Peptide-Drug Conjugates: An Emerging Direction for the Next Generation of Peptide Therapeutics

Abstract

Building on recent advances in peptide science, medicinal chemists have developed a hybrid class of bioconjugates, called peptide-drug conjugates, that demonstrate improved efficacy compared to peptides and small molecules independently. In this perspective, we discuss how the conjugation of synergistic peptides and small molecules can be used to overcome complex disease states and resistance mechanisms that have eluded contemporary therapies because of their multi-component activity. We highlight how peptide-drug conjugates display a similar multi-factor therapeutic mechanism to antibody-drug conjugates but also demonstrate improved therapeutic properties such as less-severe off-target effects and conjugation strategies with greater site-specificity. The many considerations that go into peptide-drug conjugate design and optimization, such as peptide/small-molecule pairing and chemo-selective chemistries, are discussed. We also examine several peptide-drug conjugate series that demonstrate notable activity towards complex disease states such as neurodegenerative disorders, and inflammation, as well as viral and bacterial targets with established resistance mechanisms.

Graphical Abstract

1. INTRODUCTION

The evolution of peptide therapeutics and their clinical applications.

Developments in protein chemistry have been instrumental in the identification of several protein- and peptide-based therapeutics that have entered the small molecule-dominated drug market. Protein-based therapeutics represent attractive medicinal chemistry leads, as these macromolecules have been evolutionarily optimized for activity under physiological conditions and possess inherently favorable immunogenic and metabolic properties. These attractive aspects are demonstrated by the sizable increase in approved biologics that target a broad range of disease states, including cancer and metabolic disorders.¹ Despite their successes, there are still some limitations of therapeutic proteins, including the need for intravenous, intramuscular, or subcutaneous injections. Because of the recent advances made to develop orally bioavailable peptide therapeutics, there is renewed interest in peptides as a therapeutic modality.^2,3 The aim of these peptide-based strategies is to develop a therapeutic with improved drug-like properties that also retains the desired biological activity of the parent protein.

First-generation peptide drugs such as insulin and adrenocorticotropic hormone are examples of endogenous hormones that found success as drugs when applied exogenously.⁴ In this era of drug discovery, peptides with established physiological roles were isolated from natural sources, purified, and given to patients. Hormone-based peptides were clinically successful in the mid- to late 1900s; however, few examples existed outside this class. During this time, it became apparent that poor peptide plasma half-lives and bioavailability were major hurdles to their clinical application.⁵ Natural peptides lack the intramolecular interactions necessary to replicate the higher-order structure of their parent proteins; thus, they circulate in the bloodstream as linear poly-peptide chains. These linearized peptides are efficiently digested by peptidases because they are freely recognized by enzymatic active sites. Additionally, peptides are also efficiently excreted by the same homeostatic mechanisms that regulate plasma-hormone concentrations.

The development of peptide synthesis strategies and technologies contributed greatly to the next generation of peptide therapeutics that address the major shortcomings. The introduction of automated peptide synthesizers has enabled a higher-throughput strategy for the development of highly effective and selective peptide analogs. Using this technology, peptide chemists have found that the incorporation of unnatural amino acids has had a significant effect on increasing the plasma half-life of peptides.⁵ Modifications of the peptide backbone via incorporation of amide bioisosteres yielded similar results.⁶ These two strategies have been leveraged by peptide chemists to confer secondary structure to peptides through additional non-native covalent bonds, effectively locking them into the desired conformation.^7,8 These chemically modified peptides display improved half-lives because they are less susceptible to recognition and degradation by proteolytic enzymes.

Another improvement to peptide properties is through the development of chemo-selective chemistries that permit labeling of specific amino acids. By leveraging the inherent nucleophilicity of some amino acids, particularly cysteine,⁹ chemists have designed peptide conjugates with improved solubilities and selectivities. The incorporation of unnatural amino acids that are amenable to click chemistry is also exploited for the generation of peptide conjugates.^10,11 These strategies are even more powerful when coupled with peptide synthesis as they enable site-specific labeling of peptides, which is critical for the retention of biological activity. The potential for peptide bioconjugates is exponential and includes the incorporation of biomolecules, such as fatty acids and sugars,^12,13 as well as compounds such as polyethylene glycols (PEGs).¹⁴ Peptide conjugates with increased molecular weight have also been shown to be resistant to clearance, thus staying in circulation longer.¹⁵ These chemically modified peptides have found great clinical success in the past decade and are considered the second-generation of peptide therapeutics.

The improved stability and solubility of chemically modified peptides have paved the way for novel therapeutic strategies. One therapeutic avenue that has demonstrated considerable success is the parallel application of synergistic small molecule drugs and peptides to combat complex disease states.^16,17 This strategy has led to the development of conjugates where small molecule drugs are linked to peptides, overcoming the challenge of co-administration of different compounds. Peptide-drug conjugates (PDCs) are an emergent class of therapeutics that have already found success in clinical trials (e.g. NCT03613181, NCT03486730, and NCT01767155), indicating that they will be a part of the next-generation of peptide therapeutics. In this Perspective, we will compare PDCs to other well documented bioconjugate therapies, discuss their design and optimization, as well as highlight some exciting examples in the pipeline. It is our hope that this review will give the reader insight into the current state of the field and provide inspiration for the design of their own peptide-drug conjugates.

Peptide-drug conjugates as an alternative to antibody-drug conjugates.

Before embarking on a discussion of PDCs, it is important to consider another well-studied example of bioconjugates: antibody-drug conjugates (ADCs). This class of compounds combines the highly specific target recognition activity of antibodies with the desired therapeutic effect of a small-molecule partner. These conjugates have received significant attention in the realm of cancer therapeutics where antibodies that recognize cancer-specific antigens are coupled to a cytotoxic payload.^18,19 In the case of anti-cancer ADCs, the antibody confers target-specificity to the cytotoxic small molecule with the goal of limiting off-target effects and improving the overall efficacy of non-specific cancer therapies. To date there are more than 10 ADCs approved by the U.S Food and Drug Administration (FDA), with dozens of pharmaceutical companies having disclosed the development of additional ADC therapies.¹⁹ The success of ADCs might suggest to the reader that the development of alternative bioconjugate therapeutics is redundant; however, ADC-based therapies come with a range of limitations. In the following section we will discuss some pitfalls of traditional ADC therapeutics and highlight ways that PDCs can overcome these challenges.

One of the major advantages of ADCs over traditional small-molecule therapeutics is their capacity for targeted delivery owing to antibodies that bind antigens with high affinity and specificity. This targeted delivery of ADCs represents a significant therapeutic advancement; however, the large size and molecular weight of antibodies limits the application of ADCs. As a result of these physical properties, antibodies are unable to freely diffuse through the plasma membrane with few (if any) membrane proteins capable of transporting these macromolecules into the intracellular space; therefore, ADCs are restricted to antibodies that recognize proteins that present an antigen to the extracellular space.²⁰ This therapeutic strategy requires ADCs to be endocytosed for effective delivery of the small-molecule payload to elicit the desired therapeutic effect.^21,22 In contrast, peptides are much smaller than antibodies and several membrane transport proteins exist that traffic peptides into the intracellular space;²³ thus, peptides that recognize intracellular proteins can be used in PDC design, which greatly increases the number of target proteins accessible to PDCs. PDCs can be applied to a broader range of protein targets, although peptides generally recognize binding partners with a much weaker affinity than antibodies.

The highly specific antigen recognition of ADCs has encouraged the development of bioconjugates that possess cytotoxic payloads. This therapeutic strategy is centered on the idea that ADCs elicit target-specific cell death with greatly reduced off-target cytotoxicity. These bioconjugates are particularly appealing in the area of cancer therapeutics because an antibody that recognizes a cancer-specific antigen can be leveraged for targeted killing of a cancerous mass, but the antigen-specific binding of ADCs also presents challenges as it requires that a disease-related antigen is almost exclusively present on diseased cells;²² thus, the identification of a disease-related antigen and the subsequent generation of an antibody capable of recognizing the antigen are significant components of the ADC-development process²⁴. If these conditions are not met–that is to say, if an antibody displays even slight promiscuity or the selected antigen is present on healthy cells–off-target cytotoxicity can have serious consequences for a patient. Notably, these challenges are less significant in the PDC space because peptides can be conjugated to a variety of non-cytotoxic small molecules to elicit a desired therapeutic outcome. A famous example of anti-cancer PDCs are the proteolysis targeting chimeras (PROTACs) that combine a peptide that recognizes a target protein and a small molecule that sequesters an E3 ubiquitin ligase.²⁵ These hybrid PROTACs hijack the ubiquitin-proteasome system, which results in the targeted degradation of a protein.²⁵ Compared to the off-target effects of ADCs, the directed degradation of a protein target in a healthy cell is much less severe. Unlike cytotoxic ADCs that rely upon cell death as their therapeutic mechanism, PDCs can be tailored for a wider range of therapies with less severe consequences.

