ABSTRACT
Francisella tularensis is a potential bioterrorism agent that is highly infectious at very low doses. Diagnosis of tularemia by blood culture and nucleic acid-based diagnostic tests is insufficiently sensitive. Here, we demonstrate a highly sensitive F. tularensis assay that incorporates sample processing and detection into a single cartridge suitable for point-of-care detection. The assay limit of detection (LOD) and dynamic range were determined in a filter-based cartridge run on the GeneXpert system. F. tularensis DNA in buffer or CFU of F. tularensis was spiked into human or macaque blood. To simulate detection in human disease, the assay was tested on blood drawn from macaques infected with F. tularensis Schu S4 at daily intervals. Assay detection was compared to that with a conventional quantitative PCR (qPCR) assay and blood culture. The assay LOD was 0.1 genome equivalents (GE) per reaction and 10 CFU/ml F. tularensis in both human and macaque blood. In infected macaques, the assay detected F. tularensis on days 1 to 4 postinfection in 21%, 17%, 60%, and 83% of macaques, respectively, compared to conventional qPCR positivity rates of 0%, 0%, 30%, and 100% and CFU detection of blood culture at 0%, 0%, 0%, and 10% positive, respectively. Assay specificity was 100%. The new cartridge-based assay can rapidly detect F. tularensis in bloodstream infections directly in whole blood at the early stages of infection with a sensitivity that is superior to that of other methods. The simplicity of the automated testing procedures may make this test suitable for rapid point-of-care detection.
INTRODUCTION
Francisella tularensis, the causative agent of tularemia, is classified as a tier 1 select agent by the CDC (1) because of its high infectivity at low doses and its potential for being developed as a bioweapon (2–8). Tularemia can be easily misdiagnosed due to similarities in clinical presentations with other infectious diseases (9, 10). Francisella tularensis is fastidious and slow growing, which limits its recovery in bacterial culture (11, 12), even though blood culture remains the diagnostic gold standard (13). Nucleic acid amplification (NAA)-based assays offer an alternative method to detect tularemia, and they may be more rapid and sensitive than blood culture in some cases (3, 11, 14–21). However, current NAA assays do not appear to perform sufficiently well to consistently detect F. tularensis directly from uncultured whole blood. For example, the FilmArray platform is reported to have a 250-genome equivalent (GE) limit of detection (22), which is unlikely to be sufficiently sensitive for use as a direct blood test. Furthermore, most other existing F. tularensis detection technologies require manual DNA extraction methods that are labor-intensive and can further compromise assay sensitivity (23, 24). A more sensitive NAA assay that also included integrated sample processing and target detection would permit more rapid and potentially lifesaving diagnosis of this disease, while at the same time help to address biosafety concerns by decreasing the potential for laboratory-acquired infections.
We have developed a very sensitive and completely automated system that is suitable for point-of-care detection and can identify bacterial pathogens directly from patient blood using an integrated sample processing and PCR cartridge (25). However, neither this system, nor to our knowledge any other assay, has been evaluated for the ability to detect bacteria from blood samples taken before and just after the onset of symptoms in a human-like model, even though early bacterial detection before the development of severe illness is clearly one place where NAA-based detection of bloodstream infections would have the most clinical impact. The availability of serial blood samples from cynomolgus macaques (Macaca fascicularis) infected with F. tularensis in a study performed for other purposes provided us with a unique opportunity to test the performance of a sensitive NAA blood detection system in a close-to-real-world clinical situation. Here, we demonstrate the sensitivity of our assay system to detect F. tularensis in human blood samples and investigate its ability to rapidly detect F. tularensis directly from blood at the early stages of a nonhuman primate infection.
RESULTS
Analytical LOD.Our cartridge-based F. tularensis detection assay used a three-phase double-nested-PCR approach to achieve the highest possible level of sensitivity. The GeneXpert FB cartridge has the capability to perform multiple automated rounds of reagent mixing and PCR without any human manipulation of the sample. For the purposes of this study, we monitored the F. tularensis signal generated by the assay in both of the nested (PCR2 and PCR3) phases in order to assess the additional benefit in sensitivity supplied by PCR3. Testing serial dilutions of F. tularensis Schu S4 DNA (Battelle, OH) spiked in buffer showed the assay to have a limit of detection (LOD) of 1 genomic equivalent (GE) per ml of blood sample (Fig. 1A). When F. tularensis live vaccine strain (LVS; BEI Resources, VA) CFU were spiked into human (Fig. 1B) or macaque (Fig. 1C) blood, the LOD was 10 CFU/ml of blood.
