Natural AFPs are produced by a number of different species of Bacteria, Archaea, and Eukarya isolated from natural sources (De Lucca and Walsh, 1999). Most natural AFPs have been discovered by testing their antagonistic activity in vitro against pathogenic fungi (Mania et al., 2010; Freitas and Franco, 2016; McNair et al., 2018). However, with the rise of sequencing technologies and the drop in associated costs, new strategies for prediction and discovery of new AFPs are emerging. New methods such as template-based, docking simulations, hidden Markov model and other sequence-based methods allow for novel in silico prediction of AFPs (Schneider and Fechner, 2005; Fjell et al., 2007, 2008; Robinson, 2011; Garrigues et al., 2017; Agrawal et al., 2018).

Natural AFPs are grouped in families according to their origin (Table 1). Those produced by bacteria and fungi are arguably of greatest interest with respect to medical applications (Essig et al., 2014). Although some other sources, such as plants, are a rich source of AFPs, they are mainly employed for other purposes, such as the control of phytopathogens, and thus, they are not further discussed in this review (De Lucca, 2000). Notably, the first archaeal AMP with antifungal and anti-biofilm capabilities against clinical fungal pathogens was recently reported (Roscetto et al., 2018).

Typically, natural AFPs adopt an α-helix structure, β-hairpin or sheet (containing two cysteine residues) or mixed α-helix/β-sheet structures upon interaction with membranes. Some of these peptides are rich in specific amino acids and, as a result, are classified as glycine-rich, proline-rich, arginine-rich, histidine-rich, and tryptophan-rich (Bondaryk et al., 2017). However, the structure of most AFPs has not yet been determined.

Posttranslational modifications also play a major role in the final three-dimensional structure and bioactivity of AMPs but, unfortunately, cannot be predicted with current in silico tools (Agrawal et al., 2018). The most common posttranslational modifications in natural AFPs involve glycosylation or the addition of carbohydrates, normally observed in asparagine, or serine/threonine residues (Guo et al., 2012; Bednarska et al., 2017). The latter has been shown to improve antifungal activity in some cases. Other modifications include halogenation such as chlorination (Andreu and Rivas, 1998; Shinnar et al., 2003), phosphorylation, which can increase stability (McDonald et al., 2011) or hydroxylation, mainly observed at lysine, arginine, tryptophan, and phenylalanine residues with differing effects on antifungal activity (Houwaart et al., 2014; Akkam, 2016). Finally, cyclization of AMPs, which is not considered a posttranslational modification, improves antimicrobial activity in general, reduces toxicity and improves stability against proteases (Akkam, 2016).

Antimicrobial semisynthetic and synthetic peptides are generally made with a view to improving pharmacological properties, reducing side effects and/or lowering the immunogenicity of natural peptides. Pharmaceutical formulation will also help to enhance their stability and bioavailability. As an example of synthetic transformation, it was found that the hemolytic activity of the natural echinocandin B AFP was significantly reduced by the replacement of the linoleoyl side-chain with either octyloxybenzoyl (cilofungin) or pentyloxyterphenyl (anidulafungin) side chains (Emri et al., 2013).

Structure-activity relationships (SAR) are a key element used for the design and development of synthetic peptides (Lum et al., 2015). There are several biophysical properties that can determine antifungal activity, such as net charge, stereospecificity, hydrophobicity, amphipathicity, secondary structure and peptide length, with some of these characteristics being interdependent. These properties have been extensively reviewed previously (Akkam, 2016). Most AFPs are cationic, but neutral and anionic AFPs have also been described. When present, cationic charges play a role in the electrostatic binding of AMPs to negatively charged membranes. Therefore, an increase in the positive net charge beyond a threshold might lead to a stronger activity on the membrane. However, positive charges are not a prerequisite for antimicrobial activity, and anionic peptides interact with the membrane peptide through specific amino acid distributions (Yeaman and Yount, 2003; Lakshminarayanan et al., 2014).

Most AFPs are non-stereospecific (Akkam, 2016). In the same way, there is no dominant conformation among AFPs. Thus, the main differences between peptides come from sequence and secondary structure variation (Emri et al., 2013). Hydrophobicity and amphipathicity are essential factors for peptide membrane interactions and membrane permeabilization (Lum et al., 2015), and important variables in the design of synthetic peptides. Hydrophobicity is defined as the percentage of hydrophobic residues within a peptide, ranging between 30 and 60% for most AMPs. An increase in hydrophobicity and amphipathicity usually correlates with an increase in antifungal activity, but also with higher hemolytic activity. The presence of tryptophan is also linked to hemolytic activity since it interferes with lipid polymorphism in the membranes (Schibli et al., 2002).

