TcdB or Not TcdB: A Tale Of Two Clostridium Difficile Toxins

Keywords:

• Clostridium difficile
• essential
• isogenic mutants
• large clostridial toxins
• TcdA
• TcdB
• virulence

“We consider that TcdA is an important, albeit nonessential virulence factor, but maintain that the evidence still suggests that TcdB plays a more significant role in the pathogenesis of disease…”

The large clostridial toxins (LCTs) are an important family of bacterial virulence factors that includes toxin A (TcdA) and toxin B (TcdB) from Clostridium difficile[1], lethal toxin (TcsL) and hemorrhagic toxin (TcsH) from Clostridium sordellii[2], α toxin (TcnA) from Clostridium novyi[3] and TpeL from Clostridium perfringens[4]. The LCTs are monoglycosyltransferases that inactivate Rho family GTPases, such as Rho, Rac, Ras, Ral and Cdc42 through the covalent transfer of a glucose or N-acetylglucosamine moiety. Inactivation of Rho GTPases results in disruption of the actin cytoskeleton, cell rounding and eventually apoptosis and death of the intoxicated cell [1,5].

These LCTs have considerable amino acid sequence similarity [4] and their toxicity and role in virulence is the subject of intensive research in many laboratories. For example, injection of any of the purified LCTs is lethal for mice, demonstrating the highly toxic nature of these proteins in vivo[4]. The importance of several LCTs in pathogenesis is also becoming increasing clear. For example, Dang et al. found that C. novyi was rendered avirulent when cured of the bacteriophage episome that encodes TcnA [6]. Similarly, we recently demonstrated the critical role of TcsL in the virulence of C. sordellii ATCC9714, a strain that does not produce the TcsH toxin. By constructing isogenic mutants in C. sordellii, we showed that independent tcsL mutants were avirulent in a mouse model of infection, unlike the wild-type C. sordellii strain, which caused the rapid death of all infected mice following intraperitoneal injection [7]. Whether TcsL is an essential virulence factor in strains of C. sordellii that also produce TcsH remains to be determined and current studies in our laboratory are addressing this question.

The situation in C. difficile is less clear [8–10]. For many years TcdA was thought to be the major virulence factor of C. difficile, primarily because of experimental evidence obtained using purified toxin preparations. These experiments showed that TcdA alone resulted in the onset of symptoms when administered intragastrically to hamsters, whereas purified TcdB caused no disease unless intestinal damage was already present, or the toxin was co-administered with sublethal concentrations of TcdA [11]. More recently, however, we constructed and characterized isogenic tcdA and tcdB mutants of C. difficile and demonstrated that TcdB, not TcdA, was essential for virulence using the hamster model of infection [9], a finding that is in agreement with the isolation of naturally occurring TcdA^- TcdB⁺ clinical isolates [12]. However, in 2010 a second study using isogenic tcdA and tcdB mutants was published that reported conflicting data, with both a tcdA mutant and a tcdB mutant apparently capable of causing disease in the hamster model [10].

“What is obvious from both studies, however, is that TcdA alone is not the major mediator of disease as was previously believed.”

The major difference between these two publications is the observation that tcdB mutants were unable to cause disease in the hamster model in our study [9] whilst Kuehne et al. found that tcdB mutants were still virulent [10]. Kuehne et al. suggest that this discrepancy may be explained by differences between the strains used in each study, and this is a highly plausible explanation. Although derivatives of strain 630 [13] were used by both groups, each study used a strain that had been separately passaged numerous times over a long period of time (>15 years). It is therefore likely that independent genetic differences will have arisen in these strains. In support of this hypothesis our laboratory recently found [Carter GP, Rood JI and Lyras D, Unpublished Data] that strain 630Δerm[14], used by Kuehne et al.[10], produces significantly more toxin (threefold) than the equivalent strain JIR8094 [15], which was used in our study [9]. We have previously suggested that TcdB might not be essential for virulence in strains producing higher levels of TcdA [16]; the results of Kuehne et al. support this proposition.