Analytical sensitivity of PCR2 versus PCR3 for detecting F. tularensis. (A) Positivity rate for assays spiked with the indicated numbers of F. tularensis subsp. tularensis Schu S4 genomic DNA equivalents. Ten spiked samples were tested for each concentration shown. (B and C) Positivity rates for assays spiked with the indicated numbers of F. tularensis subsp. holarctica LVS CFU spiked into human (B) and macaque (C) blood. Between 12 and 87 spiked replicates were tested for each concentration of bacteria spiked into human blood, and 13 replicates were tested for each concentration of bacteria spiked into macaque blood.
Comparing the performance of single-nested- versus double-nested-cartridge-based PCR, we found that the analytic LOD of genomic DNA in buffer and CFU in human and macaque blood was the same whether the assay used signal generated from PCR2 or PCR3. However, the double-nested assay appeared to provide better sensitivity at sub-LOD target levels (Fig. 1). In testing DNA, PCR3 identified F. tularensis moderately better than PCR2 at concentrations of <1 GE/ml (P = 0.14). The value of PCR3 was more apparent in tests of CFU spiked into blood, where PCR3 identified 5 CFU/ml and 1 CFU/ml of F. tularensis spiked into either human or macaque blood better than PCR2 (P ≤ 0.05 for PCR2 versus PCR3 in human blood; P = 0.01 for PCR2 versus PCR3 in macaque blood). Thus, the three-phase double-nested-PCR appeared to offer an improved sensitivity in the more realistic setting of bacteria spiked into blood. We continued to analyze subsequent experiments solely based on the results of PCR3.
Dynamic range and differences between human and macaque blood samples.The dynamic range of the cartridge-based assay was tested to confirm that our double-nested approach did not affect detection at higher numbers of target. Ten-fold dilutions of F. tularensis cells were spiked into human and macaque blood from 105 CFU/ml through 1 CFU/ml. The assay detected F. tularensis CFU at all concentrations above the LOD and then at varied levels below the LOD in both human and macaque blood samples (Fig. 2). However, we did note that both cycle thresholds (Fig. 2C) and the endpoint fluorescence (Fig. 2D) at most CFU concentrations were significantly different (P < 0.05) when bacterial cells were spiked into human blood compared to when they were spiked into macaque blood (Fig. 2B). For example, the macaque blood containing 5 CFU/ml had an average cycle threshold (CT) value that was 6.6 cycles later than the same CFU spiked into human blood (10 ± 7.4 versus 3.4 ± 0.89, respectively; P = 0.08). At the same concentration, the endpoint fluorescence of the macaque sample was also lower than the human sample (506 ± 109.8 versus 179.5 ± 114 fluorescence units, respectively; P = 0.003). At higher concentrations of 10 to 104 CFU/ml, there was an average of 2 to 3 CT delay in macaque blood than human (P < 0.05, Fig. 2C) and significantly different endpoint fluorescence (EPF) values (P < 0.05, Fig. 2D).
Assay performance in dynamic range testing in cartridge-based PCR assay. (A and B) Real-time PCR curves for the cartridge-based assay over the range of F. tularensis LVS spiked into human (A) and macaque blood (B). (C and D) Cycle threshold (C) and endpoint florescence (D) value plots comparing the human and macaque blood at each spiked concentration. Three to six replicates of these spiked samples were tested at each concentration of human blood, and nine replicates were tested at each concentration of macaque blood. P value compares human and macaque blood. Av., average.
