharmacological effects. This comprehensive in silico study will help to understand how Tn P induces its anti-inflammatory effects and will also facilitate the identification of possible side effects, as it shows its link with multiple biologically important targets in humans.
The Tn P family invention, currently patented in several countries, refers to synthetic cyclic peptides found in the venom of the Brazilian fish Thalassophryne nattereri. Tn P, in a preclinical development stage, is known for its therapeutic potential in chronic inflammatory diseases such as multiple sclerosis (MS). Proof of concept supporting its anti-inflammatory effect was found using the animal model of experimental autoimmune encephalomyelitis (EAE). The valuable potential of the patented Tn P family for controlling neuroinflammation and preventing demyelination is due to its systemic ability to interfere in the dynamic circuit of immune cell groups, as well as locally in the central nervous system (CNS).
We found that subcutaneous treatment with Tn P successfully improves the severity of clinical signs of myelin oligodendrocyte glycoprotein (MOG)-induced EAE, slowing down by 4 days the onset of maximal symptoms and decreasing by 40% the severity of symptoms, compared with control EAE mice treated with vehicle alone. Tn P amends EAE in an IL-10-dependent way, including suppressing activation of conventional dendritic cells (DC) and providing the emergence of plasmacytoid DC and regulatory cells during the EAE induction phase; blocking the transit and infiltration of leukocytes into the CNS by suppressing matrix metalloproteinase (MMP)-9 activity and CD18 expression; blocking the reactivation and permanence of Th (T helper)1 and Th17 lymphocytes in the CNS; preventing of microglial expansion and macrophage infiltration into the CNS; favoring the localized increase in regulatory T cells; and finally, suppressing demyelination in the spinal cord of EAE mice leading to accelerated remyelination dependent on blocking leukocyte migration in a cuprizone model.
The evidence that Tn P can be efficient in crossing the blood–brain barrier (BBB), which was described in zebrafish larvae, and can act locally in the CNS blocking the migration of inflammatory cells responsible for neuronal degeneration are features that lead us to hypothesize a potential ability of Tn P to interact with membrane receptors that control leukocyte traffic to the CNS. Therefore, the identification of potential targets for Tn P is fundamental at this stage of preclinical studies.
Drug–target interactions play key roles in drug discovery and development. The identification of drug targets is useful to know the drug’s mechanism of action but also to identify possible adverse side effects and plan its repositioning. The drug–target interactions can be inferred by in silico prediction of interactions between drugs and target proteins using bioinformatics and computational tools.
In this study, we used bioinformatics and computational approaches, as well as bioactivity databases, to evaluate Tn P–target prediction for proteins that were not experimentally tested, specifically predicting the 3D structure of Tn P and its biochemical characteristics, Tn P–target protein binding and docking properties, and dynamics of Tn P competition for the protein/receptor complex interaction, construction of a network of connectivity and interactions between molecules as a result of Tn P blockade, and analysis of similarity with bioactive molecules. Our results showed promising molecular targets of Tn P, including integrins with pharmacological action on leukocyte migration. The in silico model built to predict the possible molecular targets of Tn P showed a harmonic resonance between the anti-inflammatory effects and computational modeling.
More recently, cyclic peptides have gained significant attention for use as candidate therapeutics due to the high specificity and high affinities they can achieve against a wide range of targets. The anti-inflammatory activity of Tn P was elucidated in EAE, and its treatment efficacy was recently compared with current disease-modifying therapies, but the molecular targets of interest are still unknown and remain highly challenging.
This study was conducted with the aim to extend the knowledge of the mode of action of Tn P from the understanding of its interaction with protein targets. To start the Tn P–target prediction search, we redrew its primary linear structure (Ile–Pro–Arg–Cys–Arg–Lys–Met–Pro–Gly–Val–Lys–Met–Cys) with 13 L-amino acids using the simplified molecular-input line-entry system (SMILES) code according to Weininger to a 2D chemical structure. The Protein Data Bank (PDB) structure showed a bridge between Cys-4 and Cys-13 that gives Tn P a clip-like shape; the amino acids Ile-1, Pro-2, and Val-3 point toward the outside of the circular structure, which may contribute to van der Waals and hydrogen bonding interactions when binding to target.
