Background
Several outbreaks caused by
Clostridium difficile have been reported worldwide over the last decade. The emergence of an epidemic strain of
C. difficile, NAP1/027, was associated with more severe disease, and a higher recurrence rate [
1,
2]. This rate ranges from 8-50% and the likelihood of recurrence increases with the number of CDI episodes [
3,
4], which makes the management of recurrence challenging [
5‐
7]. Disruption of the normal intestinal microbiota is associated with a higher risk of CDI and antibiotics are generally the triggering factor [
8‐
10]. Prolonged disruption of the intestinal microbiota by antibiotic treatment may also increase patients’ susceptibility to recurrent CDI. Recurrence can be caused by the persistence of
C. difficile in the intestinal tract, or re-infection by the same or a different strain [
5‐
7]. The same strain has been isolated in 50-90% of recurrence cases, which indicates that persistence of
C. difficile spores in the intestinal tract of the patient is possibly a prerequisite to this condition [
4,
11].
Spores of
C. difficile are highly resistant to harsh environments and household disinfectants and are likely responsible for efficient dissemination of
C. difficile in hospital settings [
12,
13]. In addition, they are resistant to all known antibiotics including metronidazole (MTZ) and vancomycin (VAN) [
14]. Some studies suggested that epidemic strains of
C. difficile sporulate more efficiently and to higher levels than non epidemic strains, which might explain why epidemic strains disseminate easily in hospitals [
12,
13], but this hypothesis is a matter of debate [
15,
16].
Early upon infection,
C. difficile is capable of forming spores, as suggested by the induction of sporulation-associated gene transcription as soon as 8 h post-infection [
17]. However, the factors that affect this process are not well known. Previous reports suggested that sub-inhibitory (sub-MIC) concentrations of certain antibiotics can trigger
C. difficile sporulation
in vitro[
18,
19]. More recently, sub-MIC concentrations of fidaxomicin were shown to inhibit spore formation and toxin production in
C. difficile[
20,
21]. This suggests that antibiotics can potentially influence the number of spores that are formed during CDI and as such, directly impact the treatment outcome and the risk of recurrence.
The objective of this study was to determine the impact of sub-MIC concentrations of 5 antibiotics on sporulation of 10 different isolates of C. difficile in vitro. For this, we used a microscopic method to count vegetative cells and spores present within colonies growing on agar, as well as a classical spore recovery assay after growth and sporulation in broth cultures.
Discussion and conclusions
The objective of our study was to evaluate the impact of various antibiotics on the
in vitro sporulation of a set of 10 relevant isolates of
C. difficile. Our results strongly suggest that sub-MIC concentrations of TIGE, TZP, and to some extent CIP, have inhibitory effects on sporulation of most isolates. On the other hand, VAN slightly stimulated or inhibited spore production in a few isolates only and MTZ had no significant effect, except with one isolate where we observed a slight reduction in sporulation. Most importantly, we obtained similar results in both agar and broth sporulation assays, and using a microscopic method as well as a conventional spore recovery assay after ethanol shock and growth on agar. Therefore, the inhibition of spore formation that we observed in our study was not the consequence of a growth defect or death of vegetative cells, or poor resistance to ethanol of incompletely maturated spores. Our data rather strongly suggest that spore formation is significantly impaired. A recent study reported that
C. difficile strains ATCC 43255 (
i.e. VPI 10463) and UK-14 sporulated to similar levels in the absence or presence of sub-MIC of MTZ or VAN [
20]. In that study, sporulation was performed over 10 days in Brucella broth and spores were counted on agar plates after heat shock. In addition, the antibiotics were added at the end of the logarithmic phase of growth and not at the time of inoculation. Under these conditions, the maximum spore counts were reached between ~24 and 96 h and remained relatively stable over the rest of the experiment [
20]. Two previous studies suggested that on the contrary, MTZ and VAN strongly promoted sporulation of
C. difficile when present at sub-MIC. Ochsner reported that sporulation of
C. difficile strains ATCC 43596 and MB903 increased from a basal level of 10% and 17% without antibiotic to 100% after 4 days of incubation in the presence of sub-MIC concentrations of MTZ or VAN [