Another limitation of ADC development is the narrow range of conjugation chemistries that can be leveraged to join an antibody to a small-molecule payload. As previously mentioned, peptide chemists have developed a toolbox of chemo-selective chemistries that permit labeling of specific amino acids.²⁶ These powerful conjugation reactions allow for the labeling of complex biological material beyond peptides; however, site-specific conjugation of full-length proteins is particularly challenging, as the incidence of a reactive amino acid in a large protein sequence is usually greater than one. For this reason ADCs are commonly decorated with several small molecules, and ADC samples are heterogeneous as it is impossible to control the antibody-to-payload ratio.^27,28 The variability in antibody to payload ratio, as well as the distribution of the small molecule payload(s), can have significant consequences for ADC therapeutics because they impact factors such as activity and pharmacokinetic properties.^27,29 These obstacles are less of a concern with peptide-based conjugates because of their smaller size, as well as the speed and ease of peptide synthesis using contemporary equipment. Manual peptide synthesis has enabled peptide chemists to incorporate unnatural amino acids into peptides which can be used for site-specific labeling.^10,11 Additionally, chemo-selective labeling of peptides is more easily tolerated as unwanted reactive amino acids can be modified during peptide synthesis to avoid unwanted labeling while also retaining the physicochemical properties of the native peptide.

Despite the limitations outlined in this section, ADCs are powerful therapeutic tools that mark a significant improvement over traditional chemotherapies. In addition, the consortium of researchers involved in ADC development have invested considerable time and resources into overcoming these challenges with substantial success;^30–32 however, these solutions are outside the scope of this review. This section highlighted ways that PDC development differs from ADCs and how PDCs can overcome some of the challenges associated with working with large biologics. In the section that follows, we will outline some of the considerations that go into PDC design and optimization.

2. PEPTIDE-DRUG CONJUGATE DESIGN AND OPTIMIZATION

A PDC is a class of targeted therapeutics composed of a peptide and a small molecule drug joined through a covalent linkage. Although the name is unambiguous, it provides little information about the complex considerations that go into PDC development. Designing PDCs is rarely straightforward as researchers must consider the effects and activities of several small molecules and peptides, as well as their limitations. One must also reflect upon the architecture of both the peptide and drug to determine where to join the independent molecules, in addition to the chemistry necessary to form the covalent linkage.

Designing a PDC development campaign begins with the selection of a peptide and small molecule with complementary activities. The identification of small molecules and/or peptides with demonstrated activity towards a particular target or disease state is a logical starting point. The merger of a small molecule and peptide with parallel targets and activities may lead to an ultra-potent PDC; however, conjugation may also be used to redeem once-promising molecules that were ultimately considered undrug-like. For example, a small molecule that exhibits potent anti-cancer activity but demonstrates non-specific cytotoxicity would be a strong candidate for conjugation to peptides that specifically interact with cancer markers to generate a PDC for site-specific delivery. There are countless cases of small molecules identified in high-throughput screens or through iterative rounds of structure-activity relationship (SAR) studies that display a desired activity but fall short of entering clinical trials due to unrelated imperfections. These molecules are strong candidates for adoption into PDC campaigns because their shortcomings may be addressed by a peptide partner. Peptides can be used to combat a range of undesirable effects, from inefficient cell penetration³³ to high promiscuity and off-target effects.³⁴ Consequently, a thorough characterization of a small molecule candidate is necessary for selecting the most appropriate peptide for PDC development. This strategy may be applied in reverse, where a peptide is chosen as a starting point for PDC design, and small molecules are selected to enhance activity or overcome limitations of the selected peptide.

Rudimentary PDCs consists of a small molecule covalently bound to a peptide without the introduction of a linker. These PDCs are most effective when the small molecule and peptide act on the same target, and the separate binding sites are near one another; however, linker-less PDC strategies come with limitations as they are structurally rigid, a property that can significantly restrict the peptide or drug from accessing its target in a preferred orientation. The short distance between the peptide and small molecule is another shortcoming of bipartite PDCs as the proximity of these two components may result in poor target binding due to steric hindrance. For these reasons, tripartite PDCs composed of a peptide, drug and chemical linker are preferred as they allow for a greater degree of flexibility and decrease the likelihood of steric hindrance. Chemical linkers range widely in length, chemical composition, and rotational freedom based on their identity and level of unsaturation. Linker selection is another important consideration in bioconjugate design, and we encourage the reader to consult the excellent reviews that exist on this topic.^35,36

Potentially the most important factor in PDC design is determining the optimal sites and chemistries for conjugation. Ideally, structural information about small molecule and/or peptide binding would be available to determine the preferred molecular orientation; however, this is not always the case. Unfortunately, it is impossible to predict optimal conditions a priori; thus, trial and error is often necessary. Given the inherent stability of the peptide backbone and the relatively limited reactive functionality of amino acid side chains, identifying the peptide’s reactive centers is a good starting point for designing conjugation strategies. Bioconjugation reactions developed for cysteine thiols, as well as lysine and N-terminal free amines, are highly robust and have been optimized over the course of decades.²⁶ Newer bioconjugation chemistries have also been developed for amino acids with less nucleophilic side-chains, such as tyrosine.^26,37 The restricted reactivity of amino acids was once a major limitation for site-specific labeling; however, the incorporation of unnatural amino acids amenable to bioorthogonal chemistry has allowed researchers to precisely modify peptides while retaining many of their native properties.¹⁰ Once a conjugation site has been identified, the chemistry required to modify a particular amino acid can guide small molecule or linker derivatization. SAR data establishing the small molecule pharmacophore should also be used to determine the optimal sites of conjugation.

3. APPLICATIONS OF PEPTIDE-DRUG CONJUGATES

In the sections that follow, we will outline examples of PDCs designed to treat viral infections, bacterial infections, and inflammation, as well as compounds developed for CNS targets. Of note, cancer therapeutics are another rich area of PDC development; however, we will not touch on these compounds as several excellent reviews already exist on this topic.^38,39 These sections are not meant to be a comprehensive list of PDCs; rather, our aim is to showcase these projects as case studies.

4. ANTI-VIRAL PEPTIDE-DRUG CONJUGATES

The devastating outcomes of the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV2) pandemic has highlighted the importance of antiviral therapeutics. These outcomes are compounded by the spread of vaccine-resistant SARS-CoV2 variants, which demonstrated the need for mutant-agnostic anti-viral drugs to complement traditional vaccine campaigns; however, the significant mutation rate of viruses is also a concern for antiviral drug development as drug-resistant viruses emerge at a rapid rate.⁴⁰ Anti-viral PDCs are an attractive strategy for overcoming this challenge as peptides with a high affinity for viral proteins can be used to deliver drugs to their target. In the case of drug-resistant viruses, high-affinity peptides bring the anti-viral payload near the target protein, increasing the local concentration of the antiviral compound. This apparent increase in concentration may be sufficient to overcome the decreased binding-affinity brought on by mutation.

Notably, analyses of protein structures from genetically distinct viruses revealed significant structural conservation despite a high degree of sequence deviation.^41,42 These results suggest that these proteins share a common mechanism of action; thus, antiviral strategies may be repurposed to inhibit homologous proteins across different viral families. This idea is particularly interesting when considering PDCs as highly specific peptides can be designed to tailor therapies for particular viruses. Here, we will discuss PDC strategies that have been developed to treat infections from human immunodeficiency virus (HIV) and Dengue virus (DENV), two viruses that lack effective, mutation-resistant therapies.

Human Immunodeficiency Virus entry-inhibiting PDCs.

HIV infections reached epidemic proportions in 1981 and have claimed millions of lives to date. Several antiretroviral therapies with unique HIV targets have been developed to treat infections⁴³; however, resistance to these therapies develops at a significant rate. An important target for HIV therapies is the envelope protein, gp160, which facilitates virus-cell fusion. Briefly, gp160 on the viral membrane recognizes a receptor protein, CD4, and a coreceptor (CXCR4 or CCR5) on the surface of T-cells, which prompts the envelope protein to undergo a large-scale conformational change that ultimately fuses the two membranes.^44,45 To combat this process, therapeutics that inhibit receptor recognition or prevent formation of the post-fusion conformation have been developed to prevent HIV-entry.

The most widely prescribed HIV-entry inhibitor, Enfuvirtide (T-20), is a biomimetic peptide that is used to prevent formation of the post-fusion conformation of the envelope protein.⁴⁶ Unfortunately, poor plasma half-life and the incidence of resistant mutants⁴⁷ are significant limitations of the current treatment. The success of T-20 prompted researchers to identify additional peptides with similar binding modes.^48,49 Many of these peptides displayed comparable, low nanomolar, half-maximal inhibitory concentration (IC₅₀) values compared to T-20, while also demonstrating efficacy against resistant strains and improved pharmacokinetic properties;^48,49 however, these peptides are large (>30 amino acids), which severely limits their clinical application due to poor bioavailability. Notably, shortened peptides with similar binding-modes have significantly reduced potencies,⁵⁰ and the best small molecule candidates inhibit HIV-entry at micromolar concentrations.⁵¹ These findings further demonstrated the need for next-generation cell entry inhibitors with novel modes of action.

