The host immune response to Clostridium difficile infection

Abstract

Clostridium difficile infection (CDI) is the most common infectious cause of healthcare-acquired diarrhoea. Outcomes of C. difficile colonization are varied, from asymptomatic carriage to fulminant colitis and death, due in part to the interplay between the pathogenic virulence factors of the bacterium and the counteractive immune responses of the host.

Secreted toxins A and B are the major virulence factors of C. difficile and induce a profound inflammatory response by intoxicating intestinal epithelial cells causing proinflammatory cytokine release. Host cell necrosis, vascular permeability and neutrophil infiltration lead to an elevated white cell count, profuse diarrhoea and in severe cases, dehydration, hypoalbuminaemia and toxic megacolon. Other bacterial virulence factors, including surface layer proteins and flagella proteins, are detected by host cell surface signal molecules that trigger downstream cell-mediated immune pathways.

Human studies have identified a role for serum and faecal immunoglobulin levels in protection from disease, but the recent development of a mouse model of CDI has enabled studies into the precise molecular interactions that trigger the immune response during infection. Key effector molecules have been identified that can drive towards a protective anti-inflammatory response or a damaging proinflammatory response.

The limitations of current antimicrobial therapies for CDI have led to the development of both active and passive immunotherapies, none of which have, as yet been formally approved for CDI. However, recent advances in our understanding of the molecular basis of host immune protection against CDI may provide an exciting opportunity for novel therapeutic developments in the future.

Background

Clostridium difficile infection (CDI) is the most common infectious cause of healthcare-acquired diarrhoea. Approximately 15–25% of all cases of antibiotic-associated colitis are caused by C. difficile and this likelihood increases with the severity of disease, reaching 95–100% among patients with documented antibiotic-associated pseudomembraneous colitis (PMC) [Bartlett, 1994].

C. difficile colonization can lead to asymptomatic carriage, or a wide range of symptoms, from mild diarrhoea to fulminant colitis, systemic disease and death. The interplay between the pathogenic virulence factors of the bacterium and the counteractive immune responses of the host may in part explain how colonization with C. difficile can result in a wide spectrum of outcomes and some of these features will be explained in further detail in this review.

Significant challenges have recently arisen due to changes in epidemiology, emergence of antimicrobial resistance and increasing incidence of severe disease leading to an unanticipated increase in morbidity and mortality attributed to CDI. The increase in severe disease and the propensity for recurrence of infection ensure that CDI remains a major cause of hospital-acquired infection. The limitations of standard CDI therapies and lack of novel therapies that have been approved for clinical practice ensure that CDI remains a significant healthcare burden.

Risk factors influencing outcome of colonization with Clostridium difficile

Colonization with C. difficile does not automatically lead to development of symptomatic CDI. Colonization rates in healthy humans in the community range from 0.8% to 13% and are higher in long-term care facility residents [Arvand et al. 2012; Ozaki et al. 2004]. The host immune status plays an important role in protection against symptomatic disease after colonization with C. difficile and it is thought that repeated reinfection from the environment stimulates a protective antibody response in non-hospitalized healthy hosts [Kelly et al. 1992; Sanchez-Hurtado et al. 2008; Viscidi et al. 1983].

The immune status of hospital patients is important for determining those at increased risk of CDI as the risk of developing CDI is higher in immunocompromised patients [Yolken et al. 1982]. Approximately half of hospital patients colonized with a pathogenic strain of C. difficile develop symptomatic CDI due to an inability to mount an adequate antibody response to C. difficile toxins [Kyne et al. 2000; Mulligan et al. 1993].

Other major risk factors for CDI are increasing age, prolonged hospital stay and underlying comorbidities [Bauer et al. 2009; Moshkowitz et al. 2007]. The most prominent risk factor is recent antimicrobial use within 8 weeks prior to infection, which disrupts the protective bowel microflora, leading to loss of colonization resistance [Bignardi, 1998; Dial et al. 2008].

Clostridium difficile virulence factors

Toxins A and B

The major virulence factors of toxigenic C. difficile are the large secreted glucosyltransferase protein toxins A (TcdA) and B (TcdB). The combined action of these toxins on the colonic intestinal epithelium is responsible for the profound intestinal inflammatory response seen in CDI [Kuehne et al. 2010; Thelestam and Chaves-Olarte, 2000]. TcdA and TcdB proteins share four functional domains. The first is a catalytic domain, involved in binding and inactivation of intracellular Rho GTPases in intestinal epithelial cells, mediating disruption of the cell cytoskeleton and necrosis and loss of the colonic monolayer integrity [von Eichel-Streiber et al. 1996]. The second is the cysteine protease domain that is involved in autocatalytic processing of the toxin protein in conjunction with the host cytosolic cofactor inositol hexakisphosphate (InsP₆) [Reineke et al. 2007; Pruitt et al. 2009]. The third is the translocation domain that mediates entry of the toxin into the target cell cytoplasm and the fourth is the receptor binding domain that is truncated in TcdB [von Eichel-Streiber et al. 1996; Jank and Aktories, 2008].

