Vaccines against Clostridium difficile

Abstract

Clostridium difficile infection (CDI) is recognized as a major cause of nosocomial diseases ranging from antibiotic related diarrhea to fulminant colitis. Emergence during the last 2 decades of C. difficile strains associated with high incidence, severity and lethal outcomes has increased the challenges for CDI treatment. A limited number of drugs have proven to be effective against CDI and concerns about antibiotic resistance as well as recurring disease solicited the search for novel therapeutic strategies. Active vaccination provides the attractive opportunity to prevent CDI, and intense research in recent years led to development of experimental vaccines, 3 of which are currently under clinical evaluation. This review summarizes recent achievements and remaining challenges in the field of C. difficile vaccines, and discusses future perspectives in view of newly-identified candidate antigens.

Introduction

Clostridium difficile is a gram-positive anaerobic bacterium able to infect either humans or animals¹ and commonly found in the environment.^2,3 It was isolated for the first time in 1935 from the intestinal flora of neonates and was initially considered a normal non-pathogenic resident of the gut.⁴ Only in 1970s was C. difficile identified as one of the agents responsible for antibiotic-related diarrhea and pseudomembranous colitis.⁵

C. difficile can exist as spores: metabolically inactive particles able to survive in soil, water, and on surfaces in clinical settings, due to resistance against common sterilization methods such as high temperatures, ultraviolent light, alcohol.^6,7 Spores represent the main vehicle for transmission, infection, and persistence of C. difficile. If ingested, spores can survive in the stomach of infected subjects and subsequently reach the intestine. Here, their fate strongly depends on the environment provided by the host. Spores typically switch from their dormant state to become active vegetative cells in a process termed germination. Germination can occur in response to stimuli including the bile salts, taurocholate, glycocholate and cholate,^8,9 amino acids¹⁰ and factors present in intestinal epithelial cells.¹¹ However, the action of such germination effectors (germinants) is effectively prevented by another bile acid, chenodeoxycholate, which has 10-fold higher affinity than taurocholate to the C. difficile spores.¹²

The presence of chenodeoxycholate in association with aerobic conditions likely inhibits germination and growth of the bacterium during its passage through the small intestine. In the large intestine of healthy individuals, spores can persist asymptomatically as resident commensal species and can rapidly degrade any residual bile component, preventing their germination.¹³ On the contrary, in absence or reduction of the normal commensal flora, a condition typically induced by treatment with wide-spectrum antibiotics, spores become able to germinate into vegetative cells and the absence of natural competitors for nutrients permits C. difficile to colonize empty niches in the colonic tract.

Once vegetative cells have been released from the germinant spores, the contact with host epithelial cells triggers the upregulation of bacterial genes responsible for adaptation to the new environment.^14,15 The bacterium remodels its surface by exposing adhesins, flagella, and proteolytic enzymes including Cwp84,¹⁶ which promotes the maturation of structural components of the bacterial cell wall¹⁷ and degrades elements of the host epithelium such as fibronectin, vitronectin, laminin, and fibrinogen.¹⁸ It has been suggested that the lytic action targeting host tissues induces the release of nutrients from the damaged epithelium and also promotes toxin diffusion.¹⁸ C. difficile cells can indeed cause disease by secreting 2 large enterotoxins, TcdA and TcdB, both able to severely damage the intestinal mucosa.¹⁹ These toxins have glycosyltransferase activity and modify small GTPases of the Rho protein family within the host cell, leading to alterations of cytoskeleton, activation of apoptosis and disruption of tight junctions.²⁰ The resulting impairment of intestinal barrier function leads to fluid accumulation, inflammation, and severe intestinal damage.^19,21,22 Although mechanisms that regulate toxin production are not completely elucidated, there are evidences that toxin synthesis is enhanced by several stimuli including metabolic stress,^23,24 temperature,²⁵ and sub-lethal doses of antibiotics.^26-29

Healthy individuals are generally able to mount a robust systemic immunity that limits gut damage induced by the toxins.^30,31 On the contrary, elderly or immuno-compromised subjects are prone to a series of symptoms whose severity ranges from mild diarrhea to fulminant pseudomembranous colitis.³²

In addition to TcdA and TcdB, up to 35% of C. difficile strains produce a third toxin called CDT or binary toxin,^33-36 composed by the ADP-ribosyltransferase subunit CDTa and the binding subunit CDTb.³⁷ CDT binds to the lipolysis-stimulated lipoprotein receptor protein on the host cells,³⁸ and the toxin-receptor complex is internalized into endocytotic vescicles. Subsequently, CDTa is released into the cytosol where it inhibits actin polymerization leading to profound alterations of the cell morphology,³⁹ including formation of microtubule protrusions that trap C. difficile on the surface of intestinal epithelial cells⁴⁰. The genes encoding CDTa and CDTb have been rarely found among isolates recovered from hospitalized patients,^41,42 but they are conserved in emerging strains associated with severe virulence.^43-45 It is therefore believed that CDT might play an adjunctive role in pathogenesis by enhancing the persistence of the bacterium in the colonized host.^46,47

