Bacteriocins: An Overview of Antimicrobial, Toxicity, and Biosafety Assessment by in vivo Models

1According to the WHO, diseases caused by multidrug-resistant (MDR) pathogens are a serious worldwide public health problem. The rapid spread of MDR pathogens have reduced the effectiveness of common antibiotics. Therefore, there is a particular need for the development of new antimicrobial agents, specifically those directed against MDR bacteria. Bacteriocins represent the most important group of antimicrobial peptides with applications in human health. The ability of bacteriocins to kill or inhibit relevant pathogenic bacteria (including MDR pathogens) in vitro has been well documented. However, the use of animal models is an important part in the research toward the development of new therapeutic agents. There are many animal models used for screening drugs or chemical compounds in preclinical studies. Mice are the best-known animal models for in vivo bacteriocin efficacy studies. Other animal species, less frequently used, are guinea pigs, rabbits, and hamsters. Therefore, the present review addresses the antimicrobial effects, toxicity, and biosafety of bacteriocins in in vivo systems. The bacteriocins here included are those naturally synthesized by native or recombinant producers, those chemically synthetized or bioengineered bacteriocins, those obtained from direct application of cultures with bacteriocin-producers, and those bacteriocins produced in cell-free supernatants (CFSs). Moreover, we here discuss the toxicity and biosafety in vivo assays, in different animal models, that have been reported during the exploration of antimicrobial effects.

2The death toll caused by drug-resistant infections has been dangerously rising every year within the last two decades. According to, at least 700,000 people died annually around the world because of this. In Europe, deaths raised from 25,000 in 2007 to 33,110 in 2015. Moreover, according to WHO Africa, 54,000 cases of MDR tuberculosis were detected in 42 countries, and approximately 3200 cases of extensively drug-resistant (XDR) tuberculosis were notified in eight countries from the African region from 2004 to 2011. China, one of the most populated countries, reported that between 2005 and 2017, the range of isolated drug-resistant bacteria varied from 22,774 to 190,610. During 2017 in India, a study reported that 70% of drug-resistant Gram-negative isolates corresponded to Escherichia coli, Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa. Regarding the United States, 23,000 deaths were counted in 2013; 6 years later, in 2019, numbers increased to 35,000 deaths associated with antibiotic-resistant infections. In particular, Mexico has a dramatic lack of information and epidemiological surveillance for antibiotic resistance. Recently, a 6-month study assessed the resistance rates of several bacterial pathogens in 47 Mexican centers. They included almost 23,000 strains, and their results showed that the most common drug-resistant strains were E. coli, Klebsiella sp., Enterobacter sp., P. aeruginosa, Acinetobacter sp., vancomycin-resistant Enterococcus faecium, and methicillin-resistant Staphylococcus aureus (MRSA). According to the WHO, if severe actions are not taken by 2050, it is estimated that 10 million people will die annually around the world due to diseases caused by MDR pathogens. This situation may cause a severe economic impact with consequences similar to the 2009 Global Financial Crisis.

3Since their discovery in 1928, antibiotics have effectively controlled bacterial infections. The current treatments for bacterial infections include bactericidal or bacteriostatic agents from different classes of antibiotics. The most common classes of drugs are penicillins, cephalosporins, quinolones, macrolides, tetracyclines, glycopeptides, and monobactams, among others.

4Within the last two decades, a variety of antibiotics have been approved to treat infections caused by Gram-positive and Gram-negative bacteria. Novel antibiotics against Gram-positive bacteria include β-lactams (ceftaroline and ceftobiprole), glycopeptides (dalbavancin, oritavancin, and telavancin), oxazolidinones (tedizolid phosphate), quinolones (besifloxacin, delafloxacin, and ozenoxacin), and tetracyclines (omadacycline). The main advantage of the previously mentioned antibiotics is their efficiency in the treatment of bacterial infections caused by MDR strains. Recently approved treatments to combat MDR Gram-negative bacteria include antibiotic combinations of β-lactam/β-lactamase inhibitors (ceftolozane/tazobactam, ceftazidime/avibactam, and meropenem/vaborbactam), and aminoglycosides (plazomicin) combined with tetracyclines (eravacycline). Nonetheless, misuse and overuse of these combinatorial treatments have contributed to further development of antibiotic resistance.

