Antimicrobial resistance of ocular microbes and the role of antimicrobial peptides

ABSTRACT Isolation of antimicrobial‐resistant microbes from ocular infections may be becoming more frequent. Infections caused by these microbes can be difficult to treat and lead to poor outcomes. However, new therapies are being developed which may help improve clinical outcomes. This review examines recent reports on the isolation of antibiotic‐resistant microbes from ocular infections. In addition, an overview of the development of some new antibiotic therapies is given. The recent literature regarding antibiotic use and resistance, isolation of antibiotic‐resistant microbes from ocular infections and the development of potential new antibiotics that can be used to treat these infections was reviewed. Ocular microbial infections are a global public health issue as they can result in vision loss which compromises quality of life. Approximately 70 per cent of ocular infections are caused by bacteria including Chlamydia trachomatis, Staphylococcus aureus, and Pseudomonas aeruginosa and fungi such as Candida albicans, Aspergillus spp. and Fusarium spp. Resistance to first‐line antibiotics such as fluoroquinolones and azoles has increased, with resistance of S. aureus isolates from the USA to fluoroquinolones reaching 32 per cent of isolates and 35 per cent being methicillin‐resistant (MRSA). Lower levels of MRSA (seven per cent) were isolated by an Australian study. Antimicrobial peptides, which are broad‐spectrum alternatives to antibiotics, have been tested as possible new drugs. Several have shown promise in animal models of keratitis, especially treating P. aeruginosa, S. aureus or C. albicans infections. Reports of increasing resistance of ocular isolates to mainstay antibiotics are a concern, and there is evidence that for ocular surface disease this resistance translates into worse clinical outcomes. New antibiotics are being developed, but not by large pharmaceutical companies and mostly in university research laboratories and smaller biotech companies. Antimicrobial peptides show promise in treating keratitis.

According to the World Health Organization (WHO) antimicrobial resistance arises when the micro-organisms such as bacteria, viruses, fungi or parasites survive exposure to a medicine that is commonly used to kill them or stop their growth. 1 These micro-organisms are often referred to as superbugs because of this resistance to antimicrobials. 1 The increasing prevalence of antimicrobial resistance is one of the largest global public health concerns of the 21st century. Antibiotic resistance contributes to the increased cost of health care, and morbidity and mortality rates of patients. 2 Approximately 700,000 deaths each year are due to antimicrobial-resistant microbes, and this number is projected to reach 10 million deaths with economic losses of US$100 trillion annually by 2050. 3 There have been significant advances in the development of antiviral drugs which continue to be effective in treating current viral infections although there is the emergence of some antiviral resistance. [4][5][6][7] The major breakthrough in the field of antiviral drug research was the development of effective antiviral drugs in treating life-threating diseases such as human immunodeficiency virus (HIV), hepatitis C virus (HCV), hepatitis B virus and herpes viruses. 4,8,9 However, most viral infections are controlled by prophylactic vaccines or the host immune system, without the need of any antiviral drugs. Thus, viral adaptations to antiviral drugs are rare, 10 although not without precedent for viruses such as the human immunodeficiency virus (HIV).