Since the early 2000s cases of CDI-associated disease dramatically increased in the United States,^44,46,48 Canada,^49-51 and Europe,^52-54 accompanied by increase of case-fatality rates.^48,55 In the US, the incidence of CDI in acute care hospitals increased from 3.82 per 1000 discharges in the year 2000, to 8.75 per 1000 discharges in 2008.⁵⁶ A rise of CDI cases has also been observed in Europe. Data from the Communicable Diseases Surveillance Centre (CDSC) reported in England and Wales an increase from 1000 cases of CDI-associated disease in 1990, to 35 500 cases in 2003.⁵⁴ A survey supported by the European Centre for Disease Prevention and Control (ECDC) revealed that the mean incidence of CDI in Europe passed from 2.45 cases per 10 000 patient days in 2005, to 4.1 cases per 10 000 patient days in 2008.^52,53

A hyper-virulent strain called NAP1/027/BI was identified as being partly responsible for the increasing incidence of CDI in healthcare settings, although the reasons for its virulence are under discussion and include increased toxin production,⁵⁷ higher sporulation efficiency,⁵⁸ and greater antibiotic resistance.^50,59 Moreover, increased antibiotic use, an aging population with more comorbidities, as well as more frequent and accurate testing have been indicated as additional factors contributing to the increased CDI incidence.⁶⁰

Today C. difficile is the most common pathogen associated with nosocomial infectious diarrhea in hospitalized patients.^61,62 The Centers for Disease Control and Prevention (CDC) reported in 2012 that in the US among CDI cases with onset in healthcare facilities, approximately one half had onset in acute care units and the other half in long-term care facilities (LTCF).⁶³ Recovery is complicated by the tendency of disease to relapse. After the initial episode of CDI, a significant minority of patients is subjected to recurrences, resulting either from relapses due to persistent infection by the same strain or re-acquisition of a new strain from the environment.⁶⁴ The main factors responsible for recurrent CDI are persistence of alterations of intestinal microflora and inability to elicit an effective immune response. Antibiotic therapy for patients with mild to moderate CDI typically consists of oral metronidazole, while vancomycin is recommended for more severe cases or subjects with intolerance to metronidazole.^65-67 As antibiotics commonly used for CDI treatment cause alterations of fecal microbiota, their protracted administration prevents the re-establishment of natural resistance to C. difficile and might thereby predispose the host to recurrent infections.⁶⁸ After the initial episode, up to 33% of patients experience recurrent CDI^69-72 and recurrences can reach 45% after a second episode⁷³. This scenario is particularly frequent in healthcare settings, where the major part of the high-risk population is concentrated. A novel antibiotic, fidaxomycin, has recently been approved for CDI treatment.⁶⁹ Fidaxomicin appears to be more effective than vancomycin to prevent relapses, likely due to higher capability to preserve the intestinal microflora and favor beneficial re-colonization.⁶⁹

CDI is not limited to healthcare settings.^74,75 It has been estimated that 10–37% of all CDI cases are community associated (CA-CDI), with a population incidence of 20–30 cases per 100 000.^76,77

Diffusion of CDI imposes a considerable burden on patients in terms of morbidity, mortality, and prolonged hospitalization. Moreover, huge economical costs appear to weight on healthcare systems of developed countries. In Europe, potential costs associated to CDI management have been calculated at around €3000 million per year,⁵⁴ while costs directly attributable to CDI in US acute healthcare facilities in 2008 have been estimated at between 1 and $4.8 billion USD.⁷⁸ Additional costs may derive from recurrent CDI, adverse effects on elderly individuals, as well as from infection treatment and management in LTCFs.

Vaccination can represent a valuable strategy to prevent CDI. In the next sections we describe the efforts to identify and characterize suitable vaccine candidates against C. difficile and the state-of-the-art of preventive vaccines currently under clinical evaluation.

C. difficile Vaccines: From Toxoids to Recombinant Peptides

TcdA and TcdB are the determinants of CDI; however, the relative role of immunity against the 2 toxins in preventing the C. difficile associated disease has been long discussed. Early studies on purified proteins indicated that, differently from TcdB, TcdA alone was able to reproduce C. difficile mediated disease in animal models^79-81 and that anti-TcdA antibodies were necessary and sufficient to ensure complete protection against the clinical signs of infection.^30,31 More recently, the role of TcdB in C. difficile virulence⁸² and the relevance of anti-TcdB antibodies to prevent gastro-intestinal disease^83,84 have been re-evaluated, together with the identification of TcdA-negative, TcdB-positive strains responsible for severe clinical symptoms.^85-87 Collectively, such evidences support the conclusion that TcdB is a key factor in C. difficile disease and clearly suggests that neither toxin can be ignored in the development of an effective vaccine.⁸⁸

A further debate in the literature focused on the question of whether a mucosal or systemic antibody response is more important in protection against CDI. Several studies were driven by the assumption that local anti-toxin immunity was necessary to confer protection, since CDI in humans is clearly confined to intestinal tract. However, passive immunization studies demonstrated that circulating anti-toxin antibodies are effective in the treatment of severe CDI,^70,89,90 strongly suggesting that antibody-mediated toxin neutralization has an effect at the level of the epithelial barrier.

The protective effect of circulating antibodies at the site of damage is likely due to the ability of toxins to subvert the epithelial permeability. C. difficile toxins alter the epithelial barrier by disrupting the cytoskeleton organization and destabilizing the tight junctions.^19,20 This peculiar cytotoxic activity suggests that early toxin production is sufficient to initiate limited damage to the vasculature, which allows permeation of the immune serum containing neutralizing antibodies.

