Different drugs for bad bugs: Antivirulence strategies in the age of antibiotic resistance

Abstract

The rapid evolution and dissemination of antibiotic resistance among bacterial pathogens is outpacing the development of new antibiotics. Antivirulence is an alternative strategy that attempts to circumvent antibiotic resistance by disarming pathogens of factors that facilitate human disease, while leaving bacterial growth pathways – the target of traditional antibiotics - intact. Either as stand-alone medication or together with antibiotics in a combinatorial approach, these drugs are intended to treat bacterial infections in a largely pathogen-specific manner. Notably, developing antivirulence drugs requires an in-depth understanding of the roles diverse virulence factors play in disease processes. In this review, we outline the theory behind antivirulence strategies and provide examples of bacterial features best targeted by antivirulence approaches. Furthermore, we discuss the paradigm’s recent successes, failures, and new developments currently in the pipeline.

Introduction

The antibiotic resistance crisis and the threat it poses to future generations need little introduction. However, despite increasing awareness^1–3, only meager progress has been achieved to address the crisis. Meanwhile antibiotic resistance continues to fester.

Throughout the antibiotic era, antibiotic resistance has arisen and spread quickly, often within years after an antibiotic class was introduced into clinical use. The reasons for this are complex and multifactorial: (i) like most antibiotics, antibiotic-resistance mechanisms are ancient and diverse and existed, albeit at lower levels, prior to antibiotic use^{4, 5}, (ii) bacteria exist in such large numbers that “rare” mutations frequently occur to permit selection, (iii) the genetic elements encoding antibiotic resistance are often mobile and readily transferred even across bacterial species^{6, 7}, and (iv) antibiotics target central growth pathways that are shared among diverse bacteria, and the disruption of these pathways imposes severe selective pressures favoring resistance. Presently, these phenomena are driving the disturbing spread of resistance to last-resort antibiotics such as colistin^{8, 9} and vancomycin^{10, 11}.

Earnestly addressing the antibiotic resistance threat requires a multi-prong approach including improved stewardship of known antibiotics, development of novel antibiotic classes or effective vaccines, and alternate strategies to treat or prevent bacterial infections¹². Herein, we discuss the antivirulence paradigm, which focuses on “disarming” pathogenic bacteria by neutralizing their virulence factors¹³. This strategy possesses distinct advantages and limitations (TABLE 1) when compared to conventional antibiotics. We will first layout the principles that underlie antivirulence, followed by highlighting the paradigm’s most promising examples and recent efforts against high-priority pathogens (TABLE 2).

Table 1.

Advantages and disadvantages for antivirulence development and application

Advantages	Disadvantages
May impose less evolutionary pressure for the development of resistance	May require combination therapy to account for multiple, redundant, or strain-variable virulence factors
Less impact to host commensal flora	May not be effective in all forms of disease by the same pathogen
Can inactivate target rapidly	May have reduced therapeutic effects compared to antibiotics
Can supplement antibiotics to increase efficacy	May require rapid identification of a bacterial pathogen
Can provide protection in immunocompromised, unvaccinated individuals	May exhibit a therapeutic delay for regulatory or biosynthetic targets
Quorum sensing inhibitors may prevent production of multiple virulence factors	Bacteria may persist and cause damage after therapy is withdrawn

Table 2.

Antivirulence drugs that are approved or in development^§

Pathogen Target(s)	Compound Name	Subtype/Chemistry	Molecular Target(s)	Clinical Development Stage						Comments	Refs.
				P	A	I	II	III	FDA*
C. botulinum	BabyBIG®	Human, mostly IgG, plasma-derived Immune globulin	BoNT serotypes A and B	-	-	-	-	-	✓		32
	BAT	Equine, Fab and F(ab)2, Plasma-derived Immune globulin	BoNT serotypes A-G	-	✓	✓	-	-	✓	Approved under Animal Efficacy Rule	NCT00360737
B. anthracis	Raxibacumab	Human, mAb IgG1	Protective Antigen of anthrax toxin	✓	✓	-	-	-	✓	Approved under Animal Efficacy Rule	38
	Obiltoxaximab	Human, mAb IgG1	Protective Antigen of anthrax toxin	✓	✓	-	-	-	✓	Approved under Animal Efficacy Rule	39, 40
	11D and 4C	Chimpanzee, mAbs IgG1 and IgG3	PGA	✓	✓	-	-	-	-	Provided protection against pre- and post-exposure to spore challenge in mice	42
	F26G3, F24F2, F26G4	Mouse, mAb IgG3	PGA	✓	✓	-	-	-	-		43
C. difficile	Bezlotoxumab (MDX-1388)	Human, mAb IgG	TcdB	✓	✓	✓	✓	✓	✓	Bezlotoxumab used in combination with actoxumab in Phase II trial	50, 184
	Ebselen	Organoselenium	TcdA and TcdB	✓	✓	-	-	-	-	Also has antioxidant effects that may obscure mechanism of action	51
	Tolevamer	High-molecular-weight polymer	TcdA and TcdB	✓	✓	✓	✓	✗	-		49
S. aureus	MEDI4893	Human, mAb IgG1	Alpha toxin	✓	✓	✓	†	-	-		64
	Salvecin™	Human, mAb IgG1	Alpha toxin	✓	✓	†	†	-	-		NCT01589185
	ASN-100 (a combination of ASN-1 and ASN-2)	Human, mAb IgG1	Alpha toxin, PVL, LukED, LukGH, gamma-hemolysin	✓	✓	†	-	-	-	Inhibits pore-forming activity of 5 S. aureus toxins	66, 67
	6e	3,6-disubstituted triazolothiadiazole compounds	SrtA	✓	✓	-	-	-	-	Protected mice against S. aureus sepsis	71
	Savirin	(3-(4-propan-2-ylphenyl) sulfonyl-1H-triazolo [1,5-a] quinazolin-5-one)	AgrA	✓	✓	-	-	-	-	Reduced dermonecrosis in mouse model, S. aureus specific AgrA inhibitor	74
P. aeruginosa	MEDI3902 (BiS4αPa)	Engineered, bispecific antibody	Psl and PcrV	✓	✓	†	†	-	-		89 NCT02255760 NCT02696902
	RS2–1G9	Mouse, mAb IgG	3-oxo-C12-AHL	✓	-	-	-	-	-		185
	KB001A	Pegylated-Fab	PcrV	✓	✓	✓	✗	-	-		186 NCT00638365 NCT00691587 NCT01695343
	INP1855	hydroxyquinoline	Possibly through ATPase of T3SS (PscN) and flagellum	✓	✓	-	-	-	-		187
	M64	benzamide-benzimidazole	PqsR	✓	✓	-	-	-	-		86
	Compounds 37a and 37b	phenoxyacetamide	PscF	✓	-	-	-	-	-		188, 189
	E22	phenoxyacetyl homoserine lactone (AHL mimic)	RhlR	✓	-	-	-	-	-		190
	mBTL	halogenated thiolactone	RhlR	✓	§	-	-	-	-		191
	itc-12 and itc13	isothiocyanate-AHL	LasR	✓	-	-	-	-	-	Covalent inhibitor	192
	C30	halogenated furanone	LasR	✓	✓	-	-	-	-	May have off target effects, resistance mutants have been discovered	193, 194
	3-oxo-C12-(2-aminocyclohexanone)	acyl-aminocyclohexanone (AHL mimic)	LasR	✓	-	-	-	-	-		195
STEC	Shigamab™ (cαStx2)	human-mouse chimeric mAb	Stx2	✓	✓	✓	-	-	-	Protects mice when given 4 days after Stx intoxication	104, 107
	SYNSORB-Pk	Silicon-dioxide conjugated with Gb3 receptor mimic polymer	Stx1 and Stx2	✓	✓	✓	✗	-	-		101, 102
UPEC	mAb926	Mouse, mAb IgG	FimK	✓	✓	-	-	-	-	Bacteria pre-incubated with mAb before infecting mice	116
	Compound 22	Biarylmannoside	FimK	✓	✓	-	-	-	-	Improved potency and pharmacokinetics in mice over previous biarylmannosides	111, 114
	ec240	Ring-fused 2-pyridones	Unknown, but decreases piliation	✓	-	-	-	-	-		112
K. pneumoniae	KP3	Engineered, scFv-Fc antibody	MrkA	✓	✓	-	-	-	-		117
A. baumannii	LpxC-1	Methylsulfone hydroxamate	LpxC	✓	✓	-	-	-	-		125
	Deferiprone	3-hydroxy-1,2-dimethylpyridin-4(1H)-one	Ferric Iron	✓	-	-	-	-	-	FDA-approved treatment for iron overload disorder in patients with thalassemia	134
	Virstatin	(4-[N-(1,8-naphthalimide)]-n-butyric acid	Unknown, but decreases piliation	✓	-	-	-	-	-		132
E. faecalis	ZBzl-YAA5911	Cyclic peptide	FsrC	✓	✓	-	-	-	-	Prevented bacterial translocation in a endophthalmitis model	143
Enterobacter spp.	Stx1e1 – 4	Mouse, mAbs IgG1 and IgG2a	Stx1e	✓	-	-	-	-	-	mAbs target A subunit	157
M. tuberculosis	SMITB14	Mouse, mAb IgG1	Lipoarabinomannan	✓	✓	-	-	-	-	Bacteria pre-incubated with mAb before infecting mice	168
	2E9IgA1	Human, mAb IgA1	α-crystallin	✓	✓	-	-	-	-	Reduced lung pathology in combination with IFNγ	169
	4057	Mouse, mAb IgG3	Heparin-binding haemagglutinin	✓	✓	-	-	-	-	Bacteria pre-incubated with mAb before infecting mice	170
	Compound 7	N-(5-(azepan-1-ylsulfonyl)-2-methoxyphenyl)-2- (4-oxo-3,4-dihydrophthalazin-1-yl)acetamide	Fumarate hydratase	✓	-	-	-	-	-	Specific for M. tuberculosis TCA cycle enzyme	196