The toxins are encoded on a pathogenicity locus (PaLoc) [Braun et al. 1996; Rupnik et al. 2005] and variation in the toxin genes across different strains has led to different toxinotypes, of which 31 have been identified so far [Rupnik, 2008, 2011]. Some strains of C. difficile only produce functional TcdB, due to a truncation in the 3′ region of the repetitive domain of tcdA [von Eichel-Streiber et al. 1999]. TcdB from these strains is able to modify more substrates than wild-type toxin B from TcdA⁺TcdB⁺ strains due to alterations in the active site, giving rise to altered glucosylation of Rho proteins and inducing a differential cytopathic effect on cultured cells [Chaves-Olarte et al. 1999]. A number of TcdA⁻TcdB⁺ toxinotypes have been identified to date, but the most clinically significant is toxinotype VIII which has been responsible for outbreaks worldwide [Al-Barrak et al. 1999; Alfa et al. 2000; Kuijper et al. 2001; Drudy et al. 2007].

Once the intestinal epithelium has been breached, TcdA and TcdB are able to access the underlying lamina propria and act directly on resident macrophages and peripheral blood mononuclear cells to cause cell damage and perpetuation of the inflammatory response [Chaves-Olarte et al. 1996; Pothoulakis, 2000]. TcdA has also been shown to directly bind and induce apoptosis in monocytes, thereby inactivating key effector cells and directly influencing the host protective immune response [Modi et al. 2011; Solomon et al. 2005].

Binary toxin

In addition to TcdA and TcdB, some C. difficile strains produce an adenosine diphosphate ribosyltransferase binary toxin, Cdt, a member of the AB binary toxin group made up of an enzymatic (A) and a transport (B) component [Holbourn et al. 2006; Barth, 2004]. The precise role for binary toxin in pathogenesis is unclear; however, it has been shown that it is toxic to Vero cells and may increase adherence of C. difficile to target cells, by the formation of microtubule protrusions [Schwan et al. 2009; Sundriyal et al. 2010].

Intrinsic antigens

The C. difficile bacterium exists in two physical forms: the metabolically active vegetative cell and the inert spore. In addition to secreting the major virulence factors TcdA and TcdB, the vegetative cell expresses other intrinsic immunogenic virulence factors including flagella, surface layer proteins (SLPs) and cell wall proteins (Cwp66, Cwp84 and CwpV) [Ausiello et al. 2006; Calabi et al. 2002; Drudy et al. 2004; Pechine et al. 2005b; Tasteyre et al. 2001; Janoir et al. 2007; Emerson et al. 2009]. Surface layer proteins, especially SlpA, can be highly variable across different strains, suggesting that the variable regions of SlpA may have a role in evasion of the host immune response [Bianco et al. 2011; Spigaglia et al. 2011a, 2011b].

Flagella mediate chemotaxis of the vegetative cell and penetration of the mucus layer enabling adherence of the bacterium to the epithelial cell surface. There is less variation in flagella proteins (FliC and FliD) across C. difficile strains than other surface-associated proteins but they are highly immunogenic and are recognized by epithelial cells as part of the host cell pathogen-sensing pathway and can trigger downstream inflammatory responses [Pechine et al. 2005b; Jarchum et al. 2011].

Whilst immunogenicity and variability of the surface proteins of the vegetative cell is well known, there is little information regarding the host immune response to the metabolically inactive spore. C. difficile spores are comprised of a inner cytoplasm containing the genetic material, which is protected by a peptidoglycan inner core, a thick cortical layer and in some cases a loose fitting exosporium [Permpoonpattana et al. 2011]. Immune responses to C. difficile spore coat proteins would be expected to prevent germination and reduce the infectious burden of spores in the colon and subsequent disease by promoting phagocytosis of the spore as has been shown in Bacillus subtilis infection models [Yu and Cutting, 2009]. In vitro assays of C. difficile spore phagocytosis by macrophages has however shown intracellular survival of spores, resulting in macrophage cell intoxication and death [Paredes-Sabja et al. 2012].

Role of the innate immune response in Clostridium difficile infection pathology

The pathogenesis of CDI early (0–12 h) in the course of infection is predominantly characterized by acute intestinal inflammation mediated by the inducible innate immune response.

Host antimicrobial peptides and toxin attenuation

The first step for the host in defence against CDI is preventing or delaying the binding of either the bacterium or the secreted toxins to the intestinal epithelial cells where they can cause damage. A mucus layer covering the epithelium creates a dynamic defence barrier against luminal bacteria; however, many pathogens including C. difficile have evolved mechanisms to reduce mucin production during infection [Branka et al. 1997; Corfield et al. 2000]. A further defence at the epithelial surface is the presence of antimicrobial peptides, defensins and cathelicidins, secreted by specialist cells in the intestinal crypt called Paneth cells and by leukocytes [Bevins et al. 1999]. Cathelicidin and defensins have been shown to significantly reduce tissue damage and inflammation caused by TcdA and TcdB and reduce inflammatory cytokine production [Hing et al. 2012; Giesemann et al. 2008]. Another mechanism by which the host is able to affect the cytotoxicity of TcdA and TcdB is by direct chemical attenuation through S-nitrosylation that inhibits toxin cleavage and cell entry [Savidge et al. 2011]. Attenuation of toxin activity was shown to be enhanced by addition of InsP₆, highlighting a role for this cofactor in augmenting the antitoxin activity of S-nitrosylation [Savidge et al. 2011].

