Occupying the intermediary size space between small molecules and large biologicals, cyclic peptides possess the structural compactness to intervene in sophisticated protein-protein interactions, many of which have been historically considered "undruggable". The inherent rigidity resulting from the constrained backbone macrocycle is believed to play a critical role in affinity maturation of cyclic peptides, though it also places additional demands on chemistry and formulation efforts for pharmacokinetic optimization. The latter is particularly complicated by the typically high polarity and consequent susceptibility to proteolytic degradation of cyclic peptides which must be reconciled with membrane permeability and metabolic stability requirements for in vivo use. Methods to alter the physicochemical properties and the pharmacokinetic profiles of cyclic peptides can be based on different strategies, such as backbone engineering, side chain modification or attachment to other carrier moieties in a covalent manner. Modification of the peptide backbone by the introduction of N-methylation or non-proteinogenic residues, or through the attachment to lipid or polyethylene glycol linkers, have been shown to modulate physicochemical properties and pharmacokinetic profile.
As with all peptides, cyclic peptides have challenges associated with their ADME and pharmacokinetics which include low oral bioavailability and rapid clearance, however in cyclic peptides this is also exaggerated by their intermediate size which may lead to poor absorption from the gastrointestinal tract. Resistance to degradation by peptidases is less of an issue than in linear peptides due to their cyclization which protects the N-terminus and C-terminus from degradation by aminopeptidases and carboxypeptidases, but there is still vulnerability to endopeptidases and chemical degradation. Achieving high affinity with good drug properties is also challenging, as strong binding to a target can be achieved by making many hydrogen bonds and having large polar surfaces, but both of these can also reduce passive diffusion. Additionally the cyclic peptide may also be cleared rapidly by renal filtration and/or hepatobiliary clearance pathways, thus further reducing the effective circulating half-life. Design and optimization strategies often try to mitigate as many of these pharmacokinetic issues as possible.
It has been widely reported that cyclic peptides are proteolytically more stable than their linear analogues. Peptide bonds present in termini of linear peptides are accessible to proteases while in cyclic peptides, they are usually not. Thus, cyclic peptides are less susceptible to enzymatic degradation and therefore have longer half-life. Additionally, since many cyclic peptides are composed of unnatural amino acids, one could even argue that it is the unnatural structural properties of these cyclic peptides, rather than just their cyclic structure that is responsible for the proteolytic stability. As with any peptide, cyclic peptides can be unstable not only from a proteolytic standpoint, but can also be prone to chemical and thermal degradation. The most common chemical degradations in cyclic peptides are deamidation, oxidation, and hydrolysis. The chemical and thermal stability of cyclic peptides depends on several factors such as the type of buffer used and pH levels together with the addition of antioxidants and stabilizers in formulations.
Peptide cyclization is one of the most effective ways to protect a peptide drug from enzymatic degradation. This is because the macrocyclization generally prevents enzymatic access to peptide bonds. Peptide drug candidates with free termini are especially vulnerable to enzymatic degradation from aminopeptidases, carboxypeptidases, and, in the case of oral drugs, the various enzymes in the gastrointestinal tract. By removing these termini, peptide cyclization can help to improve the overall survival time of peptide drugs in the bloodstream and gastrointestinal tract. Peptides, even cyclic peptides, can be attacked by endopeptidases. The introduction of cyclical formation within peptides creates a strong 3D structure that limits flexibility and reduces endopeptidase accessibility to cleavage sites when the resulting macrocycle structure maintains rigidity. The use of unnatural amino acids such as D-amino acids and N-methyl amino acids can also contribute to increased enzymatic stability due to the steric bulk and altered hydrogen bonding behavior. For example, use of D-amino acids in the peptide sequence will prevent recognition by proteases, as well as decrease conformational flexibility of the peptide. Uncommon amino acids such as β-amino acids or aza-amino acids can also confer increased enzymatic stability.
