Global Reemergence of Tuberculosis: Are Host Defense Peptides an Option to Ameliorate Disease Burden?

Introduction

New Alternatives for Treating Antibiotic-Resistant Infections: Host Defense Peptides

Close to 70% of humans who have been in contact with Mtb never develop an overt disease, thus leading to the hypothesis that host defense peptide (HDP) function might explain, at least partially, this resistance as part of several determinants of host defense. We can also envision that HDPs could be one of the many host effectors helping to maintain (if not promote) a latent infection. Considering these ideas, we decided to explore how the currently proposed therapies based on HDPs could be applied to battle the worldwide reemergence of TB. A recent review describing the participation of β-defensins during TB infection has recently been published. ³⁸ In this work, we will focus on the alternative of employing the other family of major human neutrophil peptides (HNPs), the cathelicidins.

Antimicrobial activity from secreted factors present in blood, urine, leukocytes, and lymphatic tissues have been recognized since the early XIX century, leading to the isolation of many microbicides during 1920–1950. ⁶⁰ Today, there are more than 900 antimicrobial peptides that have been either isolated or predicted from available genome sequences, which come from different sources such as bacteria, plants, insects, amphibians, and mammals. ³ These molecules have been classified according to structural motifs or amino acid composition. In vertebrates, two main families have been extensively characterized: cathelicidins and defensins.

Cathelicidins are cationic and amphipathic, are synthesized as propeptides, and contain a conserved amino terminal region (cathelin domain), with the antimicrobial part of the protein, which is variable in sequence and located at the carboxy terminus.^41,43 Their size varies between 12 and 80 amino acids, and they could be subdivided into three groups: (a) those forming amphipathic alpha helices (human LL-37), (b) those containing intramolecular disulfide bridges forming beta sheets (porcine protegrins), and (c) those highly enriched in proline or arginine (bovine bactenecins). ³ In contrast, defensins are more uniform given that they are cysteine-rich peptides with beta-sheet structures, stabilized by three or four disulfide bridges. In vertebrates, there are three subfamilies, alpha, beta, and theta defensins.

In recent years, increased numbers of antibiotic-resistant microorganisms, including incurable diseases provoked during infection with XDR-TB, vancomycin-resistant Enterococcus faecalis, or methycillin-resistant Staphylococcus aureus (MRSA), have shown the need for new and improved antibiotics. Perhaps we should not overlook that because on a daily basis humans are exposed and ingest millions of microbes via airways, with many of them reaching the lungs, where they are trapped within mucous secretions that have plenty of innate immunity effectors (including antimicrobial peptides), thus constituting the first line of defense, and often hindering the progress of an infection. ⁵³

Focusing on the particular problem presented by TB, it is worth considering the use of host-derived peptides, particularly those of the cathelicidin family, and the rather modest in vivo effect of HNPs ⁵⁸ in treating this infection. Cathelicidins have shown microbicidal effects against clinical isolates of S. aureus, E. faecalis, or Pseudomonas aeruginosa, ⁶⁸ and they protect against urinary tract infections caused by uropathogenic Escherichia coli. ⁷ In fact, by utilizing a chemoinformatics approach with a neural network, Cherkasov et al. and Fjell et al.^5,11 developed in silico a number of synthetic cathelicidins. Among them, the one named HHC-10 showed improved in vitro activity over Bac2A against MDR P. aeruginosa (15–60-fold minimum inhibitory concentration [MIC] decrease), MRSA (16-fold MIC decrease), vancomycin-resistant Enterococus faecalis (2–32-fold MIC decrease), and other clinically relevant bacteria.^5,11 As a matter of fact, results showed that HHC-10 had lower MIC values than MX-226, a peptide already in phase IIIa clinical trials. ¹⁶ Further, HHC-10–treated mice decreased almost 3 log the colony-forming unit (CFU) burden of S. aureus without toxic side effects, whereas MX-226 did not alter bacterial replication. ⁵ In light of these results, it might be worth analyzing the potential of such newly designed derivatives against a number of Mtb isolates, using in vitro and in vivo models.