The peptide length of AFPs is important for the secondary structure and mode of action. Most AFPs have 11–40 residues. It has been described that 7–8 amino acids are needed to form amphipathic structures in AMPs (Bahar and Ren, 2013), while <20 amino acids limit the ability of a peptide to form transmembrane structures in the fungal membrane (Rothstein et al., 2001; Akkam, 2016). Nevertheless, longer lengths may also affect cytotoxicity, stability and manufacturing costs. In order to overcome these hurdles, short antimicrobial peptides (SAMPs), containing 2–10 amino acids, are attracting attention as less toxic and more stable alternatives. Overall, SAMPs have simpler amino acid composition and they are easier to synthesize and modify chemically with a view to improving toxicity, stability, half-life or specificity. Moreover, they are also less immunogenic (Duncan and O'Neil, 2013; Fox, 2013).

Some of the semisynthetic and synthetic peptides studied with a view to clinical applications are summarized in Table 2. Considering all of the factors mentioned above, different strategies have been employed for synthetic peptide design. Combinatorial libraries generate model peptides suitable for SAR studies (Blondelle and Lohner, 2000). The template-based strategy, known as de novo design, uses peptides with known antifungal activity as scaffolds (Lum et al., 2015). As an example, Agrawal et al., performed in silico studies by machine learning techniques based on AFP and non-AFP sequences revealing a higher frequency of certain residues (C, G, H, K, R, and Y) and the prevalence of R, V, K at N-terminus and C and H at the C-terminus of peptides (Agrawal et al., 2018).

Another strategy is to target virulence traits (Bondaryk et al., 2017). However, this approach is not efficient enough due to the expression of different virulence factors in fungi cells. For this reason, the generation of molecules with multiple ligands like dendrimers, or tree-like molecules, that have radially distributed layers of branches named generations, have been explored (Esfand and Tomalia, 2001).

Most AMPs affect a number of organisms including bacteria, fungi, and envelope-containing viruses. The most representative non-specific AMPs can be grouped as linear peptides (e.g., cecropins, magainins, cathelicidins, bombinins, lactoferrin-derived peptides) or cyclic peptides (e.g., mammalian defensins, poultry gallinacins, macrocyclic peptides, syringomycins, tachystatins) (Matejuk et al., 2010). Membrane disruption by the formation of toroidal pore is common to most peptides, resulting in leakage of essential molecules and ions with general loss of membrane functionality. Additional mechanisms include membrane thinning, formation of non-bilayer intermediates by interchelation of peptide-membrane, formation of anionic lipid-peptide domains, lipid flip-flop, demixing and clustering, and alteration of membrane potential (Shai, 2002; Bechinger and Lohner, 2006; Huang, 2006; Melo et al., 2009; Nguyen et al., 2011; Rautenbach et al., 2016). In addition, some of the peptides may have other effects such as DNA damage, apoptosis induction, inhibition of DNA replication and RNA or protein synthesis.

For supplementary reviews on the mechanisms of action and classification of antimicrobial peptides see reviews by Matejuk (Matejuk et al., 2010), van der Weerden (van der Weerden et al., 2013), and Rautenbach (Rautenbach et al., 2016).

Existing antifungal agents exhibit a diversity of drawbacks that reduce their efficiency as therapeutic tools against fungal infections. The existence of only a few approved classes of antifungal drugs and the increasing antifungal resistance further complicates the selection of an appropriate antifungal therapy (Sanguinetti et al., 2015; Pappas et al., 2018). Ergosterol synthesis, ergosterol in membranes and cell wall synthesis are the targets of azoles, polyenes and echinocandis, respectively. The low diversity of mechanisms of action of current treatments represents a problematic situation when fungal pathogens are multi-resistant to antifungals (Rautenbach et al., 2016). Furthermore, existing antifungal drugs are associated with adverse drug reactions such as hypokalemia, infusion reaction, nephrotoxicity, hepatotoxicity and gastrointestinal affections, among others (Chen et al., 2011; de Souza et al., 2016; Kyriakidis et al., 2017; Pappas et al., 2018). Some antifungal drugs affect common eukaryotic targets present both in pathogenic fungi and human cells (Rautenbach et al., 2016). This makes the development of novel efficient and non-toxic antifungal therapies more difficult than that of antibacterial ones.