Despite many discoveries of newer classes of antibiotics or modifications of existing antibiotics that target bacteria, resistance evolves in bacteria. 11,12 Factors that contribute to antibiotic resistance are overuse or misuse of antibiotics and imprudent use of antibiotics in agriculture, poultry and livestock as growth promoters. [13][14][15] The first case of Staphylococcus pyogenes (a Gram-positive bacterium) infection with a penicillin-resistant strain was observed in 1947, just four years after the drug became commercially available. 16 Ever since, resistance has developed to all classes of antibiotics (cephalosporins, fluoroquinolones, aminoglycosides, tetracyclines and so on) including the last-resort antibiotics such as vancomycin, posing a serious threat to global public health. 11,[17][18][19][20][21] The clinical impact of antimicrobial resistance has led to serious infections associated with poor prognosis, high mortality rates and prolonged hospitalisation. 22 Reports suggest that these infections are caused mostly by the Gram-negative bacteria Enterobacteriaceae, Pseudomonas aeruginosa, Klebsiella pneumonia, Acinetobacter baumannii and Stenotrophomonas maltophilia that are resistant to almost all antibiotics. 23,24 These serious infections are mostly nosocomial, for example, ventilator-associated pneumonia, bloodstream infections and implant-associated urinary tract infections, that often occur in immune-compromised patients. P. aeruginosa isolates frequently develop multidrug resistance in patients with cystic fibrosis contributing to a high mortality rate. 25 P. aeruginosa strains can also be highly resistant to the last-resort drugs colistin and carbapenems. 26,27 Another highly resistant Gramnegative bacteria, Helicobacter pylori, colonises the stomach mucosa causing chronic gastritis, dyspepsia, peptic ulcers, and gastric cancer. 28 Approximately half of the world's population has been infected by this pathogen. The most effective therapy for this bacterium is continually changing due to the rapid development of drug resistance. 28,29 Methicillin-resistant Staphylococcus aureus (MRSA) causes bloodstream, skin and soft tissue infections and pneumonia in hospitalised and community-based patients. 30 These MRSA strains are highly resistant to many antibiotics, making treatment very difficult. 24,31 Therefore, to alleviate the complications related to antimicrobial resistance, prudent antimicrobial prescribing approaches, and management based on the knowledge of local antibiotic resistance patterns is essential. Many of these microbes such as the bacteria P. aeruginosa and S. aureus and the virus herpes simplex cause ocular surface infections which can be difficult to treat if they are caused by strains resistant to antibiotics.
The overall purpose of the review was to examine the literature to determine the types of antimicrobials used to treat ocular surface infection, rate of antimicrobial resistance of ocular microbes over the past 10-years and the development of cationic peptides as potential therapies for keratitis. This paper reviews mostly data from the past 15-years (January 2005-March 2020) with a focus on pathogens associated with ocular infection, the prevalence of antimicrobial resistance in these pathogens, and the development of novel therapies for keratitis based upon cationic antimicrobial peptides. This review was initiated by performing a systematic literature search in Pubmed-Medline and Scopus using the search terms (ocular infections) AND (antimicrobial resistance) AND (ocular antimicrobial resistance) AND (antibiotic resistance), and another search with the terms (antimicrobial peptide) AND (cornea) OR (tears) OR (keratitis). Four hundred and seventy-four articles were found and those that fit with the aim of the search and were written in English were downloaded and reviewed. In addition to this, references contained in any article were examined (26 articles) for other possible references that had been missed in the original searches. None of these articles were excluded from reviewing as there were no textbook chapters or letters or non-randomised study designs in our literature search results. Furthermore, references that were required to elucidate points of interest when writing the article were also added.

Ocular microbial infections
Ocular microbial infections are a global public health issue as they can result in vision loss which compromises quality of life. 32 Ocular microbial infections are predominately caused by bacteria, followed by viruses, fungi and parasites, although differences in individual countries and societies can change this ranking. 33 Microbial keratitis is one of the most common causes of vision loss worldwide. 34 Trachoma, caused by infection of the conjunctiva by the bacterium Chlamydia trachomatis, 35 is estimated to have caused visual impairment in approximately 2.2 million people, of whom 1.2 million are irreversibly blind, and is believed to be the leading cause of infectious blindness worldwide. Onchocerciasis is another leading cause of infectious blindness, and the WHO estimates that approximately 17 million people are infected with the filarial nematode Onchocerca volvulus that causes the disease. 36 For the non-endemic diseases (that is, not trachoma or onchocerciasis), the most common risk factors of ocular infections are ocular or systemic disease, trauma or surgery, age, and improper contact lens use. [37][38][39][40] Approximately 70 per cent of ocular infections are caused by Gram-positive bacteria such as Staphylococcus epidermidis (also called coagulase-negative staphylococci; CoNS), Staphylococcus aureus, and Streptococcus pneumoniae, Gram-negative bacteria such as Pseudomonas aeruginosa and Moraxella spp. and fungi such as Candida albicans, Aspergillus spp. and Fusarium spp. 41,42 Several studies have shown that, among the bacterial pathogens, Gram-positive pathogens, particularly CoNS, are the leading causes of ocular infections (ranging from 63 to 74 per cent), and are more prevalent but less diverse than Gram-negative bacteria. [42][43][44][45][46][47][48] Viruses are the most common cause of infectious conjunctivitis, with adenovirus accounting for approximately 65-90 per cent of all infections. 49 Herpes viruses are an important cause of keratitis in developed countries. 50

Ocular antimicrobial therapies
Penicillin was commercialised in 1942, almost 14-years after its discovery. 51 Later, in 1943, penicillin was used as drops to treat experimentally induced Staphylococcus aureus and pneumococcal corneal infections in rabbits. 52 In humans as early as 1946, penicillin was used both topically and via intraocular injection for treating a variety of ocular infections including conjunctivitis to endophthalmitis. 53 Subsequently, many antimicrobial agents have been added to the armamentarium to treat ocular infections. 54 Table 1 highlights some of the most commonly used therapies to treat ocular infections.