Proinflammatory cytokine release

C. difficile toxins act quickly on the intestinal epithelial cells, causing loss of tight junctions, cell detachment and necrosis in order to breach the protective mucosal barrier [Sun et al. 2009; Genth et al. 2008]. The host needs to respond rapidly to prevent further cellular damage and dissemination of toxins into the blood stream. Epithelial cells intoxicated by C. difficile release proinflammatory mediators, including interleukin 8 (IL-8) and macrophage inflammatory protein 2 into the lamina propria to initiate an acute intestinal inflammatory response [Mahida et al. 1996; Johal et al. 2004b; Kim et al. 2002]. The intestinal epithelium is also able to sense other C. difficile proteins, including SLPs, cell-wall proteins Cwp66 and Cwp84 and flagellin via specific cell surface toll-like receptors (TLRs) including TLR4 and TLR5 [Jarchum et al. 2011; Ryan et al. 2011]. Once the protective epithelial barrier has been breached, TcdA, TcdB and other C. difficile proteins come into contact with submucosal macrophages, monocytes and dentritic cells and trigger dissemination of the inflammatory cascade via further release of proinflammatory cytokines IL-1α, IL-1β, IL-6, IL-8 and tumour necrosis factor-α [Linevsky et al. 1997; Pothoulakis, 1996; Flegel et al. 1991; Melo-Filho et al. 1997; Bianco et al. 2011; Pechine et al. 2005a; Ryan et al. 2011; Vohra and Poxton, 2012].

Peripheral blood mononuclear cell infiltration

Proinflammatory cytokines, predominantly the neutrophil chemotactic factor, IL-8 and expression of monocyte and leukocyte adhesion molecules including CD18 integrins and selectins enable prominent neutrophil and monocyte infiltration from the peripheral blood [Savidge et al. 2003; Flegel et al. 1991; Mahida et al. 1998]. Localized mast cell degranulation promotes histamine release thereby increasing the permeability of the vascular endothelium leading to fluid loss into the intestinal lumen and profuse watery diarrhoea. Clinical symptoms that reflect the innate inflammatory response include elevated white cell count and fever, whilst profuse diarrhoea can lead to abdominal cramps, dehydration, hypoalbuminaemia as a result of protein losing enteropathy, and in severe cases, toxic megacolon [Cloud et al. 2009; Sunenshine and McDonald, 2006].

Damage to the colonic mucosa and severe disease

Localized damage to the colonic epithelial cell surface leads to the formation of pseudomembranes, consisting of necrotic cell debris, fibrin, mucous, neutrophils and other polymorphonuclear cells that have infiltrated from the peripheral circulation [Kelly et al. 1994a; Pothoulakis, 1996]. In severe cases, ulceration of the colonic submucosa can lead to perforation, septicaemia and death. The shift from an inflammatory response at a local mucosal level to a general systemic reaction leads to symptoms of systemic disease, including rise in serum creatinine indicating renal failure [Cunney et al. 1998], acute respiratory distress [Jacob et al. 2004] and multiple organ dysfunction [Dobson et al. 2003].

Protective role of regulatory cytokines

In addition to perpetuating the inflammatory response and causing high white cell count, fever, diarrhoea and colonic mucosal injury, dendritic cells of the innate immune system also respond to C. difficile antigens including SLPs and toxins by promoting the release of regulatory and anti-inflammatory cytokines such as IL-10, IL-23 and IL-4 that initiate cellular repair processes, diminution of inflammation and activate regulatory T cells (Tregs) and B cells to promote the protective adaptive antibody response [Steele et al. 2011; Ryan et al. 2011; Bianco et al. 2011].

Role of the adaptive immune response in disease progression and outcome

The importance of the adaptive immune response in influencing the outcome of colonization has been appreciated for many years [Aronsson et al. 1985; Johnson et al. 1992; Warny et al. 1994], but the individual relationships between immunoglobulin (Ig) classes and bacterial virulence factors in CDI have only recently been defined.

An initial challenge with C. difficile promotes the release of specific cytokines including IL-10 and IL-4 that stimulate the maturation of naive B cells into Ig-producing plasma cells and specific memory B cells [Ryan et al. 2011]. This initial immune challenge is thought to occur in infancy, as approximately 50–70% of healthy adults have detectable serum IgG and IgA antibodies to C. difficile toxins [Kelly et al. 1992; Sanchez-Hurtado et al. 2008; Viscidi et al. 1983; Warny et al. 1994].

