Multiple cross displacement amplification-a more applicable technique in detecting Pseudomonas aeruginosa of ventilator-associated pneumonia (VAP)

Background Early and rapid identification of Pseudomonas aeruginosa (P. aeruginosa) in patients with suspected ventilator-associated pneumonia (VAP) provides theoretical clinical advantages in therapeutic optimization strategies. Methods The P. aeruginosa-multiple cross displacement amplification (PA-MCDA) assay was conducted at an isothermal temperature during the amplification stage, and products were visually detected by color changes. The entire process was completed within 1 h. A total of 77 strains, including P. aeruginosa species and various other species of non-P. aeruginosa, were used to evaluate PA-MCDA assays. Bronchoalveolar lavage fluid (BALF) of suspected VAP patients was examined by the MCDA assay. Results The MCDA assay exhibited a 100% analytical specificity in detecting PA from all 77 strains, and the limit of detection was as low as 100 fg DNA per reaction. A temperature of 65 °C was recommended as standard during the amplification stage. The agreement between PA-MCDA and bacteria culture was 91.18% (κ = 0.787; p = 0.000) in the identification of P. aeruginosa in BALF from suspected VAP. The PA-MCDA assay showed values of 92.31%, 90.78%, 77.41%, and 97.18% for sensitivity, specificity, positive predictive value, and negative predictive value, respectively. PA-MCDA had a higher detective rate of P. aeruginosa than bacteria culture in patients with antipseudomonal therapy. Conclusions The instrument-free platform of the MCDA assay makes it a simple, rapid, and applicable procedure for “on-site” diagnosis and point-of-care testing for the presence of P. aeruginosa without the need for specific bacterial culture.


Background
Ventilator-associated pneumonia (VAP) develops in intensive care unit (ICU) patients that have been mechanically ventilated for at least 48 h [1]. The infection rate was related to disease severity and the degree of organ failure [2]. Further, the EU-VAP study [3] identified that the overall incidence of VAP was 18.3 episodes per 1000 ventilator-days. Moreover, Pseudomonas aeruginosa (P. aeruginosa) and Staphylococcus aureus were the most frequently isolated pathogens in patients with VAP. In a multicenter study [4], P. aeruginosa, Acinetobacter baumannii, and Klebsiella pneumonia were found as the most frequent bacteria (97/212, 45.8%) in VAP patients. In addition, VAP induced a more prolonged need for mechanical ventilation, number of ICU stays and hospitalization, far worse outcomes, and increased associated healthcare costs [5,6].
Rapid completion of antibiotic administration might decrease the incidence of subsequent organ dysfunction and could be associated with a lower risk-adjusted inhospital mortality rate [7]. The conventional detection of P. aeruginosa in the clinical setting was generally achieved by growing the target pathogen on agar plate surfaces and cultures [8]. Although the culture-based technique is reliable, the time required for conducting it is at least for a period of 48 h [9]. Furthermore, published guidelines recommended an initial empiric combinatorial coverage with antibiotics targeted to Gramnegative and Gram-positive bacteria methicillin-resistant Staphylococcus aureus (MRSA) in the setting of highrisk VAP patients prior to obtaining culture results [1]. Inappropriate initial anti-microbial therapy and antibiotic exposure attributed to antibiotic resistance [10,11] were associated with increased in-hospital mortality rates [12]. Rapid and accurate identification of the suspected pathogens was thus warranted to provide the potential to maximize the administration of appropriate specific antibiotic and possibly avoid a need for empiric broad-spectrum antibiotic therapy. Rapidly excluding some specific and common pathogens would be possible to avoid unnecessary antibiotic exposure and minimize some undesirable consequences [13].
More recently, multiple cross displacement amplification (MCDA, Chinese IP Office Patent Application CN201510280765.X), auto-cycling, and strand displacement DNA synthesis were devised and validated as a possible replacement for PCR-based assays and applicability in the detection of specific nucleic-acid sequences [14,15]. The assay employed an isothermal temperature during amplification, and the products were visually detected by color changes. The entire process is completed in 1 h and benefits from being an instrument-free, simple, and practical procedure for "on-site" diagnosis and point-of-care testing. The current study is the first to report the application of the novel MCDA assay to rapidly detect the target pathogen, P. aeruginosa.