Inclusivity and exclusivity.We tested that ability of the cartridge-based assay to detect DNA from different F. tularensis organisms: Schu S4 (n = 2), LVS (n = 3), F. tularensis subsp. holarctica (n = 3), and F. tularensis subsp. novicida (n = 4), which contain 53, 63, 61, and 1 copy of the ISFtu1 assay target per genome, respectively (26). All tests were performed with 100 GE per reaction in replicates of 2 to 4. The assay detected F. tularensis in each sample tested. Assay exclusivity (specificity) was then investigated by testing 106 to 108 CFU/ml (or 106 GE when cultures were not available) of a broad panel of Gram-positive and Gram-negative bacteria (Table 1). The assay did not detect any of the nontarget pathogens, showing a specificity of 100% (Table 1).
Cartridge-based assay performance in a cynomolgus macaque infection model.The availability of blood from two cohorts of an F. tularensis macaque infection study allowed us to study the performance of our assay in a more real-world setting. We examined the sensitivity of our cartridge-based assay as well as the time to a first positive test relative to the initial F. tularensis infection, as well as the results of other F. tularensis detection methods. Blood samples from a total of 29 infected macaques and 3 uninfected controls over two cohorts were available to study. These same blood samples were also tested by blood culture and by a manual quantitative PCR (qPCR) for F. tularensis, as discussed in Materials and Methods. Our cartridge-based assay performed quite well compared to these other two identification methods, even though the blood samples had been stored from 1 to 18 months before testing. The cartridge-based assay detected F. tularensis in 4/19 (21%) macaques on the 1st day postinfection and 3/18 (16.6%) macaques on day 2 of postinfection, even though blood culture and qPCR were negative on all samples (Fig. 3). By day 3, the cartridge-based assay was able to detect F. tularensis infection in 60% of the samples, compared to 30% for qPCR (P < 0.05). Most macaques became symptomatic of their infection (fever, swollen lymph nodes, inflamed eyes, and/or pneumonia) between days 2 and 3 (Fig. 3). The fever was monitored for 7 days, and after day 3, 100% of the macaques were febrile. After 7 days, all the macaques succumbed to infection, or some were euthanized. All samples remained negative by culture until day 4, and culture never reached the positivity rate of either PCR detection method (Fig. 3). A very small number of cartridge-based tests were not positive at the later time points postinfection. The cartridge-based assay also appeared to be quite specific. No cartridge-based tests were positive by this assay in the day −45 and day −14 preinfection time points or in samples taken from the three uninfected control macaques at any time point. Seven percent of the samples tested as invalid due to a late internal control result, and one day 1 sample produced an error, although this sample tested negative upon repeat testing.
Detection of F. tularensis in the blood of infected macaques. Blood samples from macaques infected with F. tularensis Schu S4 were tested by the cartridge-based F. tularensis assay, a manual qPCR reference F. tularensis assay, and blood culture for the presence of this organism. Samples were obtained from macaques at the indicated days after infection. The number of macaques (as shown in parentheses) that tested positive as a percentage of the number of macaques tested on that day is shown for each test. Percent febrile includes the cumulative number of macaques showing fever on the plotted days. The adjoining table shows the numbers of positive macaques, with the number of macaques tested in parentheses.
DISCUSSION
Francisella tularensis is a potential bioterrorism threat due to its high virulence and low infective dose. Tularemia is the most commonly reported laboratory-acquired bacterial infection (1, 8, 27, 28). Francisella tularensis is difficult to culture, and serological methods fail to detect acute infection. NAA assays have provided promising alternatives for diagnosing F. tularensis infection (29, 30); however, these assays usually require technically demanding and time-consuming sample processing steps. Here, we demonstrate the first automated system suitable for point-of-care detection of F. tularensis directly from a clinical blood sample. This was accomplished by combining a very sensitive nested-PCR assay with a filter-based cartridge that performs hands-free sample processing and PCR.
