Pneumolysin as a target for new therapies against pneumococcal infections: A systematic review

Background This systematic review evaluates pneumolysin (PLY) as a target for new treatments against pneumococcal infections. Pneumolysin is one of the main virulence factors produced by all types of pneumococci. This toxin (53 kDa) is a highly conserved protein that binds to cholesterol in eukaryotic cells, forming pores that lead to cell destruction. Methods The databases consulted were MEDLINE, Web of Science, and Scopus. Articles were independently screened by title, abstract, and full text by two researchers, and using consensus to resolve any disagreements that occurred. Articles in other languages different from English, patents, cases report, notes, chapter books and reviews were excluded. Searches were restricted to the years 2000 to 2021. Methodological quality was evaluated using OHAT framework. Results Forty-one articles describing the effects of different molecules that inhibit PLY were reviewed. Briefly, the inhibitory molecules found were classified into three main groups: those exerting a direct effect by binding and/or blocking PLY, those acting indirectly by preventing its effects on host cells, and those whose mechanisms are unknown. Although many molecules are proposed as toxin blockers, only some of them, such as antibiotics, peptides, sterols, and statins, have the probability of being implemented as clinical treatment. In contrast, for other molecules, there are limited studies that demonstrate efficacy in animal models with sufficient reliability. Discussion Most of the studies reviewed has a good level of confidence. However, one of the limitations of this systematic review is the lack of homogeneity of the studies, what prevented to carry out a statistical comparison of the results or meta-analysis. Conclusion A panel of molecules blocking PLY activity are associated with the improvement of the inflammatory process triggered by the pneumococcal infection. Some molecules have already been used in humans for other purposes, so they could be safe for use in patients with pneumococcal infections. These patients might benefit from a second line treatment during the initial stages of the infection preventing acute respiratory distress syndrome and invasive pneumococcal diseases. Additional research using the presented set of compounds might further improve the clinical management of these patients.

Introduction [18]. (ST217). This serotype is a major cause of invasive pneumococcal disease globally, especially in Africa, South America, and Asia, with geographically distinct sequence types (STs) that form three genetic clusters designated as lineage A, B, and C [19].
Structurally, PLY consists of 471 amino acids (53 kDa) and has a three-dimensional conformation with 4 different domains; three of them (domains 1, 2, and 3) have structural importance, conferring stability to the PLY molecule, and are essential for oligomerization and pore formation in eukaryotic cell membranes. Domain 4 (hydrophobic region, loops L1-L3) forms the C-terminal region and is what promotes cholesterol binding, favoring the insertion of an undecapeptide sequence in the membrane, which allows for the oligomerization of toxin monomers, and the further formation of pores. The PLY pore is a 400 Å ring of 42 membraneinserted monomers [20]. (S1a Fig) [21,22].
This toxin can activate different cell death pathways such as apoptosis [23], pyroptosis [24], or necroptosis [25], which releases membrane-derived vesicles (microvesicles and exosomes). The pores of PLY cause a strong mitochondrial calcium influx which triggers mitochondrial morphological alterations with the release of mtDNA through microvesicles [26] and could regulate innate immune responses [27] (S1a Fig). During the inflammatory process, PLY activates several signal transduction pathways such as the nuclear factor kappa-B (NF-κB), mitogen-activated protein kinase (MAPK), and the NOD-like receptor pyrin domain-containing 3 (NLRP3) inflammsone [28,29]. The interaction between PLY and TLR4 remains controversial. Some authors showed that the interaction of PLY with TLR4 is involved in induce cytokine production and apoptosis [30,31], while other research showed that PLY activates the NLRP3/ACS inflammasome to enhance the secretion of pro-inflammatory cytokines IL-1β and IL-18 from macrophages and dendritic cells and contributes to the protection of the host from pneumococcal infection independent of TLR-4 and mediated by K + influx [24,32].
Therefore, PLY is able to activate and regulate a huge number of genes for chemokines, cytokines, and other molecules whose expression is involved in the recruitment of inflammatory cells via neutrophils, macrophages, and phagocytes activation (such as IL-8, MCP-3, MIP-1β, lysozymes, caspases, TNF, IL-1 and IL-6) [33,34]. Overproduction of early cytokines has been associated with tissue injury, organ dysfunction, morbidity, and mortality [35].
On the other hand, PLY is the main cause of pulmonary permeability edema due to its ability to alter both endothelial and epithelial barrier function [36]. This effect is a consequence of reducing dynamic and stable microtubule content in the endothelial monolayer and influencing VE-cadherin expression [37]. NLRP3 protect the alveolar barrier againts PLY injury [38]. Several cellular channels or transporters of the alveolar epithelium are implicated, such as the epithelial sodium channel (ENaC), the Na + / K + -ATPase and K + channels [39] (S1a Fig). Moreover, PLY play a key role in human nasopharynx colonization and in the transmission of S. pneumoniae from host to host since by promoting inflammation there is an increase in elimination [40]. In addition, S. pneumoniae can invade other parts of the body via bloodstream dissemination, and it can gain access to normally 'sterile' sites such as the lower airways or meninges [41]. (S1b Fig). When bacteria reach the cerebrospinal fluid, produce meninges inflammation, resulting in hyperemia and ischemia, and eventually permanent brain injury [42].