5The rapid spread of MDR and XDR bacteria in both hospitals and community settings has reduced the effectiveness of antibiotics. One of the main concerns regarding antibiotic resistance is that some bacteria have become resistant to almost all currently available antibiotics. Therefore, these bacteria represent a severe public health problem worldwide. Of particular interest are the following strains: MRSA, vancomycin-intermediate and -resistant; E. faecium, vancomycin-resistant; Enterobacteriaceae, carbapenem-resistant, extended-spectrum beta-lactamase (ESBL)-producing; A. baumannii, carbapenem-resistant; and P. aeruginosa, carbapenem-resistant.

6Preventing and controlling the spread of antibiotic resistance is necessary to invest in research and development of new agents with therapeutic potential to treat bacterial infections. According to data from the WHO, there is an arsenal of 50 antimicrobial agents and their combinations in clinical development. Thirty-two are antibiotics active against the WHO priority pathogens, ten are biological agents, two are classified as innovative agents, and two are active against MDR Gram-negative bacteria.

7There is a particular need for new antimicrobial agents, specifically those directed against antibiotic-resistant bacteria. Defensins and bacteriocins represent the most important groups of antimicrobial peptides with applications in human health. Defensins are small cysteine-rich (forming three to six disulfide bonds) cationic antimicrobial peptides ubiquitous among eukaryotes that form an essential element of innate immunity. They consist of two analogous superfamilies and an extensive convergent evolution is the source of their similarities. Bacteriocins are ribosomally synthesized peptides secreted by a variety of bacteria for the purpose of killing other bacteria. Thereby, whereas defensins are important components of the host immune response against infection in eukaryotes, bacteriocins participate in removing microbial competition in prokaryotes.

8The way to classify defensins and bacteriocins has been based on their biochemical (net charge) and/or structural features (linear/circular/amino acid composition) and there is a current search for common patterns that might help to distinguish them. To determine whether all molecules are homologous or have independently evolved similar features, the best evidence lies in the structure. The antimicrobial peptides can be differentiated from their overall three-dimensional structure and the spacing of half-cystine residues involved in intrachain disulfide bonds. To date, only two bacteriocins (bactofencin and laterosporulin) have been expressed as defensin-like bacteriocins. Bactofencin is a disulfide bond-containing bacteriocin with highly conserved cysteine residues and structurally related to eukaryotic defensins due to their highly cationic nature. The laterosporulin has been previously identified from a Brevibacillus laterosporus strain GI-9 and contain disulfide bonds in positions homologous to eukaryotic defensins. The presence of disulfide connectivity suggests its similarity to β-defensins while its architectural similarity is related to α-defensins. Recently, laterosporuli10, a novel defensin-like bacteriocin (class II bacteriocin) from the Brevibacillus sp. strain SKDU10, was characterized. This bacteriocin showed 57.6% homology with laterosporulin and differences in the molecular weight and the number of cationic amino acids. It was highly efficient in killing S. aureus (Gram-positive bacteria) and the Mycobacterium tuberculosis H37Rv strain when compared to laterosporulin. Interestingly, Class II bacteriocins are easily manipulated by genetic engineering techniques because they do not have large post-translational modifications and it is possible to obtain variants according to the technological requirements and needs. Therefore, this technique could be useful to develop new antimicrobial peptides to be used as an alternative to common antibiotics.

9As stated above, bacteriocins represent an interesting solution to reduce the development of resistance. Besides, bacteriocins are continuously evolving with high potential against clinically relevant pathogens. Bacteriocins can be easily manipulated by bioengineering techniques. Unlikely to antibiotics, bacteriocins can be engineered to attach anywhere on the cellular outer membrane because they do not have a specific receptor and they can be produced in situ by probiotics. Consequently, this represents a new path in bacteriocin research that will undoubtedly lead to the development of new therapeutic strategies with highly relevant clinical applications.