The GeneXpert system is well suited for use in point-of-care diagnosis of bloodstream infections. The system integrates sample processing, PCR, and assay analysis in a completely automated system, providing rapid time to results. Furthermore, the programmable fluidics of the assay cartridge enabled us to perform double nested-PCR amplification, substantially improving assay sensitivity. The Xpert system has been used successfully to detect Mycobacterium tuberculosis and Staphylococcus aureus directly from clinical blood samples (25, 31, 32). The current study demonstrates that in the case of F. tularensis bloodstream infection, our approach can diagnose bacteremia several days earlier than conventional PCR or blood culture, with the first cases diagnosed at the time of first symptom development. To our knowledge, this is the only study so far to investigate the time course for PCR positivity compared to culture in a nonhuman primate sepsis model. Our results demonstrate that a sensitive PCR assay can perform better than blood culture by diagnosing more infections earlier in the course of disease. Combined with the more rapid time to results that is made possible by an automated PCR system such as the GeneXpert, PCR diagnosis of bacterial bloodstream infections may enable earlier and life-saving treatment.
Our assay performed very well in the macaque model. The specificity of the assay was 100% in tests of blood samples collected from unexposed macaques. The assay performed better than bacterial cell culture and qPCR for detecting F. tularensis at days 1 to 3 postinfection, despite the fact that all samples had been stored for at least 1 month before testing and had undergone at least one freeze-thaw cycle. A number of blood samples had large clots requiring them to be discarded without testing. However, we did test samples containing small clots. These clots almost certainly contributed to the assay error rate, and they may also have adversely affected performance in assays that produced results without errors. We expect that actual clinical use of our assay would likely show even greater sensitivity than seen in our current study, which had to rely on stored samples.
In conclusion, we have developed a detection system which can be used to diagnose F. tularensis bacteremia at the early stages of disease using a simple blood test and an automated detection instrument. Our test requires only one manual step after the blood draw, making it suitable for point-of-care detection. This system could be useful in the setting of a large-scale bioterrorism attack. Other point-of-care tests with similar design parameters could be easily developed to detect common causes of bacterial sepsis and other potential bioterrorism agents. This type of testing panel could provide a medical response to bioterrorism and also significantly improve the detection and treatment of more routine causes of bacterial sepsis.
MATERIALS AND METHODS
Ethics statement.The use of blood samples obtained from collections of nonhuman primates (NHP) was approved by the Institutional Animal Care and Use Committees (IACUC) at Battelle Memorial Institute, Columbus, OH, IACUC protocol BRC2741. The testing of NHP blood at Rutgers New Jersey Medical School was also reviewed by the Rutgers-New Jersey Medical School IACUC and deemed not to require a separate IACUC protocol review at this institution (letter of 21 February 2014 available upon request). Testing of discarded and deidentified human blood bank blood was approved by the Rutgers-Newark institutional review board (IRB) protocol 0120100406.
Preparation of bacterial cells and genomic DNA for analytical experiments.The F. tularensis subsp. holarctica live vaccine stain (LVS) NR-646 obtained from the Biodefense and Emerging Infections Research Resources Repository (BEI Resources, Manassas, VA) was used for all analytic studies. An initial inoculum was prepared by growing the bacteria at 37°C in Mueller-Hinton II (MH) broth (BD, Sparks, MD) supplemented with 10% dextrose (Sigma, St. Louis, MO), 2.5% ferric pyrophosphate (Sigma), 2% IsoVitaleX (BD), and 2.5% fetal bovine serum (Sigma-Aldrich, St. Louis, MO) for 16 to 18 h. Serial dilutions were made in MH broth and plated on chocolate agar (BD) to evaluate the CFU per milliliter, and concurrent dilutions were used in analytical experiments. The F. tularensis Schu S4 strain (obtained through BEI Resources, NIAID, NIH as Francisella tularensis subsp. tularensis Schu S4 Submaster Cell Bank, NR-10492) was used in all macaque infections.
Genomic DNA from the F. tularensis Schu S4 strain was a gift from Carl H Gelhaus (Battelle, OH), and genomic DNA of other F. tularensis subspecies (F. tularensis subsp. holarctica, F. tularensis subsp. novicida, and Schu S4) strains were obtained from BEI Resources, Manassas, VA (Table 1). Genomic equivalents (GE) were calculated based on the molecular weight of the base pair, length of the genome, and the amount of the DNA in our stock concentration. We used the online copy number calculator to determine our target copy numbers (33). Serial dilutions of F. tularensis Schu4 genomic DNA ranging from 0.1 to 1 copy/reaction were spiked to the 1st PCR mixture and tested for assay sensitivity using the cartridge-based assay, as described below.