Antibiotics, the correct term for all drugs that kill living things (anti = against; bios = life), have a variety of actions that inhibit the growth or kill microbes. Antibiotics can conveniently be classified as antibacterials, antifungals, antivirals and antiparasitics (with parasites in this case referring most often to small protozoa or nemotodes). The general mechanism of action of these drugs is inhibition of DNA, RNA, protein or cell wall synthesis or membrane disruption, and Table 2 elucidates mechanisms of actions for these drugs.

Ocular microbial resistance
Despite the undoubted usefulness of antibiotics to treat ocular infection, rapid emergence of resistant ocular microbes is occurring, compromising the efficacy of antibiotics. Selection of antibiotic therapies could be challenging if ocular infections are caused by resistant pathogens, as these infections can result in severe complications due to poor drug efficacy or treatment failure. 93 Moreover, the severe ocular surface infections that usually respond to high-dose antibiotic therapies may result in poor prognosis due to the rapid emergence of antibiotic resistance. 93 One of the major factors contributing to the emergence of antibiotic resistance among ocular pathogens is the empirical method of prescribing antibiotics instead of a rational stepwise diagnostic approach with specimen culture to identify the causative pathogen followed by sensitivity/resistance testing. 44,94 Intraocular antibiotic prophylaxis, such as is used to prevent endophthalmitis in patients undergoing cataract surgery, may also induce antibiotic resistance. 54,95 The Steroids for Corneal Ulcers Trial (SCUT) has demonstrated the importance of antimicrobial resistance in ocular isolates. For every two-fold increase in the minimum inhibitory concentration (MIC) of the prescribed antibiotic there was an increase in average diameter of the infiltrate or scar by 0.05-mm, and a 0.02 logarithm of the minimum angle of resolution decrease (logMAR; approximately one letter of vision loss) in visual acuity three weeks after the first diagnosis. A higher MIC (after controlling for the effect of microorganism) was associated with a slower time to reepithelialisation. 96 There were also significant linear associations between worse clinical outcomes and higher MIC for ulcers caused by S. aureus, P. aeruginosa and other bacteria. 97 Antibiotic resistance in ocular microbes was first reported in the 1950s when strains of staphylococci were found to be resistant to penicillin. 98,99 In 1959, methicillin, a beta-lactam antibiotic, was introduced to treat penicillin-resistant staphylococci. 100 However, the emergence of multidrug-resistant forms of MRSA 101 and methicillin-resistant CoNS (MRSE) 101-103 has led to severe forms of ocular infections. 101-105 Early (1975101-105 Early ( -2000 and recent (2002-2017) studies have shown that ocular infections with MRSA and MRSE do respond to topical and intravitreal vancomycin. [101][102][103][104][105][106][107][108] Although there have been reports of the emergence of resistance to vancomycin in strains isolated from endophthalmitis. 109-111 P. aeruginosa is the leading cause of ocular infections during contact lens wear, 112,113 and can be isolated from contact lenses. 39,113-115 P. aeruginosa isolates resistant to the fluoroquinolone ciprofloxacin were isolated from scleral buckle infection in 1998, 116 and this was followed by resistant isolates from corneal infection in 1999. 117 A previous review article reported there are increasing rates of resistance to fluoroquinolones and beta-lactams in ocular isolates of both P. aeruginosa and S. aureus. 118 Data collected from 2009 to 2018 in a longitudinal ongoing nationwide USA study called the Antibiotic Resistance Monitoring in Ocular Micro-organisms (ARMOR), has reported on the rates of antimicrobial resistance among the 6,091 ocular isolates. 119 S. aureus had a resistance rate to the fluoroquinolones of between 25 per cent for moxifloxacin to 32.2 per cent for ciprofloxacin; 58.6 per cent of strains were resistant to azithromycin, 0.6 per cent to chloramphenicol, 13.7 per cent to clindamycin, 34.9 per cent to oxacillin/methicillin (that is, 34.9 per cent were MRSA), 5.7 per cent to tetracycline, 14.6 per cent to tobramycin, 4.3 per cent to trimethoprim and zero per cent to vancomycin. 