18]. A strong increase in sporulation of strain RMA 18386 was also observed in the presence of sub-MIC of MTZ (from 0.8% to 53%) [
18]. In a similar study, Mathur et al. showed that sub-MIC concentrations of MTZ and VAN also stimulated sporulation of
C. difficile strains ATCC 43596 and PGI 1 to very high levels after 6 days of incubation (from 8.8-11% without antibiotic to 75-100% with antibiotic) [
19]. In the two later studies, sporulation was assessed over 4 to 6 days on Brucella agar and the spores were counted after ethanol shock and recovery on agar plates [
18,
19]. We also observed increased sporulation in the presence of sub-MIC concentrations of VAN, but only with a few isolates. In addition, no difference was observed with VAN or with MTZ with most isolates. This suggests that depending on the strain of
C. difficile under study, the impact of antibiotics on sporulation can vary and thus, a larger set of strains might be necessary to observe general trends rather than effects that might be strain-specific.
The presence of TIGE at sub-MIC concentrations strongly inhibited sporulation of 9/10
C. difficile isolates tested and moreover, the effect was concentration-dependent. A similar trend was also observed with TZP, although a concentration dependence was seen with only 1/3 isolates tested. Fidaxomicin was recently found to strongly inhibit the formation of spores of
C. difficile ATCC 43255 and UK-14 when present at sub-MIC, and a lower transcription of sporulation-specific mRNAs was observed in strains CD196 and UK-1 [
20]. The synthetic methionyl-tRNA synthetase inhibitor REP3123 was found to strongly inhibit sporulation of
C. difficile ATCC 43596, MB903 and RMA 18383 in a dose-dependent manner when present at sub-MIC concentrations [
18]. Similarly, the biaryl oxazolidinone molecule RBx11760 completely inhibited sporulation of
C. difficile strains ATCC 43596 and PGI 1 at 0.5x-MIC [
19]. Together, these studies also suggest that certain antibiotics have a marked inhibitory effect on sporulation of
C. difficile when present at sub-MIC concentrations. In our study, the effect of TIGE and TZP could be observed with several isolates, thus minimizing a possible strain-specific effect. The only isolate that did not respond like the others was ATCC 9689, but the number of spores formed by this strain was already very low, making it more difficult to see inhibition of sporulation by antibiotics. On the contrary, VAN, TZP and MTZ tended to increase spore formation by this isolate.
Babakhani reported that fidaxomicin affected spore formation possibly by blocking the accumulation of
spoIIR and
spoIIID mRNAs, which are expressed specifically during sporulation [
20]. This is thought to be due to the anti-RNA polymerase activity of fidaxomicin, but whether the inhibition applies to all bacterial genes or only to sporulation genes is unknown [
20]. We noted a decrease in sporulation in the presence of CIP, especially with ribotype 027 strains. For those particular isolates, higher CIP concentrations had to be incorporated into agar plates to reach 0.5x MIC, because of their intrinsic resistance to CIP. Thus, although growth of these strains was not affected, high CIP concentrations seemed to alter sporulation somehow. Whether CIP interferes with transcription of sporulation-specific genes like fidaxomicin remains to be determined, but since CIP interferes with DNA synthesis and replication, this would not, a priori, be logical. However, a number of studies have shown that transcription of virulence-associated genes in
C. difficile can be affected by antibiotics that are not inhibitors of transcription per se. For example, Gerber
et al. observed that 0.5x MIC of MTZ, VAN, and linezolid increased transcription of
tcdA and
tcdB toxin genes in 4 different strains of
C. difficile, including VPI 10463 [
27]. Toxin gene transcription and production were also increased in high-level CIP-resistant isolates of
C. difficile, and a dose-dependent response was observed [
28]. Likewise, sub-MIC concentrations of ampicillin and clindamycin strongly increased transcription of genes coding for the colonization factors Cwp84 and the surface layer protein SlpA in NAP1/027 isolates [
29]. In addition, ofloxacin and moxifloxacin increased transcription of
cwp84 and
slpA, but only in ofloxacin and moxifloxacin-resistant isolates [
29]. Altogether, current data from the literature suggest that sub-inhibitory concentrations of certain antibiotics can affect virulence-associated phenotypes and sporulation in resistant strains, in part via modulation of transcription.