Early HIV-entry PDCs were designed to overcome the loss of potency that occurs when working with shorter peptides. In one example, this effect was achieved by substituting a critical gp160 motif, termed WWI, with N-(carboxyphenyl)pyrroles, which are known to bind to the same hydrophobic cavity.⁵² Previous research into N-(carboxyphenyl)pyrrole derivatives revealed weak HIV-entry inhibition, and SAR analysis highlighted the necessary structural features for inhibition.⁵³ N-(carboxyphenyl)pyrrole derivatives with key inhibitory features were conjugated to the N-terminus of a shortened HIV entry-peptide inhibitor, termed P26, through a β-alanine linker. Notably, this series of PDCs demonstrated a greater than 8-fold increase in potency compared to P26 alone (2540 nM), with the most potent PDC, Aoc-βAla-P26, (1, Figure 1A) reaching an IC₅₀ of 14.9 nM. Compound 1 rivaled T-20, which displayed an IC₅₀ of 10.1 nM. Remarkably, the two most potent PDCs tested in this series also demonstrated inhibition of T-20-resistant HIV strains with IC₅₀ values around 100-200 nM. Pharmacokinetic studies of 1 displayed decreased proteinase-K digestion and an increased in vitro half-life compared to T-20, although the PDC had a 4-fold shorter half-life in vivo.

Figure 1: Chemical Structures of Human Immunodeficiency Virus Cell Entry Inhibitors.

Peptide-drug conjugate structures are displayed such that peptides are in blue, linkers in black and small molecules in red. Cartoon representations of protein targets are displayed on the right with target names below. (A) HIV entry inhibitor peptide P26 is N-terminally conjugated to the N-(carboxyphenyl)pyrrole, Aoc, through a β-alanine linker.⁵² (B) P26 is conjugated to the antiviral triterpene sapogenin through a chemical linker that contains a triazole moiety covalently attached to the side-chain of a C-terminal lysine and an additional β-alanine residue.⁵⁴ (C) HIV entry inhibitor peptide SP22 is conjugated to the CCR5 inhibitor, TAK, via a PEG-12 linker covalently attached to the C-terminal lysine side-chain.⁵⁵

A similar strategy was employed using sapogenin, an antiviral triterpene, conjugated to P26 via a triazole linker generated with click-chemistry.⁵⁴ In the sapogenin series of inhibitors, the small molecule was connected to P26 at the N-terminus through a modified β-alanine, or the C-terminus via reaction with a modified lysine side-chain. Surprisingly, the C-terminally modified P26 conjugates exhibited the greatest potency, with one of the most potent PDCs, P26-BApc, (2, Figure 1B) inhibiting cell entry at 3.94 nM, which was 3-fold more potent than T-20. Compound 2 also demonstrated inhibition of T-20-resistant HIV strains and displayed improved pharmacokinetic properties compared to T-20, including a longer in vivo half-life.

Recently, alternative HIV cell entry PDC inhibitors targeting the receptor/co-receptor recognition event have shown considerable success.^55,56 For example, the gp160 peptide-mimetic, P52, was C-terminally conjugated to CCR5 antagonists that block HIV receptor recognition^57–59 to generate a dual-target inhibitor.⁵⁵ PEG linkers of varying length were selected for this conjugation strategy, and inhibition studies revealed that PEG-4- and PEG-8-conjugated PDCs exhibited inhibition at 1.11 nM and 1.52 nM, respectively, while PEG-12 and PEG-24 linkers resulted in sub-nanomolar inhibition at 0.35 nM and 0.12 nM. Although all of the PDCs mentioned were more potent than T-20 (4.02 nM), these results demonstrate the importance of linker optimization in PDC design. After determining the optimal PEG linker length, P-52 was exchanged for a modified stapled peptide, SP22, with demonstrated anti-HIV entry activity,⁶⁰ and a C-terminal lysine was introduced for site specific labeling at the side-chain amine. A potent PDC from this series, SP12T (3), is displayed in Figure 1C. Remarkably, this series of PEG-12-linked PDCs was found to inhibit HIV strains that recognize the alternative co-receptor, CXCR4, with greater potency than T-20 against all strains tested. This approach demonstrates the many ways that PDCs can be designed to address therapeutic challenges from different angles. There are even examples of HIV entry PDC inhibitors that use peptide mimetics of the HIV receptor CD4 with reported success.⁶¹

Dengue Virus protease-inhibiting PDCs.

The World Health Organization reports that Dengue virus (DENV) infections have grown rapidly over the past decades from 500,000 infections in 2000 to a staggering 5.2 million infections in 2019. Additionally, these infections were once localized to tropical and subtropical regions but have recently been reported in non-tropical climates.⁶² The rapid spread of DENV has prompted some to consider it a global health concern, as severe infections may be fatal. Worryingly, no therapies are available to treat DENV infections, although drug targets for DENV therapies have been characterized. An appealing target for anti-DENV therapies is inhibition of the viral protease activity required to digest the large DENV polyprotein into its active components.⁶³ This proteolytic activity of DENV is performed by a catalytic complex composed of the multi-domain protein NS3, and the cofactor protein NS2B.⁶⁴

Although the NS3-NS2B complex demonstrates serine protease activity, known serine protease inhibitors are largely ineffective at inhibiting DENV protease activity.⁶⁵ This gap prompted researchers to test tetrapeptide libraries for inhibitors, which revealed a preference for peptides complementary to the di-basic recognition motif of NS3-NS2B.⁶⁶ Following identification of effective tetrapeptide inhibitors, electrophilic warheads were introduced at the C-terminus to improve potency, as demonstrated for other serine protease inhibitors.⁶⁷ This tetrapeptide-warhead strategy proved effective, with the most potent inhibitor, Bz-Nle-Lys-Arg-Arg-B(OH)₂ (4, Figure 2A), achieving an inhibition constant of 43 nM; however, the intrinsic reactivity of these warhead-peptides and the poor pharmacokinetic properties associated with the dibasic motif limit the clinical application of these inhibitors.

Figure 2: Chemical Structures of Dengue Virus Protease Inhibitors.

Peptide-drug conjugate structures are displayed such that peptides are in blue and small molecules in red. Cartoon representations of protein targets are displayed on the right with target names below. (A) A tribasic tetrapeptide complementary to the DENV protease recognition motif is covalently linked to a C-terminal boronic acid moiety.⁶⁷ (B) A dibasic retro-tripeptide complementary to the DENV protease recognition motif was covalently attached to an arylcyanoacrylamide at the N-terminus.⁷⁰ (C) The N-benzoyl capped dibasic tripeptide contains an unnatural 4-benzyloxy-D-phenylglycine residue at the C-terminus with a β-lactam ring covalently attached to the terminal nitrogen in the S configuration.⁷⁴

Building on the success of the warhead-based inhibitors, researchers sought to identify small-molecules with affinity for the DENV NS3-NS2B complex. These efforts resulted in the identification of arylcyanoacrylamide,⁶⁸ rhodanine and thiazolidinedione⁶⁹ scaffolds with modest micromolar affinity for the DENV protease. Having identified inhibitory peptides and small molecules, a series of DENV PDCs were developed which situated arylcyanoacrylamide derivatives as an N-terminal cap of various peptide sequences.⁷⁰ From this series of PDCs, the N-capped retro-tripeptide Arg-Lys-Nle-NH₂ (5, Figure 2B) exhibited the best potency, with an inhibition constant of 4.9 µM, as well as high specificity relative to other serine proteases, including the closely related West-Nile virus protease. Despite the in vitro success of the arylcyanoacrylamide PDCs, these compounds demonstrated weak anti-viral activity in cell culture, prompting researchers to explore rhodanine and thiazolidinedione as peptide capping moieties.⁷⁰ Cyanoacrylamide, rhodanine and thiazolidinedione scaffolds are challenging medicinal chemistry leads due to their high promiscuity;^69–72 nonetheless, these anti-DENV PDCs demonstrated low micromolar inhibitory concentrations with a site-specific binding mode as demonstrated by competition assays.⁷¹ Interestingly, rhodanine-capped peptides with hydrophobic substituents demonstrated considerable membrane permeability. This trend was also noted in cell-based assays where peptides capped with hydrophobic rhodanines demonstrated greater antiviral activity. Recently, N-sulfonyl capped peptides have emerged as novel anti-DENV PDCs, with the most potent derivatives achieving double-digit nanomolar inhibition constants.⁷³ Remarkably, these compounds have demonstrated high DENV protease selectivity, limited cytotoxicity and significant metabolic stability, indicating that they may represent strong candidates for clinical applications.