Formulation and storage conditions such as pH, oxygen presence and temperature determine the degradation mechanisms impacting cyclic peptides.), including pH-dependent chemical reactions, oxidation, and temperature-induced conformational changes leading to aggregation or backbone degradation. In low pH conditions, aspartimide formation can occur in sequences with Asp-Gly residues, which can cause ring opening or isomerization, while methionine and cysteine residues can be prone to oxidation of their side chains in the presence of peroxides. Tryptophan residues are sensitive to photo-oxidation if exposed to light during processing or storage. Temperature can affect both chemical degradation reactions and non-covalent aggregation pathways, such as hydrophobic collapse of partially unfolded peptide species. Buffer systems (e.g. phosphate, citrate, acetate) can also have an impact on stability, for example, phosphate buffers can potentially promote metal-catalyzed oxidation. Excipients such as antioxidants, chelating agents and cryoprotectants may be used to stabilize cyclic peptides during formulation and storage, although the utility of such excipients is sequence- and route-of-administration-dependent. Lyophilization can offer improved long-term stability relative to liquid formulations, but the freeze-thaw process may also introduce structural changes that can be controlled only by optimization of cooling and annealing rates. Stability-indicating assays will need to be developed early in the process to detect low-level degradants and to establish reliable shelf-life data.
Absorption and bioavailability of cyclic peptides in vivo are difficult as a compound must have very specific physicochemical properties in order to be a molecule with sufficient in vivo exposure. Cyclic peptides are, by nature, more polar and less lipophilic than small molecule drugs and are, in general, less likely to passively diffuse across cellular membranes and also are too large to passively diffuse through. This means many small cyclic peptides have zero or very low oral bioavailability and are, when given parenterally, typically rapidly cleared by the kidneys. Properties that affect absorption and bioavailability are strongly correlated to the physicochemical properties of the macrocycle. The physicochemical properties are highly dependent on size, polarity and lipophilicity. Polarity is a measure of the ratio of polar (hydrophilic) and non-polar (hydrophobic) groups in a molecule. In peptides, the peptide bond supplies the hydrogen bond donors and acceptors that contribute to a large fraction of the total surface area of the macrocycle. In general, it is energetically unfavorable for a polar molecule with a large number of hydrogen bond donors and acceptors to go from the aqueous environment into the non-polar membrane, and the translocation of a polar cyclic peptide across the membrane epithelium is a rate-limiting step. Polarity of peptides is also highly impacted by intramolecular hydrogen bonding patterns in a macrocycle that are sometimes required for bioactive conformations. Molecular weight is another highly important component in bioavailability: while the ring size of the macrocycle often exceeds the size that is required to be passively diffused transcellularly, it also stays under the molecular weight threshold that would typically be filtered by the kidneys.
The membrane permeability of cyclic peptides is strongly related to the ring size, conformational freedom, and the distribution of hydrophobic and hydrophilic groups on the macrocycle. For larger rings, conformational flexibility can increase adaptability to membrane environment, but the entropy of desolvation and penetration is also higher. On the other hand, smaller, more constrained cyclic peptides can become diffusively permeable by intramolecular hydrogen bonds that protect the backbone polar amides from the solvent, which decreases the energy barrier for translocation. One of the most effective approaches to increase diffusive permeability is through N-methylation. As each methyl group removes a hydrogen-bond donor on the backbone, this reduces the polar surface area. The creation of solvent-excluded volumes by breaking the hydrogen bonding of the peptide backbone can promote permeability, although too much N-methylation of cyclic peptides can negatively affect their aqueous solubility and binding affinity. The self-segregation of hydrophobic and polar groups into separate faces of the macrocycle produces amphipathic structures that self-promote permeation, allowing the hydrophobic face of the peptide to interact with the lipid bilayer, while the polar face can remain hydrated. The ability to predict this permeability with computation has recently been demonstrated by mapping three-dimensional conformational ensembles to transcellular flux, without the need for active transporters or perturbing penetration enhancers.
Fig. 1 Comparison of cyclic peptide therapeutics to classical small molecules and biologics.1,2
Peptides with cyclic topology are usually only available as injectables, since oral bioavailability is challenging due to intestinal absorption issues (flux limited by poor permeability, P-gp efflux and intestinal metabolism) and chemical stability (acid-catalyzed degradation in stomach). Because oral dosage forms are convenient and safe, much effort has been directed toward the development of oral formulation and chemical modifications to achieve oral activity for cyclic peptides. Delivery approaches include the use of self-emulsifying formulations, the use of permeation enhancers (such as those transiently opening tight junctions), or the use of chemical modifications to improve peptide absorption (such as incorporation of N-methylation, use of D-amino acids, or development of hydrogen-bond surrogate backbones). For example, modifications to the structure of cyclosporine have shown that high N-methylation of the peptide backbone in combination with a small lipophilic macrocycle can lead to orally active analogs, although there can be major food effects and PK variability with such agents. New formulation technologies such as those based on nanoparticulate carriers, which offer enzymatic protection and promote transcellular transport, have shown preclinical promise for oral dosing without modification of the pharmacophore.
