Low Prevalence of Clostridium septicum Fecal Carriage in an Adult Population

Abstract

Clostridium septicum is an uncommon cause of severe infection. Real-time PCR against the C. septicum-specific alpha-toxin gene (csa) was used to estimate the prevalence of this microbe in human stool from 161 asymptomatic community-dwelling adults and 192 hospitalized patients with diarrhea. All samples were negative, suggesting a low prevalence.

Vibrion septique, later known as Clostridium septicum, was one of the first pathogenic anaerobes to be identified [1]. It is a toxigenic, anaerobic, Gram positive, spore-forming bacillus that is an uncommon cause of human infections [2-4]. Severe infections can be accompanied by refractory toxic shock, and there is a remarkable association of such infections with the presence of malignancy, either solid tumors or hematologic [5-7]. The severe clinical syndrome that accompanies invasive C. septicum infections relates to its expression of a pore forming alpha toxin encoded by the csa gene, which appears to be carried by all strains of this bacterium [8-10]. When they occur, C. septicum infections result primarily from hematogenous spread, involving either primary bacteremia alone, or appearing as spontaneous soft tissue infections (or both) [2]. It is held that the gastrointestinal tract is a common point of access for C. septicum bacteremia, which may explain its close association with gastrointestinal tumors such as colon cancer [11-14]. Despite antibiotics and supportive care, disseminated C. septicum infections remain dangerous. Unfortunately, there are no effective preventive measures against this rare outcome, in part because there is not a clear understanding of who is at the highest risk for developing fulminant infection.

Because colonization or transient carriage of C. septicum within the gastrointestinal tract is believed to be an important predisposing factor to infection it is important to better understand risk factors for such carriage. In this light, we tested the sensitivity of a C. septicum-specific real-time PCR approach in screening the prevalence of this organism in human stool specimens. A successful detection tool would help generate data that might be helpful for identifying patients at highest risk for infection. Although, C. septicum colonization occurs in humans [15, 16], its frequency and stability is unclear.

Because the csa gene appears to be ubiquitous and unique among C. septicum strains, and there is a highly specific and sensitive real-time PCR assay for its detection [10],we applied this approach to screen for the presence of C. septicum in 192 hospitalized patients with diarrhea (who tested negative for Clostridium difficile) and 161 outpatients without symptoms of diarrhea.

This study was conducted as part of a larger study about C. difficile infection. Following Institutional Review Board approval, fecal specimens were collected from 161 outpatient adults without symptoms of gastroenteritis (age > 18 years) and 192 hospitalized patients with symptoms suggestive of C. difficile but tested negative for this pathogen in the University of Michigan Health System Clinical Microbiology Laboratory. The absence of C. difficile was verified by culture. Two reference C. septicum bacterial strains used in this study (CS0712 and CS120612) were de-identified, historical isolates collected from clinical specimens in the University of Michigan Health System Clinical Microbiology Laboratory and a third strain was Pasteur III (NCTC 547), obtained from the American Type Culture Collection (ATCC, Manassas, VA).

To establish a molecular approach for C. septicum detection, we determined the sensitivity of real-time PCR for detecting the csa gene of C. septicum using DNA prepared from reference strains. This gene is believed to be highly conserved in C. septicum and csa-negative strains have not been reported to our knowledge [17]. Each of three reference C. septicum strains was grown overnight, anaerobically in Reinforced Clostridial Media (RCM, Difco^TM; Beckton and Dickinson, Sparks, MD) broth and then subjected to DNA extraction using an Easy DNA kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The DNA samples (172 ng/μl) were serially 10-fold diluted from 10⁻¹ to 10⁻⁸ and csa real-time PCR was run on a LightCycler 96 (Roche, Mannheim, Germany), as described in [17]. The highly specific C. septicum alpha-toxin (csa) gene primers used were those published by Neumann, et al. [10]: Forward: 5’-TAGGATTTGGATGGTGCGGTGGAA-3’, Reverse: 5’-TGCACGATACCCACTTGCATAAGG-3’. Using these primers, C. septicum DNA was detected at a dilution as low as a 10⁻⁷ (Fig. 1), corresponding to a concentration of 17.2 fg/μl. Cycle threshold (Ct) values close to 36 or higher, as observed for reactions using 10⁻⁸ dilutions, were indistinguishable from values obtained from the negative control, and likely corresponded to undetectable levels of C. septicum DNA. Remarkably similar values were obtained for all three reference strains.