Finally, it has been found that CAP (community acquired pneumonia) can incite up to 30% of the cases to cardiovascular events, such as myocardial damage or pro-thrombotic effects [43]. Myocardial damage is closely related to the ability of PLY to form pores in the cell membrane: movement of Ca 2+ into the cell leads to an efflux of K + with the consequent depolarization of the membrane, contributing to myocardial contractile dysfunction [44]. Regarding prothrombotic effects, this has recently been shown not to be true. PLY does not activate platelets to form thrombus, rather it destroys them by forming pores in their membrane [45] and destroy procoagulant microvesicles impaired coagulation of blood [46]. Current pneumococcal conjugate vaccines are 13-valent conjugate vaccine (PCV13) for routine pediatric immunization and a 23-valent polysaccharide vaccine (PPSV23) for adults aged �65 years. Since 2014, PCV13 was also recommended for all adults aged �65 years [47]. In 2021, two new vaccines were approved by the FDA for use in adults �18 years (PCV- [15][16][17][18][19][20] that are presently under evaluation [48]. Although these vaccines are immunogenic and effective and prevent disease caused by the serotypes determined by their capsule types, they do not cover the full spectrum of invasive pneumococcal serotypes. The management of pneumococcal infections in clinical practice frequently involves the use of broad-spectrum antibiotics, typically a combined therapy of β-lactams and macrolides [49]. However, S. pneumoniae has developed resistance to multiple antibiotics including penicillin, macrolides, fluoroquinolone, and sulfamethoxazole-trimethoprim [50]. The emergence of non-vaccine serotypes after the introduction of PCV, together with increased antibiotic resistance in these serotypes, has become a global threat [51]. Taking together the development of new therapies is necessary for the effective treatment of pneumococcal infections. Even though PLY-neutralization strategies have been reviewed before, these are partial reviews, that approach the subject from different perspectives, and the vast majority of these reviews focus on vaccination-based strategies (see for example [3,[52][53][54][55][56]). Furthermore, several previous reviews have been carried out on small-molecule-based PLY neutralization strategies that either suffer from being focused on a single family of compounds and do not represent complete overviews of the current state of the art, such as the work of Nishimoto et al. (2020) [57] which is focused on statins, the work of Li et al. (2017) [58], focused on sterols, or the work of Anderson and Feldman (2017) [59], which although it encompasses a broader collection of compounds, it cannot be considered complete and current to date.
The aim of this review was to investigate the main molecules that directly and indirectly interfere with PLY activity in host survival, colonization, infection, and transmission. A secondary objective was to identify the mechanisms of interaction with the toxin.

Materials and methods
This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines [60] as much as possible, since it is not a medical systematic review, so part of the guidelines were not applicable (S1 Table).

Search strategy and selection criteria
The inclusion criteria of the studies for the review were articles in which PLY was examined as a therapeutic target through studies of molecules that inhibit its effects. To analyze the therapeutic role of PLY in pneumococcal infection, the terms "pneumolysin" AND "therapeutic" were searched in three databases: Web of Science, MEDLINE and Scopus. At first, we only limited the search date to November 25, 2021; however, in the exclusion criteria, we decided to refine and not include articles prior to the year 2000. Therefore, the review includes the articles found from January 1 st , 2000 to November 25 th , 2021. Reference lists of selected articles were also reviewed. Several exclusion criteria were taken to ensure the quality of the study. The following were excluded: patents, articles in other languages different from English, cases report, notes, chapter books, reviews, and articles prior to 2000 year. Potentially eligible articles were reviewed by two independent reviewers (MMG-S and MDCC). Disagreements between the two reviewers were resolved by consensus.

Selection procedure
Citations were saved in the reference manager EndNote (https://endnote.com/) and in an Excel file that also included the following items: publication type, authors, article title, abstract, source title, DOI, and publication date. EndNote was used to deduplicate the search results. Full texts of these articles were analyzed by the two authors (MMG-S and MDCC). Articles on the pathogenesis of pneumococcus in general and articles in which PLY was presented as a vaccine and as tumor therapy were excluded from the study.

Data collection process
Two authors collected data from articles in an electronic spreadsheet (S2 Table). The following data were obtained from each article: tests performed on cell lines and laboratory animals, CAS number and tested doses of the compound, route of inoculation, type of infection, and strain of S. pneumoniae used as well as other data considered of interest.

Data analysis and synthesis
Quantitative and qualitative outcome data were provided and synthesized where possible. Due to data heterogeneity a meta-analysis was not possible.

Article quality assessment
Each article was scrutinized to determine the techniques that led to each molecule being proposed as a candidate for toxin-blocking therapy. The risk assessment tool used was the Office of Health Assessment & Translation (OHAT) risk of bias tool by the US Department of Health and Human Services [61]. Risk assessment was based on study type. This screening was tabulated in S3 Table. Moreover, we also inspected the number of papers of clinical trials indexed in MEDLINE for each proposed molecule (S4 Table; Fig 2).