AFPs have diverse advantages over the current antifungal drugs, which are directly related to their mechanisms of action and molecular targets. AFPs are particularly promising because they can recognize multiple microbial targets, thus reducing the possibility of resistance development (Rautenbach et al., 2016), a topic that is discussed in more detail in the following section. These microbial targets include fungal membranes, different cell wall components and molecules related to physiological processes, such as RNA, DNA and protein synthesis and cell cycle (van der Weerden et al., 2013; Bondaryk et al., 2017).

Regarding the absence of side effects, several AFPs target specific conserved fungal molecules, such as glucosylceramide, mannosyldiinositol phosphorylceramide, enzymes related to ergosterol or β-glucan synthesis, among others. This translates into high pathogen selectivity and reduces the probability of cytotoxicity against mammalian cells (Rautenbach et al., 2016). However, it does not ensure the absence of cytotoxicity (e.g., aculeacins, a type of echinocandin that targets the fungal 1,3-β-glucan synthase, displays hemolytic activity) (Matejuk et al., 2010). Over the last 13 years, three echinocandins, anidulafungin, caspofungin, and micafungin, have been approved in Europe and USA (Pound et al., 2011; Beyda et al., 2012). These peptides are poor substrates for cytochrome P450, which translates into both lower hepatotoxicity and pooled risk for discontinuation of treatment (3.7–4.8%) compared to other antifungals (Kyriakidis et al., 2017). The synthetic peptide killer peptide (KP) is another example that has demonstrated very promising antifungal activity without cytotoxicity against peripheral mononuclear blood cells in vitro or side effects in murine trials. This peptide seems to have a specific interaction with 1,3 β glucans and only the dimeric form of the peptide is active (Magliani et al., 2011).

In addition, AFPs show reduced cytotoxicity (Matejuk et al., 2010). Two reasons may explain this phenomenon. Firstly, there is a stronger interaction between the negatively charged fungal membrane (due to the higher content of phosphatidylinositol and phosphatidic acid) and the cationic charges of the peptides, in contrast to mammalian cell membranes, which are predominantly neutral to host mammals (due to the high content of phosphatidylcholine). Secondly, some AFPs target membrane lipids unique to fungi and absent from mammalian cells, which also reduces toxicity (Nguyen et al., 2011; Rautenbach et al., 2016).

Accumulating evidence demonstrates that the therapeutic activity of AMPs is multifactorial and not mediated only by their direct antimicrobial effect. Host defense peptides (HPDs), like defensins and cathelicidins families, often exert angiogeninc, immunomodulatory and anti-inflammatory effects and may also induce the recruitment of the adaptive immune response, which has been reported in several papers and reviews (Zasloff, 2002; Hirsch et al., 2008; Steinstraesser et al., 2009; Magliani et al., 2011; Hsieh and Hartshorn, 2016; Li et al., 2017).

Membrane remodeling of C. albicans has been associated with its resistance to current non-peptide antifungal drugs, mainly ergosterol-sphingolipid-rich lipid rafts containing multi-drug resistance (MDR) proteins attached to the membrane (Mukhopadhyay et al., 2004; Pasrija et al., 2005, 2008; Shahi and Moye-Rowley, 2009). In the case of resistance to AMPs and AFPs it is important to note, as discussed previously, that these peptides frequently function through membrane interaction, but that additional modes of action for microbial inhibition have also been demonstrated (Wu et al., 1999). Microorganisms, such as fungi, evolve rapidly, and can adapt quickly when exposed to antibiotics and antifungal drugs. However, it is important to note that cell membranes are slower to evolve. The rapid and potent effect on membrane, coupled with other inhibitory mechanisms exhibited by AFPs, de novo resistance is less likely to emerge in target microorganisms (Yeung et al., 2011).

As is the case for currently used antimicrobial drugs, the overuse of AFPs could accelerate the occurrence of fungal resistance. This issue complicates the application of antifungal therapeutics. Indeed, extensive echinocandin usage in hospitals has led to an increase in the number of strains with acquired (secondary) resistance to these first-line antifungals, especially among strains of C. glabrata (Sanguinetti et al., 2015; Pappas et al., 2018). The potential for the emergence of high level resistance to AMPs has been debated, but it is likely to occur at a reduced rate relative to that observed for other antifungals, though it will depend on how the antimicrobial peptide is administered. Besides, although a lower target concentration is required, the absence or change in the specific fungal target through spontaneous mutation can naturally lead to resistance. However, such modification of conserved molecules could, in turn, result in reduced pathogen virulence. In addition, combination therapy, involving the use of AMPs with antibiotics or with other peptides, will likely reduce the development of resistance markedly (Matejuk et al., 2010; Rautenbach et al., 2016).