Bacterial Resistance Mechanisms Against HDPs

Often, laboratory strains of E. coli have been used to determine the mechanism of action of several HDPs. ³ Some of these HDPs may or may not form pores in the membranes and reach the cytoplasm, with in vitro evidence of their capacity to bind DNA or RNA, as occurs with buforin II and indolicidin,^20,35,49 or they may interfere with the chaperone DnaK as with pyrrocoricin.^25,46 In some other cases, like polyphemusin I, the mode of action remains to be determined.^54,69 Likewise, how some MRSA are more resistant to LL-37 than their methicillin-susceptible counterparts remains to be elucidated. ⁴⁷ The study of the interaction of HDPs with microorganisms has not been restricted to workhorse laboratory strains only, but also to medically relevant bacteria. Our focus resides on Mtb, a guanine and citosine (GC)-rich gram-positive organism (along with Corynebacterium and Streptomyces), ⁴⁸ for which only scarce evidence detailing the mode of action of HDPs exist, and studies on HDPs potential (either in vitro or in vivo) to attack this species are limited. Therefore, we will mention studies performed using Bacillus, Staphylococcus, Streptococcus, and Listeria (gram-positive bacteria with low GC content). ⁴⁸

Genetic screenings based on the analysis of transposon mutant libraries have allowed the mapping of bacterial molecular determinants of susceptibility and resistance to microbicidal compounds. Based on the strategy employed, mutants can be divided into two categories: (a) susceptible mutants isolated from an originally resistant species and (b) resistant mutants isolated from an originally susceptible organism. The former approach has been widely used and has helped to determine that the absence of certain bacterial cell surface modifications in S. aureus, for example, teichoic acid d-alanylation^8,51 or phosphatidylglycerol l-lisynilation,^27,50 result in greater susceptibility to cationic antimicrobial peptides in vitro, neutrophil-mediated killing, and diminished virulence in mice. In vitro results were independently confirmed using a combination of genetic screen and genome-wide transcriptional profile with microarrays, which helped in demonstrating that such surface modifications depend on a three-component system.^17,29,30 Interestingly, d-alanylation is conserved and promotes similar phenotypes in Group A Streptococcus (GAS) ²⁶ and Streptococcus pneumoniae, ²⁴ whereas phosphatidylglycerol l-lysinylation in Listeria monocytogenes is required to mediate resistance toward cationic antimicrobial peptides. ⁶³ In fact, a recent report showed evidence of phosphatidylglycerol l-lysinylation being relevant to resist cationic antibiotics and peptide challenge in Mtb. ³⁴

S. aureus, GAS, E. faecalis, and P. aeruginosa possess yet another mechanism of defense, with the secretion of HDP-degrading proteinases,^55,59 albeit their relevance during in vivo infection remains to be proven. A ponA (penicillin-binding protein) mutant of GAS was more susceptible to human beta defensin and cathelicidin-related antimicrobial peptide (CRAMP, mouse cathelicidin) and was more rapidly cleared during a pulmonary infection in a rat model. ²²

Using an approach for finding mutants resistant to HDPs in GAS allowed the identification of a GntR-family transcriptional regulator, where the mutant produced more severe skin lesions in infected mice compared with those produced by wild-type, susceptible organisms. ⁴⁴ All of these findings clearly suggest that similar approaches can be used with Mtb, aiming to molecularly define which determinants confer the unique resistance of this pathogen to the plethora of agonists the immune system mounts against it. Inactivating genes homologous to those already proven to be relevant in other species seems like a direct approach, although conservation of such genes often does not take place, perhaps reflecting the different niches and stresses they encounter.