^§Abbreviations used in Table 1: BoNT, botulinum neurotoxins; FDA, Food and Drug Administration; FDA*, FDA approved; lukED, leukocidin ED; mAb, monoclonal antibody; PGA, poly-Υ-D-glutamic acid; PVL, Panton-Valentine leukocidin; Stx, Shiga toxin; STEC, Shiga toxin producing E. coli; T3SS, type III secretion system; TCA, tricarboxylic acid; UPEC, uropathogenic E. coli; †, ongoing or results pending; §, efficacy in C. elegans model; ✗, failed; P, Pre-clinical; A, Animal models; I, Passed Phase I clinical trials; II, Passed Phase II clinical trials; III, Passed Phase III clinical trials; -, Not applicable.

Antivirulence principles

Virulence factors are bacterial products that promote disease and damage the host or circumvent and evade the host immune system¹⁴. Antivirulence seeks to interfere with these factors to treat infections, a strategy that historically precedes antibiotic use. In 1893, Emil von Behring treated children suffering from diphtheria with immune anti-serum derived against diphtheria toxin. However, the limited understanding of how pathogens cause disease hindered applying antivirulence more broadly. With the discovery of highly effective broad-spectrum antibiotics, antivirulence eventually fell to the wayside. Thus, the rise of antibiotic resistance and our increasing understanding of virulence factors and their functions have spurred the current emerging renaissance in antivirulence approaches (FIG. 1).

Figure 1: The rise of antivirulence approaches.

The number of antivirulence publications and citations is sharply increasing over time. Web of Science (Thomson Reuters) was queried with the following search terms: “antivirulence” OR “anti-virulence” OR “virulence inhibition” OR “virulence inhibitor” OR “virulence factor inhibition”.

The promises.

The most encouraging potential of antivirulence agents is the possibility to reduce, eliminate, or even possibly reverse the selective pressures that are characteristic of traditional antibiotics and have led to the development of resistance^{13, 15}. These stem from putative, and to date untested, properties of antivirulence approaches. First, blocking an ideal virulence factor would not impact the natural survival and propagation of the target bacterium and thus would not change the overall selective pressures directing its evolution. Even targeting a sub-ideal virulence factor should impose less selective pressure than antibiotics. In addition, there are other antivirulence features that add to the potential of antivirulence drugs to delay the rise of resistance. Because most virulence factors are restricted to a single or a few closely related species, little advantage would be gained by horizontally transferring genetic elements that encode resistance across bacterial species, undercutting a major mechanism responsible for the current spread of antibiotic resistance. Moreover, antivirulence drugs are likely to lack the growth promotion effects of antibiotics and therefore are less likely to be excessively used in livestock—as is the case for antibiotics. Finally, antivirulence drugs could improve antibiotic stewardship by allowing for more judicious and effective use of antibiotics.

Antivirulence can bypass another pernicious and for a long-time unappreciated side-effect of antibiotics in that antibiotics also kill many beneficial bacteria that live within and on us. These bacteria make up part of our microbiota, have co-evolved with humans¹⁶, and play key roles in human health, including promoting the developing immune system, increasing our metabolic capabilities, and resisting colonization by pathogenic bacteria^{17, 18}. Antibiotics drastically deplete and alter the composition of our microbiota, leading to complications such as Clostridium difficile colitis^{19, 20}. Commensal bacteria lack many virulence factors, especially toxin production, and should thus be unaffected by antivirulence drugs. Thus, the inherent narrow spectrum of many such drugs would more precisely target the offending pathogen.

Perhaps the most important advantage of antivirulence strategies is the large number of virulence targets. Following the “Golden Age” of antibiotic discovery, large screening campaigns to discover new broad-spectrum antibiotics had little success and emphasized the need for new chemical entities to screen^{21, 22}. However, the phenotypic assays set up to test for growth defects, which were conventionally used to discover antibiotics, would have overlooked virulence targets. Thus, virulence factors represent novel targets for which to mine existing chemical libraries that could identify distinctly antivirulent compounds. Furthermore, secreted or surface-exposed virulence factors are amenable targets for biologics such as natural and engineered antibodies²³.

The limitations.

Because no antivirulence drug is yet widely used in the clinical setting, many putative advantages of antivirulence strategies still have to be verified in the real world. Many virulence factors, despite being dispensable in in-vitro assays, may contribute to bacterial survival in vivo. Predicting the propensity of bacteria to develop resistance is also difficult because bacterial transmission and colonization dynamics are complex, differ among species, and are incompletely understood. Furthermore, the disease manifestations themselves may be critical for bacterial transmission to new hosts²⁴.

Several practical obstacles to developing effective antivirulence drugs exist. Compared to antibiotics, antivirulence drug development as a whole will likely require more labor and cost significantly more¹², primarily because several antivirulence drugs may be needed to match the spectrum of a single antibiotic. Furthermore, larger clinical trials for antivirulence drugs that show modest improvement over antibiotics as an adjunct may be needed to detect efficacy, especially in infections in which standard of care antibiotics remain mostly effective.