Network of neutralizing PLY-effects molecules and mechanisms underlying
The knowledge network harboring the molecules, the processes affected, and the cellular targets studied in this review were all generated using the tool EXTRACT 2.0 [62] to gather the terms from the articles' abstracts. From the resulting dataset, a graph was built using Cytoscape 3.9.1 [63], using the yFiles layout algorithms. Finally, some images were obtained from BioRender.com (2022).

Study selection
A total of 366 records were found using the search criteria, of which 203 were found in WOS, 161 in MEDLINE and 2 in Scopus (Fig 1). Once the duplicates were eliminated, 192 articles were reviewed by title and abstract by two independent authors (MG and MC). Several exclusion criteria were taken into account to ensure the review quality. As a result, 25 reviews, 16 articles, 3 patents, 2 cases report, 2 articles in other languages different than English, 2 chapter books, and 1 note prior to the year 2000 were excluded. The full texts of the resulting 141 articles were reviewed in detail. This refined search showed that 108 studies did not meet the criteria and were excluded (74 reports about the effect of PLY as a vaccine, 3 reports focused on tumor therapy, 31 reports studying general pathogenesis). References from these relevant articles were also screened. Forty-one (n = 41) studies met the inclusion criteria (describing molecules that inhibited the effects of PLY) were included in the systematic review.

Study characteristics
A detailed analysis of the 41 resulting articles was performed, in which the effects of PLY were studied using molecules that block it. Most of the studies were published by scientists from Southeast Asia, principally China (n = 17), followed by Europe (n = 10), and the USA (n = 9) (S4 Table). In 15 of the articles, the experiments were carried out using only cell cultures, while in 26, animal models of pneumococcal infection were used (24 articles induced a pneumonia model, of which two also induced systemic infection) [64,65], one model of meningitis [66], two of keratitis [67,68], one of atherosclerosis [69] and one of nasopharyngeal colonization [70]. The most frequently used animal models were mice for pneumoniaand rabbits for keratitis. Meningitis was tested in rats and systemic infections were tested in Zebrafish embryos.

Methodological quality
The OHAT quality scores of the included articles are described in S3 Table. The overall risk of bias rating was considered as Low or Probably Low. The 41 studies were summarized and classified into 10 groups according to chemical structures: Plant-derived compounds, sterols, statins, omega-3 fatty acids, purin-6-ones, thioethers, antibiotics, peptides, cations, and antibodies ( Table 1). Moreover, because of the high heterogeneity of the data found in the articles, it was very difficult to compare the action mechanism against the toxin from each study, so they were grouped into three categories for clarification: direct (the molecule binds to or competes with PLY for binding to the target), indirect (the molecule does not bind PLY), and unknown (molecular mechanisms are currently unknown).
On the other hand, Fig 2 shows the number of articles indexed as 'clinical trials' on MED-LINE (without time restrictions), illustrating that only 17 of 39 molecules present experimental human evidence to potentially conclude data on their efficacy, toxicity, or bioavailability. The distribution of studies regarding molecules is unequal. For some molecules, there are more than 1000 studies (like cases of calcium and cholesterol), for others, the number ranges from 50-100, and for 10 of them, there are less than 20 studies. As expected, more trials and global studies are found to specifically study sepsis and pneumonia molecules. Thus, most trials focused on sepsis are limited to molecules, exhibiting more than 1000 global studies. Contrastingly, compared to the total number of trials, there is a high number of antibiotics studies in the context of pneumonia.

Types of molecules and drugs tested
Plant-derived compounds: Polyphenols, flavonoids, tannins, quinones, phenethylamines and terpenoids. The seventeen plant-derived compounds included in this review belonged to six categories according to their chemical characteristics. Some of them present antibacterial activity [71,72] and therapeutic effects on acute lung injury since regulating the TLR4/NF-κB signaling pathway, NLRP3 inflammasome activation and the MAPK signaling pathway [73,74].
Polyphenols are a large and diverse family of naturally occurring organic compounds abundant in plants. They are generally subdivided into phenylpropanoids, flavonoids, hydrolysable tannins, and condensed tannins. In this work, flavonoids and tannins are described independently for greater clarity. An example of polyphenol is Verbascoside, a phenylpropanoid that binds to the cleft between domains 3 and 4 of PLY (Asp471, Asn470, Glu277, Tyr358, and Arg359), as was shown by the molecular dynamic's simulation, thus blocking the conformational transition from monomeric to oligomeric form and inhibiting its lytic activity. Verbascoside was tested in animals suffering from pneumonia with protective effects and reduction of levels of TNF-α and IL-1β [75]. Flavonoids possess the basic structure of a chromone (1,4-benzopyrone) moiety connected to a phenyl ring at position 2. Some examples such as Apigenin [76], Acacetin [77], Amentoflavone [78], and Morin [79] decreased the hemolytic activity of PLY by inhibiting oligomerization. In animal models, the inflammatory cytokines INF-γ and IL-1β in the lungs of mice treated with these compounds decreased significantly. Other flavonoids such as Epigallocatechin gallate (EGCG) significantly increased the survival of infected animals due to it affected PLY oligomerization by binding to the toxin residues Ser256, Glu277, Tyr358, and Arg359 [80]. In addition, this molecule interferes with the biofilm formation and bacterial adherence to cells. Another interesting flavonoid is quercetin, which has been tested in animal models of lung infection, with survival rates of 80% compared to 60% in untreated mice [81]. For other flavonoids such as Dryocrassin ABBA, data from animal models are not available, but in vitro results showed relative bactericidal activity and the ability to interfere with toxin oligomerization capacity at concentrations ranging from 0.5 μg/ml to 2 μg/ml [82]. Recently, 27 hydrolysable tannins were tested by Maatsola et al. (2020) [83]. Nanomolar concentrations of Pentagalloylglucose (PGG) and Gemin A were capable of inhibiting the PLY cytolytic capacity. Molecular modeling suggests that PGG also binds to the cleft between domains 2, 3 and 4 (Glu42, Ser256, Asp257, Glu277, and Arg359).