PA-MCDA assay primer design
Based on the mechanism of MCDA, a set of MCDA primers used for P. aeruginosa detection was designed that targeted the oprL gene, which encodes L-lipoprotein. The details of MCDA primers used in the report are shown in Fig. 1 and Table 1. The primers were commercially synthesized and purified by Tsingke (Beijing, China).

PA-MCDA reactions
MCDA reactions were performed in a one-step reaction in a 25-μl mixture containing 12.5 μl 2× the supplied buffer (Beijing Hai Tai Zheng Yuan Technology Co., Ltd.); 0.1 μl each of the displacement primers F1 and F2; 0.2 μl each of the amplification primers C1, C2, R1, R2, D1, and D2; 0.4 μl each of the cross primers CP1 and CP2; 1 μl (8 U) of Bst 2.0 DNA polymerase; 1 μl of the DNA template; and 0.8 μl of the colorimetric indicator. Moreover, negative control mixtures contained 10 ng of the Staphylococcus aureus and Klebsiella pneumoniae genomic templates, and blank control mixtures contained 1 μl of double-distilled water (DW). To evaluate the feasibility of the MCDA primer set that was designed Fig. 1 Schematic depiction of the primer sequences and positions for MCDA. The location and nucleotide sequence of the P. aeruginosa oprL gene that assisted in designing MCDA primers. The primer site sequences are underlined. Right and left arrows indicate sense and complementary sequences that were used in the assay to detect P. aeruginosa, we initially conducted the MCDA reactions at 63°C for 45 min and terminated the MCDA reaction by heating at 85°C for 5 min. Then, the optimal amplification temperature of the MCDA primer set was examined at fixed temperatures from 59 to 68°C at steps of 1°C intervals. In particular, MCDA products were detected using a colorimetric indicator and agarose gel electrophoresis.

Bacterial strains and genomic template preparation
A total of 124 bacterial strains and 14 fungi of positive culture were isolated in the clinical microorganism laboratory of the Third Hospital of Xiamen from 26 June to 26 July, 2017. The bacterial strains' list of standard culture is detailed in Additional file 1. The positive bacterial strains including 6/124 (4.84%) polymicrobial growth and 32/124 (22.81%) multidrug-resistant (MDR) strains were isolated from 118 clinical samples in which the tracheal aspirate and BALF were in 38 cases, the secretion and drainage in 33 cases, the blood in 16 cases, and urine, feces, catheter, and other samples in 31 cases. We chose top 13 bacteria strains and 77 samples to design the PA-MCDA reactions ( Table 2). The bacteria strains were identified by conventional cultivation method, automatic bacterial identification system (VITEK 2, Bio-Merieux, France) and stored in a 15% (w/ v) glycerol broth at − 70°C. After refreshing the culture three times on a nutrient agar plate at 37°C, the genomic templates were then extracted from all cultured strains using DNA extraction kits (Qiagen Co., Ltd., Beijing, China) and subsequently tested with an ultraviolet spectrophotometer and stored under − 20°C before use.

Specificity of the PA-MCDA assay
To evaluate the analytical specificity of the PA-MCDA assay, MCDA reactions were conducted under conditions that were described above with the 77 P. aeruginosa and non-P. aeruginosa pure genomic templates that were derived from all pure bacterial strains.

Sensitivity of the PA-MCDA assay
The genomic templates of P. aeruginosa were serially diluted (10 ng, 1 ng, 100 pg, 10 pg, 1 pg, 100 fg, 10 fg, and 1 fg per microliter) with the intent of verifying the limit of detection (LoD), and 1 μl of each serial dilution was then added to the MCDA reaction mixtures. The LoD of the MCDA assay was confirmed by the genomic DNA  Verification of the PA-MCDA assay This study was conducted in the 30-bed ICU of The Third Hospital of Xiamen in Fujian province, which is a 1200-bed hospital in China. A total of 102 patients enrolled in this study, who were with suspected VAP from 1 January 2018 to 14 March 2020. Patients satisfied two or more of the following criteria: fever > 38.5°C, leukocytosis > 10 9 /L or leukopenia < 4 × 10 8 /L, purulent tracheobronchial secretions, and a new or persistent infiltrate on chest radiography. The clinical data were recorded, i.e, demographic characteristics, indication(s) for ICU admission, prior antimicrobial therapy within 2 months before VAP, duration of mechanical ventilation before VAP, clinical pulmonary infection score (CPIS [16]) including temperature, blood leukocytes, tracheal secretions, oxygenation, pulmonary radiography, biochemical and hematological routine tests. The BALF were abstracted in one bottle, following which, one half (5 ml) processed for standard culture by the clinical microorganism laboratory of the Third Hospital of Xiamen and the other 5 ml stored at − 70°C until the time of DNA extraction. BALF was plated on chocolate, sheep blood, and MacConkey agar plates and incubated for 48-72 h according to routine clinical protocol. The DNA extraction from BALF method was described before. In this procedure, 1 μl of the extracted DNA template of the BALF specimen was added to the PA-MCDA assay and the reactions were performed at an optimal amplification temperature for 45 min. The products were detected by a color change and compared to the results of a standard clinical culture, which was blinded to the research investigators. This study was approved by the local ethics committee of the Third Hospital of Xiamen and performed according to the ethical standards of the latest revision of the Declaration of Helsinki. Written and informed consent was obtained from family members or the appropriate responsible parties.

