Before the invention of solid-phase peptide synthesis (SPPS) in 1963 [44], the solely chemically synthesized peptides were the endogenous human hormones vasopressin and oxytocin [45]. This leap forward, together with the advent of recombinant technology, yielded the production of a large amount of natural and nature-inspired peptides. Insulin, vasopressin, oxytocin, somatostatin, glucagon-like peptide or glucagone-like peptide-1 (GLP-1), and gonadotropin-releasing hormone (GnRH) are some of the endogenous human short peptide hormones. These molecules represent the large family of peptide hormones with a short number of amino acids [46]. GLP-1 (37 amino acids) is a physiologically active short peptide that controls insulin levels. The awfully short half-life of insulin in plasma (1.5–5 min) limits its clinical use and prompted its sequence modifications to improve its stability. Semaglutide, liraglutide, and dulaglutide are commercially available GLP-1 receptor agonists for the treatment of T2DM [47]. As further discussed in Orally available peptides, while their alanine residue substitution enabled them to avoid degradation by dipeptidyl peptidase-IV (DPP-IV); the incorporation of an alkyl chain allowed improved binding to plasma proteins.

Human neurohormone vasopressin is a naturally occurring nonapeptide, which is exclusively found in mammals and synthesized in the hypothalamus. It is a vasoconstrictor and a water homeostasis regulator [48, 49]. As its use was marred with limitations, such as short half-life, target specificity, and side effects, some efforts emerged to discover novel and improved vasopressin analogues [50]. These include the natural lypressin (half-life 5–7 min)—in clinical use as a nasal spray for the treatment of diabetes insipidus [51], and the synthetic terlipressin (half-life 4–5 h), and ornipressine (half-life 1–2 h)—useful in liver cirrhosis hepatorenal syndrome (HRS) [52]. Lypressin differs from vasopressin for having the arginine in position 8 replaced by lysine, and for showing a conformational flexibility in the disulfide portion essential for biological activity [50]. Terlipressin is considered a pro-drug due to a three glycine residue at the N-terminal that are cleaved to release the active metabolite lypressin [53]. Ornipressin, the arginine in position 2 is replaced by ornithine, and is successfully used in combination with tranexamic acid acting on microcirculation.

The use of long-term terlipressin and ornipressin, explored for the coronavirus disease-19 (COVID-19) as well. Indeed, it was demonstrated that arginine vasopressin (AVP) is produced in COVID-19 patients with high concentrations of cytokines to counterbalance the high blood viscosity. High AVP levels in COVID-19 patients, induce hyponatremia, inflammatory disorders, and other complications by activation of nuclear factor-kappaB (NF-κB) and NOD-, LRR- and pyrin domain- containing protein 3 (NLRP3) inflammasome with release of pro-inflammatory cytokines. AVP antagonists might mitigate AVP-mediated inflammatory disorders and hyponatremia and could prove a potential therapeutic modality in treating COVID-19. To this end, extensive structure-activity relationship studies of vasopressin and its analogues have been reported in some recent investigations in 2020–2021 [54].

Another class of short endogenous peptides is opioids, produced mainly in the central nervous system (CNS). They are delivered to the target tissues where they act as neuromodulators [55]. For example, endorphins and enkephalins are involved in several physiological pathways, including pain regulation. Endorphins (~20 different derivatives), ubiquitous within the body, are neurotransmitters derived from the prohormone pro-opiomelanocortin (POMC), and their opioid activity is due to the C-terminal binding to opioid receptors [56]. Enkephalins are encoded from the gene pre-proenkephalin A and encompass short peptides such as 5-methionine (Met⁵)-enkephalin and 5-leucine (Leu⁵)-enkephalin. Their main functions as opioid receptor agonists are neurotransmission and pain modulation and are involved in analgesia treatment [57]. The activity of enkephalin analogues was shown to be dependent on the C-terminal amino acid, which can furthermore influence the selectivity of opioid receptors [58].

The great therapeutic potential of endogenous peptides, attributed to their interaction and activation of G-protein-coupled receptor (GPCR) targets, has drawn immense interest from researchers, and the major challenges in their development [i.e. metabolic stability, blood-brain barrier (BBB) permeability, and bioavailability] are in high demand [59].