10In vitro antimicrobial activity assays are the first step to evaluate the biological capacity of bacteriocins against clinically relevant bacterial pathogens. However, if the conditions used in the in vitro models are not adequate to assess a specific effect, then the probabilities of success in the in vivo models will be very low. Fortunately, many bacteriocins have been successfully tested using in vitro assays against relevant bacterial pathogens (including MDR pathogens). Some examples include bacteriocin AS-48, which is active against reference and clinical strains of M. tuberculosis. Pentocin JL-1 has also been demonstrated to have antibacterial activity against Gram-positive and Gram-negative bacteria, particularly MDR S. aureus. Other examples, such as the novel entianin, which has activity against MRSA (ATCC 43300) and vancomycin-resistant Enterococcus faecalis (ATCC 51299) strains, and bacteriocins klebicins have been demonstrated to be active against MDR and carbapenem-resistant Klebsiella species. Moreover, novel enterocins DD28 and DD93 showed anti-staphylococcal activity in MRSA. It is important to mention that many current reports on the study of bacteriocins are focused on pathogens considered by WHO as a priority such as carbapenem-resistant E. coli or K. pneumoniae, vancomycin-resistant, and MDR E. faecium.

11As expected, bacteriocins have an outstanding record to kill or reduce pathogens and drug-resistant pathogens during in vitro assessments. Conventionally, bacteriocins display a non-toxic behavior at in vitro assays. Thus, the promising results obtained after in vitro assays must be extrapolated into in vivo assays. Furthermore, at this stage, internal factors of the host (pharmacokinetic parameters) should be considered as well as the potential bacteriocin-induced toxicity. Therefore, the use of animal models is mandatory in the development of a new therapeutic agent and the successful results obtained from these experiments are essential to close the translational gap to the clinic. The description of the preclinical drug discovery and development process is shown in Figure 1.

12There are many animal models used for screening drugs or chemical compounds in preclinical testing (pharmacological bioassays) and specific toxicity studies (toxicological bioassays). Rodents (rats and mice) are the most frequently used animal species. Other animal species, less frequently used, are guinea pigs, rabbits, and hamsters. Invertebrate models can be used to assess many biological activities, and these include Drosophila melanogaster (fruit fly) and Caenorhabditis elegans (a nematode worm). Finally, zebrafish is the most used rapidly developing vertebrate, and it has proven to be an excellent model for toxicity testing.

13In terms of genomics, the strategies for cloning, gene knockout, and gene or genome modifications are very well-described. Genes are very well-conserved between mice and humans since they share 90% of their total genes, and they are also available to develop spontaneous mutations. In terms of biology, mice are small and easy to handle; they can be easily transported and raised in a laboratory. Gestation times of mice are relatively short, and a large number of offspring can be obtained for in vivo purposes. Therefore, mice are the best-known animal model for in vivo bacteriocin efficacy studies.

14As mentioned earlier, the ability of bacteriocins to kill or inhibit pathogenic bacteria in vitro has been well documented. Bacteriocins represent one of the most studied microbial defense systems. They may facilitate the introduction of a producer into an established niche, directly inhibit the invasion of competing strains or pathogens, or modulate the composition of the microbiota and influence the host immune system. Therefore, understanding that bacteriocins may function in several ways, studies involving direct correlations between in vitro efficacy and in vivo protection are needed.

15To date, bacteriocins have showed a promising efficacy as antibiotic alternatives in in vivo studies. Bacteriocin application focuses in the administration at the site of the infection or susceptible areas, evading an immune response and maintaining the stability of the bacteriocin. Therefore, selection of optimal bacteriocin and delivery systems has been complicated. Two factors seem to play an important role in in vivo efficacy of bacteriocins: pharmacokinetic parameters and route of administration.