A final DENV PDC series must also be mentioned that revisits the covalent inhibition strategy. This PDC series (6, Figure 2C) incorporated β-lactam warheads at the C-terminus of anti-DENV peptides with the aim of developing PDCs with ‘soft’ reactivity⁷⁴ (i.e., reactivity with “soft” nucleophiles like the thiol group of cysteine). These PDCs demonstrated considerable success with many derivatives achieving single digit micromolar IC₅₀ values. Many of these PDCs also exhibited specificity for the NS3-NS2B complex compared to endogenous human serine proteases. Interestingly, these compounds displayed a complex inhibition mechanism where binding could result in the desired covalent inhibition of the NS3-NS2B serine, or cleavage of an internal peptide bond within the PDC. This report provides an interesting starting point for the development of covalent DENV protease inhibitors with greater specificity and lower reactivity; however, further optimization is required to mitigate inhibitor degradation.

5. ANTIBACTERIAL PEPTIDE-DRUG CONJUGATES

Since their discovery, antibiotics have been the workhorse of medical practitioners with billions of prescriptions being filled on an annual basis worldwide. In 2021, the CDC reported that 211.1 million antibiotic prescriptions were filled in the United States alone. This rise in antibiotic usage is directly proportional to the increase in antibiotic-resistant bacteria, as resistance mechanisms develop at an alarming rate due to prolonged and repeated exposure. Disturbingly, the increase in antibiotic drug resistance is accompanied by an increase in multi-drug-resistant bacteria, further emphasizing the need for novel antibiotic therapeutics. To this end, antibiotic PDC strategies have been developed that either combine synergistic antibiotic activities^75,76 or modulate existing antibiotics to escape resistance mechanisms.^77,78 Antibiotic PDCs have also been designed with significant efficacy toward gram-negative bacteria,⁷⁹ which has been a challenge for traditional small molecule antibiotics. In the following section, we will explore the development of vancomycin-based PDCs and the strategies employed to improve the vancomycin pharmacokinetic profile and overcome resistance. We also encourage the reader to consult the many reviews that discuss antibacterial peptides^80,81 and the development of antibiotic PDCs^82,83 for further reading.

Vancomycin-based peptide-drug conjugates.

Vancomycin is an antibacterial glycopeptide that was first isolated in the 1950s from the bacterial species Amycola-toposisorientalis.⁸⁴ This antibacterial compound was found to have activity towards gram-positive bacteria, including species with resistance to β-lactam antibiotics such as penicillin. The novel antibiotic exhibited a unique mechanism of action where bacterial cell wall peptidoglycan synthesis was inhibited by targeting a terminal d-Ala-d-Ala motif of the growing peptidoglycan chain.⁸⁵ Vancomycin interacts with this motif on peptidoglycan intermediates, preventing the formation of the glycan backbone by peptidoglycan polymerase. Unfortunately, vancomycin-resistant organisms were identified less than 30 years after FDA approval, with resistance brought on by single amino acid alterations to the terminal peptidoglycan intermediate motif.⁸⁶ Because of this developing resistance, vancomycin is administered only in serious and life-threatening cases of bacterial infections, particularly those with established antibiotic resistance. The rise of resistance to this last-resort treatment has motivated the development of vancomycin-based therapies that overcome these resistance mechanisms. Importantly, several vancomycin-based PDCs have been developed that demonstrate antibiotic activity across a wide range of gram-positive bacteria, including vancomycin-resistant isolates.^87,88

A convincing demonstration of vancomycin-based PDCs overcoming vancomycin resistance was reported where vancomycin was C-terminally conjugated to highly basic, lipidated membrane-targeting peptides, generating a series of PDCs called vancapticins.⁸⁹ This strategy was predicated on the idea that membrane-targeting peptides could be used to increase the concentration of vancomycin at the site of action. The first series of vancapticins contained lipidated peptides that were connected to vancomycin through a disulfide bond. Remarkably, this series of conjugates demonstrated 4- to 10-fold improvement over vancomycin towards methicillin-resistant S. aureus (MRSA), vancomycin-intermediate S. aureus (VISA) and multi-drug resistant (MDR) S. pneumoniae, as well as an impressive 100-fold improvement against vancomycin-resistant enterococci (VRE). Despite the success of the disulfide-linked vancapticin series, the compounds were found to have short half-lives and poor in vitro plasma stability, which prompted researchers to exchange the disulfide linker for covalent attachment through a lysine side-chain⁸⁹. The most potent PDCs from this series, including 7 (Figure 3A), were significantly more potent than vancomycin towards MRSA, VISA, vancomycin-resistant S. aureus (VRSA) and VRE, and rivaled the potency of contemporary glycopeptides such as oritavancin. The lysine-linked PDCs were found to be stable upon exposure to plasma and glutathione, and metabolically inert when exposed to liver microsomes. These PDCs also exhibited bactericidal activity in mouse models with an improved pharmacokinetic profile compared to vancomycin.

Figure 3: Chemical Structures of Vancomycin Peptide-Drug Conjugates.

Peptide-drug conjugate structures are displayed such that non-vancomycin peptides are in blue, linkers in black, small molecules in red, and additional chemical matter in green. Cartoon representations of protein and membrane targets are displayed on the right with target names below. (A) A compound from the vancapticin PDC series. A four-lysine peptide was covalently linked to vancomycin through the C-terminal lysine side chain and vancomycin carboxylic acid. These PDCspossess hydrophobic moieties at the peptide N-terminus.⁸⁹ (B) A hexa-arginine peptide was covalently attached to vancomycin through a C-terminal cysteine sidechain linked to the vancomycin secondary amine through a cyclohexane and pyrrolidine-2,5-dione linker.⁹⁰ (C) A lipopolysaccharide-binding peptide was conjugated to the free carboxylic acid of vancomycin via a linker that contains a pyrrolidine-2,5-dione, 2 PEG units and 2 amide moieties.⁹²

Similar studies of vancomycin-based PDCs tested the effect of polycationic hexa-arginine peptide conjugation at four unique sites on vancomycin.⁹⁰ All of the PDCs reported in this study demonstrated more potent antibacterial activity compared to vancomycin. Interestingly, the most potent PDC reported, FU002 (8, Figure 3B), had a different conjugation site than the vancapticins. This compound was generated via conjugation at the lone secondary amine in vancomycin and demonstrated a 1000-fold greater antibacterial activity than vancomycin against VRE, as well as significant activity against other vancomycin-resistant strains. The PDC also contained a unique linker that included cyclohexane and pyrrolidine-2,5-dione moieties owing to the bifunctional cross-linker used to facilitate the conjugation chemistries. In vivo studies revealed that 8 was partially excreted by the hepatobiliary route, suggesting this PDC has improved biodistribution over vancomycin, which is primarily excreted by the kidneys. Adding to the success of these PDCs, a separate study reported that poly-arginine vancomycin-based PDCs connected by an aminohexanoic acid linker eradicated MRSA biofilms and persister cells, both in vitro and in vivo.⁹¹ Taken together, these independent vancomycin PDC series demonstrate that vancomycin resistance and poor pharmacokinetic profiles can be overcome using PDC strategies.

Vancomycin-based PDCs have also been developed as gram-negative antibacterial agents through careful selection of a peptide partner. In one series of vancomycin PDCs, lipopolysaccharide (LPS)-binding proteins were conjugated to vancomycin at the four sites mentioned previously via CuAAC mediated click-chemistry or maleimide-sulfhydryl addition.⁹² LPS-binding proteins were selected because LPS constitutes a major portion of the gram-negative bacterial outer membrane (OM); thus, these vancomycin PDCs are expected to target gram-negative bacteria. Notably, the most effective PDCs from this series of antibiotics (9, Figure 3C) displayed single digit micromolar minimum inhibitory concentrations (MICs) for gram-positive bacteria and double digit micromolar MIC values for several gram-negative bacterial species. In a separate study, vancomycin was conjugated to an OM-disrupting peptide, the polymyxin E nonapeptide (PMEN), to generate a PDC series called vancomyxins.⁹³ The vancomyxin series contained PDCs with varying linker lengths and compositions; however, in all cases PMEN was conjugated to vancomycin at the C-terminus or disaccharide moiety via CuAAC click-chemistry. The vancomyxins that contained alkyl chain linkers demonstrated significantly improved antibacterial activity towards gram-negative bacteria compared to vancomycin, whereas this effect was less pronounced with PEG-3 linkers. Unfortunately, the vancomyxins did not display significant antibacterial activity towards vancomycin-resistant gram-positive bacteria. The results from these studies highlight the ways that antibacterial PDCs can be designed to repurpose antibiotics to target gram-negative bacteria with limited effective treatments.

6. PEPTIDE-DRUG CONJUGATES FOR CNS TARGETS

Peptide-Drug Conjugates for Neurodegenerative Diseases.