Figure 1.

NCTC 547 had the lowest initial concentration among the three C. septicum strains, 172 ng/μl. The arrow marks the column corresponding to the Ct value for the lowest concentration detected (10⁻⁷ dilution) by the csa real-time PCR assay.

The dotted line at cycle number 36 marks the negative control Ct value.

We next addressed the sensitivity of this PCR approach to detect DNA from bacteria spiked into human stool, using a single reference strain of C. septicum. On two separate occasions, an overnight broth culture of strain CS120612 culture was used to prepare a series of 10-fold dilutions, from undiluted to 10⁻⁶ dilution. Each of these dilutions was then spiked into C. difficile- and C. septicum-negative human stool in bead beat tubes (MOBIO, Carlsbad, CA). These samples were then subjected to DNA isolation and real- time PCR. Exact colony counts (CFU/ml) were determined for each of these broth dilutions based on overnight growth on solid medium. Stool DNA was isolated using MagNa Pure LC (Roche, Mannheim, Germany), and DNA was serially diluted, 10⁻¹ to 10⁻⁶. The real-time PCR was conducted as described above. Each PCR reaction was conducted in duplicate. The Ct values from stool spiked with a 10⁻⁶ dilution of C. septicum were undefined because C. septicum was not present, or its concentration level was too low to be detected by the PCR assay. Based upon original concentrations of the CS120612 C. septicum culture used to spike the stool in experiment 1 of 6.32 × 10⁵ CFU/ml and in experiment 2 of 6.0 × 10⁶ CFU/ml, the csa gene real-time PCR detected C. septicum with a limit of detection of 63.2 CFU/ml and 60.0 CFU/ml, respectively (Fig. 2).

Figure 2.

Csa real-time PCR assay detected C. septicum in stool spiked with 10⁻⁴ and 10- 5 dilution in experiment 1 and 2, respectively – both marked with an arrow. The dotted line at cycle number 36 marks the negative control (H₂O) Ct value.

After establishing that csa real-time PCR was a sensitive method in detecting C. septicum we screened the stool obtained from 353 patients (192 inpatient samples, 161 community samples). The results for all the samples were negative. While it is possible that we failed to detect C. septicum due to either technical errors or levels below the limit of detection, we have used similar methods successfully to detect C. difficile in human stools (data not shown). In addition, we performed stool culture for each of these 353 samples on TCCFA (taurocholate cycloserine cefoxitin fructose agar) plates and failed to detect growth of either C. difficile or C. septicum, although the latter does not grow well on TCCFA. Because of poor sensitivity for TCCFA culture for C. septicum, we took additional steps to look for this bacterium. For example, for the six original stool samples from which the DNA yielded the lowest Ct values (39, 36, 36, 36, 38, 35.5 cycles), we re-streaked the original stool onto RCM plates (which support C. septicum growth) and no detectable growth of C. septicum was observed. Lastly, we failed to detect evidence of C. septicum carriage in a Basic Local Alignment Search Tool (BLAST) search of a published C. septicum 16S rRNA gene sequence [18] against a collection of partial 16S rRNA gene sequences generated from the stool samples of 338 fecal samples (155 asymptomatic, community samples, 89 diarrheal samples negative for C. difficile, and 94 diarrheal samples positive for C. difficile) [19]. Those 16S rRNA gene sequence data were used in a separate microbiome study but included 169 of the subjects investigated in the present report [19].

In summary, the point prevalence of C. septicum carriage in our geographic region is low. These data suggest that C. septicum is not a normal commensal in humans but an opportunistic pathogen whose capacity to cause disease likely reflects the coincidental occurrence of transient carriage and an enhanced host susceptibility (due to risk factors such as malignancy or recent chemotherapy). Future research efforts with larger patient populations followed prospectively are likely to be needed if to identify risk factors for transient carriage of this anaerobic pathogen. It also remains an open question whether there would be measurable benefit to screening for C. septicum carriage in patients who are most susceptible to invasive disease, and treating such patients with antibiotics directed at this potential pathogen.