Inclusivity and exclusivity tests for cartridge-based F. tularensis assay
Cartridge-based PCR assay.The multicopy (53 to 63 copies/genome) ISFtu1 gene (26) was selected as the target for our assay. Real-time nested-PCR was optimized in a GeneXpert filter-based (FB) cartridge (25) controlled by a GeneXpert instrument. A 3-phase double-nested-PCR approach was introduced to increase the sensitivity of detection (Fig. 4). The F. tularensis PCR consisted of three sets of primers designed to sequentially amplify an outer 138-bp amplicon (PCR1), a second inner 80-bp amplicon generated by two nested primers (PCR2), and finally, a third inner 78-bp amplicon, generated by one new primer paired with the reverse primer from PCR2 in a heminested amplification (PCR3) (Table 2). Bacillus globigii spores and an internal control (IC) assay that targeted a DNA sequence within B. globigii (Table 2) were also included in the FB cartridge. The IC assay served as both a positive control for sample processing and PCR amplification. Gene-specific primers were designed using the PrimerSelect (DNAStar Lasergene version 8.1.3) and/or Primer3 (34) programs, and molecular beacons were designed using the Mfold Web server (35). The PCR mixture (MM1) was made of 1× PCR buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl, and 0.001% [wt/vol] gelatin) with 2.5 mM MgCl2, 250 μM each nucleotide (deoxynucleoside triphosphate), 0.25 μM each primer (forward and reverse), and 4 U of JumpStart Taq DNA polymerase (Sigma). The primer and probe sequences are described in Table 2. All 3 phases of the PCR used the same master mix 1 except that the 2nd- and 3rd-phase PCR mixture MgCl2 that was changed to 4 mM, and molecular beacons ISFtu1-bcn and IC-bcn were added for real-time target detection. PCR was performed in a GeneXpert assay, in accordance with the fluidics described previously (25) and as shown in Fig. 4.
Three-step double-nested-PCR approach to detect F. tularensis. A first PCR (PCR1) amplifies an outer 138-bp amplicon, and this is followed by a second (PCR2) nested amplification to create an 80-bp amplicon. Finally, a third heminested-PCR (PCR3) creates a 78-bp amplicon. An F. tularensis-specific sequence is then detected using a molecular beacon. A similar amplification and detection scheme is used in the internal control reaction. OF, outer forward primer; OR, outer reverse primer; IF, inner forward primer; IR, inner reverse primer; P1, internal primer for third PCR.
Primer and beacon sequence used in this study
Blood sample processing in the GeneXpert system.Analytic studies of the cartridge-based assay using bacteria spiked into human blood made use of expired blood from the University Hospital (UH, Newark, NJ) blood bank (anticoagulated with citrate-phosphate-dextrose-adenine [CPDA]) that would normally have been discarded. Personal identifiers were removed from each unit before transport to the research laboratory. The blood was refrigerated and then used within a month of collection from the blood bank. Hematocrit values were adjusted to 40% by diluting the banked blood (hematocrit, normally 60% to 80%) in 0.01 M phosphate-buffered saline (PBS [pH 7.4]) to simulate normal adult blood hematocrit values. Whole blood collected in acid citrate dextrose (ACD) tubes from cynomolgus macaques (Macaca fascicularis) was purchased from Bioreclamation (catalog no. CYNEBACD; Westbury, NY). Both human and macaque blood were spiked with F. tularensis (LVS strain) at the desired CFU, and 1 ml of this preparation was added to an open filter-based (FB) GeneXpert cartridge, as described previously (25). Processing of the blood sample and PCR were then performed automatically in the cartridge. The fluidics of this automated procedure were similar to an earlier published protocol (25). Briefly, B. globigii cells (the targets for an internal control assay) were mixed with buffer and added to the blood sample. The sample was then mixed 1:1 with 8% NaOH, the lysed blood was passed through the internal cartridge filter, and the filter-captured bacteria were then extensively washed with a 50 mM Tris–0.1 mM EDTA (0.1 mM)–0.1% Tween (TET) buffer (pH 8). Glass beads present in the filter chamber were then agitated by an ultrasonic horn to lyse the captured bacterial cells. Finally, approximately 40% of the total eluted DNA was moved into the PCR tube that is integrated into the FB cartridge for PCR amplification and detection. A positive test occurred when the real-time cycles versus fluorescence units for the F. tularensis-specific molecular beacon reached a value above 20. A negative test required a positive internal control (IC) reaction. All negative tests with negative IC reactions were judged to be indeterminate.