16First, pharmacokinetic parameters (e.g., bioavailability, stability, solubility in physiological conditions, and susceptibility to enzymatic proteolysis in bloodstream) are important determinants of the efficacy of bacteriocins. Bacteriocins administered orally are exposed to the hostile environment (enzymatic and pH degradation) in the gastrointestinal tract; they are highly susceptible to degradation once they reach the small intestine. On the other hand, parenteral administration may offer some means of avoiding proteolytic degradation of bacteriocins in the gastrointestinal tract. Nevertheless, the efficacy may be reduced since bacteriocins will be in contact with proteases involved in hemostasis and fibrinolysis in the bloodstream. To reduce or avoid these problems, bacteriocins can be engineered to be less susceptible to proteolytic degradation by changing D-amino acids. Also, nanotechnology seems to be a valuable strategy to improve the physicochemical properties of the bacteriocins. The nano-encapsulation [e.g., lipid-based nanoparticles (nanoliposomes and solid lipid nanoparticles), carbohydrate-based nanoparticles (chitosan/alginate and phytoglycogen nanoparticles), and conjugation with nanosized metal] of bacteriocins could protect them from enzymatic degradation, hence increasing their stability for longer periods.

17Second, the route of administration determines the onset and the duration of the pharmacological effect, the efficacy, and the adverse effects of drugs. The main routes of bacteriocin administration, such as intranasal, intragastric, intraperitoneal, subcutaneous, and topical have demonstrated excellent efficacy in murine models. However, the efficacy of the different routes of administration has not been directly compared and likely depends on the pathogen targeted. In the same way, in vivo toxicity testing has been conducted to identify possible adverse effects resulting from exposure to bacteriocins. As described below, many studies have been done to investigate toxicity and biosafety of bacteriocins using different type of applications, such as oral, intraperitoneal, nasal, and topical. In particular, topical application of bacteriocins has been reported to be successfully tested for skin infection with no toxicity effects. For example, the circular bacteriocin AS-48 was evaluated and the results indicated that this bacteriocin did not induce skin sensitization or cause allergic contact dermatitis. Topical formulation with two broad-spectrum bacteriocins: garvicin KS and micrococcin P1 was used in a murine skin infection model. The formulation had a significant antibacterial effect and animals showed no changes of behavior or obvious toxic effects. Therefore, there is a growing interest for the study of the therapeutic properties and side effects of bacteriocins using in vivo systems.

18Latest publications on in vivo assessment of bacteriocins using murine models contain as minimum parameters the measurement of the antimicrobial activity of bacteriocin on the animal model, the assessment of immunogenic response, biochemical analysis, and histopathological analysis. To support this paragraph, we have summarized some of the relevant bacteriocins with their in vivo antimicrobial (Table 1) and toxicity and biosafety (Table 2) activities in murine models over the last 20 years. In this context, we included bacteriocin studies that have measured their effects in murine models using purified or partially purified bacteriocins (including naturally synthesized bacteriocins by native and heterologous producers and chemically synthetized bacteriocins), bioengineering bacteriocins, or the bacteriocin-producer microorganism directly in the host.

19For categorization purposes, in this section, we only include studies of bacteriocins that were applied purely or partially purified in murine models. We here consider bacteriocins that were chemically synthesized and those that were naturally synthesized from a native bacteriocin-producer or by a heterologous bacteriocin-producer.

20This group is characteristic since it is a straightforward system that produces bacteriocins. Native bacteriocin-producers usually excrete bacteriocins by dedicated membrane-associated ATP-binding cassette (ABC) transporters or by the general secretion (sec) pathway of the cell.