Historically, formalin-inactivated TcdA and TcdB were the first antigen mixture proposed for vaccine use.⁹¹ Well-established experience with toxoid-based vaccines against important pathogens such as Clostridium tetani⁹² and Corynebacterium diphtheriae⁹³ provided the rationale to extend the same approach to C. difficile, in an attempt to direct the immune response toward the inclusive epitope repertoire of TcdA and TcdB. Subsequent studies provided important information about the nature and localization of toxin-neutralizing epitopes and led to investigations on the use of recombinant engineered toxin fragments as vaccine candidates.

Vaccination with Toxoids

The efficacy of anti-CDI vaccines based on the administration of formalin inactivated toxoids A and B has been described for over 3 decades. This vaccine consists of formalin-detoxified toxins A and B obtained by purification from culture of VPI 10463, which is a hyper-productive strain for both toxins.

The first evidences demonstrated that the concomitant administration of toxoid A and B protected hamsters from lethal CDI.^94,95 After these initial studies many investigations explored the efficacy of toxoid-induced immunity in animal models by the evaluation of parameters influencing the vaccine efficacy such as the level of toxoid purification, the use of adjuvants and the routes of immunization, with a particular focus on the choice of the antigen delivery system.

Torres and colleagues evaluated the efficacy of chemically detoxified culture filtrates at different antigen doses and immunization routes in the hamster model.⁹⁶ They indicated that a combination of mucosal and systemic immunization induced protection against lethal outcome and diarrhea in hamsters. The combined use of mucosal and parenteral administration of both toxoids was further investigated by the use of a partially purified toxoid preparation based on fractionation of culture filtrates with ammonium sulfate.⁸⁹ Full protection was achieved only in animals vaccinated with a combination of rectal immunization with E. coli heat-labile toxin (LT) adjuvant and intramuscular injection of alum-adjuvanted toxoids. Importantly, the vaccination conferred protection to hamsters although the absence of anti-TcdA antibodies in feces suggesting that circulating antibodies were responsible for the recovery from the disease manifestations.⁸⁹ Similarly, transcutaneous immunization with toxoid A from culture filtrates has been reported to induce in mice both systemic IgG response and mucosal IgA response in serum and stool.⁹⁷

Whereas most of the previous studies were based on vaccination with toxoid A preparations, a recent study established that parenteral immunization with purified toxoid B conferred protection to CDI in hamsters infected with a TcdA-negative, TcdB-positive strain.⁹⁸ Importantly, it indicated that immunization of hamsters with a toxoid B preparation from the conventional toxin hyper-producer strain protected animals from the challenge by phylogenetically diverse strains.

Although the debate on the efficacy of mucosal vs. the systemic antigen delivery is still open, the success of passive immunization^70,89,90 supports the relevance of the parenteral delivered vaccines in humans. The use of toxoid-based vaccines in humans has been limited for a long time, despite several studies in animal models having demonstrated the importance of toxin immunity in preventing the lethal outcome of CDI. A partially purified toxoid preparation, based on ammonium sulfate fractionation of TcdA and B was the first vaccine candidate tested in a clinical study.⁹¹ Groups of 5 healthy adults were vaccinated intramuscularly with toxoid preparations using one of 3 antigen doses in absence or presence of alum as adjuvant. This preliminary clinical trial showed that vaccination was well tolerated and generated a serum antibody response against both toxins, as measured by ELISA and toxin neutralization assay. Moreover, a mucosal antibody response was also present in half of the vaccinated subjects.^91,99 The results obtained in this trial encouraged a pilot study aimed at evaluation of this vaccine in 3 patients with recurrent CDI.¹⁰⁰ Two of three vaccinated patients developed neutralizing antibodies against TcdA and B and all discontinued antibiotic treatments with resolution of recurrent diarrhea.

Recently, the vaccine has been further optimized with a second generation formulation, based on highly purified formalin-inactivated alum-adsorbed preparations of TcdA and B.^101,102 Preclinical testing in the hamster model demonstrated that intramuscular vaccinations with this toxoid-based formulation confer protection to death and disease symptoms in a dose-dependent manner.¹⁰¹ This highly pure toxoid-based vaccine, aimed at the primary prevention in a healthy population, underwent the phase I clinical trial in healthy adults and elderly volunteers showing a good level of safety and tolerability. Immunogenicity evaluation demonstrated that the vaccine induced a complete seroconversion for TcdA which was achieved at all doses in adults and at the highest vaccine dose in elderly. The TcdB seroconversion was lower, with subjects both in adults and elderly groups reaching 75% of a seroconversion. The antibody response appeared persistent only for TcdA in adult groups whereas the TcdB response declined 6 mo after vaccination.^102,103 A phase II trial has recently concluded, which evaluated the immunogenicity, dosage, and immunization schedule in 2 target populations, namely adults at risk for CDI (NCT01230957, clinicaltrials.gov) to assess the primary prevention and infected patients (NCT00772343, clinicaltrials.gov) to estimate the prevention of recurrences.¹⁰³

To overcome the safety issues associated with the large-scale production of toxoids, such as exposure to toxins and spores, Donald and colleagues have recently proposed a novel recombinant toxoid-based candidate vaccine consisting of genetically modified TcdA and B produced in a non-sporulating strain of C. difficile lacking the genes for the native toxins.¹⁰⁴ Although site-directed mutations abrogate cytotoxicity linked to the glucosyl-transferase activity of the toxins, a residual toxicity was observed which has been prevented by formalin treatment. This genetically and chemically detoxified recombinant vaccine induced functional antibodies in the hamster model and conferred a partial protection to lethality.¹⁰⁴ A phase I clinical trial of this vaccine is currently ongoing, (NCT01706367, clinicaltrials.gov), aiming at the evaluation of the vaccine dosage in healthy adults aged 50 to 85 y.