HDPs are crucial components of innate and adaptive immunity with antimicrobial properties. Natural HDPs are synthesized both ribosomally as well as non-ribosomally, by all complex animals, insects, and plants, and generally have adequate direct activity against a broad range of microorganisms [60]. They kill bacteria in response to external stimuli and involve interconnected signaling pathways [61]. The antimicrobial activity of some HDPs could be linked to their immunomodulatory functions. However, the role of HDPs in immunity is very complex, involving various receptors, signaling pathways, and transcription factors, and remains a field of active research. The promising therapeutic potential of HDPs as antimicrobial agents against diverse pathogens are drawing attention [62]. The antimicrobial peptide (AMP) database (https://aps.unmc.edu) encompasses a large collection of natural AMPs (~3,000) [30]. Notably, only seven AMPs are clinically used: daptomycin (natural and synthetic cyclic lipopeptide, 13 amino acids, inhibition of bacterial growth), gramicidin S (natural cyclic peptide, 12 amino acids, membrane disruption, and depolarization) [63], colistin (natural cyclic peptide, 10 amino acids, membrane lysis) [64], oritavancin (nature-inspired lipoglycopeptide, 7 amino acids, membrane lysis and inhibition of cell wall synthesis) [65], dalbavancin (naturally-inspired lipoglycopeptide, 7 amino acids, inhibition of cell wall synthesis) [66], telavancin (nature-inspired, lipoglycopeptide, 7 amino acids, membrane lysis and inhibition of cell wall synthesis) [67], and vancomycin (natural, 7 amino acids, inhibition of cell wall synthesis) [68]. Colistin was first approved for acute and chronic Gram-negative bacterial infections, and despite its significant renal and neurologic toxicity, it remains the last resort treatment for some multi-drug resistant (MDR) species such as Pseudomonas aeruginosa and Acinetobacter spp. [69].

Recently, two reviews [43, 70] have summarized their classification based on their chemical structure (linear, cyclic, and lipopeptides), and depicted their potential applications, including as immunomodulator agents.

The immense scaffold complexity and diverse mechanisms of action make AMPs the potential solution to fight the huge issue and challenge of antibiotic resistance [43]. The mesmerizing ability of certain HDPs in enhancing innate immune responses to control infections as well as control inflammation makes them attractive therapeutic candidates for both anti-infective and anti-inflammatory indications. There are, however, the challenges—the ‘usual suspects’ in short peptide drug development, i.e. limited bioavailability, associated toxicity, and high manufacturing costs of cationic peptides that the scientific community needs to address urgently.

NRPs represent an important source of short peptides with therapeutic relevance produced by a different pathway to the traditional ribosomal biosynthesis [71]. They usually consist of less than 10 amino acid residues and are characterized by a complex structural diversity, such as the incorporation of non-proteinogenic amino acids, alternative linkages, and post-translational modifications [72]. Indeed, their structural complexity and diversity present a challenge for organic chemists. Interestingly, NRPs are more resistant to proteases and therefore display higher plasma stability in vivo.

The antibiotic cyclosporin A (Figure 2) and its analogues [73, 74] present a fine example of how the developments in synthetic chemistry pave routes to complex molecules endowed with a broad range of biological activities. Some natural NRPs with therapeutic applications include penicillin precursor ACV-tripeptide, daptomycin, cyclosporin, vancomycin, and bacitracin.

The NRP sequences often are decorated with non-proteinogenic amino acids. Some of the latter are shown in Figure 3. The discovery of more than 800 non-proteinogenic amino acids, unique building blocks typical of NRPs that confer them a great scaffold diversity, has revolutionized the development of therapeutic peptides, especially AMPs against MDR pathogens [75].

In general, cells internalize hydrophilic macromolecules via an energy-dependent mechanism, called endocytosis. Alternatively, macromolecules can pass across the membrane in an energy-free pathway (passive diffusion) [93]. Several energy independent mechanisms for cell-penetrating peptide (CPP) internalization are known, including the pore-formation or the barrel-stave model [94], the carpet model [95], and the toroidal-pore mode [96]. Although some larger proteins possess the ability to serve as transporter after conjugation (i.e. cargoes) with peptides, peptide nucleic acids (PNA), and oligonucleotides, the discussion here is focused on short CPP with sequences up to 45 amino acids.

