- Top of page
- Materials and methods
- Supporting Information
The anthrax attacks of 2001 and the uncovering of recent bioterror plots highlight the importance of biodetection systems that can rapidly and accurately identify a wide range of potential biothreats in environmental samples (Schmitt and Shanker 2011; Kman and Bachmann 2012). An ideal system consistently detects the target organism(s) of interest at low levels without significant false-positive or false-negative results, interrogates a single sample for multiple targets (e.g. multiplexed detection), requires limited training and is cost-effective. One of the promising approaches to meet these needs is a detection system that identifies genetic signatures of biothreat pathogens using polymerase chain reaction (PCR). In this study, we evaluated a PCR-based detection system for the analysis of Bacillus anthracis (Ba), Yersinia pestis (Yp) and Francisella tularensis (Ft).
Conventional PCR-based systems have several distinct advantages for biothreat detection. PCR is sensitive (as compared to protein-based immunoassay methods), and it is also relatively rapid (as compared to direct culture methods) (Ulrich et al. 2006). Furthermore, well-designed PCR primers can selectively amplify target organisms vs genetically related near-neighbour organisms, reducing the likelihood of false-positive results (Hoffmaster et al. 2002).
However, the use of PCR-based systems is not without drawbacks. For example, PCR sample preparation is typically cumbersome, requiring a dedicated laboratory and trained personnel, particularly for environmental samples, where dirt, dust or other debris may prove challenging. Standard PCR systems are also limited by the number of targets that can be addressed in one reaction; addressing too many targets in a single reaction can decrease the sensitivity or specificity of the overall assay (Grondahl et al. 1999). Furthermore, the time required for PCR analysis of a sample (sample to answer) is on the order of 1 h and can be several hours depending on the sample preparation method. This is typically longer than lateral flow immunoassays which can take as little as 15 min, but significantly faster than culturing, which requires hours to days.
There has been a significant amount of research focused on the development and integration of automated sample preparation with PCR analysis for the detection of pathogens (Chandler et al. 2001; Hindson et al. 2005; Lee et al. 2006; Lui et al. 2009; Zhang et al. 2011; Foudeh et al. 2012). There are now hand-portable commercially available PCR systems that include assays for pathogen detection, such as the Bio-Seeq PLUS from Smiths Detection, the T-COR 4™ from Tetracore and RAZOR® EX and FilmArray® from BioFire Diagnostics Inc. (previously Idaho Technology Inc., Salt Lake City, UT, USA). However, the FilmArray® system is currently the only commercial hand-portable PCR-based detection system for pathogen detection that includes integrated sample preparation.
The FilmArray® system utilizes a ‘Lab-in-a-Pouch’ approach for the sample-to-answer detection of 17 biothreat pathogens in a single sample in just over 1 h. The system uses pouches containing all of the lyophilized reagents required for sample preparation, PCR and end-point detection. The biological sample, once in the pouch, is subjected to lysis, followed by DNA separation, purification and two-stage nested PCR (Poritz et al. 2011). The system can process liquid samples, so any solid samples such as powders collected on a swab must first be mixed with the supplied buffer solution. Briefly, the sample analysis process includes (Poritz et al. 2011): (i) 60-s mechanical disruption by vigorous mixing with ceramic bead along with Schizosaccharomyces pombe yeast cells that are freeze-dried within the pouch and serve as the internal control (ii) total nucleic acid isolation using silica-magnetic beads (iii) 3 washes and elution of the nucleic acids from the beads (iv) reverse transcription and first stage PCR (multiplexed) (v) sample dilution and splitting into 120 wells for second-stage (single-plex) PCR and (vi) amplicon melt analysis to measure PCR product in each well. The instrument is controlled by a laptop computer, and the integrated software analyses the data from multiple reaction wells (all reactions are run in triplicate) to determine the presence of a pathogen target.
BioFire Diagnostics has recently received FDA approval for use of the FilmArray® platform with the Respiratory Panel pouch that targets a panel of 15 respiratory pathogens (Poritz et al. 2011). The Respiratory Panel has been recently evaluated by several groups that found the panel to be both sensitive and specific, and to be significantly more sensitive than the Luminex xTAG Respiratory Viral Panel (Loeffelholz et al. 2011; Rand et al. 2011; Babady et al. 2012; Hayden et al. 2012; Pierce et al. 2012; Renaud et al. 2012). In addition, the FilmArray® Blood Culture Panel has been evaluated for the rapid and accurate identification of pathogens and antimicrobial resistance directly from blood culture (Blaschke et al. 2012). The goal of this study was to evaluate the FilmArray® platform and Biothreat Panel for the selective and specific identification of three potential biothreat agents: B. anthracis, Y. pestis and F. tularensis. The FilmArray® platform detects multiple pathogens in a single reaction, requires limited sample manipulation and training, and is a rapid sample-to-answer instrument. Our results from this study indicate that the FilmArray® may be a useful tool for biodetection applications where a sample must be interrogated for a wide range of potential biothreats.
