Multiplex Real-Time PCR Method for Simultaneous Identification and Toxigenic Type Characterization of Clostridium difficile From Stool Samples.

BACKGROUND: The aim of this study was to develop and validate a multiplex real-time PCR assay for simultaneous identification and toxigenic type characterization of Clostridium difficile.
METHODS: The multiplex real-time PCR assay targeted and simultaneously detected triose phosphate isomerase (tpi) and binary toxin (cdtA) genes, and toxin A (tcdA) and B (tcdB) genes in the first and sec tubes, respectively. The results of multiplex real-time PCR were compared to those of the BD GeneOhm Cdiff assay, targeting the tcdB gene alone. The toxigenic culture was used as the reference, where toxin genes were detected by multiplex real-time PCR.
RESULTS: A total of 351 stool samples from consecutive patients were included in the study. Fifty-five stool samples (15.6%) were determined to be positive for the presence of C. difficile by using multiplex real-time PCR. Of these, 48 (87.2%) were toxigenic (46 tcdA and tcdB-positive, two positive for only tcdB) and 11 (22.9%) were cdtA-positive. The sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of the multiplex real-time PCR compared with the toxigenic culture were 95.6%, 98.6%, 91.6%, and 99.3%, respectively. The analytical sensitivity of the multiplex real-time PCR assay was determined to be 10(3) colony forming unit (CFU)/g spiked stool sample and 0.0625 pg genomic DNA from culture. Analytical specificity determined by using 15 enteric and non-clostridial reference strains was 100%.
CONCLUSIONS: The multiplex real-time PCR assay accurately detected C. difficile isolates from diarrheal stool samples and characterized its toxin genes in a single PCR run.

INTRODUCTION

Clostridium difficile is a gram-positive, spore-forming, obligate anaerobe. This bacterium is the causative organism for C. difficile infection (CDI), which presents symptoms ranging from mild diarrhea to toxic megacolon, ileus, bowel perforation, pseudomembranous colitis, and even death [1]. The major determinants of virulence in C. difficile are toxin A and toxin B, two members of the large clostridial cytotoxin family. Toxins A and B are encoded by the genes tcdA and tcdB, respectively, which are located within a 19.6-kb region of the chromosome known as the pathogenicity locus, or PaLoc [2]. Although most pathogenic strains of C. difficile produce both toxins, toxin A-negative, toxin B-positive variant strains have also been reported worldwide [3, 4]. Some studies have also reported the presence of another toxin, a binary toxin produced by certain strains of C. difficile. This toxin is encoded by the genes cdtA (enzymatic component) and cdtB (binding component), which are located outside the PaLoc, at the CDT locus in the genome [5].

Several laboratory methods have been used for diagnosis of CDI, including enzyme immunoassays (EIAs) for toxins A and/or B, EIAs for glutamate dehydrogenase (GDH), cell cytotoxin neutralization assay (CCNA), toxigenic culture, immunochromogenic assay, and real-time PCR methods. Direct stool culture and CCNA are currently the reference standard methods for the diagnosis of CDI. However, both methods are laborious, time consuming, and must be performed by qualified, trained personnel. EIAs are used for CDI diagnosis in over 90% of the laboratories in the United States because of their relative speed, ease of operation, and low-cost, compared with the reference standard assay. However, EIAs are not sensitive enough [6]. Recently, commercial and in-house-developed real-time PCR methods have been developed for the detection of C. difficile toxin genes from stool samples. These methods detect the genes with high sensitivity, specificity, and low turnaround time. Commercially available products for C. difficile diagnosis include the BD GeneOhm Cdiff (BD Diagnostics; San Diego, CA, USA), Prodesse ProGastro Cd (Gen-Probe Inc.; San Diego, CA, USA), Xpert C. difficile (Cepheid; Sunnyvale, CA, USA), and illumigene C. difficile (Meridian Biosciences; Cincinnati, OH, USA) real-time PCR tests. All the tests target the tcdB gene, except for the illumigene C. difficile test, which targets the tcdA gene. The Xpert C. difficile/Epi (Cepheid) is additionally capable of detecting the binary toxin. Compared with the CCNA and/or toxigenic culture, the PCR assays were reported with sensitivities and specificities ranging from 77% to 100% and 93% to 100%, respectively. However, none of the commercially available real-time PCR tests are capable of simultaneously targeting genes for the identification of the organism, toxin A, toxin B, and the binary toxin [6, 7, 8, 9].