A summary of toxoid-based vaccines tested either in animal models or in clinical studies, are shown in Tables 1 and 2, respectively.

Table 1. Summary of toxoid-based vaccines described in the literature.

Antigens	Animal model	Route of immunization	Adjuvant	Ref.
Toxoid A and B from culture filtrates	Hamster	subcutaneous	Freund	94, 95
Toxoid A and B from culture filtrates	Hamster	Parenteral (i.p., s.c.) + mucosal (i.n., i.g., r.)	CT/RIBI	96
Partially purified toxoid A and B (44%)	Hamster	Parenteral (i.m.) + mucosal (i.n., i.g., r.)	None/Al2O3/LT	90
Toxoid A and B from culture filtrates	Mice	t.c., s.c.	CT	97
Purified toxoid B	Hamster	i.p.	MlipidA/RIBI	98
Highly purified toxoid A and B (>90%)	Hamster	i.m.	Al(OH)3/proprietary	101
Genetically modified toxoid A and B	Hamster	i.m.	AlPO4	104

i.p., intraperitoneal; s.c., subcutaneous; i.n., intranasal; i.g., intragastric; r., rectal; i.m., intramuscular; t.c., transcutaneous.

Table 2. Summary of vaccines tested in clinical studies.

Antigens	Clinical trial	Vaccination regimen/ target population	Adjuvant	References/ study (clinicaltrial.gov)
Partially purified toxoid A and B (ACAM-CDIFFTM by Acambis)	Phase I	4 doses i.m./healthy adults	None/Al(OH)3	91, 99
	Pilot study	4 doses i.m/patients	None	100
Highly purified toxoid A and B (>90%) (Acambis/Sanofi Pasteur*)	Phase I Phase II	3 doses i.m/ healthy adults, elderly Adults at risk/patients	Al(OH)3	102, 103 NCT01230957/NCT00772343
Genetically modified full-length TcdA and B (Pfizer)	Phase I	3 doses i.m/ healthy adults, elderly	None/Al(OH)3	NCT01706367
IC84 recombinant fusion protein (Valneva)	Phase I	3 doses i.m/ healthy adults, elderly	None/Al(OH)3	NCT01296386

* Acambis has been acquired by Sanofi Pasteur in 2008.

Vaccination with Recombinant Toxin-Based Peptides

Together with the development of toxoid-based vaccines, a number of studies have focused on the use of recombinant peptide antigens. The use of recombinant toxin sub-domains is an attractive strategy for the design of a vaccine against CDI for several reasons. First, it allows definition of the main neutralizing epitopes associated with a toxin domain; as a result, a vaccine targeting only the epitopes enhancing neutralizing antibodies may maximize the protective efficacy. Second, it permits to overcome the complexity of manufacturing toxoid preparations which requires purification of large proteins and chemical detoxification with intrinsic risk of incomplete inactivation and variability among consecutive preparations. Moreover, previous experiences with toxoid-based vaccines such as the acellular pertussis vaccine, highlighted that the formalin-based inactivation may alter structural epitopes with negative consequences on immunogenicity and reduced generation of neutralizing antibodies.^105,106

Importantly, classical toxoid preparations involve the use of toxin hyper-producer strain VPI 10463 but the question on whether these toxoids are also protective against heterologous strains is still unanswered. Recently the antigenic variation of TcdB in the epidemic strain was reported,¹⁰⁷ highlighting possible concerns on the limited cross-protective efficacy of the current vaccine formulations.

Pilot studies using recombinant peptides were initially aimed at the identification of protective epitopes contained along the TcdA and B sequences. Both toxins belongs to the Large Clostridial Toxin family, which present 3 distinct functional domains: an N-terminal enzymatic domain consisting of glucosyl-transferase (GT) and cysteine protease (CP) moieties, a central translocation (T) domain that mediates import into host cells and a C-terminal receptor binding domain (RBD) with 38 tandem repeats.¹⁰⁸

The first report underlining the ability of repeating units of RBD of TcdA to induce protective immunity was presented by Lyerly and colleagues who demonstrated that subcutaneous immunization with a recombinant peptide comprising 33 repeating units partially protected hamsters from death and diarrhea.¹⁰⁹

Over the last 2 decades this concept has been further dissected with a number of papers describing the potential of RBD domain of TcdA in conferring protection against CDI. As for the toxoid formulations, the debate has been centered on whether mucosal or systemic antibody response is more efficient in promoting protection and promising vehicles for vaccine delivery and route of immunization have been explored to address this point.