119 Very similar rates for each antibiotic were reported for CoNS except these were more resistant to oxacillin/methicillin (49.3 per cent), tetracycline (11 per cent) and trimethoprim (27.9 per cent). Isolates of S. pneumoniae had no or very low levels of resistance (less than four per cent) to all antibiotics except azithromycin (35.9 per cent), penicillin (7.8 per cent) and tetracycline (8.7 per cent). Rates of resistance of P. aeruginosa were generally low, with 3.3-5.7 per cent being resistant to a fluoroquinolone. Rates of resistance for H. influenzae did not go above 1.4 per cent (which was seen with tetracycline) for any antibiotic. 119 There had been ≤2.5 per cent change of antibiotic resistance over the 10-years of the study. 119 However, there was no change in antibiotic resistance among P. aeruginosa and S. pneumoniae isolates. 119 In contrast, other reports have demonstrated an increase in resistance of Pseudomonas spp. to moxifloxacin (85.6 per cent), ceftazidime (90 per cent), tetracycline (80 per cent) and tobramycin (70 per cent) over time. 82 The variability in resistance rates might be attributed to differences in antibiotic prescribing in different countries/locations. [121][122][123]  Over the years, studies have noted a growing rate of MRSA in ocular isolates. 47,69,124 In a 20-year study (1993-2012) from the USA, MRSA represented approximately 30.7 per cent of the total bacterial isolates and showed high levels of resistance to most fluoroquinolones (74.6 per cent ofloxacin, and 73.8 per cent ciprofloxacin) but all strains were susceptible to vancomycin. 124 A similar study from Italy, with a 30-year follow up (1988-2017), reported an increasing rate of resistance of S. aureus to antibiotics, from 21 per cent in 2003 to 38 per cent in 2015, as well as for S. pneumoniae (7.4 per cent to 23.3 per cent). 125 In the same study, multidrug (resistant to at least one aminoglycoside and one fluoroquinolone) resistance rates of S. aureus were significantly increased from 8.7 per cent to 12 per cent, as was resistance in isolates of Enterobacteriaceae (24.5 per cent to 34.6 per cent). 125 Retrospective studies have reported increasing rates of MRSA. 69,126 An increasing prevalence of ocular MRSA infections from nine per cent to 38 per cent over a 10-year period (2007-2017) of study duration has been reported. 120 These MRSA isolates were all sensitive to vancomycin. 120 Low levels of MRSA isolates in a study from Sydney (seven per cent) precluded any assessment of changes over time, 127 although there were decreases in the resistance of CoNS to chloramphenicol and gentamicin over the five years (2012-2016) of the study. 127 This variation in resistance rates of MRSA could be associated with changes in weather and different geographical locations. 128,129 Latitudes closer to the equator with higher temperatures, humidity and annual precipitation are associated with higher colonisation of the MRSA. 128 Another study from Australia has reported that, while ocular isolates (P. aeruginosa and CoNS being the most common) remained sensitive to most antibiotics (84.8 per cent sensitivity), bacteria were more likely to be sensitive if isolated from paediatric patients than the general population. 130 Patients with keratitis who had previously used fluoroquinolones were more likely to have higher MIC to fluoroquinolones in the bacteria causing the disease. 131 In other studies, topical fluoroquinolone use within the past six months has been identified as a risk factor for the isolation of fluoroquinoloneresistant S. aureus from the ocular surface. 132 There is less information on resistance of fungal and viral ocular isolates. Development of resistance to azole has been noted in a multicentre randomised clinical trial. There was an increasing trend (two-fold) resistance to azole for ocular filamentous fungi during the 20-month long clinical trial. 133 Long-term therapy of herpes simplex virus (HSV) infections with acyclovir has been shown to predispose to the emergence of resistance to acyclovir, leading to treatment failure, transmission of resistant virus and recurrent herpetic keratitis. 134,135 The frequency of acyclovir resistance among ocular HSV cases from the Netherlands has been reported to be 26 per cent. 135 These studies demonstrate an increasing trend in ocular antimicrobial resistance to fluoroquinolones, especially in isolates of P. aeruginosa, S. aureus (particularly MRSA), and this can result in exacerbation of ocular infections. These resistant rates are influenced by geographic regions which may be associated with local antibiotic prescribing patterns and availability of antibiotics. Knowledge of these trends in resistance in particular countries may assist clinicians in selecting rational antimicrobial therapy within the country or for people who have an infection and have returned from that country.