Quinones are a class of organic compounds formally derived from aromatic compounds. Shikonin [84] and Juglone [85] are quinones (naphthoquinones) that inhibited pore formation by means of interfered toxin oligomerization. Aloe-Emodin is a quinone (anthraquinone) with antifungal activity. The authors have proposed this compound together with photodynamic therapy to treat superficial infections of antibiotic-resistant gram-positive bacteria (Enterococcus faecalis, Staphylococcus aureus, and S. pneumoniae). Irradiation generates reactive oxygen species (ROS), leading to the destruction of biomolecules and the killing of bacterial cells; however, the mechanism by which cytotoxins expression (such as PLY) decreases is unknown [71]. Application in vivo requires future investigation.
Ephedrine hydrochloride and Pseudoephedrine hydrochloride are phenethylamines from Ephedra sinica granules [86] and MXSGT (Ma-xing-shi-gan-tang) [87] are compounds obtained from plants used in traditional Chinese medicine. All of them seem to inhibit the oligomerization of the toxin, although more studies are necessary to specify which residues of PLY are involved. In addition, Ephedrine hydrochloride is part of the WHO model list of essential medicines and is used to prevent low blood pressure during anesthesia, and for the treatment or prevention of attacks of bronchospasm in asthma. It is a very active agonist adrenergic on the receptors of the sympathetic nervous system. Ephedrine hydrochloride increased anti-inflammatory cytokine production IL-10 and decreased the production of proinflammatory cytokines TNF-α and IL-12 in dendritic cells in Staphylococcus aureus-induced peritonitis animal models [88]. Current clinical trials are focused on the effect of this compound on sinusitis, rhinitis, etc.
Oleanolic Acid, Betulin and Hederagenin are terpenoids. Oleanolic Acid and its analogues significantly inhibited the activity of important β-lactamases like NDM-1, KPC-2, VIM-1, and OXA-1 as well as the hemolytic action of cytolysins like PLY [89]. For S. aureus α-hemolysin, the authors showed through molecular modeling how oleanolic acid can interfere with toxin oligomerization and protect S. aureus-infected mice in a combination therapy with β-lactams. More research is needed to see these effects in vivo on other toxins, including PLY. Betulin [90] and Hederagenin [91] are triterpenes extracted from plants that interfere with the PLY oligomerization process, without toxic effects against epithelial cells, but have not yet been tested in animal models of pneumococcal infection.
Although most of these above mentioned molecules inhibit or hinder the PLY oligomerization, only three articles [75,78,85] studied potential interactions between molecules and toxin by using molecular modelling and docking calculation. Thus, it was probed that the flavonoid Amentoflavone was able to interact with domain 2 (Arg359) and domain 3 (Ser254, Glu277) of the toxin; similarly, to the polyphenol Epigallocatechin gallate, which forms strong interactions with the PLY domain 3 (Ser256, Glu277) and domain 2 (Tyr358, Arg359). Furthermore, Asp 471 from domain 4 of PLY could form a strong hydrogen bond with the hydroxyl group of the benzene ring on the right side of the polyphenol Verbascoside. On the other hand, the interactions of the hydrolysable tannin PGG and PLY take place through flexible galloyl groups of PGG and the pocket formed by domains 2, 3 and 4 of PLY [83]. To our knowledge, no controlled clinical trials have addressed the issue of adjuvant plant-derived compound therapy in the clinical setting of pneumococcal infections. As exception, recent clinical trials using Quercetin (or products containing it) for treating pneumonia in patients with COVID-19 have showed a significantly reduced length of hospitalization, noninvasive oxygen therapy, and number of deaths [92], which clearly confirms its potential as an anti-inflammatory molecule.
Sterols. Two sterols, cholesterol and β-sitosterol, inhibited the binding of the toxin to the membrane. Cholesterol is the natural ligand for PLY in the eukaryotic cell membrane. Marquart et al. (2007) [67] showed that cholesterol has a bactericidal effect both in vitro and in corneas of rabbits with experimental pneumococcal keratitis. The authors suggest that cholesterol would not only bind to PLY, but also to bacteria and kill them. β-sitosterol is a phytosterol with chemical structure very similar to cholesterol, that has been assayed in murine pneumonia model with good protection results [93]. Additionally, β-sitosterol possess anti-inflammatory properties, causes a dose-dependent inhibition of IL-6 and TNF-α in endotoxin-activated human monocytes, and has beneficial effects on the immune system by increasing the number of viable peripheral blood mononuclear cells and activating the dendritic cells [94,95]. A detailed study of the binding between five sterols and PLY has confirmed that PLY interacts with these molecules through residues Tyr371, Val372, Leu460, and Tyr461. In the five natural sterols studied, the critical structure is the C22-C23-C24-C25 carbon bonds [58]. We did not find clinical trials of pneumonia related to these sterols.