Neurodegenerative diseases (NDs) are associated with deterioration of the synapse and neural network, as well as the deposition and accumulation of altered protein variants in the brain. These diseases have a broad impact on basic human functioning, impairing basic abilities like speech, movement, stability, and balance, as well as more complex tasks like cognitive abilities. As the global population ages, the prevalence of memory-, cognition-, and movement-impairing diseases will continue to increase.^94,95 According to the World Health Organization, NDs require urgent treatment as they are projected to surpass cancer as the second-leading cause of death by 2040.^96,97 Alzheimer’s disease (AD) is a leading neurodegenerative disease responsible for most cases of elderly dementia. It causes progressive memory loss, impaired learning, and behavioral decline. The deposition of amyloid plaques in the brain is believed to contribute to the disease, leading to neuronal and synaptic loss.^98–100 At present, there is no established cure for AD, and the medications employed in its management primarily focus on addressing cognitive symptoms or other manifestations.⁹⁸ The blood-brain barrier (BBB) poses a problem for the development of neurodegeneration therapeutics, where most treatments fail due to limited drug permeability. Such poor BBB-permeability hinders the availability of effective options for treating NDs. Additionally, NDs are reported to have many different pathophysiologies,^101–104 so it is crucial to prioritize the development of new therapeutic agents that directly target multiple key processes involved in the progression of AD. Several PDC classes have been studied to address these challenges.

Galantamine-peptide derivatives.

Galantamine is an alkaloid natural product that acts as an inhibitor of acetylcholinesterase (AChE),¹⁰⁵ which has been studied as a potential target for AD therapeutics. Inhibition of this enzyme causes levels of acetylcholine in the brain to increase to normal levels, which may explain why galantamine treatment is associated with reduced mortality in AD patients and prolonged decline in cognition.^106,107 AD has also been linked to the formation and accumulation of amyloid beta (Aβ) deposits.¹⁰⁸ Inhibiting Aβ release by γ-secretase is also a widely sought approach against AD. Peptides such as N-(3,4-dichlorophenyl)-d,l-Ala-OH have been shown to inhibit γ-secretase,¹⁰⁹ so researchers hypothesized that by conjugating peptides containing this moiety with galantamine, compounds would exhibit joint AChE and γ-secretase inhibition.¹¹⁰ Conjugates were designed to link several di- or tripeptides with the N-(3,4-dichlorophenyl)-D,L-Ala-OH sequence linked to positions 6 or 11 of galantamine through ester or amide bonds (10, Figure 4A).

Figure 4: Chemical Structures of Galantamine-Peptide Conjugates.

Peptide-drug conjugate structures are displayed such that peptides are in blue, linkers in black and small molecules in red. Cartoon representations of protein targets are displayed on the right with target names below. (A) Example of a galantamine-peptide drug conjugate where galantamine (shown in red) is covalently linked through an ester bond to a tripeptide comprising N-(3,4-dichlorophenyl)-d,l-Ala-OH (shown in blue) at position 6.¹¹⁰ (B) Example of a galantamine-peptide drug conjugate with a shortened sequence of OM 99–2 covalently linked at position 11 through an amide bond to an Asp residue, which is also linked to a nicotinic residue.¹¹⁵ (C) Example of a galantamine-peptide drug conjugate with an analog of Leu-Val-Phe-Phe (Aβ17-Aβ20) covalently linked through an ester bond to position 6 of galantamine.¹¹⁸

In vitro testing of the conjugates towards AChE and butyrylcholinesterase (BuCHE) showed that three of the six synthesized conjugates had lower AchE IC₅₀ values than galantamine, and all of them had significantly increased activity as BuChE inhibitors.¹¹¹ This enzyme is also of interest because BuChE also hydrolyzes acetylcholine,¹¹² meaning that both AchE and BuChE participate in the degradation of this neurotransmitter. All six conjugates were determined to be potent non-competitive inhibitors of BuChE. It is also interesting to note that the less-potent inhibitors against AchE exhibited the highest potency against BuChE, demonstrating that small structural differences in these peptides play a role in selectivity towards the enzymes. γ -Secretase activity was evaluated through cellular and in vitro assays, where one of the conjugates was active in vitro but not in vivo. The results suggest that the lack of activity in vivo might be due to permeability problems and points to a factor to consider when developing these peptide conjugates. All the other conjugates were able to reduce cellular Aβ production in vitro in a dose-dependent manner with similar IC₅₀ values.

Based on the success of these conjugates, studies were continued with an additional interest in inhibition against β-secretase activity. This enzyme is also implicated in the release of the β-amyloid precursor protein.¹¹³ Peptides with shortened sequences of Boc-Val-Asn-Leu-Ala-ОН (OM 99-2), which are known β-secretase inhibitors,¹¹⁴ conjugated to galantamine were synthesized towards this goal.¹¹⁵ To additionally optimize the permeability of the compound, which is an essential aspect for drugs that act on the central nervous system (CNS), conjugates were also linked to a nicotinic or an isonicotinic acid moiety (11, Figure 4B). Nicotinic acid has been found to improve permeability of compounds and has increased their passage through the BBB.^116,117 These conjugates were expected to exhibit AChE, BuChE, and β-secretase inhibition while ensuring crossing of the BBB. It was determined that there was no sign of toxicity or behavioral changes in male albino mice after a 7-day window following intraperitoneal administration of conjugates. Cellular cytotoxicity assays determined that conjugates had no effect on cellular viability. The conjugates effectively reversed scopolamine-induced memory-impairment in mice and exhibited improved cholinesterase inhibition and antioxidant activity compared to galantamine in ex vivo experiments.

Following these studies, another set of galantamine conjugates were studied.¹¹⁸ This set had galantamine conjugated to residues 17-20 of Aβ (Leu-Val-Phe-Phe; 12, Figure 4C) and analogs of the peptide where Val is replaced by tert-Leu (Tle) or nor-Val (Nva), and Leu is replaced by nor-Leu (Nle). Results showed that all compounds exhibited inhibitory activity against β-secretase, and four of the tested compounds were able to inhibit platelet aggregation. Most notably, this study demonstrated that the alterations that include Nva or Tle in place of Val in the original peptide, along with a free N-terminal carboxyl group, resulted in the most favorable outcomes.

Trolox-conjugated β-amyloid C-terminal peptides.

The most-studied variants of the Aβ peptides are Aβ_1-40 and Aβ_1-42, although the latter exhibits the higher aggregation propensity because of additional hydrophobic residues at the C-terminus that cause aggregation.^119–122 Peptide Aβ_1–42 is thus considered to be the most relevant peptide in AD, and inhibition of its self-aggregation has been targeted for the development of novel AD therapeutics. The Aβ peptide has also been linked to the production of reactive oxygen species (ROS) by interacting with molecular oxygen and bioactive metal ions,^123–124 promoting cell-membrane perturbation and ultimately resulting in oxidative stress and cell death. Antioxidants like Vitamin E or polyphenols have been studied as a way to counteract the cellular damage caused by oxidative stress in AD.¹²⁵ A series of peptide conjugates were designed to aid in oxidative injuries and inhibit self-aggregation of Aβ peptides in AD patients’ brains. To this end, the radical-scavenging portion of Vitamin E (trolox) was conjugated to the C-terminus of several Aβ peptide variants.¹²⁶

Evaluation of the inhibitory activity of the trolox-peptide conjugates (TxAβ_x–n, where x is the starting residue and n is the final residue of Aβ peptide) on Aβ_1–42 aggregation showed that they prevent aggregation of the peptide, while incubation with trolox alone showed no activity. The inhibitory activity of the conjugates increased with the length of the Aβ_1–42 C-terminal motif in the TxAβ_x–42 peptides. Similar inhibitory activities were also observed for TxAβ_x–43 conjugates. On the other hand, incubation with TxAβ_x–40 did not show any significant inhibitory activity. This result suggests that the additional residues in the C-termini of TxAβ_x–42 and TxAβ_x–43 may play a role in the interruption of Aβ_1–42 affinity. No inhibitory activity was reported for Aβ peptides that were not conjugated to trolox, indicating that the length of the peptide, the identity of C-terminal residues, and their conjugation to trolox all play an important role in inhibiting Aβ-peptide aggregation.

Cytotoxicity of the conjugates was evaluated through cell viability assays in human neuroblastoma SH-SY5Y cells.¹²⁶ Results showed that there was no effect in cell viability after incubating the cells with TxAβ_x–42 and TxAβ_x–43, which shows lack of cytotoxicity. To study whether the conjugates could counteract the cytotoxic effects of Aβ_1–42 peptide alone, TxAβ_x–42 and TxAβ_x–43 were incubated along with Aβ_1–42. Results show that cytotoxicity towards the cells is significantly decreased, with cell viability increasing in a dose-dependent manner, and even completely inhibited by one of the conjugates.¹²⁶ Finally, radical activity of the conjugates was assessed by comparing with controls, and results showed that TxAβ_36–42 (13, Figure 5A) was able to completely scavenge ROS induced by H₂O₂. Additionally, it was able to significantly decrease ROS generation by Aβ_1–42.