It is noteworthy to mention that we observed an inhibition of sporulation with 3 different antibiotics that have different modes of action: TIGE inhibits protein synthesis, CIP interferes with DNA synthesis and replication, and TZP interferes with cell wall synthesis. A possible mechanism explaining the inhibition of sporulation is thus rather speculative at the moment, but the fact that REP3123 and RBx11760 are both protein synthesis inhibitors led Mathur to propose that the inhibitory effect on sporulation was probably due to a general inhibition of the spore coat protein synthesis [
18,
19]. This hypothesis may apply to TIGE as well, which is also an inhibitor of protein synthesis, but does not reconcile the results obtained with TZP and CIP. However, as mentioned above, transcription of several genes can be affected by antibiotics, including quinolones to which bacteria are already resistant [
29]. Of note, bacteria grew well in the presence of sub-MIC concentrations of all antibiotics tested in our study, suggesting that the inhibition of sporulation by TZP and high CIP concentrations was not directly linked to the growth capacity or cell viability, as reported previously with other antibiotics [
27,
29,
30]. Further studies will be necessary to elucidate the molecular mechanism by which CIP, TZP and TIGE inhibit sporulation.
The biological significance of the inhibition of sporulation by sub-inhibitory concentrations of antibiotics
in vitro remains to be determined
in vivo. Although fecal concentrations of TIGE are generally much higher than the MIC for
C. difficile (range of 3.0-14.1 μg/g feces) [
31], it is possible that sub-MIC concentrations could occur early at the beginning, or at the very end of the antibiotic treatment, during which period sporulation could be affected. A recent transcriptomic analysis in mice mono-colonized with
C. difficile revealed that transcription of sporulation-associated genes was upregulated as soon as 8 h post infection, therefore suggesting that spores are formed early
in vivo. It is thus reasonable to suggest that inhibiting sporulation early after infection could possibly reduce the risk of relapse due to persistence of spores in the gut. Recent studies also suggest that
C. difficile can form biofilms [
32,
33], and it was suggested that biofilms could possibly protect
C. difficile from antibiotics, creating an environment where sub-MIC concentrations of antibiotics could be present [
34]. In the case of fidaxomicin, high fecal concentrations, way above the MIC, were also reported and recent experimental data suggest that the better performance of fidaxomicin compared to vancomycin could possibly be due to inhibition of sporulation and toxin production, as determined in the presence sub-MIC concentrations of the antibiotic [
20,
21]. Case reports have shown that TIGE, alone or in combination with other antibiotics, was able to cure patients with recurrent CDI that were refractory to MTZ and/or VAN [
35,
39]. The exact reason why TIGE seems effective in treating recurrent CDI remains to be elucidated and warrants further investigations, but our study suggests that the inhibition of sporulation could be one possible explanation, like for fidaxomicin [
20].
Competing interests
L.V. has served on advisory boards for Oryx, Iroko, Abbott and Wyeth, and has received compensation to conduct clinical trials involving antibacterials from Genzyme, Wyeth, Pfizer, BioCryst, Trius, Cempra, Optimer and Arpida. All other authors have no competing interest to declare.
Authors’ contributions
Conceived and designed the experiments: LV, LCF. Performed the experiments: JRG. Analyzed the data: JRG, LV, LCF. Wrote the paper: LV, LCF. All authors read and approved the final manuscript.