Statins. The protective effect of simvastatin and pravastatin against the cytotoxic effect of PLY in vitro was described by   [96,97]. Oral treatment with simvastatin in pneumonia mice reduced neutrophil infiltration, maintained vascular integrity, and decreased chemokine production [98]. Statins may trigger mevalonate-independent pathways partially via a calcium increase and p38 MAPK activation. Statins act as competitive inhibitors of HMG-CoA reductase (HMGCR), the rate-limiting enzyme of cholesterol synthesis. Simvastatin was also used in a mouse model of sickle cell disease, which is characterized by hemolytic anemia and chronic inflammation, and was shown to lead to a high incidence of invasive pneumococcal pneumonia. The last was due to upregulation of platelet-activating factor receptor (PAFr). In this context, Simvastatin is useful as it reduced PAFr expression and interfered with toxin pore formation and host cell bacteria invasion [99]. Statins are widely used in the treatment of cardiovascular disease in humans. In last years, statins have begun to be studied in the context of infectious diseases like tuberculosis, AIDS, and COVID-19 [100][101][102]. Several clinical trials have been carried out with simvastatin in patients with CAP with contradictory results. Some authors find that prior use of this statin improves mortality in patients admitted with CAP [103] while other authors find no difference [104][105][106]. On the other hand, the administration of 20 mg in individuals admitted with CAP does not show differences in cytokine levels [107]. Recently, clinical trial have been carried out in older adults with derivate CAP bacteremia with encouraging results [108]. Simvastatin could be a good candidate to complement antibiotic therapy. In addition to the bactericidal effect against Gram positive microorganisms, statins have pleiotropic effects on cells of the immune system (i.e., stimulate autophagy in macrophages, increase the number of NK and NKT cells, inhibit MHC-II expression, increase serum levels of IL-10) [108].
Omega-3 fatty acids. Omega-3 fatty acids are polyunsaturated fatty acids (PUFAs) characterized by the presence of a double bond, three atoms away from the terminal methyl group in their chemical structure. The structure of docosahexaenoic acid (DHA) is a carboxylic acid with a 22-carbon chain and six cis double bonds, with the first double bond located at the third carbon of the omega end. DHA decreased IL-8 PLY mediated in human neutrophils by interfering with Ca 2+ influx [109]. Also, DHA hinder the binding of toxin to cellular membrane of neutrophils by an unknown mechanism. However, it is accepted that lipid rafts are putative binding sites for cytolysins and that omega-3 polyunsaturated fatty acids exclude proteins from these lipid rafts in eukaryotic cell membranes.
Omega-3 fatty acids are widely known in the field of nutrition and their role is being investigated in neurodegenerative diseases, sleep, etc. As molecules regulating inflammation in the context of infectious diseases, they have been reviewed by Sandhaus & Swick (2021) [110]. The specialized pro-resolving mediators (SPMs) are enzymatically derived from essential fatty acids, including arachidonic acid, eicosapentaenoic acid, and DHA, and have important roles in the resolution of tissue inflammation. DHA is converted to D-series resolvins, protectins, and maresins via 12-and 15-LOX enzymes. DHA-derived SPMs have been shown to be essential for modulating the number and response of T-helper cells to resolve chronic inflammation [111]. Exogenous administration of SPMs in infectious conditions has been shown to be effective at improving infection clearance and survival in preclinical models [112]. Clinical trials carried out with omega-3 fatty acids in the treatment of acute lung injury and acute exacerbations of chronic respiratory disease, asthma, and pneumonia [113,114], did not show the expected improvement results. However, in patients with COVID-19, C-reactive protein levels, erythrocyte sedimentation rate, and some clinical symptoms were reduced [115].
Purin-6-ones. PLY has a known role in increasing the permeability of fluids from microcapillaries to alveolar spaces. Witzenrath et al. (2009) [116] demonstrated the involvement of phosphodiesterase 2 (PDE2) in the hyperpermeability of lung tissue during pneumococcal pneumonia. Purin-6-ones, 9-(6-phenyl-2-oxohex-3-yl)-2-(3,4-dimethoxybenzyl)-purin-6-one (PDP), and hydroxy-PDP are specific inhibitors of PDE2. PDP decreased PLY-induced permeability in vitro in endothelial cell cultures and in vivo in murine pneumococcal pneumonia models, improving lung damage without interfering with the recruitment of immune system cells into lung tissue. PDE2 enzymes have been also found in various tissues and cells, including pulmonary arterial smooth muscle cells, endothelial cells, platelets, and macrophages. Up to now, there have been no studies available on the use of PDE2 inhibitors in the clinical studies carried out on patients with pneumococcal pneumonia. However, nonselective inhibitors of PDEs as Theophylline has been used in the treatment of bronchial asthma and chronic obstructive pulmonary disease (COPD) for more than 50 years. Various (selective) PDE3, PDE4, and PDE5 inhibitors have also demonstrated stabilization of the pulmonary epithelialendothelial barrier and reduction the sepsis-and inflammation-increased microvascular permeability, and suppression of the production of inflammatory mediators, which finally resulted in improved oxygenation and ventilatory parameters [117].