New antimicrobial strategies
For the past two decades there has been a hiatus in developing and releasing for sale new ocular antibiotics that may be useful to treat resistant pathogens. The last ocular antibiotic approved by the US Food and Drug Administration in the USA was besifloxacin in 2009, a fourth-generation fluoroquinolone for the treatment of bacterial conjunctivitis. 136 Treatment of ciprofloxacin-resistant MRSE with besifloxacin has been effective with no adverse events. 137 Besifloxacin has improved coverage for MRSA and ciprofloxacin-resistant S. aureus, 138 and can inhibit the growth of 90 per cent of P. aeruginosa and MRSA isolates compared to other fluoroquinolones. 119 Although no breakpoints are available for besifloxacin, some authors have used the breakpoint for moxifloxacin to determine sensitivity or resistance. 138 Using this criteria, 3/38 (eight per cent) isolates of CoNS from one study on ocular infections attributed to staphylococci in the USA reported in 2020 can be described as being resistant, 137 and another study from the USA published in 2013 reported 29.6 per cent of ocular staphylococci as being resistant to besifloxacin. 138 Probably the newest antibiotic that has been approved by US Food and Drug Administration is called lefamulin which was approved in August 2019. 139 Lefamulin is a semisynthetic pleuromutilin antibiotic indicated for community-acquired bacterial pneumonia caused by S. pneumoniae, S. aureus (methicillin-susceptible isolates), Haemophilus influenzae, Legionella pneumophila, Mycoplasma pneumoniae, and Chlamydophila pneumoniae. 140 Lefamulin selectively inhibits the synthesis of bacterial protein. 140 A recent in vitro study has shown that lefamulin was highly potent against Chlamydia trachomatis, the bacterium that causes ocular trachoma. 141 However, to date, there have been no in vivo studies or reports on the use of this antibiotic to treat ocular infections in humans. Studies are needed to evaluate its efficacy and safety in human clinical trials in treating ocular trachoma and bacterial ocular infections.

Combination therapies
Fortified antibiotics, usually relatively high concentrations of a combination of a cephalosporin (often cefazolin) active against some Gram-positive bacteria with an aminoglycoside (tobramycin or amikacin) active against Gram-negative bacteria, have been used for many years to treat microbial keratitis. A recent meta-analysis comparing results of treating keratitis with fortified antibiotics versus monotherapy with a fluoroquinolone found no benefit of using the fortified preparation. 142 Indeed, the authors concluded that the monotherapy was probably better tolerated than the combination and as there was no difference in efficacy, monotherapy with a fluoroquinolone was generally preferable. 142 A combination of polymyxin B and trimethoprim provides broad antibacterial activity and this combination is commonly used for the treatment of mild bacterial conjunctivitis. 143 Polymyxin B is active against Gram-negative bacteria, including P. aeruginosa, whereas trimethoprim primarily targets Gram-positive bacteria such as S. aureus. There is a low level of resistance of ocular pathogens to these two antibiotics. 143 The combination of polymyxin B and trimethoprim with another antibiotic rifampin has increased the antimicrobial efficacy against S. aureus and P. aeruginosa compared to the current generation of fluoroquinolones. 144

Antimicrobial peptides as alternative antibiotics
Antimicrobial cationic peptides (AMPs) are a class of antibiotic that have similar mode of action to polymyxins. The polymyxins (polymyxin B and polymyxin E, the latter also known as colistin) were originally derived from bacteria and can be considered the first AMPs to be used therapeutically. 