Perhaps of greater relevance to the differences in the results are the end point determinations used in the animal trials. To obtain the most objective dataset possible in our study, death of infected animals was used as a clear, defined end point, in strict accordance with relevant animal ethics guidelines. By contrast, Kuehne et al. for legitimate animal ethics reasons pertaining to studies carried out in the UK, implemented a scoring system whereby a range of subjective parameters was scored; once a predetermined cumulative value was reached the animals were culled and scored as ‘death‘. It is not clear, however, at what cumulative total the animals were culled and the relationship between this predefined end point and fulminant disease was not defined. It is therefore difficult to determine how closely the measured parameters correlate with severe disease that would result in death of the infected hamsters. Another group in the UK, which is similarly restricted in the use of death as an end point parameter, has reported a different system for monitoring C. difficile infection in hamsters. Through a process of rigorous empirical testing, this group has delineated a reliable and reproducible end point that accurately reflects the point at which hamsters succumb to nonrecoverable disease [17]. Since it is not clear if the predefined end point used in the Kuehne et al. study accurately reflects nonrecoverable or fulminant disease [10], it is possible that some or all of the animals infected with the tcdB mutant might have recovered if they had not been culled.

“Although the findings of the two studies appear to contradict each other, upon closer examination there are some very important similarities.”

Both studies suffer from an inability to complement their toxin gene mutants. Complementation is important to rule out the possibility that unknown secondary mutations might be influencing phenotypic measures. It is well accepted genetic practice that several independent mutants should be studied if complementation is not possible. For this reason, two independently derived tcdA and tcdB mutants were analysed in parallel in our experiments. Since each independent mutant displayed identical phenotypes both in vitro and in vivo, secondary mutations are unlikely to be responsible for the differences observed when comparing the mutant strains to the wild-type. Since independent mutants of each toxin gene were not used by Kuehne et al. it is not possible to rule out the possibility that unknown secondary mutations might have contributed to their results. Note that in our study the toxin gene mutants were constructed via single crossover recombination. Although it has been suggested that these mutants may be unstable [18], if this was the case then the tcdA⁺tcdB^- mutants would have reverted to virulence with resultant disease symptoms in the animals; clearly, this was not found to be the case.

“…both TcdA and TcdB should continue to be evaluated in laboratory diagnosis and for the development of new disease control measures.”

Although the findings of the two studies appear to contradict each other, upon closer examination there are some very important similarities. As pointed out in a recent commentary [8] it is clear that in both studies the tcdA mutants were as virulent as the wild-type strain, which provides good evidence that, in the presence of TcdB, TcdA is not necessary for disease. This finding is in stark contrast to previous studies performed with purified toxin preparations. In addition, Kuehne et al. did observe that their tcdB mutant had an attenuated virulence phenotype, with a mean time to culling of 4 days, compared with 1 and 1.3 days, respectively, for the wild-type and the tcdA mutant [10]. We consider that TcdA is an important, albeit nonessential, virulence factor, but maintain that the evidence still suggests that TcdB plays a more significant role in the pathogenesis of disease, particularly since clinical strains that only produce TcdB are routinely isolated from patients with severe C. difficile infections [12].

Finally, we conclude that further studies are required to confirm the relative roles of TcdA and TcdB in the pathogenesis of C. difficile-mediated disease. It would be of interest to the C. difficile research community for the two research groups to repeat the hamster experiments using the other laboratory‘s mutants, as suggested by Ballard [8], which would determine whether experimental differences can account for the observed discrepancies. Perhaps of greater importance, however, will be the study of isogenic toxin mutants of C. difficile strains with increased virulence capacity, such as the hypervirulent BI/NAP1/027 isolates, as we have previously suggested [16]. Until such experiments are conducted the exact role that each toxin plays in fulminant disease will remain open to debate. What is obvious from both studies, however, is that TcdA alone is not the major mediator of disease as was previously believed. As we suggested in our publication [9], and our colleagues agree [10], both TcdA and TcdB should continue to be evaluated in laboratory diagnosis and for the development of new disease control measures.

Acknowledgements

Our research was supported by grants from the Australian National Health and Medical Research Council.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

Our research was supported by grants from the Australian National Health and Medical Research Council