Thioethers. Montelukast is a synthetic thioether that acts by decreasing the hyperpermeability and bronchoconstriction produced by the toxin in the lungs [118]. PLY triggers the production of cysteinyl leukotrienes and their receptor CysLT1 in bronchial smooth muscle cells; the interaction of vascular endothelial cells increased vascular permeability, edema, influx of eosinophils, and neutrophils. Montelukast is a CysLT1 antagonist that increases survival in mice with induced pneumococcal pneumonia. In addition, this compound is frequently used as a treatment in patients with asthma and allergies. Recently, montelukast treatment for pneumonia, caused by Mycoplasma pneumoniae, and Sars-CoV-2 decreased inflammation markers and contributed to improvements in respiratory insufficiency [119,120].
Antibiotics. Clarithromycin is a derivative of erythromycin and differs structurally in the substitution of an O-methyl group for the hydroxy group at position 6 of the lactone ring. Clarithromycin negatively regulates ply gene transcription and activates pneumococcal autolysis [121]. On the other hand, in animal models of pneumonia caused by multidrug-resistant S. pneumoniae, the therapy combining levofloxacin and ceftriaxone modulated the inflammatory response (decreased number of neutrophils and polymorphonuclear neutrophils and vascular leakage; decreased nitric oxide production) of pneumococcal infection at sub-MIC levels by downregulating virulence genes (ply and lytA), cyclooxygenase-2 (COX-2), and inducible nitric oxide synthase (iNOS) [122]. Generally, macrolides and macrolide-like agents have antiinflammatory activities, including modulation of cytokine production [123]. Other antibiotics with anti-inflammatory properties are quinolones such as levofloxacin, due to its ability to join to the hydrophobic region of the TLR4-MD2 complex [124]. Clarithromycin is used in the treatment of pneumococcal infections in combination with β-lactams and in some clinical trials using clarithromycin as treatment (revised by Anderson & Feldman, (2017) [59]); the results supported its beneficial role in the anti-inflammatory/immunomodulatory activities. More recently, in a clinical trial in patients with COVID-19, it appears that it could be beneficial for early control of fever and early negative conversion of PCR [125]. However, increasing macrolide resistance in some CAP ethological agents (S. pneumoniae and M. pneumoniae) and its effect on decreasing the natural T-cell immune response [126], is contributing to reconsideration of their use in the future.
Peptides. In this review, seven peptides are presented as candidates for adjuvant therapy. Three of the peptides [64,65,69] (J34 binds to the SV1 splice variant of the GHRH receptor, TIP binds PKC-α, and vasculotide binds to the Tie2 angiopeptin receptor) act on endothelium hyperpermeability thus minimizing fluid extravasation into alveolar space. Le et al. (2015) [64] showed how synthetic hybrid peptides, with antimicrobial activity improved survival and reduced the sequelae of lethal pneumococcal systemic infection in animal models. Molecular docking analysis showed that one of the peptides (DM3) had a strong affinity for PLY, also for autolysin and pneumococcal surface protein A (pspA). DM3, in combination with penicillin, showed the recovery of all the treated septicemic animals, but it was not effective in pneumonia, therefore more studies are necessary (Le et al., 2015) [64]. Actually, antimicrobial peptides account for a promising alternative to conventional antibiotics since they can be designed to act on various targets and are less likely to generate resistance [127]. The problems encountered for its clinical application are related to its low absorption and toxicity to human cells. In most cases, these drawbacks can be solved with chemical modifications and in silico design [128].
The growth hormone-releasing hormone (GHRH) agonist, JI-34, decreased PLY-mediated endothelial hyperpermeability by reducing the phosphorylation of myosin light chain and vascular endothelial (VE)-cadherin and restoring basal levels of Na + [129]. Another example is Vasculotide (a tetrameric peptide with four octapeptides (NH 3 -CHHHRHSF-COOH), which are covalently attached through NH 2 -terminal maleimide) which improved the barrier function of the endothelial epithelium during pneumococcal pneumonia (Gutbier et al., 2017) [130]. The action was mediated by its binding to the Tie2 receptor of angiopeptin (Ang/Tie) thereby protecting against increased levels of fluid invading the intra-alveolar space and further pulmonary edema. Other drugs targeting the Ang/Tie pathway are being tested for the control of pulmonary vascular disorders in clinical trials. AV-001, a Vasculotide-derived peptide (Vasomune Therapeutics) is currently in phase I. The results showed that AV-001 is safe and well tolerated in patients with COVID-19, acute respiratory distress syndrome (ARDS), or COVID-19-associated ARDS [131].