The possibility to target biological molecules intracellularly with short peptides is a very significant task that may confer a number of applications in disease diagnosis and therapy [97]. However, the low membrane permeability of peptides remains an obstacle in biomedical research, and therefore strategies to deliver them to intracellular targets are highly desired. To this aim, different delivery approaches, such as the use of liposomes [98] and nanoparticles [99], were developed to translocate drugs to the targets, but their potential immunogenicity and toxicity remain of concern [100]. CPP are short peptides with particular physicochemical properties that enable them to enter cells and serve as a cargo delivery system [101]. The online repository of CPPs (https://webs.iiitd.edu.in/raghava/cppsite/) contains 1,855 sequences (number of CPP updated in February 2023), of which 1,753 are linear, and 102 are cyclic peptides. This database provides a lot of information, including chemical modifications, in vitro/in vivo model systems, and a list of several cargos delivered by CPPs. In fact, CPPs have been extensively employed for efficient intracellular delivery of nucleic acids [102], proteins [103], imaging [104], and anti-cancer agents, as well as small molecules [105]. CPPs are usually classified based on their source, i.e. if they are protein-derived (41.73%), synthetic (54.82%), or chimeric (3.45%). The composition of short peptides including their sequence and nature of amino acid residues defines their physical/chemical properties and their ability to cross the membranes. In general, they can be broadly distinguished as hydrophobic, cationic, and amphipathic CPPs. Penetratin, a 16 amino acids peptide derived from Anthennapedia homeodomain, was the first short peptide that displayed an ability to penetrate the plasma membrane [106].

According to their therapeutic use, CPPs find important applications in anticancer therapies [107], in drug resistant infections [108], and for diabetes treatment [109]. Several CPP-based therapies are currently under preclinical development [107]. In particular, poly-arginines [110], human immunodeficiency virus (HIV)-trans-activator of transcription (TAT)-derived CPPs, and penetratin are well investigated and considered as the models for the further design of novel CPPs. The naturally occurring tat sequence i.e. RKKRRQRRR is derived from TAT and is a cationic peptide that has been extensively used in imparting cell permeability characteristics to the peptides and also is employed as a delivery system. The TAT-modified tobacco mosaic virus (TMV-TAT) was developed by decorating the TAT peptide on the exterior surface and the method was used to explore novel small interfering RNA (siRNA) carriers [111]. Introduction of CPP enhanced cell internalization and thereby TMV-TAT acquired endo/lysosomal escape capacity without inducing lysosomal damage, resulting in both high efficiency and low cytotoxicity. Several other examples are being studied for their use in cancer treatment [112], including the intracellular delivery of chemotherapeutic agents and thus improving their efficiency.

In vivo, the application of CPPs is hampered by low tissue selectivity and in vivo stability. Ideally, for CPP target delivery, the short half-life could be an advantage as they are fast degraded and cleared easily after delivery. However, various research groups have explored structural modifications of peptides to confer them a better cell entrance and more in vivo stability, for example by tuning their net charge (i.e. cationic peptides) or by the conjugation with hydrophobic moiety able to interact with membranes. In 2003, Elmquist and Langel [113] studied the cellular internalization and stability of peptide vascular endothelial-cadherin (pVEC), a CPP derived from murine vascular endothelial cadherin, containing 18 amino acids that act as a vector carrier crossing the membrane with a non-endocytotic mechanism and without any toxicity. Its half-life was determined to be 10.5 min in PBS with trypsin, and 44 min in PBS with carboxypeptidase A. The designed all-D analogue was however not degraded by these proteases, and could even maintain a comparable in vitro uptake.