The purpose of this study was to develop and evaluate a multiplex real-time PCR assay for the simultaneous identification and toxigenic type characterization of C. difficile isolates. In addition, the assay characteristics were compared with those of the reference method (toxigenic culture) and a commercially available PCR assay (BD GeneOhm Cdiff assay, targeting tcdB gene).

METHODS

1. Patients and samples

The prospective study was conducted at a university-affiliated, 800+ bed hospital in the Texas Medical Center, Houston, TX. Three hundred seventy-five consecutive liquid or soft stool specimens submitted to the Clinical Microbiology Department for C. difficile testing between February 20 and March 9, 2012 were considered for this study. Among these, only the first sample obtained from each unique patient was used. A total of 351 consecutive stool samples were tested by toxigenic culture, multiplex real-time PCR, and the standard BD PCR. The BD GeneOhm Cdiff assay (the standard diagnostic test used by the clinical microbiology laboratory) and the multiplex real-time PCR assay were performed within 24 hr after collection. The specimens were also cultured for C. difficile within 24 hr after collection of the stool samples. An aliquot of each sample was also frozen at -20℃ for any required re-culturing. This study was approved by the St. Luke's Episcopal Hospital Institutional Review Board.

2. DNA extraction and BD GeneOhm Cdiff PCR assay

The BD GeneOhm Cdiff assay was performed according to the manufacturer's instructions. Briefly, the diarrheal stool samples were vortexed at high speed. The stool specimens were transferred into the sample buffer tube containing Tris-EDTA sample preparation buffer by using sterile dacron swabs, which were broken into the tubes. The sample buffer tube containing the broken swab was tightly closed and vortexed at high speed for 1 min. Forty microliters of un-inoculated sample buffer was transferred to the lysis tube with glass beads, to which 10 µL of sample buffer containing the stool sample suspension was added. The lysis tube was vortexed for 5 min at high speed in order to ensure settling of the contents at the bottom of the tube. The DNA samples were inactivated at 95℃ in a dry block for 5 min, and immediately placed on ice. These samples were used for the BD GeneOhm Cdiff PCR and the multiplex real-time PCR assays. Three microliters of each DNA sample and positive and negative controls were added to the respective, labeled SmartCycler tube, containing 25 µL of the reconstituted master mix. The reaction tubes were centrifuged for 10 sec, run by using the SmartCycler I-CORE module (Cepheid), and analyzed with the Cepheid SmartCycler software, by using the standard BD GeneOhm Cdiff assay amplification protocol [10].

3. Design of primers and probes for the multiplex real-time PCR assay

The oligo analysis and design program (Oligoware 3.0 GMMA; Ankara, Turkey) was used to design of the tpi (species-specific internal fragment of the triose phosphate housekeeping gene), tcdA, tcdB, and cdtA primers and probes (Table 1). The sequences were evaluated by using the Basic Local Alignment Search Tool (BLAST) in order to assess the specificity of the primers and probes towards the identification of target sequences. The primers and probes were synthesized by Applied Biosystems (Foster City, CA, USA).

Table 1

The multiplex real-time PCR primers and TaqMan probes for detection of Clostridium difficile, its specific toxins A and B, and binary toxin A genes

Main target	Target genes	Sequence	GenBank accession number
C. difficile	tpi	F: 5'-gaagctactaagggtacaaa-3'	FN668944
		R: 5'-ggtctattcctacttctaatgc-3'	FN668941
		Probe: 5'-VIC ataagaggtgaaacttctcctgtaaatgctcc TAMRA-3'	FN668375
			FN665654
			FN665653
Toxin A	tcdA	F: 5'-tgataacgtatagcttgacc-3'	DQ117264
		R: 5'-atggtttacctcagatagg-3'	DQ117255
		Probe: 5'-FAM tgaatactttgcacctgctaatacggatg TAMRA-3'	DQ117249
			DQ117248
			DQ117247
Toxin B	tcdB	F: 5'-gaaggattacctgtaattgc-3'	HM062511
		R: 5'-ctgccattatacctatcttagc-3	HM062510
		Probe: 5'-VIC ctctttgattgctgcacctaaacttacacc TAMRA-3'	HM062509
			HM062508
			HM062507
Binary toxin A	cdtA	F: 5'-tatattaaagcagaagcatctgt-3'	DQ102377
		R: 5'-ctggaccatttgatattaaataatt-3'	DQ102378
		Probe: 5'-FAM tcctccacgcatataatcatttacatcagc TAMRA-3'	DQ102375
			DQ102376
			AJ238325

Abbreviations: tpi, Clostridium difficile specific triose phosphate isomerase gene; tcdA, toxin A gene; tcdB, toxin B gene; cdtA, binary toxin gene.