Currently, there are not many AMPs that are produced from their natural sources for clinical use. The purification of peptides from natural sources is laborious and expensive due to their low abundance and the multiplicity of compounds present in those sources (Vriens et al., 2014). Because of this, most industrial-scale productions of AMPs are performed by heterologous expression or chemical synthesis. Brief descriptions of two approaches for industrial-scale AMP production from natural sources are presented here: microbial fermentation and proteolysis of food proteins.

Echinocandins are exceptional examples of AMPs produced by microbial fermentation, and further chemical modification in case of the semisynthetic variants. Industrial-scale production of echinocandins is based on fermentation since they possess a high-complexity chemical structure. However, little information has been published about these processes. Echinocandin B, pneumocandin B₀ and FR901379 are the natural echinocandins produced for commercial purposes from A. rugulosus, Glarea lozoyensis and Coleophoma empetri, respectively. The optimization of the fermentation process is essential to obtain a competitive product, since the fermentation and purification costs of natural echinocandins are the main variables that influence the overall production costs of semisynthetic derivatives. The clinical applications of natural echinocandins is limited by certain undesirable properties, such as a strong hemolytic activity, as discussed previously (Emri et al., 2013).

Industrial scale AMPs can also be obtained by proteolysis of food proteins and the release of encrypted peptides. Agyei and Danquah (2011) offered a brief description of the process for manufacturing pharmaceutical-grade peptides by this approach. The process involves, firstly, the acquisition of raw materials: food protein and proteolytic enzymes/or microorganisms. By-products from dairy, fish and meat industries are suitable cheap sources for proteins (Sibel Akalin, 2014; Ryder et al., 2016). The second step involves protein hydrolysis. The use of enzymatic hydrolysis is preferred over in situ microbial fermentation, particularly in food and pharmaceutical industries due to the lower or absent output of organic solvents and toxic chemical in the process and product (Sibel Akalin, 2014; Ryder et al., 2016). Under industrial-scale conditions, the use of immobilized enzymes offers several advantages over the conventional soluble enzymes, such as milder and controlled conditions and recycling of enzymes used (Sewczyk et al., 2018). The final step is fractionation and isolation of the bioactive peptides. Ultrafiltration, precipitation with solvents and liquid chromatography techniques (e.g., ion exchange, gel filtration) have been proposed for purification of peptides, however, their current implicit high costs make them prohibitive for large scale applications. It is estimated that up to 70% of the capital and operating costs in industrial biotechnology processes may correspond to the separation and purification stages (Brady et al., 2008). Electro-membrane filtration (EMF) is being established as an alternative method for the purification of bioactive peptides. It combines electrophoresis with conventional membrane filtration, being more cost-effective than chromatographic techniques (Bazinet and Firdaous, 2013). AMPs with antifungal activity, such as casocidin-I, kapacin A and lactoferrin-derived peptides can be produced by these methods as well.

Considering the often low amounts of AMPs obtained by purification from natural sources and the high costs and difficulties that may arise from chemical synthesis (Li et al., 2010; Hou et al., 2017), recombinant production of AMPs provides a solid option to make these peptides accessible at low cost and high efficiency.

Genetically modified microorganisms facilitate the production and functional expression of any bioactive molecule but, importantly, also allow the production of bioengineered and encrypted peptides that would not be achievable otherwise (Wibowo and Zhao, 2019). Moreover, new sequencing technologies have made available a vast amount of genomic, transcriptomic and metabolic data, providing the means to further explore known and novel AMPs and the rational design of new antifungal peptides (Amaral et al., 2012; Porto et al., 2012; Tracanna et al., 2017).

However, the success of recombinant production can be highly variable. Understanding the composition and physicochemical properties of AMPs influences the selection and design of hosts and expression system. The choice of host, codon bias, protein expression vector, number of copies of the plasmid and fusion proteins can influence the correct synthesis, folding, and secretion of the recombinant peptide through the cell machinery (Deng et al., 2017). There are two other aspects to bear in mind: the toxicity of the AMP for the host and the high instability and susceptibility of peptides to degradation by proteases. To overcome this, AMPs can be synthesized in a form that is fused to another protein (fusion proteins) or in inactive forms (Kosobokova et al., 2016) initially.