The narrow-spectrum property of targeting virulence creates additional complications in applying antivirulence¹². Empiric therapy would likely require combining antivirulence drugs with other antimicrobial agents (including other antivirulence drugs) to increase bacterial coverage. Moreover, for pathogens that produce several virulence factors, combination therapy may also be necessary to overcome possible redundancies among these factors. Finally, the timing of antivirulence therapy must coincide with when the targeted factor actively promotes disease. This may be especially important for targeting biosynthetic enzymes or regulatory networks, which would require additional time to affect downstream factors. However, these limitations can be mitigated by understanding the precise contribution of each virulence factor within the course of a disease or by developing better rapid diagnostic tests with improved positive predictive values.

While antibiotics remain effective, many antivirulence therapies could be limited to prophylaxis in high-risk patients or used as an adjunct to antibiotics. By not killing the pathogenic bacteria outright, an antivirulence-only approach might achieve weaker therapeutic benefits. Furthermore, antivirulence relies on the host immune system to clear a “disarmed” pathogen from the disease site. Antibiotics may still be required to achieve a cure if pathogenic bacteria persist beyond the duration of antivirulence therapy, which may be particularly problematic in immunocompromised patients¹³.

The targets.

Antivirulence phenotypic screens are challenging to conduct and often require disease models or, along with biochemical screens, some a-priori knowledge of the candidate targets. Thus, before undertaking an antivirulence campaign, it is imperative to prioritize both the pathogenic bacterial species and the molecular virulence target.

Although developing a precise and broadly agreed-upon definition for virulence has proven troublesome^{14, 25}, the list of generally accepted virulence factors is long and diverse (http://www.mgc.ac.cn/VFs/), and includes adhesins, toxins, secretion systems, siderophores, immune evasion and modulation factors, and factors that promote biofilm formation. Additionally, any corresponding regulatory factor (e.g. quorum sensing) or biosynthetic enzyme can likewise be targeted. The choice of target will undoubtedly have a profound impact on the properties of any resulting antivirulence drug. For instance, blocking a toxin secreted by a pathogen that colonizes non-disease sites may very well impose little to no selective pressure for resistance and leave the commensal microbiota unscathed.

Among all possible pathogenic species, antivirulence agents are needed most for antibiotic-resistant pathogens as well as for treating acute, toxin-mediated diseases, and those for which no vaccines are readily available. For the remainder of this Review, we will cover a selection of recent antivirulence efforts against the antibiotic-resistant ESKAPE pathogens²⁶, the toxin-mediated diseases caused by C. difficile, C. botulinum, Escherichia coli, and Bacillus anthracis, and end with Mycobacterium tuberculosis. The diversity of these bacteria, the progress made in understanding pathogenesis, and initial translational efforts provide precedents and lessons for guiding future antivirulence strategies during preclinical and clinical studies.

Pathogenic targets with FDA-approved antivirulence drugs

We first discuss three pathogens for which Food and Drug Administration (FDA)-approved antivirulence therapies exist. Although these bacteria have not exhibited clinically relevant antibiotic resistance, the diseases caused by C. botulinum, B. anthracis, and C. difficile are acute toxin-mediated diseases that are often life-threatening and for which directly interfering with the toxins can improve outcomes. Furthermore, C. botulinum and B. anthracis pose risks as bioweapons against which antibiotics may be of limited help.

Clostridium botulinum.

Botulism is a rare illness that presents with descending flaccid paralysis and can lead to respiratory distress in infants and adults. The causative agent is the spore-forming C. botulinum, which secretes one of eight serotypes of the potent botulinum neurotoxin²⁷ (BoNT A-H). After gaining access to the cytosol, BoNT disrupts the cholinergic neurons at neuromuscular junctions, cleaving various host factors that are important for neurotransmitter release into the synaptic cleft²⁸. Among the most potent toxins known²⁹, BoNT is a potential bioweapon in part because culturing C. botulinum from the environment and purification of BoNT is relatively easy³⁰.

Current FDA-approved antivirulence treatments for botulism are immunoglobulins purified from donor plasma that neutralize BoNTs (FIG. 2) and harken back to the serum therapy approach by Emil von Behring to treat diphtheria. BabyBIG^® is purified from human donors injected with pentavalent BoNT A-E toxoids (inactivated toxins) and was approved in 2003 for treating infant botulism caused by serotypes A and B, which are the most common causes of infant botulism³¹. A decade-long clinical study with infants demonstrated that BabyBIG^® significantly reduced the duration of mechanical ventilation and hospital stay³². Notably, BabyBIG^® was given to patients with suspected botulism after initial BoNT intoxication. The intervention succeeded presumably because it prevented further intoxication of motor-neurons and serves as an example of a treatment window following initial diagnosis. The second biologic is the equine-derived heptavalent botulism antitoxin (BAT^®) that was approved under the FDA’s animal efficacy rule (Title 21 CFR 601 Subpart H) in 2013 for treatment covering serotypes A-G in adult or pediatric patients. The rare BoNT H serotype, which was discovered later²⁷and is similar to BoNT A, was effectively neutralized by BAT^®33, suggesting that BAT^® can afford protection to BoNT H. Because of its equine origin, BAT^® IgG is digested with pepsin to reduce cross-species immunogenicity, yielding a mixture of F_ab and F_(ab)2 fragments with nearly little equine F_c fragments remaining.

Figure 2: Antivirulence agents targeting secreted and surface-exposed virulence factors.

Overview of pathogenic secreted and surface-exposed virulence factors and inhibitors. Antivirulence agents highlighted yellow or red are FDA approved or in clinical trials, respectively. LF/LT = lethal factor/toxin, EF/ET = edema factor/toxin, CM = cellular membrane, MOM = mycolic outer membrane, PG = peptidoglycan, AG = arabinogalactan, LAM = lipoarabinomannan, scFv = single chain variable fragment, T3SS = type III secretion system.

Bacillus anthracis.

B. anthracis, the etiological agent of anthrax, primarily causes disease of warm-blooded livestock. However, this highly infectious Gram-positive pathogen, which naturally exists in the environment as durable and temperature-resistant endospores, is easily transmissible to humans. Even though antibiotics can be used to treat anthrax infections, failure to administer antibiotics early can result in mortality rates as high as 90%³⁴. For these reasons alone, B. anthracis is listed as a Category A bioterrorism agent.

The majority of anthrax antivirulence efforts have focused on targeting two essential virulence factors, the anti-phagocytic poly-γ-D-glutamic acid (PGA) polypeptide capsule and the tripartite anthrax toxin (FIG. 2), with small-molecule inhibitors and monoclonal antibodies (mAbs) (for detailed reviews^{35, 36}). The anthrax toxins, lethal toxin (LT) and edema toxin (ET), are composed of a common receptor-binding component, protective antigen (PA), which associates with either lethal factor (LF), a zinc metalloprotease, or edema factor (EF), a calcium- and calmodulin-dependent adenylate cyclase. LF or EF disrupt immune responses and induce cellular damage when delivered intracellularly in a PA-dependent manner. Importantly, the effects of ET and LT may persist long after the bacteria have been neutralized³⁷, justifying the need for early and late stage therapeutic intervention. Multiple doses of the PA-based anthrax vaccine (anthrax vaccine adsorbed, AVA, BioThrax) are needed to generate protective immune responses. However, the time required to induce these responses would be too long to protect against inhalational anthrax, the most severe and acute form anthrax disease, which can develop within a week after exposure.