Statistical analysis
Continuous variables of patients' characteristics were reported as the means ± standard deviations (SD) or the medians (interquartile ranges (IQR)), and categorical variables were reported as numbers (%). The accuracy of the PA-MCDA assay was compared with the microbiological culture in a cross-sectional analysis. P value < 0.05 was considered significant. Statistical analysis was performed using SPSS version 20.0 for Windows (SPSS Inc., Chicago, IL, USA).

Successful establishment of the PA-MCDA assay
To verify the feasibility of PA-MCDA primers, the MCDA reactions were initially carried out in the presence or absence of genomic DNA templates within 45 min at a constant temperature of 63°C [14]. A color shift of positive amplification in PA-MCDA tubes was directly observed to change from one of colorless to one of green, while the negative control tube remained colorless by the naked eye (Fig. 2a). The positive MCDA products were seen as ladder-like patterned bands on ethidium bromide-stained 2.5% agarose gels that were resolved by electrophoresis; however, these were not seen in the Staphylococcus aureus, Klebsiella pneumonia, or blank control (Fig. 2b). Hence, the designed MCDA primer was a good candidate to establish the MCDA methodology for detecting P. aeruginosa.
Optimizing the temperature for the PA-MCDA assay To confirm the optimal reaction temperature for the PA-MCDA assay, the P. aeruginosa strain was used as a Optimal reaction temperature for the PA-MCDA primer assay. MCDA reactions when detecting the P. aeruginosa gene were monitored by real-time measurement of turbidity. A turbidity of > 0.1 was considered positive. Nine kinetic graphs were obtained at various temperatures (59-67°C, at 1°C intervals) with P. aeruginosa DNA at a concentration of 1 ng per tube. The graphs showed that 65°C was an optimal temperature for amplification Fig. 4 Specificity for P. aeruginosa detection by MCDA assays. Of all 77 pure genomic templates, only the genomic DNAs from the P. aeruginosa strains generated positive results. The color shift in the PA-MCDA tubes (tubes 1-17) was directly observed as a green color. A gray color was seen in tubes 18-34, in which 18-21 were E. coli. Tubes 22-23 were S. aureus, of which tubes 24-25 were A. baumannii and tubes 26-34 were respectively S. epidermidis, S. capitis, Streptococcus pyogenes, S. agalactiae, E. faecalis, S. typhimurium, K. pneumoniae, E. cloacae, and S. maltophilia. A 2% agarose gel electrophoresis assay was applied to detect P. aeruginosa MCDA; lane 0, DL 100-bp DNA marker; lanes 1-17 were positive MCDA products that corresponded to P. aeruginosa tubes 1-17; and lanes 18-34 were negative and corresponded to tube 18-34 respectively positive control at a concentration of 1 ng per tube and the MCDA amplifications were monitored by a real-time turbidity technique. Performing the PA-MCDA assay at temperatures that ranged from 59 to 67°C at 1°C increments verified that 65°C was an optimal temperature for amplification-with a faster amplification procedure obtained from assay temperatures of 65°C (Fig. 3).