Analytical dynamic range and limit of detection.The analytical performance of our cartridge-based assay was determined by testing serial dilutions of genomic DNA in buffer and bacterial cells spiked into human or macaque blood, as mentioned before. The dynamic range studies in human (n = 6) and macaque (n = 6) blood used samples spiked with 1, 10, 102, 103, 104, and 105 CFU/ml of LVS. For this study, the limit of detection (LOD) was defined as the lowest number of spiked DNA or CFU at which 100% of the samples tested were positive.
Analytical specificity.The specificity of the cartridge-based assay was determined by spiking 106 to 108 CFU/ml of 12 different bacteria/yeast, including 4 Gram-positive, 7 Gram-negative, and 1 Candida species (listed in Table 1) into human blood obtained from the blood bank and then testing the assay in a GeneXpert. These bacterial strains were obtained from BEI Resources or the Newark University Hospital Microbiology lab (University Hospital, Newark, NJ).
Macaque sepsis studies. Macaca fascicularis macaques were infected with F. tularensis Schu S4 (individual doses ranged from 884 to 2,182 CFU), using a primate head-only aerosol exposure system (36). An all-gas impinger (AGI) sampler collected a sample of aerosol for quantification of bacteria inhaled (this was performed by plating the sample on agar). The animal's tidal volume (volume inhaled with each breath) was determined by plethysmography. The time of exposure was also recorded. An aerodynamic particle sizer (APS) spectrometer (APS model 3321; TSI, Inc., St. Paul, MN) was used to measure the size and distribution of the aerosol particles, which needed to be in respirable range (up to 5 μm in diameter). Exposure times ranged from 10 to 30 min, depending on various factors, including minute volume, total accumulated tidal volume, and concentration of aerosol suspension.
Samples of whole blood ranging from 0.5 to 2.0 ml were collected into acid citrate dextrose-B (ACD-B; BD, Franklin Lakes, NJ) tubes prior to and following exposure to F. tularensis Schu S4 on days 1 to 7. Concurrent samples were collected in EDTA anticoagulant tubes (BD) for immediate testing by culture plating and qPCR. The ACD tubes containing whole-blood samples were stored at −70°C in the presence of glycerol (20% [vol/vol]) prior to shipment to New Jersey Medical School, Rutgers University. Blood samples obtained in this way were only tested with our F. tularensis assay if the blood collection tubes contained >0.5 ml and did not contain any visible clots.
Blood samples were initially available from a cohort of 10 macaques that had been infected via aerosol with F. tularensis Schu S4, and from three uninfected macaques to serve as healthy controls. Due to restrictions in specimen volume collection based on animal care and use committee guidelines, extended storage of specimens prior to analysis, and identification of samples containing blood clots, the F. tularensis assay in GeneXpert was conducted on the following time points after infection (including the number of macaques at each time point): day 3 (n = 5), day 4 (n = 6), day 5 (n = 9), day 6 (n = 10), and day 7 (n = 1). Blood samples from a second cohort of infected macaques drawn day 1 (n = 19) and day 2 (n = 18) postinfection later became available, along with seven samples obtained 14 days prior to infection. Blood from the second cohort was only stored for 2 weeks at 4°C, and then for 2 weeks at −80°C, before testing.
Fever was monitored for all days among both the exposed and control macaques using surgically implanted telemetry transmitters (TL11M3-D70-PCTP; Data Sciences International) prior to study activities. Each D70 PCTP transmitter contained two pressure leads and one biopotential lead. Core body temperature, electrocardiography (ECG), activity, and cardiopulmonary function (heart rate, systolic/diastolic pressure, pulse pressure, mean pressure, and respiratory rate) were monitored at least 30 s every 15 min during the pre and postexposure monitoring period. Use and maintenance of the validated telemetry system were performed according to validated Battelle standard operating procedures.