4. Multiplex real-time PCR for C. difficile identification and toxin detection

The multiplex real-time PCR assay was designed for the targeting and detection of the tpi and cdtA, and tcdA and tcdB genes in the first and second tubes, respectively, in a single PCR run. One microliter of the DNA sample was added to each of the two tubes containing 24 µL of the reaction mixture (0.8 µM of each primer, 0.4 µM of each fluorophore probe, 6 mM MgCl2, 200 µM dNTPs, 1 IU Super-Hot Taq polymerase, and 1×PCR buffer; Applied Biosystems). The assay was run on the ViiA 7 real-time PCR System (Applied Biosystems). The cycle conditions set for the use of the hydrolysis probe were: denaturation of the pre-amplified templates at 95℃ for 15 min, followed by 45 cycles of denaturation at 95℃ for 15 sec, and annealing and extension at 60℃ for 60 sec.

5. Determination of sensitivity and specificity of the multiplex real-time PCR assay

The bacterial reference strains used in the analytical sensitivity and specificity tests are listed in Table 2. The analytical sensitivity of the assay was determined by using spiked stool specimens. A bacterial suspension, approximately 102 to 109 colony forming units (CFU)/mL, was obtained from an overnight culture of the C. difficile R20291 (BI/NAP1/027) toxigenic strain. Suspension aliquots (0.1 mL) were transferred into 0.9 mL C. difficile-negative stool samples. DNA was extracted from bacteria isolated from the stool samples at concentrations ranging from 101 to 108 CFU/g [10]. This extracted DNA was used to determine the analytical sensitivity of the multiplex real-time PCR assay. The analytical sensitivity of the multiplex real-time PCR was determined by using two-fold serial dilutions of the extracted genomic C. difficile R20291 (BI/NAP1/027) toxigenic strain (DNA concentrations ranging from 0.015625 pg to 4 pg). The assay was performed on the ViiA 7 real-time PCR System (Applied Biosystems) by using the same reactants and run conditions that were described above. The cycle of quantification (Cq) value for each dilution was recorded. The lowest concentration of DNA obtained from the spiked stool and pure C. difficile R20291 (BI/NAP1/027) samples was identified as the analytical sensitivity of the assay. The PCR efficiency for each assay was determined from the slopes of standard curves (E=2[-1/slope]). A panel of 14 non-clostridial isolates was used to determine the analytical specificity (Table 2). Pearson's chi-square and Fisher's exact tests were used to compare sensitivity, specificity, positive and negative predictive values of two tests (BD GeneOhm Cdiff assay vs. the multiplex real-time PCR assay).

Table 2

Bacterial strains used for preliminary validation of the real-time-tpi-tcdA-tcdB-cdtA PCR assay

Organisms	Multiplex PCR results*
	tpi	tcdA	tcdB	cdtA
Clostridium difficile R20291 (BI/NAP1/027), C. difficile CD196 (BI/NAP1/027), C. difficile M120 (BK/NAP7,8,9/078), C. difficile CD305 (023)	+	+	+	+
C. difficile ATCC 9689, C. difficile BI-9 (J/NAP2/001), C. difficile Liv024 (J/NAP2/001), C. difficile Liv022 (DH/NAP11/106), C. difficile TL178 (G/NAP6/002), C. difficile TL176 (Y/NAP4/014), C. difficile TL174 (015), C. difficile 630 (R/012)	+	+	+	-
C. difficile CF5 (CF/NAP9/017), C. difficile M68 (CF/NAP9/027)	+	-	+	-
Klebsiella pneumoniae ATCC 13883, Pseudomonas aeruginosa ATCC 27853, Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 29213, Campylobacter jejuni ATCC 33291, Enterococcus casseliflavus ATCC 700327, Streptococcus pyogenes ATCC 19615, Enterobacter aerogenes ATCC 13048, Acinetobacter anitratus ATCC 19606, Salmonella typhi ATCC 14028, Shigella flexneri ATCC 12022, Enterococcus faecalis ATCC 29212, Acinetobacter baumannii BAA 747, Streptococcus pneumoniae ATCC 46619	-	-	-	-

*tpi, Clostridium difficile specific triose phosphate isomerase gene; tcdA, toxin A gene; tcdB, toxin B gene; cdtA, binary toxin gene.