Specificity and sensitivity for P. aeruginosa detection by MCDA assays
When genomic templates were used in MCDA assays, only the genomic DNAs that were isolated from the P. aeruginosa strains (tubes 1 to 17) generated positive results. Genomic templates from all non-P. aeruginosa strains (tubes 18 to 34) did not provide production of detectable amplification products (Fig. 4). The color change was observed in positive MCDA tubes (tubes 1 to 17), and a ladder-like pattern was seen on an ethidium bromide-stained 2.5% agarose gel via electrophoresis resolution. These were not seen in negative tubes (Fig. 4).
Serial dilution of the P. aeruginosa genomic DNA (10 ng, 1 ng, 100 pg, 10 pg, 1 pg, 100 fg, 10 fg, and 1 fg per microliter) was used in MCDA assays. When the dilution exceeded 100 fg/uL, the green color by MG, the ladder pattern by agarose gel electrophoresis, and the positive reactions on real-time turbidity products were all directly observed ( Fig. 5a and b). It indicated that the LoD and the sensitivity of the PA-MCDA assay were 100 fg genomic templates per reaction.

Application of PA-MCDA to clinical samples
The MCDA assay was used to examine 102 BALF from patients who were suspected of presenting with VAP. Demographic and clinical characteristics of our patients are shown in Table 3. The median duration of mechanical ventilation was 7 days (IQR 4.00-32.50) before suspected VAP onset, and 97 patients had received antibiotics within 2 months before suspected VAP. The clinical diagnosis of VAP was given by physician judgment with respect to particular patients or special clinical situations combination with culture results. A total of 94 positive bacteria results were detected by conventional culture in 82/102 (80.39%) BALFs in which polymicrobial growth was 12/102 (11.76%) and polymicrobial included P. aeruginosa was 4/ 102 (3.92%). Bacterial strains list of BALF by standard culture is detailed in Additional file 2. After extracting DNA from these BALF specimens and adding 1 μl of the DNA template to the PA-MCDA assay, the reactions were carried out at 65°C for 45 min and the results were compared to those of the standard culture. The P. aeruginosa was detected in 26 samples by microbiological culture and in 31 samples by PA-MCDA. Two methods unanimously detected 24 P. aeruginosa strains while the MCDA detected 7 P. aeruginosa strains in culture-negative patients who received antipseudomonal therapy. The positive results of PA-MCDA and microbiological culture were 24/76 (31.58%) and 17/76 (22.37%) respectively in 76 patients who had received antipseudomonal therapy. In the crosssectional analysis, the agreement between the tests was 91.18% (κ = 0.787; p = 0.000), likelihood ratio positive was 10.02, and the likelihood ratio negative was 0.08. The PA-MCDA assay showed values of 92.31%, 90.78%, 77.41%, Fig. 5 Sensitivity of the MCDA assays using serially diluted P. aeruginosa genomic DNA. a P. aeruginosa genomic DNA was serially diluted to 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, 100 fg, 10 fg, and 1 fg per microliter. When the dilution was more than 100 fg/uL, the green color by MG and the laddering P. aeruginosa pattern by agarose gel electrophoresis were directly observed. The LoD of the PA MCDA assay was as low as 100 fg per microliter (white arrow). b Real-time turbidity was applied to analyze the amplification products. Genomic DNA levels > 100 fg per reaction produced positive reactions and 97.18% for sensitivity, specificity, positive predictive value, and negative predictive value, respectively.