145 Polymyxin E is one of the few cationic antimicrobial peptides commercialised in both human and veterinary medicine. Due to the emergence of antimicrobial resistance to many other antibiotics, polymyxin E has been considered the last line of defence against infections caused by multidrug-resistant Gram-negative bacteria such as Acinetobacter baumannii, P. aeruginosa and Klebsiella pneumoniae. The WHO has reclassified polymyxin E in the category of very high importance in human medicine, attempting to limit or prevent its use in animals. The most common mode of action for polymyxins and AMPs is to cause damage to microbial membranes which leads to cell death. 146 AMPs have been investigated for their antimicrobial action on a variety of microbes from ocular infections as well as for their ability to control keratitis in animal models. Importantly, several studies have shown that bacteria such as P. aeruginosa and S. aureus grown at sub-inhibitory concentrations of AMPs do not become resistant to them, whereas this drives rapid development of resistance to antibiotics such as ciprofloxacin, levofloxacin and gentamicin ( Figure  1). [147][148][149][150][151][152] AMPs are active even against microbes that are resistant to other forms of antibiotics. 153 Furthermore, AMPs can synergise with other antibiotics or disinfectants (such as those used in contact lens solutions). 20,[154][155][156][157][158] The ocular surface is protected from microbial attack by many different systems, and one of these is the ability of resident and recruited cells to produce AMPs. 159 Naturally occurring antimicrobial peptides on the ocular surface are produced by recruited neutrophils (α-defensins), corneal and conjunctival epithelial cells (for example, β-defensins, macrophage inflammatory protein [MIP]-3a, Thymosin β-4, Liver Expressed Antimicrobial Peptide [LEAP]1 and LEAP2), or neutrophils and corneal and conjunctival epithelial cells (the cathelicidin LL-37). 146,160 These peptides can be expressed under normal conditions or upregulated during infection. LL-37, lipocalin-2, LEAP1 (hepcidin), human betadefensin (hBD)1, hBD2, hBD3, S100A7, S100A8, S100A9, S100A12 but not hBD4 or RNAse7 are upregulated in the human cornea during P. aeruginosa keratitis. 161 In a mouse model of Aspergillus fumigatus keratitis, beta defensins 1, 3, 4 and 14 and cathelin-related antimicrobial peptide (CRAMP; similar to human LL-37) are upregulated, 162 whereas in a mouse model of Candida albicans keratitis CRAMP was upregulated during infection but beta defensins 1 and 2 and Thymosin-β4 were downregulated. 163 Most human defensin genes, particularly hBD2, hBD3, LEAP1, LEAP2 and RNAse7 are upregulated during Acanthamoeba interactions with human corneal limbal cells in laboratory culture. 164 LEAP1 is expressed during herpes simplex ocular surface infection in humans. 165 Studies have examined the activity of these naturally occurring ocular surface AMPs against microbes that can cause ocular surface disease. HBD2 can inhibit the growth of P. aeruginosa 166 and α-defensins from rabbits are active against P. aeruginosa and S. aureus. 167,168 Topically applying LL-37 to mouse eyes can reduce inflammation caused by the addition of the Gramnegative bacterial endotoxin lipopolysaccharide. 169 Mice made deficient of mouse β-defensins 3 or 4 or CRAMP demonstrate increased pathology during F. solanii infection, 162 implicating these as important to control the infection. Human α-defensins, LL-37 and lactoferricin (an antimicrobial peptide derived from the tear protein lactoferrin) can prevent HSV infection in vitro. 170,171 LL-37 can also inhibit adenoviral replication. 172 However, there are several issues with the naturally occurring AMPs that prohibit them from being used therapeutically. These issues include their cost of synthesis (as they tend to have fairly long amino acid sequences; LL-37 is 37 amino acids in length; Table 3), and the fact that several naturally occurring AMPs can be cytotoxic at the concentration needed for antimicrobial action, or can be degraded by peptidases. 173 Furthermore, naturally occurring AMPs can be inhibited by the high salt concentrations found in tears at the ocular surface. 166,167 Therefore efforts have been made to overcome these deficiencies by either using AMPs (modified or not) from other animals or plants, modifying naturally occurring human AMPs, or synthesising synthetic (often smaller) AMPs.