A landmark development in the anthrax therapeutics field was the first FDA approval of two high-affinity anti-PA mAbs, raxibacumab (Abthrax^®, GlaxoSmithKline)³⁸ and obiltoxaximab (Anthim^®, ETI-204, Elusys Therapeutics)³⁹ under the FDA’s animal efficacy rule. Both mAbs completely block the receptor interaction of PA and effectively prevent the translocation and, consequently, the intracellular toxin-mediated effects of LF and EF. In rabbit and monkey models of inhalational anthrax infection^{38, 40}, significant protection is attained with one prophylactic, or within a short-window, post-exposure treatment. It is recommended that both mAbs be used in conjunction with the antibiotics levofloxacin or ciprofloxacin.

Conjugation of the poorly immune-reactive PGA antigen with strong immunogenic protein carriers^{41, 42}, or with a CD40 agonist mAb⁴³, has been shown to produce highly reactive polyclonal and mouse- and chimpanzee-derived mAbs that demonstrate opsonophagocytic bactericidal activity. Although none of the PGA-specific mAbs have gone through clinical trials, prophylactic treatment with the mAbs provided extraordinary protective efficacies in murine models of pre- and post-inhalational anthrax spore challenge.

Clostridium difficile.

C. difficile is a spore-forming, toxin-producing, Gram-positive pathogen that causes diarrhea and colitis^{44, 45}. The incidence and severity of C. difficile infections (CDI) has sharply increased⁴⁶ in developed countries paralleling the rise of the epidemic BI/NAP1/027 strain, which exhibits increased toxin production⁴⁷. Although clinically relevant antibiotic resistance has not occurred, antibiotic use for other illnesses is a major risk factor for developing CDI and recurrent infections following the initial resolution are common. C. difficile now causes more deaths in the United States than all other antibiotic-resistant pathogens¹.

Disease progression is driven by the action of two major C. difficile exotoxins, TcdA and TcdB, which glucosylate Rho GTPases inside epithelial enterocytes, resulting in cytoskeletal destabilization. Even though it is unclear whether one toxin is more virulent than the other⁴⁸, these toxins are the primary targets of antivirulence approaches against CDI. Initially, a toxin-binding polymer was tested to treat CDI. However, this proved inferior compared to treatment with the standard-of-care antibiotics⁴⁹. Subsequently, in two phase-III clinical trials (NCT01241552 and NCT01513239), Merck tested two mAbs, actoxumab and bezlotoxumab (FIG. 2) that target TcdA and TcdB⁵⁰, respectively, in conjunction with standard-of-care antibiotics with the primary outcome of CDI recurrence, not resolution of the ongoing CDI. Indeed, bezlotoxumab was efficacious, leading the FDA in October 2016 to approve bezlotoxumab (Zinplava^™) to reduce the recurrence of CDI in patients who are at a high risk of recurrence. Notably, bezlotoxumab has become the first FDA-approved antivirulence agent to prevent a common and increasingly prevalent bacterial infection.

Other toxin-targeting inhibitors have been investigated at the pre-clinical stage. Ebselen, a small organoselenium compound, was identified in a high-throughput screen as an inhibitor of the TcdA and TcdB cysteine protease domains⁵¹ (FIG. 2). Ebselen protected mice against TcdB-mediated killing and improved histopatholgy in a murine CDI model. Furthermore, ebselen has been studied in clinical trials for unrelated conditions and has been shown to be safe⁵². However, the antioxidant property of ebselen may reduce TcdB-mediated inflammation and thus confound its mechanism of action. Additionally, at higher concentrations, ebselen is bactericidal towards several Gram-positive bacteria by inhibiting thioredoxin reductase^{53, 54}.

Other potential C. difficile antivirulence targets are being investigated, but largely remain at the target identification stage. Interestingly, toxin production in C. difficile is under the control of two agr-like elements similar to that in S. aureus and disrupting these elements attenuates virulence in murine models^{55, 56}.

Pathogenic targets with investigational antivirulence drugs in clinical trials

Staphylococcus aureus.

Methicillin-resistant S. aureus (MRSA) causes the most deaths in the United States due to antibiotic resistance¹. In part this is because S. aureus encodes numerous virulence factors that enable it to cause multiple diseases, including invasive conditions such as severe pneumonia, sepsis, and endocarditis. While the incidence of hospital-associated invasive MRSA infections has been declining, the opposite is true for community-associated (CA-MRSA) infections⁵⁷, which are more challenging to contain using infection control practices. Moreover, the epidemiological success of CA-MRSA strains is due in part to increased virulence, driven by the presence and enhanced production of multiple immune evasion molecules and toxins⁵⁸.

The many virulence factors of S. aureus offer the possibility of numerous antivirulence approaches. Ideally, such an approach should be effective against all geographically and genetically distinct S. aureus lineages^59–62 and priority should be assigned to those virulence targets considered to be most important, as defined by distribution and contribution to disease. For instance, the genome-encoded alpha toxin (also known as α-hemolysin, Hla)⁶³ is a major virulence factor that is highly expressed among many CA-MRSA isolates and has been investigated both as a vaccine antigen and as a target for mAbs. One such mAb specifc for alpha toxin is MEDI4893⁶⁴ (FIG. 2), which can inhibit both oligomerization of alpha toxin and interaction with its cognate receptor, ADAM10. MEDI4893 displays protective efficacy against S. aureus in a variety of animal infection models when administered prophylactically. Importantly, MEDI4893 recognizes all 12 genotypic variants of alpha toxin produced from multiple S. aureus isolates⁶⁵ and is only one of two mAbs targeting alpha toxin that has entered phase I and phase II clinical trials.

Arsanis, Inc. has developed a cross-reactive mAb (ASN-1) that recognizes a conserved region of the alpha-toxin phosphocholine-binding pocket, in addition to three members of the S. aureus leukotoxin family⁶⁶. This mAb neutralized the cytolytic effects of all four toxins and decreased mortality in murine models of S. aureus pneumonia and sepsis. The company also recently developed another mAb, ASN-2, which targets LukGH, the only leukotoxin not recognized by ASN-1⁶⁷, and promotes the combined ASN-1 and ASN-2 mAbs as ASN-100 (FIG. 2). Targeting several key pore-forming toxins with one mAb cocktail has the added benefit of treating a diverse range of S. aureus strains that have different leukotoxin expression profiles⁶⁸.

In addition to toxins, S. aureus expresses a number of cell surface adherence and immune evasion proteins, of which many are anchored to the cell wall through the sortase A (SrtA) enzyme⁶⁹. Deactivating SrtA results in a significantly attenuated phenotype⁷⁰. Triazolothiadiazole compounds, which emerged from a screen for small molecule inhibitors of SrtA⁷¹ (FIG. 2), decreased mortality in a murine model of S. aureus sepsis. These compounds showed cross-reactive inhibitory activity of sortase homologues from other pathogenic bacteria, suggesting that they may have broad-spectrum applicability.