Figure 5: Chemical Structure of Anti-Neurodegenerative Peptide-Drug Conjugates.

Peptide-drug conjugate structures are displayed such that peptides are in blue, linkers in black, small molecules in red, and additional chemical matter in green. Cartoon representations of protein targets are displayed on the right with target names below. (A) A Trolox-peptide conjugate where trolox is linked through an amide bond to Aβ peptide variant Aβ_36–42.¹²⁶ (B) General structure and an example of a tripartite structure consisting of membrane anchor (raftophile), spacer and pharmacophore (inhibitor). The tripartite structure shown features GL 189 as the BACE1 inhibitor and dihydrocholesterol as the membrane anchor linked through a H-3Gl-OH spacer.^127,128 (C) Example of a shuttle-cargo fusion molecule from Hybrid Molecule Subset I. Peptidic carrier TfR-P is conjugated to fallypride directly through an amide bond.¹³¹ (D) Example of a shuttle-cargo fusion molecule from Hybrid Molecule Subset II. TfR-P is conjugated to fallypride through a PEG-2 linker.¹³¹ (E) Example of a shuttle-cargo fusion molecule from Hybrid Molecule Subset III. Angiopep-2 was conjugated to fallypride through a PEG₄ linker.¹³¹ (F) Example of of a shuttle-cargo fusion molecule from Hybrid Molecule Subset IV. Carrier peptide TfR-P is conjugated to fallypride through a PEG-3 linker and derivatized with a radiolabeling moiety (SiFA, shown in green), which is linked to the carrier peptide of the compound.¹³¹

Tripartite structures of β-secretase inhibitors to raftophilic lipid anchors.

In another approach to target β-secretase activity—specifically, the β-site APP-cleaving enzyme 1 (BACE1)—the concept of “tripartite structures” was proposed.¹²⁷ These structures consist of a pharmacophore (i.e., a BACE1 inhibitor), a lipid, and a linker to serve as an appropriate connection between them (general structure shown in Figure 5B). Lipids such as cholesterol or dihydrocholesterol partition into lipid rafts on the outer leaflet of the cell membrane. Since BACE1 inhibitors should be anchored to the cell membrane before the compound is endocytosed, coupling inhibitors to these lipids was expected to result in a decrease in Aβ production by tethering the compound to the membrane and facilitating endocytosis.¹²⁸ A sterol moiety was therefore coupled to the C-terminus of the inhibitory peptide by a polyglycol linker (14, Figure 5B). Commercially available BACE1 inhibitor III (H-Glu-Val-Asn-Sta-Val-Ala-Glu-Phe-NH₂) was used as the inhibitor. The conjugates were more potent than the free inhibitor both in vitro and in vivo. As controls, the inhibitor with the polyglycol linker was tested, along with the inhibitor directly linked to the sterol moiety. Both were inactive, which indicated the need for the tripartite structure.

As a continuation to this study, inhibitors GL 189 and a derivative of lysine were coupled to dihydrocholesterol through linkage with a series of spacers.¹²⁷ The spacers were chosen based on their low toxicity, physiological stability, solubility, and variability in lengths. These new conjugates were tested in cell-based assays, and results showed that at a concentration of 100 nM, tripartite structures reduced the secretion of Aβ40 in all cell lines, while free inhibitors had no effect even at higher concentrations. Because one of the active conjugates had the lysine pharmacophore, results suggested that the tripartite concept also showed efficacy when applied to non-peptidic inhibitors. Conjugates with spacers ranging from 18 to 123 Å were synthesized and tested for their ability to reduce Aβ-peptide secretion to identify optimal length of the spacers. Best results were obtained with spacer lengths ranging from 35 – 89 Å, with a 75–90% reduction in Aβ-peptide secretion. Some flexibility may therefore be allowed in terms of the distance to the membrane. Several other analogs with tripartite structures have also been synthesized to further evaluate the structure-activity relationship.¹²⁹

Shuttle−cargo fusion molecules of transport peptides and the hD_2/3 receptor antagonist fallypride.

Many efforts have been employed to improve blood-brain-barrier permeability of CNS drugs. The most widely studied approach to overcome this limitation is commonly called the molecular Trojan horse approach,¹³⁰ where a “shuttle” is used to facilitate the crossing of the BBB of a “cargo,” which is the poorly permeable therapeutic; however, the use of large bioactive molecules as shuttles, which, although highly efficient and specific, results in poor pharmacokinetic properties along with other undesirable side effects. A proposed solution is to use small peptides as shuttles that are transported across the BBB by endogenous transporters.¹³¹ The use of these small peptides is favorable because they exhibit fewer undesired side effects and better pharmacokinetic properties. They are also easier to synthesize and derivatize with the cargo drug. Several conjugates were synthesized consisting of the hD_2/3 receptor-binding-drug fallypride and a peptide carrier. Compound fallypride can be used in vivo with positron emission tomography (PET) imaging and in vitro with hD_2/3 receptor binding assays. For the shuttle moiety, peptides with different lengths, structure, polarity, and geometry that enter the brain parenchyma by receptor-mediated transcytosis were chosen. The structural variations of the peptides would give insight into the tolerance of peptide identity while retaining biological activity and receptor selectivity and affinity.

Four different sets of conjugates were designed and synthesized for this first study. The first subset tested the influence of the complexity of the carrier peptide (15, Figure 5C). The second subset was conjugated to a transferrin receptor binding peptide HAIYPRH (TfR-P) by linkers of different lengths to explore the extent to which steric hindrance exerted by the peptide on the drug affects binding affinity of fallypride (16, Figure 5D). The third subset consisted of the conjugation of AngioPep to fallypride through a PEG linker to test the BBB transport capacity of AngioPep in the conjugates (17, Figure 5E). Finally, the fourth subset contained a moiety for radiolabeling with ¹⁸F or ⁶⁸Ga to be applicable in in vivo PET imaging or combined PET/optical imaging of hD₂ receptors (18, Figure 5F). The binding affinity for all compounds was in the nanomolar range, which was comparable to that of controls raclopride and chlorpromazine. Results showed that complexity, length, and individual structure of the peptides had no correlation to affinity. Best results were obtained with “medium-sized” linkers, while shorter or longer linkers resulted in lower affinities; however, with the third-subset conjugates, different results were observed: the conjugate with highest binding affinity had no linker at all.

Docking studies based on homology modeling of the receptor more closely relate the influence of the different structural elements of the fusion molecules on hD₂ receptor binding. Results showed that the linker, but not the peptide carrier most exposed to solvent, played a role in stabilizing conjugate binding to the receptor. It also confirmed that fallypride can still bind to the receptor without being blocked by the other constituents.

Peptide-Drug Conjugates for Pain Management.

Pain represents a significant burden to healthcare systems as individual treatment plans are often required to manage the frequency and severity of pain in individual cases. Medicinal chemists have addressed this issue with the development of highly potent, small molecule analgesics that typically interact with G protein-coupled receptors (GPCRs) to relieve pain. Despite the drug-likeness of analgesic compounds, these molecules often display target promiscuity and produce harmful off-target effects¹³². Opioids, such as morphine or fentanyl, are well-known examples of highly potent analgesic compounds with a high potential for abuse and instances of fatal overdose¹³³. According to the CDC, opioid overdose was responsible for over 645,000 deaths between 1999 to 2021. The most potent opioids developed target the GPCR called the μ-opioid receptor (MOR); however, molecules that target this receptor often result in severe and undesirable side effects, including addiction¹³⁴. To overcome the unfavorable and severe side-effect profiles of MOR targeted ligands, medicinal chemists have designed small molecules that target the κ-opioid receptor (KOR), which represents an alternate target with less significant side-effects. Despite the identification of this parallel target, ligands developed for KOR activation continued to produce unfavorable side effect profiles, albeit with significantly less severe side-effects¹³⁵. The gradual improvement towards KOR-related non-addictive analgesics has prompted researchers to develop novel KOR ligands with improved side-effect profiles.

Recently, PDCs have been designed to target the KOR to generate compounds with high receptor specificity, while also reducing analgesic side effects¹³⁶. In this study, the crystal structure of KOR complexed with a ligand, MP1104¹³⁷, was used to computationally design a series of PDCs that link KOR ligands based on MP1104 to thioether cyclized peptides designed to interact with the extracellular loops of KOR. Before small molecule conjugation, the computationally designed peptides were assessed in cell-based assays which demonstrated that the compounds were unable to displace an orthosteric ligand at concentrations as high as 10 μM. These results suggested that the peptides did not bind the common small molecule KOR binding-pocket, The C-terminal carboxylic acid of the cyclic peptides was then conjugated to the free amine of an MP1104 derivative, β-NalA, to generate this series of PDCs. Remarkably, all four of the PDCs demonstrated low nanomolar affinities, with the best binding PDC, DNCP-β-NalA (19, Figure 6A), producing a K_i of 3.9 nM. This PDC marked a substantial improvement over β-NalA which gave a K_i of 72 nM.