Binding of cyclic peptides to plasma proteins (albumin) can affect their distribution and clearance profiles. In general, plasma protein binding is associated with reduced renal clearance due to the exclusion of the bound drug from glomerular filtration. Plasma protein binding may also restrict the tissue distribution of cyclic peptides. Plasma protein binding can be affected by physicochemical properties of the cyclic peptide (e.g., hydrophobicity and charge), but can also be engineered through the incorporation of specific protein binding domains or moieties. However, plasma protein binding can also decrease the free fraction of a cyclic peptide in plasma and thus limit its therapeutic effect. Distribution to tissues is another key factor that will determine the pharmacokinetic profile of cyclic peptides and can also be affected by their physicochemical properties (e.g., size, charge and lipophilicity). The intermediate size of cyclic peptides may also limit their extravasation from the blood vessels, while the rigidity of cyclic peptides can also impact their cell membrane penetration. The charge of the cyclic peptide can also impact distribution, with cationic peptides often showing a higher affinity for negatively charged cell surfaces. Targeted delivery of cyclic peptides through the conjugation of targeting ligands is one approach to enhance tissue penetration. This involves the conjugation of a targeting moiety to the cyclic peptide, which can selectively bind to a desired tissue or cell type. Common targeting moieties include homing peptides, carbohydrates, and small molecule ligands. The resulting conjugate can then be recognized by specific receptors on the target tissue, which can mediate uptake and retention of the cyclic peptide in the tissue.
Cyclic peptides have two primary metabolic pathways, hepatic metabolism and renal excretion. These two primary mechanisms for metabolic processing of cyclic peptides provide opportunities for design challenges. In hepatic metabolism, cytochrome P450 monooxygenases are usually responsible for oxidation reactions, which can be followed by conjugations. If a compound is able to passively absorb into the liver, it will most likely also be metabolized. Cyclic peptides are more resistant to peptidases than linear peptides. However, depending on the location and the amino acid, this resistance may not be enough to avoid metabolism and certain residues in a cyclic peptide can be modified. If oxidation reactions occur, the cyclic peptide itself can be oxidized. This will also result in an inactive form. Renal excretion can occur with glomerular filtration, which typically restricts circulating compounds of a molecular weight of less than 20 kDa. As a result, small cyclic peptides are more likely to be rapidly excreted in the urine. In addition, the many proteases found on the brush border membrane can cleave filtered peptides. As mentioned, cyclization can increase the stability of a peptide. PEGylation prevents renal filtration due to its increase in molecular weight by adding the bulky PEG groups. It also sterically hinders active sites on enzymes. PEGylation creates a hydration shell around the peptide, which increases the peptide's circulation half-life. However, if the PEG group is conjugated to a site near the pharmacophore of the peptide, this process could reduce the peptide's binding affinity. The primary benefit of lipidation is to reversibly attach a small molecule to a protein serum albumin. This forms a depot in circulation, which slowly releases the active drug. Fatty acids can also increase a peptide's lipophilicity, thus increasing its ability to partition into membranes. In addition, fatty acid groups can increase uptake through the lymphatic system following a subcutaneous injection. The two modifications must also be optimized for linker length and conjugation chemistry to avoid immunogenicity and loss of activity. In most cases, the decision between PEGylation and lipidation is based on the indication and desired route of administration.
Optimization of the PK properties of cyclic peptides has been influenced by two main approaches: rational chemical modification and the development of computer-aided methods to predict ADME/Tox properties of drug candidates. As previously mentioned, several properties need to be optimized at the same time to increase the likelihood that a cyclic peptide will be drug-like. However, it is well known that there are tradeoffs, or negative correlations, between several of these properties. For example, it has been observed that modifications to increase permeability can decrease affinity for the target, and changes that improve metabolic stability can lead to a decrease in water solubility. As a result, several different modification strategies, along with an accompanying workflow, have been developed to systematically improve PK properties of cyclic peptides, taking into account possible tradeoffs. First, chemical modifications based on structure-activity relationships are made in order to address some of the most common problems in cyclic peptide optimization: susceptibility to proteolytic degradation and lack of membrane permeability. Next, models are used to prioritize which analogs are likely to have better ADME properties. By doing so, one can focus on a limited set of analogs to test in the lab rather than testing all possible chemical modifications. Model predictions, as well as experimental PK data from these analogs, can then be used to make further chemical modifications, thus continuing the cycle. The next two sections will discuss in more detail the chemical approaches to tuning peptide properties as well as computational methods to streamline the optimization process.