With the aim to induce anti-TcdA immunity in the intestinal tract, RBD sub-domains have been introduced in live attenuated vector strains used as delivery vehicle for inducing mucosal immunity. A large RBD domain fused to a secretion signal was introduced in an attenuated Vibrio cholerae strain; oral inoculation of rabbits with this construct evoked a systemic and mucosal immunity against TcdA and prevented fluid secretions and histological changes in an ileal loop challenge assay.¹¹⁰

With a similar approach, a recombinant fusion protein comprising 14 repeat units of TcdA and the immunogenic fragment C of tetanus toxin, was introduced in an attenuated Salmonella typhimurium strain. Intragastric and intranasal administration of this strain generated a significant anti-TcdA IgG serum response as well as an IgA response in intestinal and pulmonary mucosa.¹¹¹ The same domain in the purified form was tested after the direct administration to the mice nasal mucosa. This recombinant domain either as histidine tagged protein or fusion protein with the fragment C of tetanus toxin, generated anti-TcdA serum antibodies and, when combined to the mucosal adjuvants LT and LTR72, also a strong mucosal response in the pulmonary and nasal lavages but not at the intestinal surface.¹¹² More recently, Bacillus subtilis spores have been demonstrated to be an alternative oral vector able to deliver toxin fragments to the intestinal mucosa.¹¹³ RBD subdomains of both toxins expressed on the outer layer of spore coat induced neutralizing serum IgG and fecal IgA in mice and partially protected hamsters from fatal outcome; strikingly, the authors also presented evidences that antibodies raised against TcdA domain are cross-reactive to TcdB, which is in discrepancy with evidences reported by other groups.^114-116

Interestingly, since this modular organization in repeating units presents structural homology with the most powerful mucosal adjuvants such as cholera toxin (CT) and E. coli LT, Castagliuolo and colleagues studied the adjuvant activity of the RBD of TcdA¹¹⁷ (US Patent 5919463). The authors found that a peptide corresponding to 14 repeating units had an adjuvant effect on poorly immunogenic peptides, stimulating systemic response as well as mucosal IgA response after oral and nasal administration.¹¹⁷ More specifically, the profile of IgG subclasses and cytokine release suggests the induction of a mixed Th1/Th2 response with a predominance of the Th1 component.

DNA vaccine technology, known to provide humoral and cell-mediated immunity, has been also evaluated as proof of concept for a safe and easily-manufactured vaccine against CDI. A DNA vaccine encoding 92% of RBD of TcdA provided high antibodies titers and protected mice from death after parenteral inoculation of TcdA.¹¹⁸ Similarly, an adenovirus-based vaccination directed against a region spanning 85% of RBD of TcdA was demonstrated to generate a robust humoral and T cellular immune response and to provide protection in the mouse model after challenge with lethal doses of TcdA.¹¹⁹

Although all these studies have been informative in the search for the key attributes for an efficient anti CDI vaccination based on recombinant peptides, an optimal vaccine strategy may also need to be redirected for the inclusion of TcdB fragments.

Recently, the search of an efficacious vaccine targeting both toxins prompted novel studies on the use of combined recombinant peptides. Tian and colleagues proposed a fusion protein containing 19 of the 38 repetitive units of TcdA and 23 of the 24 repetitive units of TcdB.¹²⁰ In mice vaccinated intramuscularly, this fusion protein generated an IgG response against both toxins, with the TcdB portion appearing generally less immunogenic than TcdA. Functional antibodies were analyzed both by in vitro cytotoxicity assay and mouse toxin challenge model, demonstrating that a full protection against TcdA was achieved at all tested doses and in absence of adjuvant, whereas the presence of alum hydroxide was necessary to elicit protection against TcdB. The vaccination of hamsters with the adjuvanted fusion protein protected animals from lethality and reduced the severity of the disease. Finally, when co-administered with alum hydroxide this vaccine was also efficacious in a non-human primate model generating functional antibodies against both toxins. This recombinant fusion protein, named IC84, has been tested in a phase I clinical trial in healthy adults and elderly subjects (Table 2). The IC84 vaccine, administered at different dosages in absence of adjuvants, showed in both study populations favorable safety and tolerability, and induced antibodies against TcdA and B (NCT01296386, clinicaltrials.gov).

With a more traditional approach, the use of RBDs of both toxins were also tested in combination with the immune-adjuvant flagellin of Salmonella typhimurium, with the rationale to stimulate the toll-like receptor 5,¹²¹ known to protect mice against C. difficile colitis. Although the adjuvant activity of flagellin contributed to an enhanced systemic and intestinal IgA response against TcdA, the vaccination was not equally efficient against TcdB; moreover, flagellin did not confer an advantage in the animal model since the protective effect of the toxin domains promoted mice survival from lethal challenge even in absence of adjuvants. Recently, flagellin of C. difficile has also been demonstrated to stimulate toll-like receptor 5.¹²²

In contrast to the previous assumption that only binding domain regions induce protective antibodies, recent data^115,116 demonstrated that alternative neutralizing epitopes within TcdB are promising vaccine candidates. The complete screening of each subdomain of TcdA and B were recently analyzed for the ability to induce neutralizing antibodies against both toxins. This systematic analysis further confirmed that RBDs of both toxins are protective in an in vitro neutralization assay, but while a minimal protective sub-region of TcdA was identified, the entire RBD of TcdB was necessary to neutralize the TcdB toxicity. Importantly, this screening investigated the presence of protective epitopes within the enzymatic domains, establishing that whereas CP is not immunogenic, the GT domain of TcdB elicited functional antibodies.¹¹⁵ These evidences were confirmed in the hamster model where the co-immunization with the RBD of TcdA and the GT domain of TcdB was demonstrated to be the minimal combination necessary and sufficient to confer protection to lethality and recovery from diarrhea and tissue damage. As reported in previous studies, parenteral immunization resulted in the presence of anti-toxin IgG also in the intestinal tract, likely as consequence of toxin-induced alteration of gut mucosa which allows the permeation of circulating neutralizing antibodies.¹¹⁵