Figure 6: Chemical Structure of Analgesic and Anti-inflammatory Peptide-drug conjugates.

Peptide-drug conjugate structures are displayed such that peptides are in blue, linkers in black, small molecules in red, and additional chemical matter in green. Cartoon representations of protein targets are displayed on the right with target names below. (A) The cyclic peptide DNCP was linked to β-NalA through an l-Thr-d-Phe linker.¹³⁶ (B) Example of a betulinic acid peptide conjugate where betulinic acid is linked to an azido peptide directly through a C-C bond.¹⁴² (C) Example of a Methotrexate-Hyaluronic acid conjugate where methotrexate is linked to hyaluronic acid through a cleavage susceptible peptide and a PEG-3 linker.¹⁴⁸ (D) Methotrexate-cIBR conjugate where methotrexate is linked by an amide bond to the N-terminus of cIBR peptide.¹⁵¹ (E) β-carboline peptide conjugate where a Β-carboline alkaloid is linked to ARPAK pentapeptide through an amide bond.¹⁶²

Pharmacokinetic profiling of 19 revealed full agonism and demonstrated a preference for GPCR binding compared to β-arrestin. This finding is notable as compounds that interact with β-arrestin are associated with a greater incidence of side effects. Mice models also demonstrated the antinocecptive efficacy of 19 while also highlighting a lack of side-effects associated with KOR-specific ligands. Notably, opioid receptor selectivity studies demonstrated that 19 binds KOR and MOR in an equipotent manner; thus, the results of the mouse model studies could be the result of dual KOR/MOR agonism. Another important component of this study was the validation of the binding site of 19 site via cryo-electron microscopy which revealed that β-NalA bound to the common small-molecule binding site of KOR, while the cyclic peptide portion associated with extracellular loops and transmembrane domains of KOR. In summation, this PDC series demonstrates how small molecules with significant side-effects can be leveraged in PDC design to generate potent compounds with reduced off-target effects.

7. ANTI-INFLAMMATORY PEPTIDE-DRUG CONJUGATES

Inflammation is the body’s defense against injury or infection. Acute inflammation is a short-term response wherein the immune system releases white blood cells to defend the area, causing visible redness and swelling. Although inflammation is crucial for fighting infections, chronic inflammation leads to various inflammatory disorders. PDCs have a broad range of applications, including peptides serving as carriers for a drug. Conjugation of anti-inflammatory drugs with peptidic components can result in targeted delivery of drugs to inflammation sites, such as bones or tumor cells.

Betulinic acid-peptide conjugates.

Betulinic acid is a triterpenoid natural product that has been reported to have antiviral, anticancer, anti-bacterial, antimalarial, and anti-inflammatory activities.^138–140 Various strategies to increase its solubility and bioavailability have been employed in the past, and integration of amino acids in various positions of the natural product was demonstrated to increase water solubility.¹⁴¹ A series of betulinic acid derivatives with 1,2,3-triazole peptide fragments were developed as peptide-drug conjugates with anti-inflammatory activity (20, Figure 6B).¹⁴² The 1,2,3-triazole was introduced as a bioisostere for an amide bond with improved stability in physiological conditions. Anti-inflammatory activity for these conjugates was first tested in vivo through a histamine-induced paw edema model, where analogs conjugated with histidine, alanine, tryptophan and isoleucine residues exhibited anti-inflammatory effects. Comparison between the conjugates’ activity and a synthetic intermediate in their synthesis, a betulinic acid derivative with C-3 position bound to an alkyne group, showed inferior activity to its peptide-conjugated analogues.

Docking studies were employed to evaluate affinity of the peptide conjugates to the Kelch domain of the KEAP1 protein. The broad-complex, tramtrack, bric-à-brac and intervening region domains of KEAP1 are known molecular targets for triterpene scaffolds.¹⁴³ KEAP1 is involved in antioxidant/anti-inflammatory activity through the NRF2/KEAP1/ARE pathway. Activation of this pathway has been reported to promote the suppression of pro-inflammatory enzymes.¹⁴⁴ Although the conjugate’s affinity towards KEAP1’s Kelch domain was less than that of the native ligand, structural analysis revealed that the π-system of the dimethoxyphenyl group plays a role in the interaction with Kelch-domain amino acid residues. The absence of a correlation between the docking studies and the conjugates’ inflammatory activity may suggest that its activity involves a more complex mechanism that gives rise to their anti-inflammatory effects.

Peptide conjugates of the structurally related betulonic acid have also been studied for their anti-inflammatory activities after research showed that 1,2,3-triazole derivatives of the acid exhibited high hepatoprotective and anti-inflammatory activity.¹⁴⁵ This prompted the synthesis of betulonic acid-conjugates that were further derivatized with 1,2,3-triazole peptide fragments.¹⁴⁶ Six tripeptides were incorporated into the betulonic acid conjugates to increase aqueous solubility and enhance bioactivity. The peptide fragment and betulonic acid structure were linked through the 1,2,3-triazole moiety that was previously conjugated to the acid. Six conjugates were evaluated for their anti-inflammatory activity, but none of them showed promising activity.

Hyaluronic acid – methotrexate conjugates.

Arthritis is a chronic autoimmune disease that affects approximately 0.5% to 1% of the worldwide population.¹⁴⁷ It leads to joint stiffness, deformity, and damage. Arthritis is characterized by recurring episodes of joint swelling, pain, and resultant disability, which significantly impact physical health, quality of life, employment, and expenses. Hyaluronic acid (HA) and methotrexate (MTX) conjugates have been studied as potential treatments for osteoarthritis (OA), based on HA’s pain-eliminating effect and MTX’s anti-inflammatory effect.¹⁴⁸ These conjugates were proposed because of HA’s reported use as a carrier for various other medicinal agents and because of the need for MTX derivatives that retain anti-inflammatory activity but lack harmful adverse effects.^149,150

Conjugation of the two clinically evaluated agents was done through a peptide chain sensitive to lysosomal enzymes. Upon cell uptake and peptide chain degradation, the conjugate would release the active form of MTX into the cell. Linkage of this peptide was studied using PEG-13 to eliminate potential steric hindrance related to interaction with proteases. This linker would further separate the HA backbone from the peptide chain, allowing the peptides to be accessible for cleavage in the lysosome. Conjugates with and without the peptide chain were synthesized to validate this reasoning (21, Figure 6C). The anti-proliferative effect of the conjugates was evaluated using human fibroblast-like synoviocytes (HFLS) stimulated by TNF-ɑ, and results showed that conjugates with the peptide chain inhibited proliferation, while linker-only conjugates had no anti-proliferative activity. The anti-inflammatory effects of these conjugates were also tested in vivo in an antigen-induced-osteoarthritis rat model. Two peptidic conjugates with different amino acid composition were tested alongside HA in this model. Neither intra-articularly injected MTX or HA had any effect on swelling. On the other hand, one of the conjugates showed significant reduction of swelling, and the other exhibited inhibition of swelling. This suggests that the composition of peptide chains plays a role in its lysosomal cleavage.

This compound was then further modified starting with the composition of its peptide chain.¹⁵¹ Peptide sequences were designed based on recognition sites for cathepsins B, D, or L, because they are the most abundant lysosomal enzymes in synovial tissues. Fragmentation resulting from cathepsin cleavage was studied through mass spectrometry. Only peptides with α-connected compounds were recognized and cleaved by cathepsins and released free MTX. A Phe-Phe sequence was deemed to be important in cathepsin recognition and cleavage because of the tendency of cathepsin D to hydrolyze inside the sequence and cathepsin B and L to hydrolyze after the sequence. Linker optimization was done by comparing different linker lengths, and although it was hypothesized that the most important factor was separation of MTX compound and HA, hydrophilicity turned out to be the leading factor in linker efficacy. In vitro, the three best linkers were PEG-8, PEG-5, and Alkyl2, which had the lowest Log P values. This correlated with in vivo efficacy of the conjugates and thus Alkyl2 was selected as the optimal linker because of its efficacy in both in vitro and in vivo studies and its low molecular weight. Molecular weight was also studied as a potential determining factor for conjugate activity. Low molecular-weight HA yielded low or no efficacy in vivo. On the other hand, conjugates with HA with higher molecular weight exhibited the same effects as free HA.

Methotrexate – cIBR conjugates.