Further, Mtb, a much successful pathogen, relies on a variety of active and passive resistance mechanisms ranging from nonreplicating persistence to limited antibiotic diffusion. The use of antimicrobial peptides as novel drug candidates would abrogate most of the resistance mechanisms. For instance, many HDPs target and lyses the cell membrane, a compartment that is the most conserved portion of the bacterial cell. ³ The bacteria would have to dramatically alter its cell membrane, thus likely intimately affecting its entire life cycle. This ability to act against any persistent or resistant state is not true for most of the small molecules that depend on the bacteria actively replicating to be highly susceptible to them. ¹⁴

As the mycobacterial cell is surrounded by a highly complex and mycolated cell wall, which is implicated in drug resistance and resistance to innate immune system, HDPs ability to potentially permeabilize the bacterial cells to small molecules could be one of the most interesting applications. Given that the known targets of these two drug classes (conventional antibiotics and HDPs) are different, the utilization of HDPs could probably increase the potency of classical small molecule drugs several fold.

HDPs and Their Activity Against Mtb

Despite their demonstrated capacity to eliminate several pathogens both in vitro and during in vivo infection, few groups have endeavored using HDPs to fight Mtb. The defensin-class HNP1 and rabbits as well as porcine protegrin-1 were shown to be bactericidal in vitro (ranging from 86% to 99% killing) against an avirulent Mtb laboratory strain (H37Ra) at ∼15 mM (50 μg/ml) and 23 mM, respectively; moreover, the killing capacity was equal to that observed when clinical isolates were tested (range, 86–99%). ³⁹ Another defensin, porcine protegrin PR-39, eliminated 80% of the virulent laboratory strain H37Rv in vitro, but only killing 39% and 49% of two clinical isolates. ³¹ Other in vitro studies showed that using ∼30 mM of a HNP1-3 mixture resulted in a 4-log drop in H37Rv CFUs compared with nontreated cells, whereas 22 mM of the cathelicidin LL-37 resulted in less than a 1-log decrease.^32,36

In the murine macrophage cell line J744A.1, ∼12 mM (40 μg/ml) of HNP1 resulted in ∼98% H37Rv killing, ⁵⁷ thus demonstrating that peptides of this family are able to reach as well as effective to kill intracellular bacteria. Whether activity of other peptides for intracellular and particularly intraphagosomally harbored bacteria exists remains to be experimentally determined. Further, after 4 weeks of treatment, the use of 5 μg of HNP-1 per mouse on a weekly basis resulted in more than a 1-log-unit decrease in CFU numbers compared with nontreated controls. ⁵⁸ This led the authors to speculate over the reported capacity of HNP-1 to recruit neutrophils, ⁶⁵ T cells, ⁶ and monocytes, ⁶² as well as its capacity to induce interleukin-8 in airway epithelium ⁶⁴ as potential explanations for the need for a low dose for eliminating H37Rv as opposed to what is needed in vitro. The same group demonstrated synergism of HNP-1 with RIF and INH, decreasing the concentration of each bactericidal required for eliminating H37Rv from infected organs in a murine model. ²³ Other than these results, no more investigation on the use of HNPs to treat TB infection has been published recently, although according to Méndez-Samperio, ³⁸ a multidisciplinary effort to develop new defensins has already begun. A summary of HDPs' activities against Mtb is shown in Table 1.

Would Synthetic Cathelicidins be An Alternative to Treat TB?

In view of the aforementioned results, what would be an ideal drug candidate against TB? The following points listed below could guide any new drug development (best described as target product profile):

Exhibit a novel mechanism of action.

As HDPs satisfy all of the above criteria, they can be considered as prime candidates for drug development against TB.

A relevant starting point for considering the use of cathelicidins as an alternative treatment for TB cases is that human beings showing a diminished capacity to produce LL-37 present a greater risk for developing overt disease,^32,36 although the peptide was not potently microbicidal against Mtb in vitro or when it is expressed in cell culture, resulting in elimination of 50–75% of the infecting bacteria.^33,37 Further, native LL-37 is susceptible to degradation by bacterial proteases,^55,59 and toxic effects such as mast cell degranulation ⁴² and apoptosis induction in epithelial cells ²⁸ have been reported. An approach designed to circumvent these disadvantages relies on applying amino acid substitutions to native sequences, as has been applied to EFK17, a LL-37–derived peptide. Some tryptophan substitutions rendered molecules with greater resistance to proteolysis by human neutrophil elastase, S. aureus aureolysin, and V8 protease. ⁶¹