Immunoglobulin A

Certain immunoglobulin classes are more important than others in mediating immunity to C. difficile. IgA plays a critical role in mucosal immunity and would be expected to provide a protective role against C. difficile antigens in the intestinal lumen. Low levels of faecal IgA and reduction in colonic IgA-producing cells associated with the gut mucosa have been shown to be associated with prolonged CDI symptoms and recurrence of infection [Johal et al. 2004a; Warny et al. 1994].

Immunoglobulin M

IgM is the earliest serum antibody to respond to infection and it has been shown that patients who had high concentrations of serum IgM antibody against TcdA, TcdB and non-toxin antigens at day 3 post diarrhoea onset were significantly more likely to have a single episode of infection [Kyne et al. 2001]. Patients with a low IgM response against SLP at day 3 post diarrhoea onset have also been shown to exhibit a 25-fold increased risk of recurrence compared with patients with a high anti-SLP IgM response [Drudy et al. 2004].

Immunoglobulin G

A high IgG response to C. difficile toxins and surface proteins at the time of colonization is thought to protect against development of CDI [Pechine et al. 2005a; Kyne et al. 2000; Mulligan et al. 1993]. Once a patient has developed symptomatic CDI, serum IgG responses to TcdA, B and nontoxin antigens on day 12 post onset of diarrhoea have been shown to be higher in patients who experience a single episode of CDI compared with patients with recurrent CDI [Kyne et al. 2001]. However, a specific IgG response to SLP was not found to be protective against recurrence [Drudy et al. 2004]. C. difficile flagellar proteins FliC and FliD and surface-associated proteins Cwp66 and Cwp84 have also been shown to be immunogenic. Antibodies against these proteins were detected after CDI diagnosis and for the following 2 weeks [Pechine et al. 2005b]. Deficiency in certain subclasses of IgG (IgG2 and IgG3) has also been found to be related to recurrence of disease [Katchar et al. 2007].

A high natural antitoxin antibody response does not always protect from symptomatic CDI. Patients who are critically ill with high comorbidities and severe or systemic CDI are less likely to be protected than those who are less severely ill, despite similar antibody levels [Kyne et al. 2000]. This suggests that other host factors are also important in immune protection and should be considered when predicting outcome of colonization on host immune status alone.

Role of protective commensal microflora and immune homeostasis

The role of the intestinal microflora in protecting from CDI has long been thought to be due to competition for nutrients, physical occupation of mucosal cell surface adhesion sites and production of antimicrobial peptides [Guarner and Malagelada, 2003]. The precise mechanisms by which the intestinal commensal microflora is able to regulate host immunity during invasion by C. difficile are only recently being elucidated. Innate immunity signalling modules including the TLRs and adaptor molecules including MyD88 present on the surface of epithelial cells and dendritic cells are able to continually sample the luminal contents, detecting bacterial-derived molecules including lipopolysaccharide (LPS) and flagellin [Chieppa et al. 2006]. Continuous detection of commensal bacterial molecules is able to promote ‘intestinal homeostasis’ by inducing the accumulation of microbe-induced regulatory T cells (iTregs) and suppressing the accumulation of proinflammatory T helper 17 (Th17) T cells in the lamina propria, thereby maintaining the iTreg:Th17 balance (Figure 1(a)]. Release of anti-inflammatory cytokines transforming growth factor-β (TGF-β) and IL-10 further restrains inflammation [Curotto de Lafaille and Lafaille, 2009]. Commensals have also been shown to induce IgA secretion from plasma cells, protecting the mucosa from invasion by pathogens and reducing pro-inflammatory signals [MacPherson and Uhr, 2004]. During antibiotic treatment, the loss of commensal flora and their protective TLR ligand molecules (dysbiosis) downregulates release of IL-10 and TGFβ, thereby altering the iTreg:Th17 balance [Figure 1(b)]. The numbers of iTreg cells are reduced and Th17 and Th1 cells accumulate in the lamina propria, initiating a mild inflammatory response and changes in epithelial permeability [Vaishnava et al. 2008; Littman and Pamer, 2011]. This effect has been shown to be reversed by addition of single bacterial-derived molecules, including LPS and flagellin, as substitutes for whole commensal bacterial cells, thereby maintaining the protective TLR signals [Jarchum et al. 2011; Rakoff-Nahoum et al. 2004]. Subsequent infection with C. difficile after dysbiosis may therefore result in intense inflammation, as the colonic epithelium is already predisposed to inflammation due to disruption of the iTreg:Th17 balance (Figure 1(c)]. This coupled with the direct action of toxins on epithelial cells and lamina propria immune cells serves to perpetuate a pro-inflammatory response.

Figure 1.

