Development and evaluation of a real‐time polymerase chain reaction assay for the rapid detection of Talaromyces marneffei MP1 gene in human plasma

Summary Penicilliosis caused by Talaromyces marneffei is a common AIDS‐defining illness in South and Southeast Asia. Diagnosis is based on culture which can take up to 14 days for identification, leading to treatment delay and increased mortality. We developed a TaqMan real‐time PCR assay targeting the MP1 gene encoding an abundant cell wall protein specific to T. marneffei. The assay's performance was evaluated in MP1‐containing plasmids, clinical isolates, and plasma from HIV‐infected patients with and without penicilliosis. The assay consistently detected 10 copies of MP1‐containing plasmids per reaction and 100 T. marneffei yeast cells per millilitre plasma. There were no amplification with seven other Penicillium species and six other HIV‐associated fungal pathogens tested. The assay was evaluated in 70 patients with AIDS: 50 patients with culture‐confirmed penicilliosis and 20 patients with opportunistic infections other than penicilliosis. The diagnostic sensitivity was 70.4% (19/27, 95% CI: 51.5–84.1%) and 52.2% (12/23, 95% CI: 33.0–70.8%) in plasma samples collected prior to and within 48 h of antifungal therapy respectively. The diagnostic specificity was 100% (20/20, 95% CI: 83.9–100%). This assay provides a useful tool for the rapid diagnosis of T. marneffei infection and has the potential to improve the management of patients with penicilliosis.


Introduction
Talaromyces marneffei (previously named Penicillium marneffei) is an emerging fungal pathogen that causes a systemic mycosis in immunocompromised and, less commonly, immunocompetent residents and travellers to Southeast Asia, India and China. [1][2][3] The HIV epidemic has transformed T. marneffei from a rare infection to a leading AIDS-defining diagnosis in this region, 1,4-8 trailing only tuberculosis, cryptococcosis or Pneumocystis jiroveci pneumonia (PJP) in incidence. 1,7 Mortality is up to 20% despite antifungal therapy. 5,8 The diagnosis is often delayed as it can take up to 14 days to isolate and identify the pathogen from clinical specimens. Typical skin lesions develop in 50% to 70% of patients 1,5 which enable a presumptive diagnosis to be made based on Gram's or Giemsa staining of skin smear. Patients without skin lesions experience a significant delay in antifungal therapy initiation. 5 These patients have significantly higher mortality and frequently die before the diagnosis is made. 1,5 Development of non-culture-based assays for rapid detection of T. marneffei infection has the potential to improve treatment outcomes.
Several serological assays detecting antigen and antibody to T. marneffei have been developed [9][10][11] ; however, none has been evaluated in prospective clinical studies to diagnose active infections. Polymerase chain reaction (PCR) technique offers a sensitive and specific tool to detect several invasive fungal pathogens. 12,13 PCR-based assays have been developed to detect T. marneffei in clinical isolates 14,15 and in a small number of clinical specimens. 16,17 TaqMan and SYBR Green real-time PCR assays evaluated in 20 and 23 patients with penicilliosis reported diagnostic sensitivities of 60% and 77% respectively; however, diagnostic specificity was not assessed. 17,18 To date, the molecular targets of these PCR assays are based on the fungal ribosomal DNA, a highly conserved region shared by many groups of fungi, potentially compromising specificity. Here, we report the development and assessment of a TaqMan real-time PCR assay to detect T. marneffei MP1, a novel gene encoding T. marneffei cell wall manoprotein. 19 MP1 is a unique gene without homologues in sequence databases, thus cross-reaction to related fungal pathogens is minimised.

Statement of ethics
The study was approved by the Scientific and Ethical committee of the HTD. All patients gave informed consent for blood drawn for this diagnostic study.