On the other hand, bactenecins (bovine cathelicidins) are small cyclic molecules (12 amino acids), which were isolated in 1988 by Romeo et al. from neutrophil granules. ⁵⁴ This molecule and its derivatives have been synthesized in vitro since the early 1990s, and some of the synthetic variants have shown enhanced microbicidal capacity against gram-positive bacteria. One of these molecules, Bac2A, had an MIC of 1 μg/ml against Corynebacterium xerosis (an actinomycete related to Mtb), and some of its variants were active even at 0.25 μg/ml, with no hemolytic or hemagglutination activity. ⁶⁷ Bac2A was chemotactic for human macrophage-like THP-1 cells but did not inhibit tumor necrosis factor-alpha production in response to lipopolysaccharide stimulation. ² Recently, Bac2A was used as a starting molecule for a combinatorial approach, and some of its evolved derivatives showed MIC values 9–20 times lower than the parental molecule against Staphylococcus and Enterococcus, ¹⁸ a process that has been improved to diminish the cost and time of production to an estimated 95%. ¹⁹ Unfortunately, the activity of synthetic, improved molecules against C. xerosis (or better yet, a mycobacteria) was not determined by Hilpert et al., ¹⁸ and also the safety for human cells was not assessed. A question that certainly deserves attention is whether Mtb is susceptible to Bac2A or any of its derivatives, as well as determining the mechanism of action of these molecules or the mechanisms of bacterial defense. Moreover, evaluating the in vivo efficacy of the synthetic variants for controlling or eradicating infectious agents is worth testing too. Other HDPs with chemotactic activity on immune cells can serve as a bridge between the innate and cellular immune systems. ¹³

Concluding Remarks and Perspectives

Synthetic peptides showing improved bactericidal effects and/or capacity to stimulate protective immune responses^40,56 could provide an alternative for shortening long-lasting drug regimens and help reduce the global MDR- and XDR-TB burden. The current drugs prescribed for MDR cases include a number of injectable drugs such as kanamycin, amikacin, and/or capreomycin, which makes the hospitalization of patients compulsory, an event which in the long run is not logistically feasible. However, there is a strong demand for molecules to be given as inhaled antibiotics because they are able to act much more potently at the site of infection, which are most commonly the lungs. This route of administration, when perfected, would massively increase the spectrum of HDP utilization.

We must consider that even though there are many mechanisms of resistance against HDPs, ³ their resistance frequency is in the order of ∼10⁻¹¹−10⁻¹². Considering that an active pulmonary TB case has around 10⁸ cfu/ml, ⁴ the chances of generating a resistant mutant become next to impossible, once a microbicidal molecule has been found, isolated, and tested to be effective. In comparison, the resistance frequency for INH and RIF is ∼10⁻⁷–10⁻⁸; thus, at the start of the multidrug therapy, there could be ∼1–1,000 bacteria already resistant to the drugs present in the population. ¹ After the administration of these drugs, these resistant bacteria would be amplified, possibly leading to increased relapses of infection, and thus increasing the spread of MDR-TB. ¹⁴ Perhaps, the use of improved synthetic HDPs would thwart this trend.

On the other hand, Mtb is a sui generis organism, which in many instances lacks genes homologous to those required in other bacteria to resist HDPs' challenge. Further, it has a peculiar cell wall, enriched with complex lipids, some of them conserved in actinomycetes, particularly in Corynebacterium, ¹⁰ and others restricted to the Mtb complex. These complex lipids are at least partially responsible for the limited accessibility of many conventional antibiotics. ⁶⁸ In summary, it might be well worthwhile to consider the use of synthetic cathelicidins to combat TB. If they prove to be unsuccessful, then the aim should be unraveling the molecular mechanisms governing resistance to these molecules.