The accessory gene regulator (Agr) quorum-sensing (QS) system (FIG. 3, BOX 1) has long been considered an appealing antivirulence target of S. aureus because it is the master regulator for a number of major staphylococcal virulence factors. It uses a secreted autoinducing peptide (AIP) as cell-density signal. However, approaches that sequester free AIP or block AIP binding to the extracellular domain of the cognate membrane-bound histidine kinase receptor, AgrC, are not always effective at preventing the downstream effects activated by the response regulator, AgrA, across all four Agr subgroups in S. aureus^{72, 73}. Nevertheless, targeting AgrA directly with the small molecule inhibitor savirin⁷⁴ (FIG. 3) efficiently blocked Agr-regulated gene expression among all Agr subgroups and provided protection against S. aureus-induced dermonecrosis. Importantly, savirin demonstrated species-specificity to S. aureus in vitro, while it had no effect on Agr of S. epidermidis, which is considered to be part of the beneficial microbiota on human skin.

Figure 3: Quorum sensing and inhibition in model Gram-positive and Gram-negative pathogens.

QS pathways and inhibition in the model Gram-positive and Gram-negative organisms S. aureus (A) and P. aeruginosa (B), respectively (adapted from Papenfort and Bassler¹⁷⁶, Figure 3). Synthases and exporters (black filling) produce autoinducer (AI) that signal through receptors (gray filling). Activated receptors globally modulate gene expression including many virulence factors. Selected examples of QS inhibitors that block receptors are shown. Note that P. aeruginosa produces homoserine lactone (HSL) and quinolone-based AI and S. aureus produces cyclic peptide-based AI. QS feedback loops and crosstalk between pathways is omitted for simplicity. AIP = autoinducing peptide, OM = outer membrane, PG = peptidoglycan, IM = inner membrane.

Box 1: Taking aim at Quorum sensing.

Quorum sensing (QS) is widely conserved in bacteria and enables a population of cells to coordinate gene expression based on the accumulation of autoinducer (AI) signals¹⁷⁴. A complete QS circuit requires AI biosynthesis, secretion, diffusion, sensing, and usually a feed-forward loop that amplifies the QS signal. Although Gram-negative and Gram-positive bacteria have evolved multiple and distinct QS networks (reviewed elsewhere^{175, 176}), virulence traits frequently fall under their control. The presumed rationale for this is that below a threshold number of cells within a hostile environment, bacteria express adhesins, pro-biofilm and immune evasion factors to establish an infection while avoiding activating and confronting immune cells; but above that same threshold bacteria switch production to defensive, pro-inflammatory factors to defend themselves and to acquire diminishing resources. For these reasons alone, QS has merited the interest it has received as an attractive and potentially broad-spectrum antivirulence target. QS inhibitiors (QSI) have been developed to target each step of the QS pathway.

However, QSI may have adverse effects. In S. aureus, for example, because the Agr QS system negatively controls biofilm formation and suppresses colonizing factors, QSI may actually increase total biofilm and colonization. Therefore, Agr inhibitors may only be appropriate for acute types of S. aureus infection¹⁷⁷. Even so, acute S. aureus bacteremia is often caused by Agr-deficient strains¹⁷⁸, casting doubt that QSI can be therapeutic for this common and serious S. aureus infection.

Compared to antibiotics, whether QSI can reduce the selective pressures favoring drug resistance remains controversial^{15, 179}, but serves as a good case study to appreciate the challenges inherent in predicting the likelihood of developing antivirulence drug resistance in general. Two models with in-vitro support help explain why QSI drug resistance may not develop. The first is a numbers game. In the case in which AI production is blocked in a well-mixed population, QSI-resistant clones would emerge in such few numbers that they would likely fail to produce sufficient AI to activate QS pathways¹⁸⁰. This, however, requires potent QS inhibitors capable of achieving near complete shutdown of AI production. The second model derives from sociomicrobiology. In P. aeruginosa, QS-deficient mutants are social “cheaters”: they benefit from publicly derived goods (e.g. extracellular nutrients liberated by secreted proteases) without contributing to their production^{180, 181}. When goods are public, cheaters prevent the rise of QSI-resistant cells because the latter expend energy to derive nutrients that the former exploits^{180, 181}. As a corollary however, QSI suppresses the emergence of natural, less-virulent QS mutants by pheno-copying cheating⁸², potentially leading to higher virulence when QSI is withdrawn. Nonetheless, the sociomicrobiology model may not apply to other pathogens and a physical structure that constrains AI diffusion may allow a single cell to self-activate¹⁸². Furthermore, QSI-resistant mutants have been derived in vitro and identified in clinical isolates¹⁸³, although these may be selected for by nonspecific antibiotic-like effects that incidentally confer resistance to QS inhibitors.

Pseudomonas aeruginosa.

Frequently antibiotic-resistant, P. aeruginosa is a highly adaptable Gram-negative pathogen with numerous virulence factors, capable of causing acute disease in severely burned or immunocompromised patients. It also causes chronic disease by colonizing the airways of cystic fibrosis patients. Because much research has been invested in P. aeruginosa, it has served as a model for targeting conserved virulence pathways in Gram-negative bacteria such as QS, type-III secretion systems (T3SS), and biofilm formation.

The four P. aeruginosa QS pathways (FIG 3, BOX 1), las, rhl, IQS, and pqs (also termed mvf) control biofilm formation, regulate the expression of many virulence factors and are required for full virulence in disease models^75–77. Therefore, QS factors may be promising targets for attenuating P. aeruginosa infectivity and a plethora of QS inhibitors (QSI)⁷⁸ have been developed and validated using reporter strains. There remains a need to develop more potent and specific inhibitors⁷⁹, but also to more rigorously evaluate QSIs as potential treatments. Currently, few QSIs have been tested in mammalian models. Only the antibiotic azithromycin, which below growth-inhibitory concentrations acts as a QSI against P. aeruginosa^{80, 81}, has been tested in a small number of patients. In those, it did reduce QS-gene expression⁸². Furthermore, QSIs should be tested against multiple P. aeruginosa clinical isolates because there may be heterogeneous QSI-resistance⁸³.

QSIs targeting the las pathway may only be effective in acute infections because this pathway is often inactivated during chronic infections in cystic fibrosis patients^{84, 85}. Interestingly, the recently discovered IQS quorum sensing signal bypasses las control in phosphate-limiting conditions to produce the PQS signal and permit virulence expression in las null mutants⁷⁷. Therefore, compounds for chronic infections should be directed against the three remaining pathways. One example is the benzamide-benzimidazole M64 that blocks the PqsR QS receptor (FIG. 3). Moreover, M64 inhibited QS across a series of clinical P. aeruginosa isolates and promoted mouse survival in burn and lung models of P. aeruginosa infections⁸⁶.

A series of recent publications illustrate the potential that antibody engineering provides to target surface-associated virulence factors. Two distinct antibody derivatives have been developed: a phenotypic screen yielded a single-chain variable fragment (scFv) that bound the exopolysacchride Psl⁸⁷ and a targeted screen produced a mAb that bound to a T3SS subunit, PcrV⁸⁸. Both agents conferred protection in animal disease models, although with different mechanisms of action. The antigen-binding sites were then combined into a single bispecific antibody derivative, MEDI3902 (FIG. 2), which further increased protection of mice in multiple disease models⁸⁹. In addition, MEDI3902 was synergistic with antibiotic treatment even against an antibiotic-resistant P. aeruginosa isolate, supporting the use of antivirulence as an adjunct to antibacterial treatment regimens. Currently, a phase II trial (NCT02696902) is ongoing to test the safety and efficacy of MEDI3902 in nosocomial pneumonia patients caused by P. aeruginosa. However, a phase-II clinical trial (NCT01695343) for another, less potent antibody derivative, the anti-PcrV pegylated-F_ab KB001A, in a phase II clinical trial for cystic fibrosis has failed.