The potential of the GT domain of TcdB to induce neutralizing antibodies has been also evidenced by an alternative approach based on the use of a genetically detoxified chimeric protein comprising the full-length TcdB with the original RBD domain replaced by the corresponding portion of TcdA.¹¹⁶ This chimeric protein comprises the major protective epitopes of TcdA and B, residing in the RBD and GT domains, respectively, and induced neutralizing antibodies against both toxins. Mice and hamsters vaccinated by parenteral route develop long-term immunity against both toxins and are protected from primary and recurrent CDI, although non-stringent sub-lethal challenge conditions were applied. Notably, this chimeric vaccine, generated with toxin sequences of VPI10463 strain, protects experimental animals from challenge by BI/NAP1/027. This last point evidenced the potential of the GT domain to confer broad protection across diverse strains. Indeed, several studies indicated that the RBD region of TcdB appears variable between different toxinotypes^123,124 and recent characterization of TcdB in the epidemic strain 027/BI/NAP1 demonstrated that the RBD region is antigenically variable¹⁰⁷ from other strains. The prevalence of immunogenic epitopes in the binding domains of TcdA and enzymatic domain of TcdB has been reported also by Jin and colleagues in a DNA-based vaccination study.¹²⁵ Moreover, the importance of the GT domain of TcdB in inducing protective antibodies was also revealed by epitope mapping studies on humanized monoclonal antibodies.⁹⁰

Overall, these evidences suggest a potential advantage in the use of a more conserved region such as the GT domain as vaccine candidate. The mechanism by which anti-GT antibodies mediate protection remains to be determined. However, the extracellular cytotoxic activity of TcdB has been recently described, opening new perspectives in the understanding of its mechanism of action.^126,127

Table 3 shows a comprehensive summary of the vaccines based on the use of recombinant peptides.

Table 3. Summary of vaccines based on recombinant toxin peptides described in the literature.

Recombinant toxin-based peptides
Antigens	Animal model	Route/vehicles	Adjuvant	Ref.
RBD TcdA- 33 out 38 repeating units	Hamster	subcutaneous	Freund	109
RBD TcdA- 720 aa C-terminal peptide	Rabbit	Oral/V. cholerae vector	None/Cholera toxin	110
RBD TcdA- peptide 2387-2708	Mice	i.n., i.g./ S. typhimurium vector	None	111
RBD TcdA- peptide 2387-2708	Mice	i.n.	None/LT/LTR72	112
RBD TcdA/B peptides 388-2706 and 2137–2366)	Mice, hamster	o./B. subtilis spores	None	113
RBD TcdA- peptide 1839-2710	Mice	DNA vaccine, i.m.	/	118
RBD TcdA- peptide 1870-2680	Mice	adenovirus vaccine, i.m.	/	119
Fusion protein RBD TcdA+Ba	Mice, hamster, monkey	i.m.	None/Al(OH)3	120
RBD TcdA+B- full lenght	Mice	i.n.	FliCb /None/ Al(OH)3/LT192	121
ToxinA/B chimeric proteinc	Mice, hamster	i.m., i.p.		116
RBD, CP, GT TcdA+B	Mice, rabbit	DNA vaccine, i.m.	/	125
RBD, CP, GT TcdA+B	Mice, hamster	i.p.	Al(OH)3/MF59	115

i.p., intraperitoneal; o., oral; i.n., intranasal; i.g., intragastric; i.m., intramuscular. ^aFusion protein contains 19 of the 38 repetitive units of toxin A and 23 of the 24 repetitive units of toxin B. ^bFlagellin from Salmonella typhimurium tested as adjuvant both in fusion with toxin domains or as recombinant protein. ^cGenetically detoxified chimeric protein comprising the full-length toxin B in which the original RBD domain was replaced by the corresponding portion of toxin A.

Surface-Associated Antigens as Vaccine Candidates

Antibodies to surface protein antigens have also been associated with reduced CDI recurrence, although the correlation was not as statistically significant as in the case of an anti-toxin response.^128-130 These evidences suggest the possibility of adding other vaccine targets such as surface associated proteins and polysaccharides to toxin combinations, in an attempt to reduce gastrointestinal colonization and transmission.

Surface proteins

Toxin-based vaccinations, although effective in preventing the disease manifestations, are likely unable to prevent C. difficile colonization. A vaccine targeting surface-associated antigens could confer a potential advantage in the control of the disease offering preventive measures against carriage and transmission. Indeed, prevention or reduction of the colonization could limit the transmission in the healthcare facilities where the population at highest risk for CDI resides, and thereby reverse the recently emerging increase in community-acquired infections.⁷⁵ Moreover, several studies indicate that colonization with non-toxigenic strains is associated with a decreased disease incidence,¹³¹ suggesting that a vaccination with non-toxin components could be of added value in the prevention of the disease manifestations.