The effect of antimicrobial therapy on the immune homeostasis of the gut, predisposing to Clostridium difficile infection. a.) Immune tolerance to commensal gut microflora is mediated by TLR receptors. These detect bacterial-derived molecules and promote accumulation of protective iTreg cells in the lamina propria and release of anti-inflammatory cytokines. b.) During antimicrobial therapy, dysbiosis results in loss of protective TLR signalling, accumulation of proinflammatory Th17 cells and renders the mucosa more permeable. c.) Subsequent infection with C. difficile results in additional toxin-mediated injury to the epithelium, causing necrosis, proinflammatory cytokine release and neutrophil migration from the peripheral circulation.

Mice infected with C. difficile in the absence of antimicrobial treatment have been shown to be more likely to become asymptomatically colonized unless they were deficient in the MyD88 TLR adaptor molecule responsible for disseminating the protective signals, upon which they were highly susceptible to infection and developed severe disease [Lawley et al. 2009; Ryan et al. 2011]. The mechanism by which antibiotic treatment predisposes to CDI may therefore involve the disruption of the iTreg:Th17 balance, leading to a mild inflammation, which results in an increase in the permeability of the epithelial barrier. In individuals who are not colonized with C. difficile, this may be a short-term self-limiting effect, but during C. difficile infection, the additional injury caused by the direct action of TcdA and TcdB on the epithelial cells enables both toxins to access the immune cells in the underlying lamina propria, triggering the release of potent proinflammatory cytokines, infiltration of neutrophils, inflammation and disease.

Treatment strategies for Clostridium difficile infection

In addition to the treatments recommended for CDI, there a number of treatment strategies under development (Table 1). However, few have passed the rigorous clinical trial process to become accepted treatment options, and as yet, there is still no effective prophylactic therapy to reduce the incidence of disease for those most at risk.

Table 1.

Therapeutic approaches for treating Clostridium difficile infection.

	Currently used	Under development
Antimicrobial agents	Metronidazole	Ramplanin
	Vancomycin
	Nitazoxanide
	Teicoplanin
	Fidaxomicin
Biotherapy	Saccharomyces boulardii	Non-toxigenic C. difficile
	Faecal microbiota  transplantation
Localized intraluminal therapy	Cholestyramine	Bovine immunoglobulin concentrate
		‘Mucomilk’
		Tolevamer
		SYNSORB
Parenteral passive immunotherapy	Intravenous immunoglobulin	Human monoclonal antibodies
Active immunotherapy/vaccination		Toxoid vaccines (toxins A and B)
		Recombinant toxin polypeptides
		Multivalent vaccine
Innate immunity-mediated agents		Anti-CD18 antibodies
		Anti-interleukin-8 antibodies

Antimicrobial therapy

Despite the implication for antimicrobial therapy in predisposing for disease, there is still a reliance on antimicrobial agents for treatment of CDI. Non-severe CDI can be treated with metronidazole for 10–14 days [ASHP, 1998], and for severe disease, vancomycin treatment for 10–14 days is indicated [Bauer et al. 2009]. Recurrence of infection can occur in 25% of cases [Kelly et al. 1994b] and is treated with the same antimicrobial regimen as for a first episode [Bauer et al. 2009]. Multiple recurrences can occur and the recommended treatment is pulsed or tapered vancomycin therapy to allow spore germination between doses to improve efficacy [McFarland et al. 2002]. Recently the use of fidaxomycin has been indicated in multiple recurrent CDI [de Lalla et al. 1992; Louie et al. 2011; Musher et al. 2009] and has been shown to reduce the incidence of recurrence through the preservation of the intestinal microflora during and after therapy [Louie et al. 2012].

There is therefore a strong reliance on antimicrobial therapeutics to treat initial episode, severe and recurrent CDI. The deleterious effect on the protective gut flora may explain why standard antimicrobial treatment can fail to resolve recurrent CDI in up to 23% of cases [Surawicz, 2004].

Faecal transplantation

A potential non-antimicrobial approach to management of CDI is the use of faecal microbiota transplantation (FMT), which involves infusing intestinal microorganisms (in a suspension of healthy donor stool) into the intestine of a patient with CDI to restore the protective microbiota. Studies assessing the efficacy of FMT have shown resolution in up to 93% of CDI cases [Gough et al. 2011; Postigo and Kim, 2012] and the microbiota has been shown to resemble that of the donor after infusion [Khoruts et al. 2010]. The use of FMT is becoming increasingly widespread; however, additional prospective studies are needed to determine the best route of administration, optimal preparation and volume of donor faeces and to determine any risks associated with treatment.

Probiotics

Disruption of the healthy gut microflora by antibiotics is well known as the risk factor for primary CDI and recurrent disease. However, it is impossible to avoid antimicrobial therapy entirely. Administration of live beneficial bacteria (Saccharomyces boulardii, Lactobacillus spp.) or probiotics simultaneously with antimicrobial therapy would be considered to mitigate the impact of antimicrobials on the gut flora, but this has proven controversial for treatment of recurrent CDI [McFarland, 2009; Johnson et al. 2012b] and is not a recommended treatment option [Cohen et al. 2010]. Probiotics may have greater efficacy in preventing primary CDI after initial colonization with C. difficile rather than recurrence. Probiotic bacteria administered during exposure to antibiotics may be able to provide consistent anti-inflammatory signals to the gut immune system, thereby perpetuating protection throughout the treatment course. Once an infection with C. difficile is established, however, the extensive tissue damage and proinflammatory response as a result of the effects of TcdA and TcdB would be expected to outweigh the protective effects of probiotics [Johnson et al. 2012b].