Escherichia coli.

E. coli is remarkable for its numerous disease-causing variants, which can largely be explained by a distinct set of virulence factors⁹⁰. Serious and potentially fatal infections can be caused by Shiga-toxin producing E. coli (STEC) and the extraintestinal uropathogenic E. coli (UPEC). In addition, E. coli has increasingly been found to exhibit antibiotic resistance. Nearly pandrug-resistant strains are emerging as a result of plasmid-borne colistin resistance that has been found in many recent UPEC isolates^91–93. Furthermore, gut-resident E. coli has the potential to share antibiotic resistance elements with other enteric pathogens such as K. pneumoniae⁹⁴ and Enterobacter spp⁹⁵.

Shiga toxins (Stxs) Stx1 and Stx2 are the main virulence factors produced by STEC. Produced in the gut, Stxs disseminate systemically and damage renal endothelial cells, which contain high levels of the Stx receptor⁹⁶, globotriaosylceramide (Gb3). This damage may progress to the potentially fatal hemolytic uremic syndrome (HUS). Antibiotics are contraindicated for STEC infections and may in fact increase the risk for developing HUS⁹⁷ by activating the lysogenic bacteriophage that encodes Stxs and lyses STEC to release a bolus of Stx^{98, 99}. Therefore, much effort has gone towards blocking Stxs directly. An early approach was to develop receptor analogues to neutralize Stxs in the gastrointestinal tract before entering systemic circulation. SYNSORB-Pk is a polymer with the Gb3 trisaccharide moiety covalently linked to silicon-dioxide. While SYNSORB-Pk neutralized Stx in vitro¹⁰⁰ and in Stx-positive stool samples from patients¹⁰¹, it lacked efficacy in a small trial with 145 children diagnosed with diarrhea-associated HUS¹⁰². The authors speculated that treatment may have started too late in the disease process; and while that remains unproven, it again highlights an important point: initiating antivirulence therapy must precede or overlap with the timing that the targeted factor is actively promoting disease.

Several antibodies have been developed that neutralize Stx1 or Stx2 and protect mice from lethal challenges^{103, 104} and several clinical trials showed that these mAbs can be well tolerated^{105, 106}. Bellus Health is developing the chimeric cαStx2 mAb under the brand name Shigamab^™ that targets Stx2 (FIG. 2). A recent press release¹⁰⁷ highlighted conference data that showed cαStx2 given 4 days after the initial Stx2 injection in mice reduced biomarkers of renal damage and improved histopathology. While this suggests that Stx can be neutralized after initial production and may thus be a viable target in patients suffering from STEC-associated HUS, no evidence was given that it protected in a lethal challenge model and human efficacy data are still lacking.

UPEC is a major cause of urinary tract infections (UTIs) and can potentially gain access to the bloodstream¹⁰⁸ by ascending into the kidneys. UTIs are commonly treated with antibiotics, which is promoting increased antibiotic resistance in UPEC¹⁰⁹. Adhesins, especially fimbriae (also termed pili), are well-characterized virulence factors that promote UTIs by facilitating UPEC adherence to and invasion of epithelial cells. Inhibitors have been developed that interfere with adhesin binding, such as the potent biarylmannoside receptor analogues^{110, 111} (FIG. 2), or with adhesin biogenesis, such as the ring-fused 2-pyridones^{112, 113}. The biarylmannosides showed urine bioavailability, prevented UTIs, potentiated prophylactic antibiotics in a mouse UTI model^{111, 114}, and were effective against a notorious drug-resistant UPEC strain¹¹⁵. In addition, a mouse mAb noncompetitively inhibited FimK binding to mannose, detached preformed UPEC biofilms, and prevented UPEC attachment to epithelial cells in a mouse model¹¹⁶. However, no antivirulence agent against UPEC has been tested in patients.

Pathogenic targets with virulence inhibitors in preclinical development

Klebsiella pneumoniae.

The recent emergence of the plasmid-borne colistin-resistance mcr-1 gene in hypervirulent strains and the global rise of carbepenem resistance makes K. pneumoniae a high-priority pathogen for antivirulence development. The few virulence factors that have been described for this Gram-negative pathogen support bacterial survival within a hostile host and include bacterial lipopolysaccharide (LPS), a polysaccharide capsule, fimbrial adhesins, and iron-chelating siderophores. The capsule is the best-studied virulence factor and protects K. pneumoniae from phagocytic killing. However, achieving broad strain coverage by targeting the capsule is complicated by the existence of >78 serotypes.

To overcome this problem, Wang et al¹¹⁷ used a capsule- and LPS-deficient whole-cell phenotypic screen to identify novel virulence factors that could be blocked by mAbs in a high-throughput phagocytic killing assay or bound by scFvs using phage display. Both approaches identified MrkA, a type-III fimbrial subunit, as a common target on the bacterial cell surface¹¹⁷. The product of further engineering, scFv-Fc KP3 (FIG. 2), bound cells in 62% of 700 tested clinical isolates that included 509 multidrug resistant strains. KP3 had pleiotropic effects on virulence: it inhibited biofilm formation, increased opsonophagocytosis, and reduced bacterial burden in a mouse lung infection model.

The capsule may still prove a useful virulence target against emerging hypervirulent K. pneumoniae (kvKP) strains that are mostly K1 or K2 types. There is also some indication that antibiotic resistance is rising among these hypervirulent strains¹¹² although current kvKP isolates tend to be sensitive to antibiotics. These isolates have caused invasive disease in the community, are more resistant to phagocytic killing than non-K1/K2 strains and exhibit a hypermucoviscous phenotype^{118, 119}.

Acinetobacter baumannii.

Primarily causing severe nosocomial infections in critically ill patients, A. baumannii is a Gram-negative pathogen with alarming antibiotic resistance trends. Recent isolates are more frequently resistant to clinically relevant antibiotics such as carbepenems^{120, 121}. Some isolates are pandrug resistant to all FDA-approved antibiotics including colistin and tigecycline^121–124. Furthermore, A. baumannii has a remarkable ability to persist on abiotic surfaces in hospital environments and form biofilms on medical devices such as catheters and breathing tubes.

Attempts to interfere with virulence factors remain at an early investigational stage and only one study, which focused on blocking lipid A biosynthesis, has demonstrated substantial efficacy in a mouse model¹²⁵. Lipid A, the major immune-stimulatory factor that leads to septic shock, is the lipid component of LPS (also referred to as endotoxin) found in the outer membrane of Gram-negative bacteria. While LPS is essential in most Gram-negative bacteria, surprisingly A. baumannii strains lacking lipid A and LPS have been cultured in vitro and isolated from patients¹²⁶. Accordingly, the small molecule LpxC-1, which inhibits the lipid A biosynthetic enzyme LpxC, did not affect A. baumannii growth in vitro. Thus, in the special case of A. baumannii, LpxC inhibitors are potential antivirulence agents. Indeed, LpxC-1 did reduce LPS production and strongly attenuated virulence in mouse septicemia models¹²⁵. Moreover, LpxC inhibition may prove useful as an adjunct to antibiotics because LpxC inhibition increased cell permeability and augmented killing of A. baumannii by multiple antibiotics¹²⁷.