- Top of page
- Materials and methods
- Supporting Information
Advances in PCR technology over the last few years have led to sensitive and rapid methods for detection; however, there are few fully automated systems for highly multiplexed detection of biothreat agents. Sample-to-answer systems incorporate sample preparation, which can simplify the assay process and reduce the overall analysis time. These systems require less manual sample manipulation, which reduces human error that can occur during sample transfer and pipetting. Simple sample-to-answer systems can also decrease the amount of training required for instrument use. Unlike other sample-to-answer PCR systems that have been developed, the FilmArray® is the first highly multiplexed sample-to-answer PCR biothreat detection system.
Overall, we observed that the FilmArray® consistently detected 250–25 000 GEs of Ba, Ft or Yp genomic DNA in a sample and that detection in 50% or more of the samples occurred at lower concentrations (25 and 12·5 GEs per sample). As we noted in the Materials and Methods section and Table 2, the Biothreat Panel pouch analyses less than half of the available sample (approximately 250 μl of the 550 μl diluted sample is drawn into the pouch). Therefore, it is conceivable that the sensitivity could be improved even further by optimizing the method for introducing the sample into the system.
We observed one of the Yp inclusivity strains, Y. pestis Java 9, was consistently negative for the YpT3 signature, even in the most concentrated samples. This strain, unlike the other Yp inclusivity strains tested, lacks the pMT1 plasmid (Tomaso et al. 2003). We therefore infer that the YpT3 signature is likely derived from the pMT1 plasmid. Regardless of the specific assay signatures, as only one Yp signature is required for a call of Yp, the instrument software still identified these samples appropriately as Yp in all cases.
As noted above, the Biothreat Panel is unable to currently distinguish between F. novicida and F. tularensis; however, this is not surprising, as there has been some considerable disagreement in the literature recently regarding the classification of F. novicida as a separate species and not as a subspecies of F. tularensis (Larsson et al. 2009; Busse et al. 2010; Huber et al. 2010). Further development of the Ft assays used in the Biothreat Panel pouch would be needed to distinguish these two near neighbours using the FilmArray® system.
It is also notable that we observed no pathogen false positives in the exclusivity genomic DNA samples, and we observed only two pathogen false positives in all 224 samples analysed (0·9% error rate). One of these false positives was easily determined to be due to a pouch failure of a blank sample, and the other was a Francisella tularensis sample that correctly called Francisella tularensis (true positive) but incorrectly called S. aureus (false positive) in the same sample. We also observed four samples with a positive Ba marker when testing Yersinia spp.; however, these did not result in Ba-positive calls, because all 3 Ba signatures are required for a call of Ba. These results demonstrate the value of having multiple signatures to increase the confidence of an analysis. Finally, it is also worth noting that of 226 total runs, only two runs did not report results, leaving us with a data set of 224 genomic samples. These two cases were due to software crashes (data not shown), and the analysis was repeated with a replicate sample.
Finally, although it is impossible to draw significant conclusions about the utility of the system for spore analysis from the limited number of spore sample replicates in our preliminary spore sample study, it is worth noting that the FilmArray® was able to detect spore samples at extremely low levels (25 CFU per sample). The intended use of this instrument is to provide a sample-to-answer result, which requires both efficient lysis and nucleic acid purification. Spore testing provides a more real-world test case for environmental sample biodetection scenarios, as these samples (e.g. white powder samples) are more likely to be in spore and/or vegetative cell form than purified DNA. The positive results for spore analysis demonstrate the utility of the system for a simulated sample-to-answer situation. Future testing with complex surface and powder samples will provide valuable information about the utility of this system for test case samples.
In conclusion, our testing found that the FilmArray® Biothreat Panel provided both sensitive and selective detection of B. anthracis, F. tularensis and Y. pestis genomic DNA. The assays were highly selective, even at very high concentrations of near-neighbour genomic DNA. Our initial evaluation suggests the FilmArray® Biothreat Panel offers a highly multiplexed detection system that meets many of the essential needs in environmental sample biodetection situations, including, but not limited to: (i) minimal hands-on manipulation of the sample (ii) integrated sample processing (iii) multiplexed detection and (iv) easy interpretation of the results. We are continuing our evaluation of the platform with additional live spore samples in the presence of environmental matrices and potential interferents.