21Mersacidin is a lantibiotic-type bacteriocin that was isolated and purified from Bacillus spp. HIL Y-85 54728. It has been tested in female BALB/cA mice infected by S. aureus 99308. Results showed a decrease in the inflammatory response of the host. Mutacin B-Ny266 is naturally produced from Streptococcus mutans Ny266. Its antibacterial effect was proved in mice infected with methicillin-susceptible S. aureus (MSSA) strain. Moreover, no mortality was observed in mutacin B-Ny266-treated mice. Another bacteriocin is nisin; it has been tested along with clausin and the two components (α- and β-peptides) bacteriocin amyloliquecidin (AmyA) from Bacillus amyloliquefaciens against a bioluminescent strain of S. aureus Xen 36 in adult female nude mice in a wound skin infection model. Interestingly, all antimicrobial treatments reduced the bacterial load after 7 days of treatment. Penisin, from Paenibacillus sp. Strain A3, was used to effectively protect mice from a MRSA infection. Penisin-treated infected mice had a significant higher survival rate than untreated infected mice. TSU4 is a bacteriocin recovered from Lactobacillus animalis TSU4. This bacteriocin was used in BALB/c mice to evaluate the acute and sub-chronic toxicity tests. Biochemical and histopathological analysis was performed. The bacteriocin demonstrated to be safe in a sub-chronic toxicity test. No antimicrobial in vivo test was performed. Finally, AS-48 bacteriocin is produced by E. faecalis strain UGRA10. The immunogenic response and biochemical and histopathological effects were analyzed in BALB/c mice. Later, its activity as antiprotozoal peptide was tested in BALB/c mice challenged with Trypanosoma cruzi strain Arequipa (Chaga’s disease etiological agent). Results demonstrated that this bacteriocin reduced the acute infection in mice.

22Occasionally, the internal mechanisms of the native bacteriocin-producers to produce and excrete the bacteriocins are not enough. The objective to produce the bacteriocin in a heterologous expression system is to increase the bacteriocin production yield from native producers by facilitating the control of gene expression or increasing the production levels.

23Lacticin 3147 is a two-peptide bacteriocin that is heterologous-produced in the recombinant strain Lactococcus lactis subsp. cremoris MG1363. This bacteriocin has been tested in an animal model, and the results have showed that it was able to reduce in vivo infection with S. aureus Xen 29 in mice. A very well-known example of bacteriocins from Gram-negative bacteria are pyocins, which are believed to be produced in 90% of P. aeruginosa strains. The heterologous-produced pyocins S2, S5, AP41, and L1 were used to study their protective function in an acute P. aeruginosa lung infection in C57/BL6 mice. Among all the pyocins, S5 was the best because no pathogenic bacteria were recovered from any of the S5 treated mice. The remaining pyocins were able to reduce the bacterial count in their respective treated mice. Shortly after, from the heterologous-produced pyocins SD1, SD2, and SD3, pyocin SD2 exerted the best performance among the other pyocins when it was tested in previously challenged C57/BL6 mice with P. aeruginosa PAO1. Treated mice were able to survive, and no signs of illness were reported. On the other hand, plantaricin E/F are two bacteriocins (plantaricin E and F) that have been heterologously produced in L. lactis NZ3900. The in vivo effects of both plantaricins were tested independently in a murine model infection. The favorable results obtained in antibacterial and toxicological tests suggest that plantaricin E or F are unharmful compounds that can be considered as a strong antibiotic candidate.

24Chemically synthesized bacteriocins are bacteriocins that were previously reported on non-modified bacteriocin-producer strains, and their antimicrobial effect have been measured on in vivo assays, but they have been synthesized chemically; some examples are lysostaphin and epidermicin NI01. First, lysostaphin was formulated on a petroleum-based cream, and it was able to eradicate S. aureus strain MBT 5040 in cotton rats after one single application. Second, the efficacy of epidermicin NI01, for eradicating the nasal burden of MRSA strain ATCC 43300 in a cotton rat model, was carried out. Results showed that a single dose of topical epidermicin NI01 was effective in eradication of MRSA from the nares of rats.

25Bioengineered bacteriocins are a group of bacteriocins whose characteristics have been modified to upgrade their properties. These modifications consist of the generation of novel bacteriocin variants that enhance the antimicrobial activity or expand the antibacterial spectrum and anti-biofilm efficacy or improve their physicochemical properties. Some examples of this type of bacteriocins are described below.