Examples of some of the many animal-derived AMPs include indolicidin (a cathelicidin family member) from bovine neutrophils, brevinin-1, magainin-2 and dermaseptin-S4, all three of which are derived from frog skin, which can inhibit HSV-1 and/or HSV-2. 170 Indolicidin also inhibits HIV. 170 Tachyplesin derived from the horseshoe crab is viricidal and cecropin originally from insects inhibits virus-cell surface interactions toward HSV.
Changing the amino acids in naturally occurring AMPs from αamino acids to β-amino acids or L-isomers to D-isomers can give some protection of the AMPs toward peptidases and proteolytic degradation without compromising their antimicrobial action. 173,174 Also, peptides manufactured to contain peptide linkages between ε-amino groups, such as are carried on the amino acid lysine, rather than α-amino groups, through which normal peptide bond formation occurs, can retain activity while reducing cytotoxicity and susceptibility to peptidases. 175,176 Polyε-lysine has been formed into bandage contact lenses and combined with penicillin G. These bandage lenses had significant antimicrobial activity against planktonic and attached S. aureus. 177 When combined with amphotericin B, poly-ε-lysine bandage lenses also had good activity against C. albicans. 178 Peptide FK16 (Table 3) is a short 16 amino acid cathelicidin-derived synthetic peptide which can increase the susceptibility of P. aeruginosa to vancomycin, is not affected by tear salt concentration and is not toxic to mammalian cells at its active concentration. 198 Peptide 120-146 WH (Table 3), a short 26 amino acid AMP derived from the human AMP CAP37 (also called azurocidin), retains antimicrobial activity and is non-toxic. 199 Esculentin-1 (Table 3) is an AMP derived from the skin of frogs which has excellent antimicrobial activity against Gram-negative bacteria (including P. aeruginosa) and C. albicans but poorer activity against S. aureus. 200 A shorter version of Esculentin-1, Esc-1a (1-21) (Table 3), retains its activity and can reduce biofilm formation by P. aeruginosa, 201 as do short peptides derived from another amphibian-derived AMP Kunitzin-RE. 148 AMPs of insects such as trialysin and gomesin, can be reduced in size and retain activity. For example, peptide P5 from the N-terminus  of trialysin can permeabilise trophozoites of A. castellanii and can reduce their growth. 202 Shiva-11, a peptide derived from cecropin (originally from insects), is active against ocular isolates of P. aeruginosa and S. aureus. 203 Chimeric peptides (melimine and Mel4; Table 3) derived from the AMPs protamine from salmon sperm and melittin from bee sting are active against P. aeruginosa, Serratia marcescens, S. aureus, Streptococcus pneumoniae, and other bacteria as well as F. solanii, C. albicans, A. castellanii either in solution or when bound to surfaces including contact lenses. 152,153,190,[204][205][206][207][208][209][210][211][212][213] Another chimeric peptide CA(1-7)-M (5-9) ( Table 3) derived from the AMPs cecropin and melittin is active against P. aeruginosa. 214 The peptide OH-CATH30, derived from the venom of the king cobra, is active against P. aeruginosa including antibiotic-resistant strains. 147 A derivative of LyeTx I, an AMP from the venom of the spider, called LyeTxI-b (which differs from its parent AMP in the loss of one histidine amino acid) is active against planktonic and biofilm forms of S. aureus. 215

Evidence for effectiveness of antimicrobial peptides to treat ocular infections
Cell culture, ex vivo and animal models of ocular surface infections have been used to test whether several AMPs or derivatives are effective at preventing ocular surface microbial infections. The AMP OH-CATH30 alone or in combination with levofloxacin, significantly improved the clinical outcomes of rabbit keratitis induced by antibiotic-resistant P. aeruginosa. 147 Evidence was also presented to show that this AMP also possessed anti-inflammatory activity. 147 Esc-1a(1-21)NH 2 lowered clinical scores and reduced bacterial numbers in corneas in a mouse model of P. aeruginosa keratitis. 189 The branched AMP B2088 (Table 3) combined with gatifloxacin reduced numbers of P. aeruginosa infecting corneas compared to gatifloxacin or B2088 alone in a mouse keratitis model. 