Further improvement of the cell-penetration ability may be achieved either by masking the backbone amide groups, by incorporating a guanidinium pattern, or by stapling the side-chains [114]. Enhancing the lipophilicity of peptides (by conjugation with hydrophobic moieties) may help the passive membrane permeability [113–115]. Cyclosporin A (Figure 2), a fungal metabolite with immunosuppressive activity used in organ-transplant patients, is a highly N-methylated cyclic short peptide that is orally available [116], and that has inspired the design of passively penetrating peptides [117].

Somatostatin analogues are short peptides that can be used in peptide receptor radionuclide therapy (PRRT). Somatostatin receptors (1–5 types) are over-expressed in tumoral tissues [118–120]. Radiolabelled somatostatin analogues [labelled with radioactive substance like Yttrium-90 or Lutetium-177] have been employed in targeted radiopeptide therapy and proven to significantly improve survival and quality of life in patients suffering from metastatic or unresectable neuroendocrine tumors (NETs) [121]. In a recent work, in vivo, experiments were performed wherein tumors were targeted by the chelator 1,4,7,10-tetraazacyclododecane-1,4,7,10- tetraacetic acid (DOTA) coupled with highly potent somatostatin receptors agonists or antagonists [122]. The somatostatin receptor subtype-3 (sst₃) antagonist DOTA-sst₃-ODN-8 (Figure 4) showed no internalization but high binding affinity and selectivity, preventing agonist-stimulated internalization. Although the agonist ^natIn-DOTA-NOC can induce sst₃ internalization, all antagonists labelled 76 times more sst₃ sites in cultured HEK-sst₃ cells than the agonist, plausibly owing to their higher hydrophobic content in antagonists than agonists, and thus, a stabilization in the lipidic bilayer of receptors, and a longer time of action.

Many AMPs respond to the offensive environment by adopting mechanisms that render them permeable to bacteria [94], and they are found both in vertebrates and invertebrates. In these peptides, the secondary structure and conformational changes lead them to lyse the microbial membrane. It was demonstrated that the α-helix content is not necessary for selective membrane lysis [123]. A non-cell selective lytic natural short AMP from bee venom, melittin (26 residues), possesses a broad-spectrum bactericidal activity against several clinical strains [124].

Recently, the use of CPP-conjugated drugs is emerging as a novel strategy to combat antibiotic resistant infections, drug-resistant tumors, and other pathologies. For instance, conjugation of AMPs with CPPs enhances the antibacterial activity [108]. Lee et al. [125] described the generation of arginine rich magainin (R9-megainin, RRRRRRRRR-GGGGIGKWLHSAKKFGKAFVGEIMNS) that showed higher antimicrobial activity against gram-negative bacteria than magainin. The same effect was observed in R9-M15 (RRRRRRRRR-GGGKWKKLLKKPLKLLKK). Notably, R9 alone does not possess antimicrobial activity.

Conjugation of CPP with glycopeptide antibiotics is also interesting example reported in the literature [126]. Indeed, in 2018, Antonopolis and co-workers reported the conjugation of vancomycin with octa arginine, conjugation of vancomycin-D-octaarginine (V-r8) (Figure 5), that exhibited stronger activity and superior efficacy than vancomycin alone against gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA) biofilms and persister cells [126]. Persisters-active peptide based drugs are generally polycations that unselectively disrupt the bacterial cell membrane curvature. The eptide-nucleoside Relacin [127, 128] and its analogues have been the first example of a tailored approach against a specific bacterial target (Rel proteins) [129] that unfortunately showed limited activities and cell penetrating properties. In this regard, we have been active in the targeting of Rel proteins with a structure-based approach [130–132] that led to interesting amino acid-derived compounds.

In 2019, the synthesis of vancomycin analogues with the addition of single amino acid residues was reported demonstrating that only one arginine, V-r, led to significantly improved efficacy against gram-positive bacteria, including MRSA [133].

CPP-conjugation with polymers or with supramolecular nanostructures is an alternative shielding strategy to protect CPPs from proteases attack [134]. Remarkably, D-JNKI-1, a conjugated CPP, is found in vivo application in Alzheimer’s disease since it is able to cross to the BBB and penetrate the brain parenchyma till its target in brain mitochondria [135].

