Adaptive immune constraints on C. difficile vaccination

1. Introduction

C. difficile has emerged as the leading cause of nosocomial infections in the US and there is currently no effective vaccine in use. New highly virulent strains have emerged that pose a growing threat to public health. Promising candidate vaccine antigens have emerged that may stimulate a protective immune response to C. difficile. However, the mechanistic aspects of the adaptive immune response to C. difficile are understudied, and there are several challenges currently being addressed, or that need to be overcome to bring an effective prophylactic vaccine to fruition. Herein, we summarize those challenges that pertain to adaptive immunity, discuss some recent promising vaccine studies, and suggest approaches to mitigate the remaining challenges.

2. An efficacious C. difficile vaccine

Arguably an ideal C. difficile vaccine will prevent gastrointestinal colonization and infection, block cellular intoxication by secreted toxins, and provide coverage against several pathogenic strains. The vaccine will cause minimal inflammation, but generate a protective immune response in over 90% of aging individuals. Protection will be long-lived and sustained by robust T cell and B cell memory and therefore a rapidly reactivated cellular and humoral response. The humoral response will provide immediate protection through the ongoing secretion of pathogen- or toxin-neutralizing antibody (Ab). Additional protection will be provided by stimulation of memory B and T cells to cause differentiation of new Ab-secreting plasma cells.

However, these goals have not yet been realized. One reason for this is significant constraints imposed by the target vaccine antigens (Ags) and the limitations of the host immune response to them. These constraints are compounded by other factors such as genetic variability, immune-senescence, and too few licensed vaccine adjuvants. However, there are possibilities for maximizing efficacy in the short- and longer-term.

3. Vaccine candidates and constraints on adaptive immunity to them

Several vaccine candidate Ags have been identified and tested in animal models. Secreted toxins A and B (TcdA, TcdB), cell wall protein family members, peptidoglycans, and polysaccharides all have potential for inclusion in a vaccine (reviewed in [1]). TcdA and TcdB are amongst the most intensely studied vaccine antigens largely because they are major virulence factors. TcdB is essential for local and systemic pathology in animal models [2], and in patients with C. difficile infection (CDI), TcdA− / TcdB+ strains are becoming more prevalent [3]. The presence of TcdA- and TcdB-specific serum IgG also correlates with less severe disease [4]. Clinical trials using a TcdB-blocking mAb reduced disease recurrence [5], suggesting that programming an anti-TcdB response by vaccination could be valuable in limiting reinfection. Although TcdA- and TcdB-specific Ab could arguably result in a population of asymptomatic but infectious individuals, immunogens that stimulate these responses could be very useful as a first generation vaccine. It should be noted that Hong and co-authors recently observed that oral vaccines consisting of B. subtilis expressing TcdA fragments induced TcdA-reactive mucosal Abs that cross-reacted with the vegetative cell surface and spore coat of C. difficile and limited colonization [6]. This work suggests that selection of toxin-derived vaccine epitopes that stimulate cross-reactivity against the C. difficile bacterium could be promising.

The emergence of ‘hyper-virulent’ strains of C. difficile has highlighted significant challenges for vaccination. The 027 C. difficile ribotype is associated with more severe infection than ‘historical’ strains, and produces a hyper-virulent TcdB (designated TcdB2) that is more toxigenic than its historical counterpart (designated TcdB1). Studies by the Ballard laboratory have revealed that while TcdB1 and TcdB2 share 92% sequence identity and are similarly immunogenic in rabbits, TcdB2-specific Ab is poorly neutralizing and TcdB1-specific Ab neutralizes TcdB1 but fails to cross-neutralize TcdB2 [7]. This work provides compelling evidence that a single TcdB-based immunogen may not sufficiently protect against multiple C. difficile strains. Indeed, vaccines may also need to account for the same ribotypes (including 027) expressing a third toxin known as binary toxin since there is some indication that TcdA⁻/TcdB⁻ strains expressing binary toxin could be pathogenic [8]. A study by Secore and colleagues reinforce this point since a tetravalent vaccine that included TcdA, TcdB, and binary toxin proteins CDTa and CDTb provided better protection against a 027 ribotype strain than a bivalent vaccine including TcdA and TcdB immunogens [9].

Recurrent C. difficile infection is a growing problem that demonstrates that natural infection does not adequately stimulate a protective immune response. This could be because people are re-infected with distinct C. difficile strains each time they are infected. Thus protective adaptive immunity to one strain may not protect against another strain. This is consistent with the lack of toxin cross-neutralizing Abs, but does not account for other Ags common to multiple strains. There could also be intrinsically different adaptive immune responses to similar Ags produced by different strains. The Lang laboratory recently reported that TcdB1 and TcdB2 were similarly immunogenic in mice, resulting in very similar Ab profiles with regard to Ig sub-class [10]. In that study, while long-term memory B-cell driven Ab recall responses were observed against TcdB1 and TcdB2, neutralization of TcdB2 and resistance to in vivo challenge with toxin was vastly inferior to that against TcdB1 [10]. This suggests that the memory B cell compartment could essentially be mis-directed against TcdB from some strains, resulting in high titers of non-neutralizing Ab. Indeed, while Bauer and colleagues observed a good inverse correlation between TcdB-specific serum IgA and IgG and disease recurrence [11], their observation that 12/16 recurrences were in patients infected with the 027 ribotype is consistent with mis-directed B cell memory against TcdB2.