Compared with linear peptides, Tn P is stable and less prone to protease degradation such as trypsin and pepsin. The low toxicity found in preclinical toxicology studies in zebrafish suggests transdermal absorption and a rapid in vivo clearance, which could indicate the need for a large number of such peptides. Furthermore, during the pharmacotechnical phase, the use of pharmacological and biotechnological alternatives, such as the design of delivery-directed systems, can be applied to improve the Tn P capacity for longer blood circulation.
Bioinformatics and computational approaches have been fostered as valid alternatives to experimental procedures for the prediction of absorption, distribution, metabolism, and excretion (ADME) parameters, especially during the initial stages of drug development. Figure 1C illustrates the radar of oral bioavailability of Tn P, where the pink-colored zone indicates the ideal physicochemical properties for the compound to perform a pharmacological activity in the human organism, while the red indicator marks the specific index of each property for the tested compound. We observed that red indicators of flexibility and polarity show Tn P as too flexible and polar with a high number of rotatable bonds and a topological polar surface area (TPSA) higher than 130 Å 2. The predicted value of the size (higher than 500 g/mL) corresponds to the actual molecular weight (MW) of 1514.8 Da. In contrast, the indicators of lipophilicity, solubility, and saturation for Tn P are in the pink zone, indicative of oral bioavailability. We showed earlier that the high solubility of Tn P allows it to be quickly absorbed by the subcutaneous route of administration.
Longer cyclic peptides can form secondary structures upon binding, offering low nanomolar values of binding interactions (hydrophobic, van der Waals, polar, and hydrogen bonding) that are able to access intracellular proteins involved in disease complexes. McAllister et al. analyzed cyclic peptides obtained by various technologies against a protein target (phage-display, mRNA-display, and split-intein circular ligation of peptides and proteins—SICLOPPS) between 2015 and 2019 in more than 40 publications, and interestingly, the MWs of the identified cyclic peptides cover a relatively wide range, from 443 to 2717 Da.
Figure 1D shows the prediction of BBB penetration (yellow area) and gastrointestinal absorption (white area) of Tn P using the BOILED-EGG method based on Log P and TPSA values. The analysis of Tn P pharmacokinetics revealed that neither the white area of gastrointestinal absorption nor the yellow area of BBB penetration was found in the gray zone, indicative of high gastrointestinal absorption and permeability through the BBB, which corroborates our in vivo data. An effective neurologic drug should be able to permeate the BBB so as to bind to specific receptors and initiate signaling pathways.
We searched for possible protein targets of Tn P using the SwissTargetPrediction, a web interface able to predict the targets of bioactive small molecules by combining 2D and 3D similarity measures. There are many types of drug targets, such as G protein-coupled receptors (GPCRs), protein kinases, enzymes, ion channels, and transporters.
In Figure 2A, we observe that the most prevalent top targets for Tn P interaction are enzymes as Eraser-associated proteins, which remove specific post-translational modifications (PTMs) from histone substrates with 32% and kinases (8%), and membrane receptor proteins (24%) and Family A of the GPCRs or rhodopsin family (20%). Enzymes account for 40% of the preferred targets of Tn P, while membrane receptors together represent 44% of predicted targets. These data support the finding that, beyond its capacity to inhibit serine proteases, Tn P can offer other intrinsic properties that contribute to the termination of the inflammatory process, including modulation of integrin expression, signal transduction, and tissue remodeling.
Then, we tried to identify the high-affinity bindings that occur between Tn P and the pockets of these target proteins in static modeled 3D structures by calculating the free energy of the binding complex score, which is often used to determine the affinity of biomolecular interactions and the efficacy of drugs. Figure 2B shows highly negative scores of free energy of binding, ranging from −215 to −200 kcal·mol−1, which signifies very strong binding to all targets, including integrins of beta subunit such as ITB1 and ITB7 or alpha subunit as alpha 4 (ITGA4 or ITA4) and alpha-IIb (ITA2B); furin, a type 1 membrane-bound protease member of the subtilisin-like proprotein convertase family; and two members of melanocortin receptor family (MCR)-4 and MCR-3), with central roles in weight regulation. Interestingly, we recently demonstrated that the improvement in the clinical score induced by Tn P treatment in EAE mice corresponded to an increase in body weight immediately after disease induction, consistent with its role in controlling weight regulation.
The root-mean-square deviation (RMSD) was performed to find the simulation result stabilities confirmed low scores within an acceptable range for all