Linares et al. [136] in an interesting recent research described the synthesis of cell-penetrating peptide conjugate (C-POC), a novel peptide-platinum (IV) CPP with a high anticancer efficacy and a reduced impact on non-tumoral cells compared to the standard platinum-based therapy.

As the oral bioavailability of a molecule is determined by its PK properties, several factors need to be considered in designing orally available peptides. After oral administration, only the fraction that passes through epithelial cells, blood circulation, and reaches the target, is the amount that correlates with the efficacy of the drug [137]. In general, natural linear peptides are not orally available since they are rapidly attacked by proteases, leading to their degradation and clearance. Among the first strategies towards improved oral availability involved using constrained conformations of peptides which make them less prone to proteolytic degradation [138]. Indeed, the constrained conformation induced by cyclization can prevent endopeptidase cleavage of peptides, while other modifications can avoid exopeptidase deactivation [139].

Subsequently, several strategies including prodrug approaches, PEGylation [140], cyclization [138], lipidation [141], and other chemical modifications (i.e. the N-methylation and the use of D-amino acids mentioned above), have been extensively used to improve peptide stability and bioavailability.

Interestingly, the “Lipinski rule of 5” prescribed for orally available small molecule drugs [142, 143], cannot be applied to peptides. An exhaustive review collecting 125 cyclic peptides reported that only a few peptides are marketed drugs, and even fewer are orally absorbed [144]. The commercially available orally administered peptides are all short and cyclic peptides (e.g., cyclosporine—the immunosuppressant, desmopressin—the antidiabetic, linaclotide—for chronic idiopathic constipation, vancomycin hydrochloride and colistin sulfate—antibiotics, taltirelin hydrate—for spinocerebellar degeneration, glutathione—for AIDS-related cachexia and cystic fibrosis and tyrothricin—for pharyngitis) [145]. In 2019 FDA approved the Novo-Nordisk insulin, semaglutide, as the first orally available GLP-1 receptor agonist. Other oral insulins are in phase I–III trial studies [30].

The scientific data of orally available molecules suggest to reducing hydrogen bond donors (HBDs) by the replacement of amide backbone with N-alkyl groups or other surrogates in peptides and the incorporation of nitrogen into heterocycles appended around the backbone. In 2008, Biron et al. [117] reported the synthesis of 31 orally available somatostatin analogues by N-methylation on the cyclic hexapeptide named Veber-Hirschmann peptide c(PFwKTF) [146] (Figure 6A), selective for somatostatin receptors 2 and 5. Indeed, the short half-life of somatostatin (< 2 min) prevents its use as a drug. Its fast degradation is attributed to the secondary structure (conformation) and to the presence of natural unmodified amino acids in the sequence. Further modifications on somatostatin structure (Figure 6B) based on conformational rigidity, led to the development of the improved drug octreotide.

Octreotide (Sandostatin^®) is used in acromegalia treatment and in intestinal endocrine tumors [147]. Its higher metabolic stability (half-life 110 min) is mostly owed to the β-turn stabilization induced by the incorporation of substitution with a D-amino acid [148]. Although this synthetic somatostatin analogue possesses a longer half-life, it is still not orally bioavailable. This demonstrates how demanding the task of taming peptides into oral drugs and calls for further research.

Lipopeptides, i.e. oligopeptides linked to a fatty acid portion, are a promising family of drugs as they encompass key advantages, including an extended half-life, the ability to cross the BBB easier than non-lipidized counterparts, and to bind plasma proteins [149, 150]. In particular, their binding to plasma proteins (such as albumin) [151] confers protease stability. Daptomycin (Cubicin^®, Figure 7), a branched cyclic lipopeptide synthesized by non-ribosomal bacterial synthetases, is the only antibiotic lipopeptide approved for clinical use [152]. Its calcium-dependent mechanism of action results in the depolarization of the bacterial membrane in anaerobic and aerobic Gram-positive strains [153]. Daptomycin is intravenously administered once daily (half-life 8–9 h) to treat MDR infections.

In 2022, Wan et al. [154] described the advances in the Iturins family, the cyclic lipopeptides extracted from Bacillus subtilis with a broad-spectrum activity against fungi and bacteria, and they are mostly used in plant infections and pesticides production [155]. The iturin family is characterized by the peculiar presence of β-amino acids and fatty acid chains, but their clinical use is limited as they can cause hemolysis in human erythrocytes [156].