26In vivo activity of nisin A and nisin V against Listeria monocytogenes was evaluated in mice. Nisin V (a modified version of Nisin A) was more effective than Nisin A to controlling infection. Mutacin 1140 (MU1140) is a lantibiotic produced by S. mutans. A study to identify a lead compound for the treatment of Clostridium difficile-associated diarrhea was carried out. The variant OG253 emerged as the lead compound based on superior in vivo efficacy along with an apparent lack of relapse in a hamster model of infection. In vivo testing of another MU1140-derived variant (OG716) conferred 100% survival and no relapse at 3 weeks post C. difficile infection. Also, the effect of OG716 is determined using an in vivo hamster model of C. difficile-associated disease. Results demonstrated that OG716 was an excellent compound to treat C. difficile enteritis in hamsters.

27Bacteriocin-producers act similarly to probiotics because both can be consumed and exert a health benefit to the host. As long as they stay in the host, they may act as colonizing peptides, killing peptides, or serve as signaling peptides (signaling other bacteria or the immune system). It has been shown that bacteriocin production by bacteriocin-producers in the gut of the host can modulate the niche competition by preventing the intestinal colonization of MDR bacteria without disturbing the natural microbiota, therefore limiting the infection.

28A study demonstrated that Lactobacillus salivarius UCC118 (a sequenced and genetically tractable probiotic strain of human origin) significantly protected mice against infection with the invasive foodborne pathogen L. monocytogenes. In other study, rats were administered daily with Lactobacillus plantarum 423 and Enterococcus mundtii ST4SA. Then, they were challenged by infection with Salmonella enterica serovar Typhimurium. Results showed that L. plantarum 423 was more effective than E. mundtii ST4SA. performed the safety assessment and evaluation of probiotic potential of bacteriocinogenic E. faecium KH 24 strain using an in vivo model. Mice were challenged with Salmonella enteritidis MTCC 3291 and fed with E. faecium KH 24 strain. Results showed beneficial intestinal results (decreased counts of bacteria and coliform, and enhanced growth of lactobacilli). The protective effects of L. plantarum B7 on diarrhea in mice induced by Salmonella typhimurium ATCC 13311 was evaluated. Results demonstrated that L. plantarum B7 could inhibit growth of S. typhimurium, decrease levels of proinflammatory cytokines, and attenuate symptoms of diarrhea induced in mice by S. typhimurium. Another study evaluated the effects of the bacteriocin-producing Lactobacillus acidophilus strain JCM1132 and its non-producing spontaneous mutant, L. acidophilus CCFM720, on the physiological statuses and gut microbiota of healthy mice. The results showed that both strains can have different effects on the host such as prevention of metabolic diseases and reduced inflammatory response.

29The use of alternative models has gained great popularity among the scientific community since these models are simple, fast, and cheaper than current murine models. As we previously mentioned, some of these alternative models include D. melanogaster (fruit fly), Danio rerio (zebrafish embryos), C. elegans (roundworm larvae), Galleria mellonella (greater wax moth larvae), and Artemia salina (brine shrimp larvae). These alternative models allow evaluating bacteriocins and their potential effects on a living organism (such as antimicrobial activity, antibiofilm effect, immunogenic response, and toxicity). Also, murine models and alternative models do not share the same ethical considerations since the first ones have more restrictions when it comes to conducting experiments.

30A compilation of bacteriocin studies with their in vivo antimicrobial activity (Table 3) and/or toxicity and biosafety (Table 4) activity using alternative models over the last 10 years is shown below. In this context, we included studies that have measured bacteriocin effects in alternative models using purified or partially purified bacteriocins (including naturally synthetized bacteriocins by native and heterologous producers and chemically synthesized bacteriocins) or CFS with bacteriocin-like substance or with a bacteriocin-producer directly in the host.

31A fruit fly model (D. melanogaster) was used to evaluate the acute toxicity of antimicrobial peptide LR14. The results showed that the peptide had a dose-dependent toxicity property. Another study used the same in vivo model to evaluate the efficacy of peptide NAI-107 as treatment in MRSA infections. The authors reported that this peptide was able to rescue adult flies from fatal infection with efficacy equivalent to that of reference antibiotic (vancomycin). The antibacterial spectrum and cytotoxicity of a bacteriocin produced by Lactobacillus lactis strain in A. salina nauplii. The antibacterial activity of bacteriocin showed a broad range against foodborne pathogens. Also, the purified bacteriocin showed cytotoxicity in brine shrimps. identified a novel bacteriocin produced by Bacillus licheniformis strain 146 (lichenicin 146) with a high in vivo antimicrobial activity in liquid C. elegans–S. aureus assay.