Escherichia coli, yeasts (mainly Pichia pastoris) and plants are the most common recombinant expression platforms for biopharmaceutical proteins. Comprehensive reviews have previously reviewed the different heterologous production platforms available for AMPs (Sanchez-Garcia et al., 2016; Deng et al., 2017). Most AFPs produced by recombinant platforms target plant phytopathogens rather than human fungal pathogens. However, many of these AFPs may represent underutilized resources whose antifungal activity against human fungal pathogens is waiting to be discovered. In the next subsection, a brief overview of the production of AFPs in different hosts with clinical applications is provided.

Chemical synthesis of peptides can be divided in two types: solid- (SPPS) or liquid (solution) phase peptide synthesis (LPPS). In general terms, LPPS is suitable for large-scale manufacture of short peptides or structures that are not easily prepared by SPPS. SPPS is generally used for lower scales or to provide mechanistic insights about peptides and offers the potential for the creation and production of more cost-effective antifungal therapies (Matejuk et al., 2010). Currently, Fmoc SPPS is the preferred method for peptide chemical synthesis due to the versatility and low cost of very high-quality building blocks. The diversity of synthetic peptides entering clinical trials has increased over the last 13 years, stimulating advances in Fmoc SPPS technologies in response to the growing demand for medicinal chemistry and pharmacology (Behrendt et al., 2016). Several groups have synthesized linear (Tran et al., 2008; Magliani et al., 2011; Konno et al., 2015; Cools et al., 2017; Park et al., 2018) and cyclic peptides (Mosca et al., 2000; Schaaper et al., 2001; Konno et al., 2015; Ng-Choi et al., 2019) using Fmoc SPPS.

The long process of isolation and characterization of new natural AMPs delays their clinical use. In this regard, Fmoc SPPS is at the forefront in the design of therapeutic peptides since it permits the easy alteration of features such as hydrophobicity, polarity, charge, structure, and it may also enhance activity and overcome the limitations of natural peptides (Freitas and Franco, 2016). The rational design of synthetic sequences is a new approach of relevance, and results from optimizing the sequence and chemical characteristics shared by different AMPs (pharmacophoric patron) (Freitas and Franco, 2016). Ideally, an antifungal peptide agent should be short, as mentioned previously. De novo peptide design may help reducing production costs, potential toxicity and lability, as well as increasing the in vivo activity (Steckbeck et al., 2014).

Unfortunately, Fmoc SPPS is currently far from meeting its potential and still cannot compete with the template-based process of expression for the large-scale demand of therapeutics (Behrendt et al., 2016). However, it should be noted that companies specialized in the large-scale manufacture of peptides are currently being established, claiming productions from multi-10 kg/lot (SPPS) to multi-100 kg/lot (LPPS)¹. This suggests that the versatility of chemical synthesis is on its way to reach the cost-efficiency and scale of peptide expression.

A recurrence of disease and establishment of chronic fungal infections may result when antifungal treatments are not sufficiently effective. Thevissen (2016) proposed that more efforts should focus on combination therapy in addition to screening for novel antifungal compounds. Synergy between the combined compounds is the main objective of combination therapy, thereby increasing their activity and diminishing their toxicity on the host. Another option is combining an antimycotic (either currently used or novel AMP) with an enhancer molecule that, for example, weakens one or multiple pathogen tolerance mechanisms, such as biofilms, without direct antifungal activity. Either way, it is essential to demonstrate efficacy and safety of the combination before proceeding to clinical trials.

Limited data from clinical trials are available in this regard. A study performed by Candoni et al. showed for the first time that early mortality of patients with invasive aspergillosis was reduced by combined treatment with two antifungal agents (Candoni et al., 2014). In a retrospective study published in 2017, Lee et al. analyzed records from a pediatric department in South Korea and described how the combined therapy of voriconazole and caspofungin was an effective and safe treatment for children with leukemia (Lee et al., 2017). In vitro and in vivo studies suggests that AFPs are excellent candidates for this type of approaches and have a great potential for clinical success. Lactoferrin-derived peptides such as Lf(1-11) and bLfcin have shown synergy with azole antifungal drugs or amphotericin B, greatly reducing the minimal inhibitory concentrations (MICs) against C. albicans, C. glabrata, C. krusei, C. parapsilosis, C. tropicalis, and fluconazole-resistant strains of C. albicans (Wakabayashi et al., 1996, 1998; Lupetti et al., 2003; Fernandes and Carter, 2017). Some other examples of these studies are summarized in Table 6.