Blood culture.A 100-μl sample of the fresh whole blood was collected from the macaques at the indicated time points in EDTA tubes. The blood samples were then serially diluted in sterile phosphate-buffered saline (PBS [pH 7.4]) without added Mg2+ or Ca2+ and containing 0.01% gelatin. Each dilution was plated on the same day onto chocolate agar (catalog no. 221267; BD) in triplicate. A range of 3 or more dilutions were plated to ensure that the dilution containing 250 to 2,500 CFU/ml could be counted. The actual number of dilutions plated varied depending on the nature of the sample. The plates were incubated at 37 ± 2°C for ∼72 h. Colonies were counted, dilution factors applied, and the results reported as positive for culture. Samples were reported as negative for culture if no growth was seen in any plate after 72 h.
Nucleic acid isolates and qPCR assay.The EDTA whole blood collected from infected and uninfected macaques was used on the same day as the blood draw for total nucleic acid isolation and qPCR detection. The qPCR assay targeted the tul4 gene from Francisella tularensis using FDA/International Conference on Harmonisation (ICH) guidelines (37) for method validation. Total nucleic acid was isolated from whole-blood samples (100 μl) using the Specific B protocol on a NucliSENS easyMAG instrument (bioMérieux, l'Etoile, France), according to the manufacturer's methods, and total nucleic acid was eluted in 40 μl of NucliSENS easyMAG extraction buffer 3.
Each nucleic acid sample was assayed in duplicate by qPCR for detection of a portion of the tul4 gene using a 7900HT real-time PCR system (Applied Biosystems, Life Technologies Corp., Carlsbad, CA). Each 25-μl reaction mixture contained 5 μl of sample, with the remaining volume consisting of 2× TaqMan gene expression master mix (Applied Biosystems), sterile water, and a custom gene expression assay (Applied Biosystems) consisting of primers and probe specific for a portion of the tul4 gene (Table 2). A plasmid containing a cloned insert of the primer and probe-specific portion of the tul4 gene (Retrogen, Inc.) was run as a 10-fold serial dilution (5 × 107 to 5 copies/reaction) reference standard curve with all samples to determine the concentration of gene copies in each sample. The final concentration for each sample was reported as gene copies per milliliter of original blood sample. The validated assay limit of quantitation (LOQ) is 1.84 × 103 gene copies/ml of blood, and the limit of detection (LOD) is 9.36 × 102 gene copies/ml of blood.
Statistics.Standard statistical analysis (average, standard deviation, and t test) were performed using Microsoft Excel 2000 for Windows. Gompertz regression curve fits were performed using SigmaPlot (version 8.0). Fisher's exact test was run to derive P values between the sample groups.
ACKNOWLEDGMENTS
This research was supported in part by the National Institute of Allergy and Infectious Disease of the National Institutes of Health under awards AI098713 (D.A.), HHSN27220100003I (Battelle Memorial Institute, Columbus, OH), and HHSN266200400095I (Lovelace Respiratory Research Institute, Albuquerque, NM). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
We acknowledge the blood bank staff at University Hospital, Newark, NJ, for providing expired blood. We thank Jessica McCormick-Ell, Biosafety Officer at Rutgers Biomedical and Health Sciences, Newark, NJ.
D.A. is one of a group of investigators who invented molecular beacon technology and who receive income from licensees, including Cepheid, which licenses the molecular beacon technology in the Xpert MTB/RIF assay. To manage potential conflicts of interest, D.A. has irrevocably limited the fees that can accrue to him from the Xpert MTB/RIF assay to $5,000/year. M.J. is an employee of Cepheid.
FOOTNOTES
- Received 25 May 2016.
- Returned for modification 14 June 2016.
- Accepted 2 November 2016.
- Accepted manuscript posted online 9 November 2016.
- Copyright © 2016 American Society for Microbiology.