The TIP peptide (TNF-derived tonoplast intrinsic protein) and the Ro 32-0432 peptide (bisindolylmaleimide XI hydrochloride) are Protein Kinase C-α (PKC-α) inhibitors. PKC-α is stimulated by the Ca 2+ influx mediated through PLY-induced pores which increased RhoA / RKK; arginase I reduces NO generation in vivo and consequently, microvessel leakage. TIP peptide diminished PKC-α activation and therefore, reduced PLY-induced endothelial hyperpermeability [129]. AP301 (Solnatide) is a circularized derivative of TIP peptide and has a Phase IIB clinical trial in ARDS [132] and is also being tested in patients with COVID [132]. The expected results, based on the previous studies, are activation of the pulmonary sodium ion channel (ENaC) to directly activate alveolar liquid clearance and reduce the leakage of blood and fluids from the capillaries in the airspace, i.e., accelerate the resolution of alveolar edema and reduce barrier injury in the lung.
Among the therapeutic peptides studied, some of them were designed to form hydrogen bridges with the cholesterol-interaction loop of different CDCs (PLY, LLO, and SLO) in a way that inhibits their hemolysis capacity and their pro-inflammatory activity. Thus, the named MRC-1 peptides bind PLY in domain 4 and inhibit binding to the human mannose receptor (MRC-1/CD206, a phagocytic receptor for bacteria, fungi, and other pathogens) [65]. Therefore, they reduced PLY-induced damage to the integrity of the epithelial barrier, the release of the chemokine IL-8, and the pro-inflammatory cytokines TNF-α and IL-12 as well as blocked bacterial invasion into the epithelium. Another example includes a peptide generated with the 70 C-terminal amino acids of PLY (C70PLY), which competed with MD2 to bind to TLR-4, inhibiting the effect of lipopolysaccharide on the phosphorylation of the ERK1/2 and NFκB-p65 subunit. TLR-4 is a member of the toll-like receptor family that senses pathogen-associated molecular patterns, stimulating cytokine production and apoptosis. Also, its activation leads to an intracellular signaling pathway NF-κB and the activation of the innate immune system [69].
As mentioned above, the innate immune system recognizes the pathogen components through TLR receptors, which trigger the host's first defense system. The inflammatory action of PLY is mediated by TLR-4 and induces caspase 3-dependent apoptosis in macrophages. zVAD (carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone) is a modified peptide caspase inhibitor used to block apoptosis [31,133]. However, apoptosis appears to be a defense mechanism against invasion by pneumococcus; therefore, inhibition of apoptosis results in increased mortality in animal models of pneumococcal pneumonia. Yet, cells can recover from PLY-induced apoptotic cell death by controlling mitochondrial [Ca 2+ ] m flux which contributes to the preservation of membrane potential (ΔCm). PLY binding cholesterol in rich cellular membranes leads to large pores (250-350 Å) which contributes to a rapid influx of Ca 2+ into the cytoplasm. Ca 2+ induces repair mechanisms for protection of the host cell [134]. One of them is sequestering Ca 2+ into the mitochondria. However, exceeding the mitochondrial capacity can result into release of pro-apoptotic factors such as cytochrome C (cytC) or apoptosis inducing factor (AIF) [26]. We did not find clinical trials related to with the treatment of diseases with zVAD.
Cations. Cations like Ca 2+ , Mg 2+ , and Zn 2+ hinder pore formation thereby decreasing the cytolytic capacity of PLY. Hupp et al. (2017) [66] showed that Mg 2+ decreases the pore-forming capacity of PLY in primary glial cells, improving survival in animal models of meningitis. Cations are known to influence membrane fluidity or inhibit toxin binding, but these hypotheses did not hold for Mg +2 since it does not prevent the binding of the toxin to the membrane, nor does it seem to change its characteristics. Another divalent cation, Zn 2+ [135], protected rat cochlear hair cells from the toxic action of PLY by preventing its binding to the membrane. Recently, the deficiency of Zn 2+ has been involved in a higher risk of acute infection in nasopharyngeal S. pneumoniae carriers. Therefore, supplementation with this element would improve immune system performance and consequently reduce pneumonia [136], Recent research on the relationship between micronutrients and some infections has shown improved immune function [137,138]. Unlike the other cations, Wippel et al. (2011) [139], using brain tissue, postulated that Ca 2+ could vary the fluidity of the membrane rather than having a direct action on the toxin because reductions in Ca 2+ concentrations improve the binding of the PLY membrane thus increasing its lytic capacity. However, S. pneumoniae is capable of inducing apoptosis thanks to the pores formed by PLY which produce rapid inflows of mitochondrial calcium [Ca 2+ ] m thereby favoring fragmentation, loss of motility, and membrane potential in addition to activation of caspase 3. Nerlich et al., (2021) [140] showed that a significant number of alveolar epithelial cells survived caspase activation after being challenged with PLY, critically regulating [Ca 2+ ] m to control cell fate after attack by PLY. Therefore, calcium ions were postulated to be a useful candidate in therapeutic intervention during pneumococcal infection.