Discussion
PA-MCDA reaction was performed with a set of 10 oligonucleotide primers, which specifically recognized 10 distinct sites on the target sequence, wherein the optimal temperature for amplification was 65°C. In particular, a colorimetric indicator (malachite green, MG) had been applied and the color changed from colorless to a light green color when the reaction was positive. The specificity of the PA-MCDA assay was 100%, and the sensitivity achieved a level as low as 100 fg of the template. The entire procedure, including that of specimen processing (15 min), the isothermal reaction, and result reporting (45 min), could be completed in approximately 1 h. The agreement between PA-MCDA and bacteria culture was 91.18% (κ = 0.787; p = 0.000) in the identification of P. aeruginosa in BALF from suspected VAP. PA-MCDA had a higher detective rate of P. aeruginosa than bacteria culture in patients who had received antipseudomonal therapy.
A rapid, simple, and accurate detection method of pathogenic microorganisms was necessary for the timely administration of appropriate therapy and arriving at a time to discontinue unnecessary antibiotic(s). Delayed receipt of an appropriate antibiotic(s) was independently associated with poorer clinical and economic outcomes in patients with serious Gram-negative bacterial infections, regardless of any resistance status [9]. Molecular diagnostic assays, such as PCR-based methods (e.g., conventional PCR, real-time PCR [17], and PCR-Electro Spray Ionization MS (PCR/ESI-MS) [18]), permit more rapid detection of targeted bacterium by nucleic acid amplification and have been established and applied in the clinic. However, these PCR-based techniques have some shortcomings, which include the following: (i) the instrument used is extremely expensive, (ii) the diagnostic specificity is highly affected by the amplification conditions and the primer design, and (iii) use of these techniques indicates that PCR results require gel electrophoretic analysis or real-time analytical apparatus. Matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF) [19] consistently decreased the time for successful identification of the pathogenic organism, shorten the period of time for administering an effective and optimal antibiotic, and thus arrive at improved patient outcomes. However, it also needs a bacterial culture and expensive instruments.
For diagnosis of VAP, not the standard culture but the MCDA assays are intended to supplant physician judgment respect to particular patients or special clinical situations. The PA-MCDA assay just represents a rapid, sensitive, and nearly instrument-free method of P. aeruginosa detection. Only a water bath or heat block was needed during the reaction stage, and the cost is about $8 per sample compared to standard culture which is $10 per sample. Due to MCDA being able to provide results within only 1 h, the clinician can save time to provide targeted therapy for patients, especially in the context of severe sepsis or septic shock patients. The positive test of PA-MCDA means there were P. aeruginosa in BALF; an initial narrow antibiotic therapy with antipseudomonal activity should be given in stable patients with suspected VAP in order to provide targeted therapy and reduce antibiotic exposure. The broadspectrum empiric therapy with antipseudomonal activity needed to be given to patients with severe acute respiratory distress syndrome and profound (unstable) septic shock when the PA-MCDA were positive because the high sensitivity for PA-MCDA assays and the polymicrobial growth always existed. The negative test indicated that there were no P. aeruginosa growth or the number of P. aeruginosa was very below the 10 4 cfu/ml in the BALFs as the high sensitivity of PA-MCDA assay [20]. The value was that empiric combination coverage for P. aeruginosa might not be necessary and the discontinuation of antibiotic therapy with antipseudomonal activity needed to be considered by clinicians [13]. Although the bacterium specific-MCDA assay cannot differentiate the infection from colonization for it cannot quantified the bacteria strains of BALF that is the methodology to diagnose VAP, the different time of color change of positive MCDA assay may relate to the amount of DNA templates of bacteria because its reaction products could be detected by real-time fluorescence and less time of positive reactions were produced in more specific-DNA templates [20]. Whether a threshold time of color change for MCDA reflects the quantitative or semiquantitative bacterium for clinical application needs a precise design and analysis. The PA-MCDA assay also detected 7 positive P. aeruginosa, while standard culture was negative in 76 patients who had received antipseudomonal therapy. The interpretation might suggest the MCDA approach had a higher sensitivity than standard culture [15,21], or that the culture negativity might reflect the presence of active culture inhibitors in the samples [22], or that the P. aeruginosa had become non-viable before or even between the standard culture periods since the growth conditions had changed. Clinical factors should also be taken into account because they might alter the decision of whether to withhold or continue antibiotics. The ultimate clinical determination of VAP and pathogens, and regarding antibiotics application, was made by the physician in the light of each patient's individual circumstances [13]. In order to use MCDA method to clinical application, this study also has limitations. The first is the bacterium-specific MCDA assay cannot differentiate the infection from colonization because it cannot provide the results of the quantitative bacterium of BALF. A more precise study will be designed recently to explore the relationship between a cutoff time for MCDA color change and the quantitative or semiquantitative bacterium in clinical samples. In addition, whether or not the assay could detect the target pathogen in other specimens such as blood, urine, or serous effusion needs further study. Consequently, both the negative and positive results of PA-MCDA cannot rule out the presence of other pathogens, Gram-positive or Gram-negative bacteria, or fungi, for polymicrobial growth always existed. MCDA cannot provide precise information to prescribe or withdrew pathogen-specific therapy [23] until more pathogen-specific MCDA and resistance-associated genes are designed [20]. We propose to assign some microorganism-specific MCDA assays and resistanceassociated genes (e.g., the nfxB gene and the blaPER-1 gene) [18,19,24,25] that are aligned to common pathogens in ICU, which would include Staphylococcus aureus [26], Acinetobacter baumannii, Escherichia coli, fungal species, and others in one template.