Another approach on the development of MTX-peptide conjugates was explored through its conjugation to (cyclo(1,12)PenPRGGSVLVTGC) peptide (cIBR).¹⁵² Peptide cIBR mimics Domain-1 of the intracellular adhesion molecule-1 (ICAM-1), a protein that interacts with leukocyte function-associated antigen-1 (LFA-1) on T-cells.¹⁵³ This interaction has been identified as a target for rheumatoid arthritis, because of its role in the formation of inflammatory cellular aggregation in synovial tissues, which would make the delivery of MTX more targeted to leukocytes and avoid side effects.¹⁵⁴ cIBR was linked to MTX through an amide bond between the γ-carboxylic acid group of MTX and the N-terminus of the peptide (22, Figure 6D). In vivo efficacy of this conjugate was evaluated using the rat adjuvant arthritis model. Arthritic rats and normal control rats were treated with injections of different doses of MTX-cIBR and were monitored for ankle diameter, paw weight, inflammation, and bone resorption. Treatment showed reduced body weight and suppressed the progress of arthritis, with moderate synovium and no bone resorption at the physis. On the other hand, arthritic rats treated with vehicle exhibited severe periarticular inflammation, synovitis, and bone resorption. Uptake of MTX-cIBR by LFA-1 was evaluated in Molt-3 T cells by evaluating how anti-LFA-1 monoclonal antibody (mAb) and CIBR peptide would modulate the toxicity of the conjugate on the cells. Results showed that increasing anti-LFA-1 mAb and cIBR peptide concentrations increased cell viability, which suggests that MTX-cIBR uptake is mediated by LFA-1. Stability of the conjugate was studied at pH 1, 4, 5, 6, 7, 8, 10, and 12. Maximal stability was observed at pH 6, with the conjugate exhibiting a half life (t_1/2) of 770 min, which suggests good stability and longer half-lives close to physiological pH (7.2). In vitro stability of MTX-cIBR was evaluated in rat plasma and homogenized rat heart tissue and was found to be 43.8 and 38.1 min, respectively.

β-Carboline-peptide conjugates.

A new class of β-carboline-peptide conjugates was studied for the treatment of ischemia/reperfusion (I/R) injuries, a condition caused by low/blocked blood-flow to tissues, leading to tissue and organ death.^155,156 I/R injuries have been associated with an increase in ROS, leading to attack on cell-membrane constituents resulting in lipid peroxidation, DNA damage, release of proinflammatory substances, and increased formation of ROS, which leads to cell death.^157,158 These events activate the inflammatory cascade, and ultimately result in organ failure and death. Existing therapies focus on anti-inflammatory, anti-leukocyte, and antioxidant agents. β-Carbolines are naturally occurring alkaloids that have been studied for their antioxidant properties, and some have been reported to be even more potent than ascorbic acid.^158–161 Short peptide sequences such as Ala-Arg-Pro-Ala-Lys (ARPAK), Gly-Arg-Pro-Ala-Lys (GRPAK) and Gln-Arg-Pro-Ala-Lys (QRPAK), have thrombolytic activity and, in combination with β-carbolines, it was hypothesized that their combination would exhibit both anti-oxidant and thrombolytic activities that may have therapeutic potential in treating I/R injury.

β-Carboline alkaloids were first synthesized along with the unconjugated pentapeptides and were then coupled through an amide bond to produce nine conjugates.¹⁶² In vivo thrombolytic activity of the β-carboline-peptide conjugates was evaluated through comparison of thrombus mass reduction. Conjugates exhibited the same significant activity in reducing thrombus mass. The conjugates’ ability for free radical-scavenging was studied in vitro through a cell survival assay in rat pheochromocytoma (PC 12) cells, which are particularly sensitive to oxidative stress. The scavenging capacity of the conjugates compared to their parent species was comparable in increasing cell survival after exposure of cells to NO, H₂O₂, and HO⁻. Of the nine conjugates, only one (23, Figure 6E) was studied for its effect on lipid peroxidation-malondialdehyde (MDA), glutathione (GSH), and glutathione disulfide (GSSG) levels in the cell. 23 increased MDA in muscle tissues, which indicates its role in suppressing the lipid peroxidation cascade that results in inflammatory response. It was also found to reverse decreased GSH levels in tissue and blood by increasing the GSH/GSSG ratio, which demonstrates it lowers oxidative stress.

8. CONCLUSIONS

The rapid advancements in peptide chemistry have enabled the development of the next generation of peptide therapeutics that incorporate the potency of small molecules with the specificity of peptides in a class of compounds called PDCs. These multi-component molecules have had success overcoming drug-resistance mechanisms and treating complex disease states that present significant challenges to both peptides and small molecules independently. Unlike other bioconjugate therapeutics, such as ADCs, PDCs can be tailored to address a wide range of therapeutic challenges because of their smaller size and their ability to accommodate a wide array of small molecules. This perspective highlights the many considerations that go into PDC design and optimization, while also emphasizing several PDC series that demonstrate significant effectiveness against infections and disorders that lack adequate treatment. As drug resistance mechanisms develop proportionally with the increased prescription of modern medicines, new drugs will be needed to address these clinical challenges. It is our belief that PDCs represent an important part of the next generation of peptide therapeutics that will be required to overcome the rise in drug resistance and treat complex disease states.

SIGNIFICANCE.

The significance, impact, and innovation of this article:

• Juxtaposition of peptide-drug conjugates and antibody-drug conjugates with a focus on challenges that can be overcome with peptide-drug conjugate strategies.

• Considerations for the design and optimization of peptide-drug conjugates.

• Analysis of several peptide-drug conjugate series that demonstrate effectiveness towards drug resistant infections and complex disease states.

ABBREVIATIONS USED

Aβ

amyloid beta

AChE

acetylcholinesterase

AD

Alzheimer’s disease

ADC

antibody-drug conjugate

BACE1

β-site APP-cleaving enzyme 1

BBB

blood-brain barrier

BuChE

butyrylcholinesterase

CNS

central nervous system

DENV

dengue virus

FDA

U.S Food and Drug Administration

GK

glucokinase

GSH

glutathione

HA

hyaluronic acid

HFLS

human fibroblast-like synoviocytes

hIAPP

human islet amyloid polypeptide

HIV

human immunodeficiency virus

IC₅₀

half-maximal inhibitory concentration

ICAM-1

intracellular adhesion molecule-1

IR

insulin receptor

IRS

insulin receptor substrate

I/R

ischemia/reperfusion

LFA-1

leukocyte function-associated antigen-1

LPS

lipopolysaccharide

mAb

monoclonal antibody

MDA

malondialdehyde

MDR

multi-drug resistant

MRSA

methicillin-resistant S. aureus

MIC

minimum inhibitory concentration

MTX

methotrexate

ND

neurodegenerative diseases

Nva

norvaline

OA

osteoarthritis

OM

outer membrane

PC

pheochromocytoma

PDC

peptide-drug conjugate

PEG

polyethylene glycol

PET

positron emission tomography

PMEN

polymyxin E nonapeptide

PROTAC

proteolysis-targeting chimera

PTP1B

protein tyrosine phosphatase 1B

RA

rheumatoid arthritis

ROS

reactive oxygen species

SAR

structure-activity relationship

SARS-CoV2

Severe Acutre Respiratory Syndrome Coronavirus 2

SPP

similarity property principle

TfR-P

transferrin receptor binding peptide

Tle

tert-lecuine

VISA

vancomycin-intermediate S. aureus

VRE

vancomycin-resistant enterocci

VRSA

vancomycin-resistant S. aureus

KOR

κ-opioid receptor

MOR

μ-opioid receptor

Biographies

Biographies

Trevor T. Dean is a Pharmaceutical Sciences graduate student in the Chemistry and Drug Discovery concentration at UIC. He completed his H.BSc in biochemistry at McGill University in 2019. He later completed his MSc at the University of Toronto in 2022 where he used molecular and structural biology techniques to study viral proteins. In the Moore lab, Trevor develops binding assays to study antimicrobial peptides such as apidaecin.

Juliet Jelu-Reyes is a Pharmaceutical Sciences graduate student in the Chemistry in Drug Discovery concentration at UIC. She received her B.S. in Chemistry from the University of Puerto Rico Rio Piedras in 2022. Her research focuses on the development of stapled peptides that modulate the estrogen receptor to address breast cancer.

A’Lester C. Allen is an IRACDA postdoc affiliated with the Moore lab at UIC. He graduated with a PhD in chemistry in 2022 from the University of California Santa Cruz, where he studied plasmonic nanoparticles to detect protein and lipid biomarkers using surface-enhanced Raman scattering. He holds an MS in materials engineering from San Jose State University, and a BS in Chemistry from Stanford University. He joined the Moore lab in 2022 and contributes to the apidaecin project.

Terry W. Moore is an Associate Professor of Pharmaceutical Sciences at the University of Illinois Chicago. He holds a PhD in organic chemistry from the University of Illinois at Urbana-Champaign, and a BA in biochemistry from Abilene Christian University. His research interests include small molecules and peptides active against a number of different molecular targets.