192 The cecropin-melittin chimeric peptide CA (1-7)-M (5-9) ( Table 3) reduced the clinical score during infection with P. aeruginosa to the same degree as gentamicin in a rabbit model. 214 Peptide 120-146 WH (Table 3) derived from CAP37 can reduce the numbers of P. aeruginosa infecting mouse corneas. 199 A synthetic peptide mimic of human defence peptides called RP444 (Table 3) was able to reduce the clinical score and numbers of infecting P. aeruginosa in a mouse model of keratitis. 151 Poly-ε-lysine can reduce bacterial numbers of P. aeruginosa or S. aureus in a rabbit model of keratitis. 176 Minimal substitution of α-lysine with ε-lysine in melittin resulted in a peptide with improved cytotoxicity which could reduce numbers of S. aureus in a mouse keratitis model. 175 LyeTxI-b can also reduce numbers of S. aureus this time in a rabbit model of keratitis, and also reduce slitlamp scores of inflammation. 215 Synthetic derivatives of α-mangostin, designed to mimic AMPs, reduced the numbers of infecting S. aureus in a mouse keratitis model. 149 Also, synthetic derivatives of icaritin, also designed to mimic AMPs, could reduce numbers of S. aureus in a mouse keratitis model. 150 Another non-peptide analogue of AMPs (mimicking the structural properties of a naturally occurring defensin) was also able to reduce S. aureus numbers in a mouse keratitis model. 216 Two small AMPs, peptides 1 and 2 (Table 3), were as effective as amphotericin B when treating mouse keratitis caused by C. albicans. 217 Two other small peptides, peptides 2 and 3 (Table 3), were able to reduce the severity of keratitis and C. albicans 'mass' in corneas of infected mice. 218 Peptide  W8, derived from Kunitzin-RE, reduced clinical score and fungal numbers similar in amount to amphotericin B in a C. albicans mouse keratitis model. 148 A cationic steroid antibiotic (CSA)-13 that is proposed to mimic the activity of endogenous AMPs can inhibit Acanthamoeba trophozoite growth without significant toxicity to mouse fibroblast cells. 219 The cationic Peptide TAT-Cd0 can reduce the ability of HSV-1 to infect human corneal epithelial cells 220 and reduce the severity of keratitis, delay the onset of vascularisation and stromal keratitis, reduce the percentage of mice presenting with disease and viral titres. 221 However, treatment needed to be early in the disease process (four hours after initiation of infection) for these effects to be seen. 221 A combination of the AMP G2 (Table 3) and acyclovir was able to reduce viral loads in an ex vivo model using porcine eyes. 195 However, not all AMPs or mimics have been found to be effective for treating ocular surface disease. The peptide COL-1 (Table 3) was not able to control P. aeruginosa keratitis in a rabbit model, despite being effective in vitro. 196 A synthetic theta-defensin called retrocyclin ( Table 3) that reduces HSV infectivity without being directly viricidal, was not able to treat HSV keratitis in a mouse model. 222 In conclusion, AMPs are produced at the ocular surface to help protect it from infection. Studies have shown that use of AMPs (modified or not) from animals or plants or synthetic (often smaller) AMPs can be used to reduce keratitis in animal models. Further studies are needed in this area to define pharmacokinetics and pharmacodynamics, as well as safety profiles of these AMPs or mimics.

Conclusion
Considering the rates of ocular antimicrobial resistance and the gradual decline in the efficacy of fluoroquinolones, clinicians may need to re-evaluate the empirical methods of treating common ocular infections and the outcomes of studies examining the resistance characteristics of ocular microbes will help in this regard. While large pharmaceutical companies have virtually ceased all antimicrobial development, 223 it is important that research and development in this area is continuing. The development of AMPs is ongoing and showing encouraging results for their use to treat ocular surface infections. Future research should focus on ways to decrease the resistance development and the evaluation in vitro and in vivo of new antimicrobial agents.
[Correction added on 5 Oct 2020 after first online publication: Reference numbers starting from the "Ocular microbial resistance" heading to the end of the article, including the figure and tables, have been updated.]