32AS-48 is a bacteriocin produced by Enterococcus strains. The toxicity of this bacteriocin has been evaluated in in vivo models. In zebrafish embryos, the AS-48 was highly toxic; however, in a murine model, no toxicity was observed.

33A nonvertebrate host, the G. mellonella caterpillar, was used to evaluate the activity of pyocin S2 against P. aeruginosa YHP14 biofilms. Results showed a potent antibiofilm activity in vivo, representing a potential therapeutic option. The antimicrobial activity of peocin, a bacteriocin produced by the probiotic Paenibacillus ehimensis NPUST1, was demonstrated in aquatic, food spoilage, clinical, and antibiotic-resistant pathogens. For example, a significant increase in survival rates was observed in peocin-treated zebrafish after Aeromonas hydrophila challenge.

34A study reported that epidermicin NI01 had a protective effect of G. mellonella larvae from infection with clinically isolated MRSA. The authors reported that epidermicin NI01 did not induce toxicity and did not trigger the larvae immune system.

35The in vivo assessment of bacteriocins in alternative models has been evaluated using a CFS that contains a bacteriocin like-substance (BLIS). For example, A. salina brine shrimp showed no toxicity of CFS with BLIS from Enterococcus hirae. Recently, CFS from Lactobacillus curvatus P99 cultures showed no toxic effect in D. melanogaster flies.

36The evaluation of the probiotic effect of Bacillus sp. LT3 was performed in brine shrimp Artemia franciscana larvae challenged with Vibrio campbellii LMG 21363. Bacillus sp. LT3 was able to colonize the brine shrimp gastrointestinal tract and therefore increased their survival. A C. elegans model was used to evaluate the functionality of Lactobacillus fermentum strain JDFM216 (a novel probiotic bacterium). Interestingly, the probiotic bacterium was found to be toxic to the C. elegans host. Therefore, it has beneficial effects on longevity and immunity of C. elegans. The effect of Pediococcus pentosaceus strain (SL001) in growth-related and immune-related genes was evaluated in grass carps. Results showed that the strain was able to enhance immunity and promoter growth of grass carps. A recent study evaluated the effect of probiotic Chromobacterium aquaticum on zebrafish model. Fish was challenged with A. hydrophila and Streptococcus iniae and then treated with probiotic. The probiotic-treated fish had a higher survival rate than the non-treated fish. Finally, the effect of E. faecalis UGRA10 and its bacteriocin AS-48 (multiple baths and single dose) was tested against Lactococcus garvieae in an Oncorhynchus mykiss rainbow trout fish model. Neither the strain nor its bacteriocin showed toxic effects, displaying a protective effect against L. garvieae infection.

37The in vivo assessment models are important to evaluate bacteriocins. This review included various bacteriocins that were assessed by different in vivo models. The animal model must be chosen according on the expect effects for the bacteriocins. Below is summarized (Table 5) some advantages and disadvantages of murine, fruit flies, zebrafish, roundworm, greater wax moth, and brine shrimp for drug discovery trials. We included the limitations and strengths of each model as well as their scope of interest, time to reach the optimal developmental stage according to model, frequency of use in drug discovery studies, infrastructural and cost requirements for rearing, special qualifications, and ethical considerations, among others, which seem to be expected for the evaluation of new bacteriocins depending on the used model.