Additional early efforts to interfere with A. baumannii virulence factors focus on blocking QS or preventing iron uptake. A. baumannii, like several other Gram-negative bacteria, uses a LuxR/LuxI and acyl-homoserine lactone (AHL) QS network to regulate biofilm formation, motility, and potentially other virulence factors^{128, 129}, suggesting that QSIs could be effective therapeutic options for A. baumannii infections. Thus far, non-native AHLs, developed as QSIs for other bacteria, inhibited A. baumannii QS, with concomitant reductions in motility on soft agar and biofilm formation on a glass surface¹³⁰. Likewise, the antivirulence drug virstatin, originally developed to block the transcriptional activator ToxT in Vibrio cholerae¹³¹, decreased motility and biofilm formation in addition to inhibiting pili biosynthesis in A. baumannii¹³², possibly by interfering with QS¹³³. Finally, iron chelators, which bind free iron, or gallium nitrate (FIG. 2), which mimics ferric iron and competitively binds to bacterial siderophores, inhibited A. baumannii growth in vitro^{134, 135} and gallium nitrate decreased bacterial burden in a mouse pneumonia model¹³⁵.

Recent genome-wide transposon mutagenesis approaches have revealed genes important for A. baumannii survival in murine pneumonia¹³⁶ and bloodstream models¹³⁷. The virulence factors identified by those screens as well as others that have been described¹³⁸ represent potential novel targets for antivirulence approaches directed against A. baumannii, but more research is needed to exploit these factors.

Enterococci.

Enterococcus is a genus of Gram-positive bacteria that inhabit the gut. Enterococci, especially E. faecalis and E. faecium, cause a large proportion of healthcare-associated urinary tract infections, bacteremia, and endocarditis¹²⁰ and often exhibit high levels of antibiotic resistance, especially to clinically important antibiotics such as vancomycin. Recent research has revealed mechanistic underpinnings of enterococcal virulence factors, but few inhibitors have been identified that interfere with virulence.

Enterococci produce numerous virulence factors that may be exploited to treat disease. Several enterococcal MSCRAMMs mediate adherence to host tissue such as collagen and fibrinogen. In fact, immune serum derived against the conserved pilus subunit EbpA can be therapeutic in a mouse catheter-associated urinary tract infection model¹³⁹. After infection, the anti-EbpA serum significantly reduced bacterial burden in the kidneys. Other factors that are regulated by the virulence-control QS system fsr and promote biofilm formation, such as the gelatinase GelE and the serine protease SprE, could be directly targeted. Finally, some enterococcal strains produce a cytolysin¹⁴⁰, which can disrupt multiple cell types including erythrocytes and the presence of which in clinical isolates is associated with higher mortality¹⁴¹. However, its precise role in promoting disease remains unclear.

The few instances in which a virulence inhibitor for enterococci has been developed are limited to blocking the fsr QS system^142–144. One example is the peptide antagonist ZBzl-YAA5911 that has a 39 nM K_d for the E. faecalis receptor FsrC¹⁴³. In a rabbit endophthalmitis model, this natural ligand analogue prevented bacterial spread and improved retinal function.

Enterobacter.

Members of the genus Enterobacter are prominent opportunistic nosocomial pathogens, which have developed broad resistance to third-generation cephalosporins, penicillins, quinolones and carbapenems. Although Enterobacter pathogenesis mechanisms remain incompletely defined, the characterization of pathogenic Enterobacter isolates from human patients points towards the involvement of heat-stable enterotoxins¹⁴⁵, α-hemolysin HlyA¹⁴⁶, and the presence of at least 2 Stxs (Stx2 and Stx1e)^147–149 likely acquired by horizontal gene transfer from other Stx-producing bacteria^{150, 151}. Moreover, Enterobacter spp. produce siderophores¹⁵², proteinaceous extracellular fibers involved in adhesion and biofilm formation^{153, 154}, a T3SS¹⁵⁵, and a luxR homolog that may have a role in interspecies communication¹⁵⁶. However, the distribution of these virulence genes among clinical Enterobacter isolates is poorly understood.

Given the significance of Stxs in STEC-mediated virulence, it is tempting to speculate that Stxs in Enterobacter spp. may be required for pathogenesis. Interestingly, there is evidence that the newly characterized Stx1e plays a role in Enterobacter pathogenesis in vitro¹⁵⁷. Therefore, it may be prudent to target Enterobacter Stxs, and possibly HlyA, if a predominant role in pathogenesis can be definitively shown. mAbs that recognize the A subunit of Stx1e¹⁵⁷ modestly neutralize toxin cytotoxicity (FIG. 2). It may be advantageous to develop mAbs that cross-react with the B subunits of multiple Stx subtypes. These would block receptor interaction and the intracellular activity of the A subunit, offering a solution to protect against a broad range of Shiga toxin-producing bacterial strains in addition to pathogenic Enterobacter spp.

Mycobacterium tuberculosis.

Tuberculosis (TB), one of the leading causes of disease worldwide¹⁵⁸, is caused by the highly contagious Mycobacterium tuberculosis. TB is spread via inhaling aerosolized droplets from individuals with active bacterial infection. The bacteria can settle in the lungs causing pulmonary disease, or disseminate throughout the body via the blood or the lymph system. After infection, immune cells contain M. tuberculosis by surrounding the bacteria and forming a granuloma. In this cellular trap, the bacilli enter an asymptomatic dormant phase (latent TB infection) but can become active later in life. Lengthy antibiotic treatment regiments are significantly complicated by multi- and, more seriously, extensively-drug resistant M. tuberculosis strains.

Several new vaccines and boosters are currently at different stages of clinical development¹⁵⁹ to overcome the many problems of the traditionally used BCG (Bacillus Calmette–Guérin) vaccine, including the low efficacy against TB disease¹⁶⁰. The presence of antibodies to a variety of mycobacterial antigens after BCG vaccination^{161, 162} and in patients with active TB¹⁶³ supports the investigational use of therapeutic antibodies, especially in immunocompromised individuals who cannot receive the BCG vaccine. Early intervention with protective antibodies reduced disease manifestations by F_c receptor-mediated phagocytosis and killing of extracellular M. tuberculosis¹⁶⁴. Furthermore, mAbs have been produced towards the mycobacterial glycolipids lipomannan (LM) and lipoarabinomannan (LAM)¹⁶⁵, the surface protein heparin-binding hemagglutinin (HBHA)¹⁶⁶, and the chaperone α-crystallin (Acr; HspX)¹⁶⁷. These are key cell wall virulence antigens needed for attaching, entering and surviving within different cellular microenvironments. Importantly, mAbs to these cell surface antigens show protection in mouse models of M. tuberculosis infection^168–170 (FIG. 2).

Additionally, the recently discovered tuberculosis necrotizing toxin (TNT)¹⁷¹ and an increased understanding of T7SS efflux pumps in M. tuberculosis pathogenesis¹⁷² may serve as alternative virulence factors to target. The timing of antivirulence therapy in TB will likely be critical due to the ability of M. tuberculosis to remain dormant for decades and the difficulty of drugs to diffuse into intracellular bacteria. This may limit the utility of antivirulence to the active phase of the disease or as a prophylactic given to recently exposed individuals or those at high risk.