P113 is a 12-mer peptide developed by the company PacGen Life Science (Vancouver, Canada) as a mouth rinse formulation for the topical treatment of oral candidiasis. Clinical trials from PacGen demonstrated that oral candidiasis was effectively treated by P113, which compared favorably to the efficacy of nystatin, a standard treatments for oral candidiasis. Vaginal, dermatological and ophthalmic applications are on the list of P113 therapeutic potential (Duncan and O'Neil, 2013). Currently, the P113-containing line of products includes oral rinse solution and spray, feminine soothing spray and cleansing wash, and antibacterial hand cream, among others². NP213/Novexatin was the lead product of NovaBiotics of Aberdeen, UK. It is an arginine-rich cyclic cationic peptide based on human α and β defensins (among others). It was used for treatment of toenails stubborn fungal infections such as onychomycosis (patents PCT/GB2006/004890 and PCT/GB2005/003245). Indeed, independent podiatrist analysis determined the treatment against mild to moderate onychomycosis as being 80% clinically effective. This effectiveness rate is significantly higher than that provided by existing topical treatments for onychomycosis in different territories, including United States and Europe (Duncan and O'Neil, 2013; Fox, 2013)³. However, the clinical study did not acomplish the main goal of a Phase IIb study by not showing differences over the placebo treatment under FDA current guidelines⁴.

Despite their promising properties, only a few AFPs have reached the clinical phase. One example is hLF(1-11), a peptide that was developed for the systemic treatment of bloodstream and deep tissue infections produced by fungi and bacteria in severely immunocompromised transplant recipients. However, after favorable safety and tolerability clinical trials of hLF(1–11), no more studies have taken place putting on hold the commercialization of this peptide (van der Velden et al., 2009; Martin et al., 2015; Bruni et al., 2016). Another example is CZEN-002, a synthetic octapeptide derived from alpha-Melanocyte-Stimulating Hormone (a-MSH). CZEN-002 modulates immune and inflammatory responses, and has been shown to kill C. albicans as well (Fjell et al., 2011; Mahlapuu et al., 2016). This AMP had been in phase II clinical trials for the treatment of vulvovaginal candidiasis. However, no recent development has been reported. It was developed by Zengen, Abiogen Pharma and Lee's pharmaceuticals (Duncan and O'Neil, 2013)⁵.

In addition, there are many cases of peptides with promising properties currently being evaluated pre-clinically (Koo and Seo, 2019). Other pre-clinical examples are described in the review by Duncan and O'Neill (Duncan and O'Neil, 2013) and in the database DRAMP 2.0 (Kang et al., 2019).

As discussed, different formulations and delivery strategies for AFPs might be explored depending on the peptide properties, the potential toxicity (suitable for topical and/or systemic use) and the marketing strategies of the companies (e.g., mouth rinse, toothpaste, spray, dermal cream, etc.). Different carriers can enhance the pharmacodynamics and stability, and reduce toxicity of the active peptide, such as liposome encapsulation or the use of peptoids, the D-conformation-based peptide and β-peptides (Sajjan et al., 2001; Porter et al., 2002; da Silva Malheiros et al., 2010; Chongsiriwatana et al., 2011). Diverse pre-clinical delivery tools and formulations are being studied. In one case, Park et al. developed a pH-responsive and redox-sensitive polymer-based AmB-delivery carrier system (Park et al., 2017), by conjugation with histatin 5 acting both as a synergistic antifungal molecule and a targeting ligand against C. albicans. Other authors describe that some properties of the peptides could make them suitable for self-delivery systems. Examples include the synthetic killer peptide (KP) and the ultrashort peptide NapFFKK-OH, which, at certain pHs and concentration conditions, undergo a self-assembly process. The results are hydrogel-like aggregates that could slowly release the peptides in physiological conditions, as well as reducing the proteolytic susceptibility and increasing the storage stability of the active compound (Magliani et al., 2011; Albadr et al., 2018). Future clinical studies of these hydrogels and other delivery technologies will determine their safety and efficiency. Additional drug delivery strategies include carbon nanotubes and magnetic nanoparticles (López-Abarrategui et al., 2013; Chaudhari et al., 2016).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.