38Bacteriocins represent a potential drug alternative for replacing current antibiotics to treat diseases caused by resistant bacteria. According to the body of knowledge that has been developed in the field, in general, bacteriocins can retain their in vivo antimicrobial properties in a challenged host, while at the same time, they showed a null or reduced toxic effect. According to 2019 WHO’s Antibacterial Agent in Preclinical Development Book, 27 of 252 antimicrobial agents in preclinical revision status are considered as antimicrobial peptides. From the total antimicrobial peptides, 12 peptides are on lead optimization (LO) phase, 12 peptides are in Preclinical Candidate (PCC) phase and, three peptides are on CTA/IND-enabling studies. In an independent study, identified the current global antibacterial pipeline and found that 135 of 407 preclinical projects from 314 private and public institutions were related to producing synthetic and natural antimicrobial peptides, natural products, and LpxC inhibitors, and most of these molecules are targeting Gram-negative bacteria. Therefore, bacteriocins may have a window of opportunity to face the drug-resistant bacteria crisis since the WHO is demanding research and development of new drugs to target the most wanted drug-resistant pathogens, many of them Gram-negative bacteria.

39Although it is a fact that the current literature for bacteriocins produced from Gram-negative bacteria is dominated by bacteriocins toward Gram-positive bacteria, there is an acceptable amount of bacteriocins reported to have a strong activity against Gram-negative bacteria, including the pathogenic strains. For example, the S-type pyocin group from the P. aeruginosa or the microcin and colicin groups that are vastly reported for E. coli, and other examples in less quantity but not less relevant, are bacteriocins produced by K. pneumoniae, Citrobacter freundii, Shigella boydii, and Serratia marcescens.

40The future of bacteriocins lies not only in their discovery but also in their testing for toxicity to prove their safe use in a preclinical phase as candidates for therapeutic processes. An increasing approach that can be exploited for bacteriocins is the use of alternative models to the murine model to evaluate their in vivo antimicrobial and/or toxicity effects. According to, there is an increase in drug studies using alternative organisms to murine, since there was a rise of 909% in drug discovery from 1990 to 2015.

41Finally, we have stated that bioengineering is an important tool along with the current technologies to discover new bacteriocins, since they can improve their antimicrobial activity or change their physicochemical properties. Moreover, new strategies are being introduced in the design of bacteriocins. were the first to design the very first fully de novo bacteriocin by using a machine-learning approach. also constructed a library of the linear peptides from the membrane-interacting region of circular bacteriocin with pore-formation dynamics by selective aminoacidic substitution. On the other hand, were the first to design chimeric bacteriocins that retained the properties to kill both Gram-positive and Gram-negative bacteria. Moreover, other authors have preferred to repurpose the bacteriocins by exploiting their capability against tumor cells or by exploring synergistic effects of bacteriocins while combining with renowned antibiotics or with other bacteriocins.

42Bacteriocins are strong candidates to be used as future therapeutic agents. Recent studies have shown the antibacterial activity of bacteriocins using in vitro models. Nonetheless, the next step in the development of a new bacteriocin-based therapeutic agents involves the use of animal models. The antimicrobial and/or toxic effects of the bacteriocins have been studied in murine models and the most recent alternative animal models such as fruit flies, zebrafish embryos, roundworm, greater wax moth, or the brine shrimp. These results have demonstrated that bacteriocins can exert a variety of positive responses in the host such as modification of the immunogenic response, alteration of the inflammatory response, and the reduction of biochemical and histopathological parameters related with infection. However, approximately half of the bacteriocins tested in mice (47.4%) performed antimicrobial assay, but no toxicity assays were described. On the other hand, 20% of the studies carried out in alternative models evaluated antimicrobial activity, but no toxicity was reported. These data reveal the lack of toxicity and biosafety studies of bacteriocin in vivo models, which are crucial to advance into clinical trials. Therefore, it is imperative to use in vivo models to assess the therapeutic efficacy of bacteriocins as well as their toxic effects; both successful results will lead toward clinical research phases and the development of bacteriocin-based therapeutics to treat infections caused by antimicrobial-resistant Gram-positive and Gram-negative bacteria.

43DB-C, AL-B, and JM-R: conceptualization. DB-C, AL-B, JL-E, and JM-R: writing—original draft preparation and writing—review and editing. AL-B and JM-R: supervision. All authors contributed to the article and approved the submitted version.

44The 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.