Localized intraluminal passive immunotherapy

The important role that TcdA and TcdB play in mediating the clinical pathology of CDI has led to the development of a number of therapeutics that specifically target these proteins in the intestinal lumen and neutralize their bioactivity.

Chemical agents and polymers that bind toxins have been developed, including cholestyramine, SYNSORB and tolevamer, although none of these have progressed beyond phase II clinical trials [Heerze et al. 1994; Kurtz et al. 2001; Baines et al. 2009; Louie et al. 2006, 2007], or shown toxin-neutralizing activity in an in vitro gut model [Baines et al. 2009].

Intraluminal toxin binding and neutralization by anti-toxin antibodies has been shown to protect hamsters from CDI after they were administered immunoglobulins harvested from milk and eggs raised in immunized cows and chickens respectively [Kink and Williams, 1998; Lyerly et al. 1991; Kelly et al. 1996]. Bovine immunoglobulin concentrate consisting of antitoxoid IgG has shown toxin-neutralizing activity in the human intestinal tract [Kelly et al. 1996; Warny et al. 1999], but has not shown successful prevention and treatment in trials due to a reliance on obtaining an effective neutralizing dose in the colon after transit through the gastrointestinal tract.

Oral administration of bovine anti-C. difficile IgA in whey protein purified from immunized cow’s milk (Mucomilk) (MucoVax, BV., The Netherlands) has shown efficacy with a sustained response in human primary CDI treatment and recurrence reduction [Mattila et al. 2008; van Dissel et al. 2005]. However, studies were not continued to include larger patient cohorts.

Parenteral passive immunotherapy

To circumvent the inactivation of antibodies during passage through the gastrointestinal tract, neutralizing antibodies have been developed for parenteral administration. Human monoclonal antibodies have been developed in human immunoglobulin gene transgenic mice immunized with both inactivated TcdA and TcdB [Babcock et al. 2006]. These have shown effective protection from mortality and recurrence in hamsters and have protected from recurrence in humans [Babcock et al. 2006; Leav et al. 2010; Lowy et al. 2010]. However, they have not shown efficacy in reducing severity of symptoms [Lowy et al. 2010].

The most widely studied and currently the only available passive immunotherapy for severe and recurrent CDI is pooled intravenous immunoglobulin (IVIG) that contains neutralizing levels of antitoxin IgG [Abougergi et al. 2010; O'Horo and Safdar, 2009; Wilcox, 2004]. Initially shown to resolve recurrent CDI after administration to immunoglobulin-deficient children [Leung et al. 1991], IVIG has been used to treat severe or refractory CDI with varying success. The lack of consensus over the optimal dose and regimen and insufficient supporting data from clinical trials, however, currently limits its attractiveness as a possible therapy [Abougergi et al. 2010; O'Horo and Safdar, 2009; Wilcox, 2004].

Active immunotherapy/vaccination

Studies concerned with the development of longer-term protection from C. difficile toxins via active immunotherapies and vaccines were first carried out 30 years ago [Kim et al. 1987; Libby et al. 1982]. However, many years later there is still is no widely available vaccine recommended for use in humans.

Early vaccine candidates comprised TcdA and TcdB toxoids that induced a high serum anti-TcdA and TcdB response and protected hamsters from diarrhoea and death [Giannasca et al. 1999].

Subsequent studies suggested that TcdA was the main mediator of intestinal injury, triggering the development of a series of therapeutics targeting the single antigen. A TcdA toxoid was developed that induced anti-TcdA IgG antibodies in mice [Ghose et al. 2007]. Recombinant TcdA binding domain polypeptides were used to immunize hamsters and mice, providing partial protection against CDI and inducing a systemic neutralizing immunity respectively [Lyerly et al. 1990; Ward et al. 1999]. Other recombinant TcdA binding domain vaccines have been expressed in Vibrio cholerae that showed efficacy in rabbits [Ryan et al. 1997] and in transgenic mice that showed efficacy in humans [Gardiner et al. 2009]. The recent identification of a role for both TcdA and TcdB in mediating CDI has limited the likelihood of a single valency toxoid that is effective in humans [Kuehne et al. 2010].

More recent studies have generated toxoid vaccines that comprise preparations of both toxins. A vaccine based on adjuvanted purified TcdA and TcdB toxoids was able to protect hamsters from C. difficile challenge and induced both anti-TcdA and anti-TcdB antibodies [Anosova et al. 2012]. Recombinant TcdA and TcdB proteins expressed in Clostridium sporogenes were chemically inactivated and combined to form a toxoid A and B combination that was able to induce robust, neutralizing anti-TcdA and anti-TcdB antibodies with long-term protective activity in hamsters and nonhuman primates [Johnson et al. 2012a].