Development of the real-time PCR assay
Oligonucleotide primers and hybridisation probe All 23 nucleotide sequences of T. marneffei MP1 (accession number DQ08822 to DQ08845) were retrieved from GenBank and were aligned using BIOEDIT software (v7.0.1, Ibis Biosciences, Carlsbad, CA, USA). The primers and TaqMan probes were designed to be highly specific to a conserved region of MP1 using PRIMER EXPRESS software (v2.0, Applied Biosystem, Foster City, CA, USA) ( Table 1). The specificity of the primers and probe was confirmed by performing BLAST searches on GenBank showing no cross-reacting sequences with other fungi.
Template DNA preparation Talaromyces marneffei yeast cells were sub-cultured on Sabouraud Dextrose Agar (SDA) at 37°C for 5-7 days until single cells were created. The cells were homogenised in 5 ml PBS (pH 7.2) containing Tween 20 (5%) (Sigma, Singapore). Cell pellets were collected by centrifugation at 8000 g for 5 min, and after thoroughly washed two times with PBS, cell pellets were resuspended in 5 ml PBS. Fungal DNA extraction was performed as described previously with minor modifications. 18 Briefly, fungal cell wall was removed to form spheroplasts by treating 200 ll of resuspended cultured fungal isolates or plasma specimens with 600 ll sorbitol buffer (1 mol l À1 sorbitol, 100 mmol l À1 EDTA and 14 mmol l À1 b mercaptoethanol) supplemented with 200 U lyticase enzyme (Sigma). The mixture was incubated at 30°C for 30 min. Spheroplast DNA was extracted using QIAamp DNA blood mini kit (QIAGEN, Hilden, Germany) and was eluted in 30 ll of elution buffer.
TaqMan real-time PCR assay Real-time PCR was performed using the Light-Cycler â 480 (Roche Applied Science, Penzberg, Germany). The reaction was conducted in 25 ll volume containing 2.5 ll buffer 2.59 (QIAGEN), 500 nM MgCl 2 (QIAGEN), 400 nM dNTP (Invitrogen), 400 nM each primer, 200 nM TaqMan probe, one unit Hotstart Table 1 Primers and probe for the real-time PCR assay.  , and plasmid quantity was determined using UV-spectrophometer. To assess the analytical sensitivity of the TaqMan-PCR assay, 10-fold serial dilutions of purified plasmid DNA (10 6 to 1 copies ll À1 ) in sterile distilled water were tested in triplicate.

Assessment of sensitivity using T. marneffei yeast cells
The sub-cultured T. marneffei yeast cells prepared as above were counted on a cell counting chamber under a light microscope. Ten-fold serial dilutions of T. marneffei yeast cells (10 6 -10 1 cells ml À1 ) spiked into human plasma collected from healthy volunteers were tested in the PCR assay. A quantity of 200 ll of plasma of each dilution was used for DNA extraction.
Assessment of sensitivity and specificity using fungal isolates

Assessment of the Taqman assay in plasma samples
The diagnostic performance of the assay was assessed using plasma samples of HIV-infected patients, 50 with culture-confirmed penicilliosis and 20 with other opportunistic infections (Fig. 1). The samples came from patients who participated in a case-control study to investigate the risk factors of AIDSassociated penicilliosis between January 2011 and July 2012 at the HTD. Blood culture was performed in all patients using the automated Bactec TM 9240 culture system (Becton Dickinson, Sparks, MD, USA). Penicilliosis case patients were those with an illness in which T. marneffei was isolated from a clinical specimen, including blood, skin scrapings, bone marrow, lymph node and/or other body fluids. Control patients were diagnosed with AIDS-associated opportunistic infections other than penicilliosis. Plasma samples for PCR testing were collected at enrolment either prior to or within the first 48 h of antifungal therapy.

Statement of safety
Clinical materials presumed to be infected with T. marneffei are handled and processed according to Biosafety Level 2 (BSL-2) practices. All culture materials, whether in the yeast or the mould form, are handled in a Class II biosafety cabinet. All culture containment materials are covered with paraffin paper when transferred outside of the safety cabinet.