+, positive result; -, negative result.

6. Toxigenic culture and detection of toxin genes

The stool samples were directly cultured onto cycloserine-cefoxitin-fructose agar (CCFA) plates (Anaerobe Systems; Morgan Hill, CA, USA). The plates were incubated at 37℃ for up to 5 days in anaerobic conditions and examined daily for growth. C. difficile was identified on the basis of typical odor and colony morphology, and confirmed by a rapid latex agglutination test (C. difficile, Oxoid Ltd.; Hampshire, UK) [4]. Following identification of the isolates, a 0.5 McFarland suspension of the C. difficile isolate was prepared, and 1 µL of this suspension was used for the detection of tpi, cdtA, tcdA, and tcdB genes by multiplex real-time PCR. The assay was performed as described in the previous section. A toxigenic strain was defined as a C. difficile isolate expressing either tcdA or tcdB, or both genes.

RESULTS

The multiplex real-time PCR assay identified 48 out of the total 351 (13.6%) samples as being positive for toxigenic C. difficile (46 positive for both tcdA and tcdB, two positive for only tcdB), whereas the toxigenic culture identified 46 toxigenic C. difficile-positive (13.1%) samples (44 positive for tcdA and tcdB, two for tcdB only). The BD GeneOhm PCR assay, on the other hand, identified 52 out of 351 C. difficile-positive (14.8%) samples.

Twelve samples (3.4%) were defined as non-toxigenic C. difficile (only positive for tpi expression) by toxigenic culture, while the multiplex real-time PCR assay classified 11 (3.1%) samples as non-toxigenic C. difficile.

A total of 11 C. difficile isolates were positive for cdtA gene expression, as well as expression of tcdA and tcdB genes, by the multiplex real-time PCR assay and toxigenic culture.

The analytical sensitivity of the multiplex assay was found to be the same for all targeted genes, at 103 CFU/g stool (from stool samples) and 0.0625 pg genomic DNA (from culture genomic extraction) (Fig. 1). The multiplex real-time PCR assay yielded negative results with the 14 non-clostridial isolates, demonstrating the absence of cross reactivity (Table 2). The sensitivity, specificity, negative predictive value, and positive predictive value of the multiplex real-time PCR and the BD GeneOhm PCR compared with the toxigenic culture are listed in Table 3. The reaction efficiencies of real-time tpi, ctdA, tcdA, and tcdB PCR were found to be 97%, 90%, 105%, and 104%, respectively, with correlation coefficients of over 0.99.

Fig. 1

Amplification curves and dilution end-point standard curves of log10 genome equivalents versus threshold cycle number. The analytical sensitivity of this assay for tpi (A-B), tcdA (C-D), tcdB (E-F), and cdtA (G-H) was approximately 103 CFU/g.

Abbreviations: CFU, colony forming unit; ΔRn, normalized reporter.

Table 3

The BD GeneOhm Cdiff assay and the multiplex real-time PCR assay results compared to toxigenic culture as the reference method

Assay	Toxigenic culture		Assay performance (95% confidence interval)
	Negative	Positive	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
BD GeneOhm Cdiff
Negative	298	1	97.8 (87.0-99.8)	97.7 (95.1-98.9)	86.5 (73.5-93.9)	99.6 (97.8-99.9)
Positive	7	45
Total	305	46
Real-time PCR
Negative	301	2	95.6 (83.9-99.2)	98.6 (96.4-99.5)	91.6 (79.1-97.2)	99.3 (97.3-99.8)
Positive	4	44
Total	305	46
P value*			0.361	> 0.999	> 0.999	0.413

*P value for the comparison of sensitivity, specificity, PPV, and NPV of BD GeneOhm Cdiff and Real-time PCR assays.