Concluding Remarks

While significant progress in antivirulence approaches has been achieved, most antivirulence drugs are still at the preclinical stage. However, a few have recently achieved FDA-approval with others entering into clinical trials. The approved drugs are currently limited to ones that block exotoxins, indicating that toxins may be promising targets in some of the most refractory and deadly pathogens, such as S. aureus, for which no antivirulence drugs exist. Other investigational antivirulence agents that have been proven to work in animal models, such as those directed at QS or adhesins, now need to pass safety and efficacy evaluation in patients. Furthermore, for a number of less-studied bacterial pathogens, such as A. baumannii, Enterobacter spp., and enterococci, more research is needed to define the full complement of virulence factors and detail their mechanisms, from which improved antivirulence agents can be rationally developed. Finally, as recent phenotypic screens have largely failed to identify novel antibiotics, this publication demonstrates the benefit of incorporating virulence factors as targets in otherwise conventional screening approaches.

Altogether, the examples we have discussed support the antivirulence paradigm by demonstrating that virulence factors can be directly antagonized with drugs to prevent or treat disease. However, as antivirulence is a growing and developing discipline, the pipeline for new drugs to combat diverse bacterial infections, even considering other antimicrobial strategies, remains small¹⁷³. To fulfill its intended role to help counter the antibiotic-resistance threat, efforts need to be accelerated and investments expanded at each stage of preclinical and clinical development.

Key points.

• Antibiotics are losing efficacy in treating many bacterial diseases due to increasing drug resistance. Meanwhile, antibiotic development has lagged; the last antibiotic class was introduced in 2003.

• Antivirulence is an alternative approach that focuses on interfering with bacterial virulence factors instead of central growth pathways to treat disease.

• Antivirulence drugs have been FDA-approved for bacterial toxin-mediated diseases and investigational drugs for antibiotic-resistant bacteria have entered clinical trials with several others in preclinical development.

• Antivirulence drugs will likely have distinct properties compared to antibiotics including reducing selection pressures that lead to drug resistance and minimizing perturbation of the healthy microbiota.

• However, antivirulence poses unique challenges for drug development and clinical use and such drugs may need to be used as combination therapy or as an adjunct to antibiotics.

• Increased investments and efforts are needed to develop new virulence inhibitors and advance existing ones through the translational pipeline.

Glossary

Adhesins

Cell-surface components or appendages of bacteria that facilitate adhesion to other cells or to surfaces e.g. fimbriae (pili). Adhesins contribute to host specificity and tissue tropism.

Antivirulence

A strategy that employs small molecules or biologics that inhibit or interfere with the production of disease-causing virulence factors.

Bacterial toxins

Includes a diverse variety of molecules produced by many pathogenic bacteria that injure cells. Toxins may directly form pores in eukaryotic membranes, or transfer enzymatic entities into cells that deregulate essential intracellular pathways or induce cell death. Additionally, some toxins modulate immune cell function thwarting immune responses and facilitating immune evasion for the invading pathogens.

Biofilms

Biofilms are surface-attached agglomerations of microorganisms embedded in an extracellular matrix. Biofilm-associated infections are difficult to eradicate and represent a significant reservoir for disseminating and recurring serious infections.

Empiric therapy

The use of broad-spectrum antimicrobial therapy prior to a definitive diagnosis of the disease causing organism.

F_ab and F_(ab)2

The antigen-binding fragment of an IgG antibody after proteolytic digestion. F_ab and F_(ab)2 are monovalent and divalent, respectively, and are produced under differing digestion conditions.

Horizontal gene transfer

The exchange of genetic information between organisms in a manner other than by traditional reproduction. Horizontal gene transfer is a key mechanism for the evolution of antibiotic resistance in bacteria.

Microbiota

Collective term for all microflora that are found residing within a given environment. Many micro-organisms are considered to be part of the beneficial microbiota on human skin and gut.

Monoclonal antibody (mAb)

Antibodies produced by a single B cell clone that are specific to one epitope of an antigen.

MSCRAMMs

An acronym for microbial surface components recognizing adhesive matrix molecules, are bacterial surface proteins that mediate adherence of the microbes to components of the extracellular matrix of the host.

Pandrug resistance

Resistance to all approved antimicrobial agents with activity against a specific bacterial species

Polyclonal antibody

A pool of antibodies from different B cells that recognize multiple epitopes on the same antigen.

Quorum sensing (QS)

A phenomenon describing the regulation of gene expression in response to cell-population density and the production and secretion of signaling molecules (autoinducers) by quorum sensing bacteria. Effector functions of quorum sensing include the regulation of genes involved in virulence and biofilm formation.

Receptor decoy

Receptor decoys acts as inhibitors by competitively binding to ligand thus preventing binding to the cognate receptor.

Response regulator

One part of a two-component system that activates transcription of a specific set of genes in response to certain stimuli.

Siderophores

Low molecular weight, high-affinity metal chelating agents produced by micro-organisms produced in low-nutrient conditions. Bacteria depend on siderophores for sequestering iron within vertebrate hosts in order to replicate and cause disease.

Single-chain variable fragment (scFv)

An antibody derivative in which the variable domains from the heavy and light chains have been fused with a linker. An scFv retains capability of binding the target antigen.

Two-component system

A common bacterial regulatory network that consists of a membrane-embedded sensor protein and a cytoplasmic response regulator.

Type III secretion system (T3SS)

A needle-like apparatus found in Gram-negative bacteria that delivers substrates across the inner and outer bacterial membranes and across eukaryotic membranes to the host membrane or the cytosol.

Type VII secretion (T7SS)

Specialized secretion systems found in Gram-positive bacteria. The most studied T7SSs are in M. tuberculosis. Protein substrates are first exported across the inner membrane via an ATP-dependent multimeric protein complex. Secreted substrates can localize to the culture supernatant or remain embedded in the highly hydrophobic mycobacterial cell envelope. Homologous systems are found in other bacteria from the phyla Actinobacteria and Firmicutes.

Virulence factors

A broad term used to define molecules produced by pathogens that promote disease or damage the host and include, but not limited to, adhesins, regulators, toxins, and siderophores.

Biographies

Seth W. Dickey is a Postdoctoral Research Associate Program (PRAT) fellow in the National Institute of General Medical Sciences (NIGMS) conducting research in the Pathogen Molecular Genetics Section, Laboratory of Bacteriology at the National Institute of Allergy and Infectious Diseases (NIAID) within the National Institutes of Health (NIH). He received his Ph.D. in Molecular Biology and Genetics with Siniša Urban at the Johns Hopkins School of Medicine studying the biochemistry of intramembrane proteases. His current research focuses on membrane-embedded virulence factors in methicillin-resistant Staphylococcus aureus.

Gordon Y. C. Cheung is currently a staff scientist at the NIAID. He received a Ph.D. in 2006 from the University of Glasgow, UK with Drs. John Coote and Roger Parton, investigating the immunomodulatory effects of the adenylate cyclase toxin of Bordetella pertussis. He then joined the laboratory of Stephen Leppla (NIH) as a visiting post-doctoral fellow studying the molecular mechanisms of anthrax toxin pathogenesis. His now investigates the contribution of a series of peptide toxins, phenol soluble modulins (PSMs), produced by Staphylococcus species towards disease manifestation, their interactions with the innate host response, and is exploring and identifying anti-staphylococcal therapeutics.

Michael Otto is a senior investigator at the NIAID, where he leads the Pathogen Molecular Genetics Section of the Laboratory of Bacteriology. Dr. Otto received his Ph.D. from the University of Tübingen, Germany. He has been researching staphylococcal pathogenesis mechanisms for more than 20 years.