Chemical modifications play a crucial role in optimizing cyclic peptides for pharmacological applications. Backbone rigidification is a common strategy that involves introducing conformational constraints to reduce the flexibility of the peptide backbone. This can be achieved through various means, such as cyclization, incorporation of proline residues, or the use of non-natural amino acids. Backbone rigidification can enhance stability and bioavailability by increasing resistance to proteolytic degradation and improving target binding affinity. Charge tuning is another important modification that involves optimizing the distribution of charges within the peptide to enhance its interaction with the target protein. This can be achieved by incorporating charged amino acids or modifying existing residues to alter the overall charge of the peptide. Charge tuning can improve solubility, binding affinity, and minimize non-specific interactions. Linkers can also significantly affect the pharmacokinetic properties of cyclic peptides. Linkers can be used to connect different peptide segments or to attach targeting ligands, thereby increasing stability and directing the cyclic peptide to specific tissues. For example, PEGylation involves the conjugation of polyethylene glycol (PEG) polymers to cyclic peptides to improve their half-life and reduce immunogenicity. Lipidation, on the other hand, involves conjugating lipid moieties to cyclic peptides to enhance their membrane permeability and stability in biological environments.
Machine learning (ML) can be used to predict the PK properties of cyclic peptides, such as permeability and half-life. ML models can be trained on large datasets of known cyclic peptides and their corresponding PK profiles to identify patterns and relationships between peptide structure and PK properties. These models can then be used to predict the PK properties of new cyclic peptide candidates, allowing researchers to prioritize compounds with favorable profiles for further development. Additionally, computational tools can be used to predict the binding affinity and specificity of cyclic peptides to their targets, providing insights into their potential efficacy. By integrating computational PK prediction with experimental validation, researchers can optimize the design of cyclic peptides to achieve the desired balance between potency and drug-like properties.
In terms of the P450 system, the pharmacokinetics of cyclosporin A, a natural macrocycle, and lanreotide, a synthetic peptide macrocycle, present a clear contrast as a result of the optimization for different clinical contexts and administration routes. As a macrolide with oral availability, cyclosporin A demonstrates a unique and synergistic structure-activity relationship (SAR) that includes multiple N-methylation of backbone amides, the use of non-standard amino acids, and a significant conformational plasticity, allowing for the molecule to exist both in a lipid-compatible and in a fully solvated form. As a result, despite a relatively high molecular weight, it is able to permeate the intestinal epithelium by passive diffusion and is relatively well-protected against proteolytic degradation in the gastrointestinal tract by the high degree of methylation. The lipophilicity of the molecule allows for tissue distribution and cellular uptake, including of immune cells, in order to elicit its immunosuppressive effect, primarily via calcineurin inhibition. In contrast, lanreotide, as a somatostatin analog, is a synthetic peptide macrocycle intended for parenteral use that has been optimized for sustained release and selective activity on neuroendocrine tumors. As a result, it has a macrocyclic disulfide-constrained structure, which has been shown to have a high receptor affinity, as well as greater metabolic stability than the native peptide, but has not been designed to enhance permeability for oral administration. It is formulated as an injectable depot delivery system which provides an extended duration of action, with limited hepatic metabolism and primary renal excretion of intact drug. In this case, the former route and concomitant need for oral absorption has been directly targeted by a series of chemical modifications, while for lanreotide, the decision to administer the drug via injection allowed for the minimization of steps that would have lowered the selectivity for the target receptors and increased the risk of gastrointestinal degradation.
An extensive review of the optimization strategies for cyclic peptide drugs uncovers a series of both chemical and computational approaches. Chemical modifications such as backbone rigidification, charge tuning, and linker addition have been found to significantly improve the stability, bioavailability, and target specificity of cyclic peptides. On the other hand, computational PK prediction using machine learning has emerged as a valuable tool for predicting the pharmacokinetic properties of cyclic peptides, allowing for efficient identification of candidates with desirable drug-like properties. Case studies such as cyclosporin A and lanreotide further illustrate the real-world impact of these optimization strategies on the clinical application of cyclic peptides. Future trends in the field may include the development of oral cyclic peptides and the integration of AI-guided ADME (absorption, distribution, metabolism, and excretion) design.
Achieving optimal pharmacokinetics is one of the biggest challenges in cyclic peptide drug development. Our team specializes in engineering macrocycles with improved stability, permeability, half-life, and overall drug-like performance.
We provide:
From early optimization to preclinical preparation, we help your candidates advance with confidence.
References