After the spore germination, the adherence of C. difficile to the intestinal mucosa is the first step for the settlement of the pathogen in the gut.¹³² The interaction between bacteria and epithelial cells is a multi-factorial process and requires the involvement of surface-displayed adhesins and virulence factors. A number of evidences suggest that during the infection surface-exposed antigens are able to induce an immune response. Studies on patient sera revealed the presence of antibodies directed to the flagellar components FliC and FliD, the Cwp66 adhesin, the fibronectin binding protein Fbp68, and the cysteine protease Cwp84.^128,129 Moreover, a proteomic-based approach on cell wall proteins surrounding the bacterium, showed that S-layer proteins are immunoreactive with patient sera.¹³⁰ These evidences confirm the expression of these surface proteins during the course of the disease and suggest that they are possible good candidates for an active immunization aiming at the prevention of bacterial colonization (summary in Table 4). The S-layer protein has been tested as a surface-antigen vaccine in combination with a series of systemic and mucosal adjuvants. The antibody response was variable depending on the vaccination regimen and adjuvant adopted, but overall none of the vaccinations conferred a significant advantage in the survival of hamsters to C. difficile challenge.¹³³ The vaccination with colonization factors is expected to reduce the level of colonization rather than to protect against lethal outcome. On the basis of this consideration Pechiné and colleagues tested several antigen combinations in a human flora-associated mouse model.¹³⁴ After pilot experiments evaluating the route of choice for an efficient systemic and mucosal immune response, mice were vaccinated by rectal route with combinations of FliD, flagellar preparation, Cwp84, and cell wall extract, showing a significant lowering of the level of colonization compared with control group. More recently, the cysteine protease Cwp84 was also evaluated as vaccine candidate both administered by rectal route¹³⁵ and encapsulated in pectin beads in an oral vaccine,¹³⁶ conferring a partial protection from lethality in the hamster model (Table 4).

Table 4. Summary of surface-associated antigens described in the literature.

Surface proteins
Antigens	Animal model	Route/vehicles	Adjuvant	Ref.
Crude SLP	Mice, hamster	i.p., i.n.	Al(OH)3/ CT /RIBI/chitosan glutamate/TMC	133
SLP, FliD, Cwp84	Mice	i.n., r., i.g./ PLGA encapsulation	Freund/CT	134
Cwp84	Hamster	s.c., r., i.g	None/ Freund/CT	135
Cwp84	Hamster	i.g./ pectin beads encapsulation	None	136

Surface glycans
Glycans	Animal model	Conjugation	Adjuvant	Ref.
PSII	Mice	CRM197	MF59	139
PSII	Rabbit	LTB E. coli	None	143
PSIII	Mice, rabbit	ExoA P. aeruginosa/HSA	Freund	148

i.p., intraperitoneal; s.c., subcutaneous; i.n., intranasal; i.g., intragastric; r., rectal.

Surface carbohydrates

Polysaccharides coating the surface of bacterial pathogens represent an optimal target for eliciting carbohydrate specific antibodies. Glycans are T cell independent antigens, but they can be turned into molecules able to evoke a T cell memory response following conjugation to a carrier protein. This strategy has found application in the prevention of many deadly infectious diseases.¹³⁷ Consequently, great attention has been directed in the recent years to the structural analysis of polysaccharides on the surface of C. difficile with the result of identifying 3 glycan structures, named PSI, PSII, and PSIII.¹³⁸ Among these 3 carbohydrates, PSII was found to be the more abundantly expressed by most of C. difficile ribotypes, including the hypervirulent strain NAP1/027 and other clinical isolates belonging to ribotypes 001, 018, 027, 078, and 126.^139,140

Following the discovery of PSII, it was not clear whether PSII was part of a capsule or a surface glycoprotein, or released to the external surface of the bacterium. Antibodies against the conjugated PSII detected the polysaccharide at the surface of the bacterial vegetative cells, thus confirming this molecule as a target for a carbohydrate based vaccine.¹³⁹ However the sugar coating was not as thick and uniformly distributed as expected for a capsule. Therefore, it can be hypothesized that PSII is expressed by the bacterium either as cell wall-linked polysaccharide not bound to peptidoglycan or as a conjugate with lipoteichoic acids.¹³⁹

Intriguingly, strain 630 and the hypervirulent strain R20291 can form structured biofilms in vitro and antibodies against the synthetic phosphorylated hexaglycosyl unit detect the presence of PSII in the biofilm matrix.¹⁴¹ As biofilms protect bacteria from multiple stresses, including immune responses, this finding sustains the interest for PSII as component of a glycoconjugate vaccine.

Notably, glycoarray analysis showed that specific IgA antibodies in the stools of hospital patients infected with C. difficile can recognize the synthetic PSII hexasaccharide hapten, suggesting that under exposure to PSII the human immune system may mount antibodies against several structural epitopes of PSII.¹⁴²

The PSII polysaccharide conjugated to diphtheria toxoid CRM197 was tested in Balb/C mice formulated with adjuvant MF59, and elicited high levels of IgG.¹³⁹ Importantly, in the same study it was evidenced that one single repeating unit phosphorylated at the non-terminal sugar was able to reach IgG levels comparable to the polymer, and the charged phosphate group is critical to induce IgG antibodies able to recognize the entire native polysaccharide.¹³⁹ Conjugation of PSII to the enterotoxin B subunit (LTB) of enterotoxigenic Escherichia coli (ETEC)¹⁴³ also achieved the induction of an immunogenic response against native PSII in rabbits.