Dual valency vaccines that elicit simultaneous protection against both TcdA and TcdB have also been developed by creation of fusion constructs or ‘chimeras’. A recombinant fusion protein comprising the receptor binding domains of both TcdA and TcdB has been shown to reduce disease severity and confer antibody protection in a hamster model of CDI [Tian et al. 2012]. Immunization with a chimeric construct (cTxAB) comprising an attenuated glucosyltransferase substrate binding domain from TcdB and the receptor binding domain from TcdA was shown to protect mice and hamsters from symptomatic CDI and induced a neutralizing antibody response to both TcdA and TcdB [Wang et al. 2012]. Importantly, this response was seen after challenge with an epidemic BI/NAP-1/027 strain in addition to the laboratory strain, VPI10463, indicating cross reactivity to other C. difficile strains.

Recent identification of the important role that binary toxin may play in CDI pathogenesis [Stubbs et al. 2000] has led to development of a multivalent vaccine comprising four components: TcdA and TcdB and binary toxin A and B (CdtA and CdtB). The toxins were detoxified by mutation and were able to induce antisera of neutralizing capacity in in vitro studies [Heinrichs et al. 2012].

Other nonprotein vaccine candidates include a potential polysaccharide vaccine, based upon a highly complex C. difficile cell surface polysaccharide (PSII) that has been shown to elicit a specific IgM response in swine, although this has yet to be proven protective against CDI [Bertolo et al. 2012]. DNA vaccines comprising either the gene fragments from the glucosyltransferase substrate binding domain from TcdB or the receptor binding domain from TcdA have also been shown to provide a degree of protection in a toxin challenge model in mice [Jin et al. 2012], and were able to increase the level of protection when administered in combination with anti-TcdA passive immunotherapy.

The only candidate vaccine to progress to phase II human clinical trials is a formalin-inactivated toxoid A and B vaccine that induced a high serum-antitoxin antibody response and showed success in treating patients with recurrent CDI [Aboudola et al. 2003; Kotloff et al. 2001; Sougioultzis et al. 2005]. Recent improvements in the purity of this toxoid vaccine and coadministration with aluminium hydroxide adjuvant resulted in robust seroconversion for anti-TcdA and anti-TcdB in a phase I randomized, placebo-controlled study on healthy volunteers and an effective dose has been defined for use in further clinical trials [Greenberg et al. 2012].

There are still considerable advances to be made in the field of C. difficile vaccine research. The optimum antigen combination is still under debate, although there is sufficient evidence for both toxins to be represented in any formulation. The requirement of holotoxin as an immune activator is still unclear. Variation in toxin proteins, that is, the number of toxinotypes represented by clinical isolates, suggests that cross reactivity and protection from infection with a variety of toxinotype strains may be reduced in vaccines that are dependent on recombinant toxin fragments.

Novel therapeutic approaches to Clostridium difficile infection

Protective bacterial peptides

Commensal flora have been shown to mediate their protective effects in the gut by triggering the accumulation of iTreg cells into the mucosa, promoting an anti-inflammatory protective response via IL-10 secretion and inducing IgA secretion [Curotto de Lafaille and Lafaille, 2009; MacPherson and Uhr, 2004]. The TLR ligands that sense the presence of commensal flora have been shown to mediate this protective effect by recognizing single microbial effector molecules [Mazmanian et al. 2008]. Mice administered with flagellin purified from Salmonella enterica serovar Typhimurium were protected from CDI-induced epithelial cell damage, colitis and had a lower density of C. difficile colonization [Jarchum et al. 2011]. Therapeutics comprising single protective microbial effector molecules may therefore be just as effective as probiotics in providing intestinal homeostasis. They may also be co-administered with antimicrobial agents that may otherwise affect the growth of whole live bacterial cells. The candidate microbial effector molecules that would provide the greatest protection against CDI remain to be determined, as different microbial antigens would favour different TLRs and induce iTreg differentiation towards other effector phenotypes. The use of C. difficile peptides such as SLP may trigger a C. difficile-specific protective immune response, but these have also been shown to induce proinflammatory cytokines [Ryan et al. 2011], therefore research is needed to identify other peptides that may provide better protection without the associated inflammation.

Blocking proinflammatory cytokines

Therapies that block proinflammatory cytokines are currently used for treating gastroenterological autoimmune diseases and inflammatory bowel disease (IBD). Infliximab comprises a monoclonal anti-TNF-α antibody that neutralizes TNF-α activity and inactivates TNF-α-producing cells [Van Den Brande et al. 2002]. Use of infliximab is, however, not recommended for treating CDI and cases of CDI are actually higher in patients with IBD [Rodemann et al. 2007].