Antibodies. Antibodies against PLY have been used against pneumonia, keratitis, and nasopharyngeal colonization in animal models of infection [68,70,141]. The protective effect of mouse monoclonal antibodies in reducing PLY proinflammatory properties was clearly demonstrated as well as an increased survival rate with decreased bacteria lungs colonization, leukocyte infiltration, and lung injury [141]. Similarly, a reduction in toxin effects was observed in an induced keratitis model in rabbits using anti-PLY polyclonal antiserum [68]. Interestingly, the efficacy of human anti-PLY antibodies in reducing S. pneumoniae nasopharyngeal colonization was also demonstrated in mice [70]. Although PLY is an intracellular protein that is released during bacteria lysis, it would also be located on the cell surface, where it would play a role in the aggregation and formation of biofilms. Potential rejection reactions to repeated doses of antibodies have been solved through the humanization of monoclonal antibodies, which has facilitated their use for the treatment of various diseases [142]. This, together with the fact that immunoglobulins have been used since ancient times for the treatment of infectious diseases, makes this strategy very promising. In fact, the effect of PLY on platelets and erythrocytes can be neutralized by polyvalent human IgG and trimodulin [143]. In a clinical trial in patients with S. pneumoniae CAP, treatment with trimodulin (182.6 mg/kg, for 5 consecutive days) improved survival compared to patients treated with placebo [45]. Drugs developed against toxins from other microorganisms, including monoclonal antibodies, antibody fragments, antibody mimetics, have recently been reviewed [144]. It is important to highlight the phase I clinical trial using liposomes to capture S. pneumoniae toxins [145].
Network of compounds and related proteins. In order to evaluate the molecular interactions of the 39 compounds with neutralizing toxin properties described in the selected articles, the EXTRACT 2.0 tool was used for text mining of biomedical named entities and ontology terms such as diseases, subcellular localizations, tissues, drugs, and other small molecule compounds [62]. With the resulting dataset, a knowledge network (Fig 3) was built using Cytoscape (see Materials and methods). Analysis of the network reveals that the most affected process is the formation of the pore with 19 interactions (edges), followed by hyperpermeability and inflammation (showing nine interactions), and membrane binding (seven interactions). The relevance and complexity of pneumonia is evidenced by the high number of molecules (24) used for its treatment while other diseases show three or fewer interactions with therapeutic molecules. Regarding host cell proteins, TNF-α was the protein most affected with 6 interactions, followed by IL-6 and, IL-1β (both showing five interactions), and VE-cadherin (three interactions). The presented knowledge network visually shows the summary of the reviewed articles. In addition to this, the structure of the network reveals the molecules with the greatest interactions which can be useful in the search for new treatments for the neutralization of the toxin. Drug repurposing identifies new uses of drugs outside their original medical scope, considering that, sometimes, one molecule can act on multiple targets [146]. Recently, a network medicine platform based on systems pharmacology was used to reposition antiviral drugs against SARS-CoV-2 [147]. This approach aims to reduce both development costs and the time it takes for already approved drugs to reach the market.

Current limitations and future perspectives
Despite the effective development of antibiotics and more recently pneumococcal vaccines, pneumococcal pneumonia continues to cause high morbidity, even though actually the mortality caused to invasive pneumococcal disease is declining due to childhood vaccination and herd protection [148]. The aging of the population and the combination of pneumococcal disease with other important viruses such as HIV, influenza and, more recently, SARS-CoV-2, could be behind these high levels of mortality and morbidity. To proceed in the discovery of new strategies to combat this disease, we decided to focus our work on PLY and its role in the performance of pneumococcal infections.
The combination of antibiotics with other molecules that can somehow block the action of PLY, was presented as a hopeful therapy aimed at significantly reducing the severity of pneumococcal pneumonia and therefore, diminishing the high mortality that is still associated with it. The search strategy for the review was broad to locate all studies in which some type of molecule was used as an alternative treatment or blockade of the toxin. The articles reviewed focused on experimental and laboratory studies and provide reduced information about the behavior with other doses or routes of inoculation. Most studies do not provide information on survival rates and limit themselves to studying the characteristics of the organs or tissues involved. One of the limitations of this systematic review is the lack of homogeneity of the studies. We did not find comparable studies in molecules, doses, inoculation routes, etc. that allow a statistical comparison of the results or meta-analysis. However, most of the studies reviewed has a good level of confidence. The clinical efficacy of the molecules in the treatment of pneumococcal infections will depend on several factors such as the pharmacokinetic/pharmacodynamic parameters of the molecule used for treating the infection, and the infection site (i.e. Central Nervous System is a difficult place for the molecules to penetrate). Indeed, some compounds have a good perspective for inhibition of pore formation and apoptosis, such as polyphenols and sterols, which also exhibit anti-inflammatory properties. With respect to hyperpermeability, it appears that it can be reduced by thioethers and some antimicrobial peptides. However, although some of these compounds have a great potential to be implemented as adjuvants in the treatment of pneumococcal infections, others (like peptides revised in this study), require future clinical trials using different combinations and deeper investigation. Hopefully, recent research related to the SARS-CoV-2 pandemic will help to understand the effect of these molecules with respect to ARDS in pneumococcal pneumonia.
Other limitation of this review was that surprisingly, for some molecules, (like cations and cholesterol), of which we did not find PLY-related studies in animal models, there are a considerable number of clinical trials studies related with pneumonia. We considered that this need exhaustive testing and are out of score of this review.
In summary, although the clinical relevance of this review is limited due to the lack of clinical tests in humans, some molecules have already been clinically tested and therefore could be safe for use in pneumococcal infections.