Glycan structures other than PSII have been also detected on the surface of vegetative C. difficile cells. PSI has been detected in ribotype 027, but not in strains MOH900, and MOH718.¹⁴⁴ Subsequent NMR analysis or purification of carbohydrates from a larger collection of clinical isolates¹³⁹ has not led to identification of this polysaccharide, so PSI appears to be much less conserved than PSII. A glycoarray study to assess the presence of antibodies against PSI in CDI patients revealed that IgA levels to both PSI and PSII were higher in patients with less severe disease compared with asymptomatic controls, indicating that higher antibody levels to these antigens correlate with milder forms of CDI.¹⁴⁵ Additionally, sera from healthy horses have been demonstrated to contain natural anti-PSI IgG antibodies detecting both the synthetic non-phosphorylated repeating unit and the native polysaccharide, with a slightly higher recognition of the native PSI polysaccharide. This result suggests that the glycosyl phosphate and the polymeric nature of PSI could be immunologically important to develop a vaccine.¹⁴⁶

Antibodies against the third isolated glycan structure, named PSIII, have been detected in the blood of infected patients.¹⁴⁷ Intraperitoneal or subcutaneous immunizations of Balc/C mice or rabbits, respectively, with either intact or de-O-acylated LTA fraction of PSIII conjugated to the genetically inactivated P. aeruginosa exoTcdA protein (ExoA) or HSA revealed that it was possible to elicit IgG antibodies recognizing PSIII on C. difficile cells¹⁴⁸ rendering this molecule a potential target for vaccine development. It is worthy of note that to date no evidence of protective activity of specific antibodies against C. difficile carbohydrates has been reported. However, glycans such as PSII, which are well exposed on the bacterial surface and sufficiently conserved among the different strains, undoubtedly represent an attractive target for a preventive vaccine. It is reasonable to assume that these carbohydrates, either alone or in combination to surface layer antigens or proteins from flagella, could find application in the control of bacterial gut colonization.

Conclusions and Perspectives

The prevalence of CDI among elderly and immunocompromised individuals, together with the marked recurrence of the infection, poses substantial challenges to future C. difficile vaccines. Pivotal studies in humans¹⁰⁰ indicated that intra-muscular 4 doses immunization with the toxoid mixture after cessation of vancomycin can lead to resolution of recurrent C. difficile associated diarrhea, suggesting that vaccination could be highly recommendable to prevent recurrences in subjects that treated a first episode with antibiotics. However, the advent of fidaxomicin represents an important advance in treatment and prevention of recurrent CDI, enforcing the rationale to focus the vaccine therapy against primary CDI in adults. Three experimental vaccines against C. difficile are currently under clinical evaluation, all of them having as objective the primary prevention of CDI in adults and elderly (Table 2). Both toxoid based and recombinant vaccines have proven to be highly immunogenic in healthy adults, including subjects with age ≥65 y (Ref. 102, NCT01706367 and NCT01296386 studies in clinicaltrials.gov). This suggests the possibility to prevent by vaccination the insurgence of CDI in a high-risk population which includes the elderly, adults with planned hospitalization, LTCF residents and patients with co-morbidity requiring prolonged use of antibiotics. Challenges related to such a vaccination strategy will reside mostly in the ability of inducing in elderly and immuno-compromised individuals a rapid, long lasting, and protective immunity. For these reasons, further efforts can still be pursued to optimize vaccine efficacy, including regular booster immunizations in adulthood, development of innovative adjuvants¹⁴⁹ and design of accelerated schedules compatible with planned hospitalization and surgery.

Despite recent surveys which indicated that effective measurements for infection control resulted in beneficial effects on reducing the CDI transmission in nosocomial environments,^150,151 healthcare facilities continue to represent a reservoir for C. difficile re-infection and transmission. This emphasizes the importance of developing multi-component vaccines able to neutralize the effects of both toxins and to reduce the bacterial persistence within the host. Colonization factors such as adhesins, flagellar proteins, S-layer components, and the binary toxin CDT represent excellent auxiliary candidates for next generation vaccines, although the sequence variability among different isolates raises the question of their ability to confer broad cross-protection.

Besides the high risk patients in the healthcare facilities, cases of community acquired CDI are increasing.¹⁵² To explain the phenomenon, new potential sources of transmission have been considered. Although direct C. difficile transmission from animals to humans has not been definitely proven, direct or indirect contact with colonized animals and foodborne transmission have been indicated as potentials sources of community acquired CDI.¹⁵³ For this reason, domestic pets and production animals represent a potential novel target population for preventive vaccines.

Overall, changing epidemiology poses new challenges to CDI treatment. The promise of new vaccines able to eliminate the distressing diarrheal symptoms paves the way to novel treatment opportunities, as well as to have a potential beneficial impact in reducing environmental contamination. Future optimizations of toxin-based vaccines should include development and use of novel adjuvants able to enhance the immune response in elderly. Moreover, the inclusion of additional structural bacterial antigens could help to limit the C. difficile survival into the host, further reducing the potential sources of infection and relapses in hospital environments.

10.4161/hv.28428