Liraglutide, a palmitoylated GLP-1 possesses a very long half-life after subcutaneous administration (11–15 h) [157], attributed to albumin binding [158], allowed its marketing for the treatment of T2DM and obesity.

Lipidated hormones are a distinguished class of lipopeptides. Ghrelin (28 amino acids and an octanoyl acid moiety)—the hunger hormone, is the most representative example of a natural lipopeptide, produced in the human stomach and is among the very few peptides able to penetrate the BBB [159]. Ghrelin is active only if modified with octanoic acid and it finds application in eating disorders, obesity, and in growth-hormone deficiencies.

Development of chemical ligation methods towards chemoselectively linking unprotected amino peptide fragments in an aqueous medium was a major leap in the synthetic methodology of peptides [163]. The peptide ligation methods encompass reactions that involve the reactivity of the α-nitrogen, or the participation of the side chain (Figure 8A–C) [187].

Saxon et al. [168] and Saxon and Bertozzi [188] reported modifications of the Staudinger ligation, the chemoselective formation of an amide group from a functionalized phosphine and an azide (Figure 8D). They introduced a new reaction (called “traceless” Staudinger ligation), that uses a phenol ester and N-acylimidazole phosphines, to rapidly form amide bonds. Staudinger-Bertozzi ligation has found many applications in peptide synthesis as well as in enabling the study and manipulation of biological processes. For instance, in 2013, the Staudinger-Bertozzi ligation was found to be ideal for the incorporation of molecular photoswitches such as azobenzenes into biomolecules [189].

In general, the Staudinger ligation is a widely used approach to conjugate biomolecules for in vitro therapeutic targeting, but some drawbacks regarding the biocompatibility, such as the incomplete removal of unreacted phosphane-biotin that interferes with background fluorescence [190], had been reported [191, 192]. More biocompatible techniques were recently explored, introducing the so-called click chemistry that employs copper(I) [3 + 2] N₃ cycloaddition reactions [193–195].

In 2006, de Araújo et al. [166] reported a novel strategy for a site-specific chemoselective ligation of peptides and proteins by applying Diels-Alder cycloaddition reaction executed between a dienyl ester and a maleimide dienophile under mild conditions (Figure 8F). Other selective ligations have explored reactions of hydroxylamines and hydrazines with aldehyde-ketones to form oximes and hydrazones [196, 197], and by reactions between aldehyde with cysteine (or serine), via imine ligation, that rearrange to form a stable pseudoproline (Figure 8G) [198, 199]. An important evolution of NCL is presented by the approach involving N-terminal serine and threonine residues (Ser-Thr peptide ligation) [200, 201]. Recently, this method was expanded in an elegant one-pot ligation-desulfurization reaction by the exploration of C-terminal proline salicylaldehyde esters [202]. The resulting reaction between N-terminal penicillamine and C-terminal proline salicylaldehyde esters was proposed as an example of cysteine/penicillamine ligation, useful for chemical protein synthesis [202].

Post-synthesis covalent modifications with an orthogonal double crosslinking of genipin and sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (Sulfo-SMCC; through the formation of a maleimide) were described by Ciulla et al. [179] (Figure 9) to increase the stiffness in linear short peptides and in order to induce a higher β-sheet content capable of mimicking physiological microenvironments, such as the extra-cellular matrix (ECM), and with constructs closer to stiffnesses of cardiac tissue, osteoids, and skin [179].

Structural modifications can improve the physicochemical properties of peptides. Modifications that confer rigid conformations, as well as proteolytic resistance are the desired ones. To this end, backbone modification and cyclization are two classical strategies applied to stabilize peptides in their bioactive conformation and improve resistance to biodegradation. Cyclization aims to introduce a conformational constraint into the recognition sequence and the use of non-standard amino acids and/or small molecule fragments etc. enhance their stability into the bioactive conformation, decreasing the entropy loss due to receptor binding and, therefore, increasing their affinity for the target [203].