Abbreviations: PPV, positive predictive value; NPV, negative predictive value.

DISCUSSION

The gold-standard methods for diagnosis of CDI include the toxigenic culture and CCNA methods [11]. Both assays are labor-intensive and time-consuming. EIAs that detect toxins A and B are used worldwide, but these assays lack sensitivity. Commercially available real-time PCR assays, with improved sensitivity and faster diagnosis time, have been recently developed. These assays generally target either toxin A or B, and are therefore unable to detect variant strains that produce toxin B only, with or without the binary toxin [12]. In this study, a multiplex real-time PCR assay was developed and validated for the simultaneous detection of genes coding for toxin A, toxin B, binary toxin, and a housekeeping gene for the C. difficile organism, from stool samples. The primary advantage of this assay was the simultaneous detection and identification of the tpi, cdtA, tcdA, and tcdB genes within two reaction tubes in one PCR run. Performance of the assay was comparable to that of a reference method (toxigenic culture) and a commercially available real-time PCR diagnostic test.

Using the toxigenic culture as reference method, the performance characteristics of the multiplex real-time PCR assay were observed to agree with those seen in previous studies using real-time PCR assays. In this study, four samples determined to be negative in the toxigenic culture were observed to be positive with the real-time PCR assay. Potential reasons for these false-positive PCR results could be a low number of microorganisms in a very heterogeneous stool sample or growth inhibition due to concomitant anti-C. difficile treatment. Two samples determined to be negative with the multiplex real-time PCR assay were positive in the toxigenic culture. These samples were repeat-analyzed by both methods, with no change in results. This suggests that some stool samples contained a low bacterial concentration, below the analytical sensitivity of the multiplex real-time PCR. Another reason could be that the modification of the primer and/or hydrolysis probe-binding site sequence within the toxin genes could result in a negative PCR result. The analytical sensitivity of the PCR assay was found to be 103 CFU/g stool and 0.0625 pg genomic DNA, results that were 1,000-fold more sensitive than the conventional PCR assay described by Guilbault et al. [13], and comparable to other published real-time PCR assays targeting toxin genes of C. difficile in stool samples [14, 15, 16, 17].These results suggested that the concentration of C. difficile in patient stool samples may be too low. Future research should assess C. difficile isolated from patients with CDI at the time of diagnosis, in order to determine the limit of detection required for these emerging assays.

Variant pathogenic C. difficile strains containing toxin B but not toxin A have been described, although the clinical characteristics of these strains have not been well studied. This could be due to the lack of identification of these variant strains using assays that detect toxins A or B [18, 19]. The incidence of variant strains ranges from 3% to 92% worldwide [20, 21, 22, 23, 24, 25, 26] and from 0.2% to 1.3% in the United States [27, 28]. In this study, two (4.1%) out of 48 toxigenic isolates were discovered to belong to the tcdA-/tcdB+ variant strain by multiplex real-time PCR. Likewise, the binary toxin is associated with the epidemic BI/NAP1/O27 strain, and has recently been shown to be associated with increased mortality [29, 30, 31]. The incidence of binary toxin in clinical C. difficile isolates varies from 1.6% to 34.6% [5, 32, 33, 34, 35, 36, 37]. In this study, the binary toxin gene (cdtA) was found in 11 (22.9%) of the 48 toxigenic isolates. This assay provides investigators and clinicians with the ability to simultaneously evaluate these variant strains of C. difficile.

This study has certain limitations. The assay was not designed to detect deletions in the tcdC gene, a known marker of organisms displaying hypervirulence. Likewise, other known virulence genes in the PaLoc (tcdE and tcdR) were not targeted. In the future, we aim to target these genes in a confirmatory assay that could be performed on all isolates, after identification using the currently described assay. In addition, we did not confirm the binary toxin results with those of any other published assay.

In conclusion, we developed and validated a multiplex real-time PCR stool sample assay for the simultaneous identification and toxigenic type characterization of C. difficile isolates. Assay performance characteristics were similar to those for toxigenic culture and a commercially available real-time PCR assay. This assay provides investigators and clinicians with the ability to evaluate variant strains of C. difficile simultaneously, including toxin B+A- strains with or without the binary toxin.