A better outcome has been obtained with inhibition of IL-8 action and production with a neutralizing anti-IL-8 antibody. IL-8 has a pivotal role in the neutrophil-mediated acute inflammatory response seen in CDI and neutralizing its activity has been shown to prevent neutrophil-mediated tissue injury and infiltration in several animal disease models [Mukaida et al. 1998]. The lack of an exact IL-8 homologue in mice may however have hampered research into this approach in current animal models of CDI. Antibody blocking of the leukocyte recruitment molecule CD18 inhibited neutrophil migration, reduced fluid secretion and epithelial cell damage in a rabbit model of CDI, suggesting that perturbation of the proinflammatory cascade was sufficient to reduce CDI symptoms [Kelly et al. 1994a], although the direct effect of TcdA and TcdB on the epithelium remained.

Antagonists of the adenosine A_2A receptor have also been used to suppress the inflammatory response in animal models of CDI, by preventing neutrophil infiltration, inhibiting inflammatory cytokine production [Warren et al. 2012].

Inducing anti-inflammatory cytokines

Patients with IBD in a phase I trial responded well to treatment with probiotic bacteria that produce recombinant IL-10 at the mucosa surface. This counteracts the hyperactive immune response and promotes mucosal tolerance by programming induction of anti-inflammatory iTreg cells [Braat et al. 2006]. Administration of anti-inflammatory cytokines, such as IL-10 directly at the mucosa surface may also suppress the acute immune response in CDI or be an effective therapy to prevent CDI after colonization with C. difficile. This approach, however, requires further research.

Suppression of the inflammatory response may reduce inflammation-mediated damage to the host, but it has been postulated that this may serve to increase the bacterial burden in the colon [Madan and Petri, 2012]. Therapies that protect the host from inflammation may therefore require adjunctive therapy that promotes clearance of C. difficile from the gut, to reduce the risk of symptom recurrence once anti-inflammatory therapy has ceased.

Future considerations for design of novel Clostridium difficile infection therapies

Novel therapies and newly developed vaccines for CDI will need to withstand the future challenges posed by continuous changes in both CDI patient demographics and bacterial evolution. An increasing number of cases of CDI (20–41%) involve patients who do not conform to the traditional risk factors [Khanna and Pardi, 2010; Lessa et al. 2012]. The majority of these cases are community acquired and involve significantly younger patients with lower comorbidity scores and lower rates of antibiotic exposure [Khanna et al. 2011]. Vaccination strategies currently under consideration to prevent primary and recurrent CDI would target adults most at risk, that is, those with planned hospitalization, long-term care/nursing home residents and adults with comorbidities requiring frequent/prolonged antibiotic use [Foglia et al. 2012]. As the CDI patient demographics change, those individuals at risk may become harder to identify for vaccination purposes and therapies designed to provide short-term resolution of symptoms may be required for initial CDI in these cases, followed by vaccination to prevent recurrence.

Novel vaccines and biotherapeutic agents have, in the most part, been designed to prevent the proinflammatory action of TcdA and TcdB by promoting or augmenting a specific neutralizing immunoglobulin response. Although a degree of cross reactivity has been observed for particular vaccine candidates [Wang et al. 2012], future therapies may have to take into account the continual evolution of the C. difficile species and emergence of new genetically distinct clades [Cairns et al. 2012; Stabler et al. 2006, 2012; He et al. 2010]. Monitoring both the genetic diversity and epidemiology of circulating C. difficile strains may enable a prediction of newly emerging genetically distinct strain types that are able to cause severe disease or epidemics. The continual identification of new toxinotypes highlights the genetic evolution of the PaLoc region in particular and new emerging strains with significant changes in the protein structure of TcdA and TcdB may reduce the future effectiveness of vaccines for CDI currently in development.

Conclusions

The outcome of C. difficile colonization and severity of CDI symptoms is mediated at every stage by the host immune system. The innate immune response maintains intestinal immune homeostasis in the healthy gut by continually sampling the gut flora and signalling anti-inflammatory messages to the downstream effector cells. Antimicrobial treatment predisposes to CDI by removing the anti-inflammatory signals, triggering a dysbiosis effect that makes the mucosa more susceptible to CDI-induced inflammation and tissue injury. The adaptive immune response mediates long-term antibody protection either from initial infection or subsequent recurrence.

The most effective therapies against C. difficile would therefore augment a specific immunoglobulin response whilst suppressing the innate proinflammatory response that mediates mucosal damage.

Therapeutic approaches that mimic the modulating effect of healthy gut commensals may also reduce the incidence of CDI, by perpetuating anti-inflammatory signals during antimicrobial administration and maintaining host intestinal immune homeostasis.

Vaccine targets developed to protect the at-risk populations have shown initial efficacy, however further trials are needed to determine whether they have the widespread cross reactivity needed to protect patients from all possible strains of C. difficile. It also remains to be seen whether the majority of older patients will be able to mount an immune response of a sufficient magnitude to the vaccine, to provide adequate protection.

Until these conditions are met, antimicrobials (metronidazole, vancomycin and fidaxomicin, which is newly approved by the US Food and Drug Administration) remain the mainstay of CDI treatment despite their limitations.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest statement

The author declares no conflicts of interest in preparing this article.