Efforts are increasingly focused on C. difficile surface polysaccharides as vaccine Ags. There are three major surface polysaccharides known as PSI, PSII, and PSIII [12]. PSII may exert several influences on C. difficile including virulence and data thus far indicates that it is expressed by all pathogenic strains [13]. PSII is a phosphate-containing poly-hexa-saccharide molecule that stimulates B cell responses. One limitation of polysaccharides is their poor capability for inducing B cell memory, Ab isotype class switch, and affinity maturation. However, glyco-conjugates such as the pneumococcal and meningococcal vaccines in which the polysaccharide is conjugated to a T cell-dependent protein Ag are used in the clinic. Although the B cell antigen receptor is specific for the polysaccharide, internalization, processing and presentation of carrier-protein derived peptides on MHCII lead to the necessary T cell help for class switch, affinity maturation, and memory [14]. Some reports have been published recently indicating that conjugation of PSII to carrier proteins such as tetanus toxoid can induce T cell help and promote desirable IgG responses [1].

At present there are few licensed vaccine adjuvants and we remain reliant on Alum or oil-inwater suspensions. Alum is heavily used in the US and results in reasonably good humoral immunity, but stimulates poor cellular immunity [15]. However, Alum formulations substantially enhance the anti-TcdB IgG response in mouse models [10], but the benefit of adjuvant inclusion is less clear in human volunteers [16]. Another major challenge is the inherent variation in human immune responses. There is substantial variability in immune cell representation, phenotype, and function in the human population, which is subject to genetic and environmental influences (reviewed in [17]). To this challenge, one might also add immuno-senescence in which aging results in increasing difficulty in stimulating protective adaptive immune responses. This is evident in influenza immunization, where the dose of the major vaccine ingredient, hemagglutinin is four times higher for individuals aged over 65 than it is for younger individuals (https://www.cdc.gov/flu/about/disease/65over.htm). Immuno-senescence may be a problem for controlling natural infection since, in older individuals with CDI bacterial burden inversely correlates with toxin-neutralizing Ab [18]. Collectively, these issues make successful vaccination against C. difficile more challenging.

4. Minimization of constraints on successful vaccination

It appears likely that a vaccine will need to be formulated to contain immunogenic fragments of TcdA and TcdB and possibly binary toxin, enzymatically inactive toxins, or inactivated toxoids. Several ribotypes and strains may need to be represented with regard to TcdB, and even then coverage against ribotypes that produce TcdB2 may be poor or incomplete. Conjugation of protein carriers to the PSII polysaccharide is a promising avenue for vaccination [12]. The inclusion of a glyco-conjugate using common vaccine Ags such as diphtheria toxin or tetanus toxoid may harness and boost-pre-existing T cell memory. In this regard, conjugation of PSII to toxin fragments may also be worth considering, with the caveat that TcdB2 may not be an optimal choice of carrier protein. These approaches appear likely to be at least moderately efficacious, and will perhaps be boosted by inclusion of adjuvants and/or increasing dosage in older individuals. Good tracking of ribotype and/or strain in infected individuals may also offer opportunities to understand which are most prevalent in a given region and allow better matching with the vaccine formulation.

These practical efforts could offer short term gains, but basic research is needed to fundamentally understand the immune response to C. difficile natural infection and vaccination. It is known that individuals with a prior C. difficile infection have circulating memory B cells reactive to TcdA and TcdB [10,19]. A single cell analysis of the Ab repertoire may offer important insights into responses that correlate with toxin neutralization and/or limited disease recurrence. A systems biology approach to understanding immuno-senescence and how to circumvent it may also be complementary to these efforts. More basic research on the mechanisms of action of combination adjuvants may lead to development of adjuvant platforms to direct the immune response as required. This type of approach may benefit vaccination against several pathogens including C. difficile.

5. Concluding Remarks

C. difficile represents a growing threat to public health and the need for a vaccine is becoming a matter of some urgency. A combination of practical steps using existing knowledge will likely result in a good first generation vaccine formulation and this seems to be borne out by ongoing clinical trials [20]. However, more basic research to fundamentally understand the adaptive immune response to C. difficile may be needed to develop highly efficacious vaccines that program a desired immune response.

Abbreviations

Ab

Antibody

Ag

Antigen

CDI

C. difficile infection

PSII

Poly-saccharide II

TcdA

C. difficile toxin A

TcdB

C. difficile toxin B

CDTab

C. difficile binary toxin