Results
Sensitivity for detecting MP1 gene copy number The TaqMan-PCR assay was able to detect MP1-containing plasmids at one copy per reaction (or 30 copies ml À1 ), but detection was consistent at 10 copies per reaction (or 300 copies ml À1 ). Figure 2 shows the quantification of MP1-containing plasmids that were serially diluted to concentrations from 10 6 to 10 1 and the linear regression of the Cp values vs. the log 10 concentrations of MP1 plasmids. Linearity was achieved within a range plasmid DNA from 10 6 to 10 1 copies per reaction, and the correlation coefficient Sensitivity for detecting T. marneffei yeast spiked into human plasma The TaqMan-PCR assay was able to detect T. marneffei yeast cells at a concentration of 100 cells ml À1 . Linearity was achieved within a range of yeast concentrations from 10 6 to 10 2 cells ml À1 , R 2 = 0.997. The calculated number of plasmids at the limit of detection of 100 cells ml À1 is 30 copies ml À1 , indicating significant DNA loss during extraction.

Specificity for detection of T. marneffei among fungal isolates
The primers and probe successfully detected all seven T. marneffei clinical isolates. No significant elevated signal was detected in the 13 non-T. marneffei isolates. The absence of DNA was confirmed by agarose gel electrophoresis of TaqMan-PCR products derived from these non-T. marneffei isolates, indicating that the assay is highly specific.

Diagnostic performance in patient plasma
The assay was evaluated in plasma samples from 50 HIV-infected patients with culture-confirmed penicilliosis and from 20 patients with other HIV-associated opportunistic infections.  Table 2 shows the Cp values in the 50 plasma samples of the patients with culture-confirmed penicilliosis, timing of blood collection with respect to antifungal therapy, and the duration of the Bactec culture system to detect fungal growth from the hospital microbiology laboratory. The median time to detect a growth signal in blood culture bottles using the Bactec system was 5 days (inter quartile range: 5-7). Among the 16 patients who did not have skin lesions, the PCR assay was positive in only seven (43.8%); however, sensitivity was higher, 66.7%, when blood samples were collected prior to antifungal therapy ( Table 2).

Discussion
We developed a highly specific PCR assay for T. marneffei detection with no amplification detected from the 13 different fungal isolates in the Cryptococcal, Aspergillus, Candida and Penicillium genera, and no amplification was observed in plasma specimens of 20 AIDS patients infected with other HIV-associated fungal and opportunistic pathogens. The diagnostic sensitivity was 70.4% in samples collected prior to antifungal therapy, which decreased to 52.2% in samples collected within 48 h after antifungal drugs had been started, suggesting that the assay can be useful to monitor fungal clearance on therapy. With respect to the analytical sensitivity, the assay had a limit of detection of one copy of MP1 per reaction, but it consistently detected 10 copies per reaction (or 300 copies per millilitre). In T. marneffei spiked human plasma, the assay consistently detected 100 cells ml À1 . The analytical sensitivity of our PCR assay appears to be  Time on antifungal therapy prior to blood collection for real-time PCR assay, time 0 = prior to antifungal therapy. 2 Time to detect a growth signal in blood culture bottles using the automated Bactec culture system; median: 5 days (interquartile range: 5-7). All reactions were performed in triplicate. Compared to the Bactec culturing method, our assay was less sensitive (70% vs. 88%), although the difference did not reach statistical significance. A major advantage of culturing is the larger volume of sample being used (approximately 5 ml of blood compared to 200 ll used in this assay). Using larger DNA extraction volume and further concentrating the DNA used in PCR reaction might increase the sensitivity of this assay.
Recently, a monoclonal-based immunoassay has been developed that detected T. marneffei Mp1p antigen in patient plasma with a sensitivity of 75% (15/ 20) and specificity of 99.4% (537/540). 10 The assay has been shown to detect Mp1p antigenemia in 9.4% of more than 8000 HIV-infected patients in outpatient clinics from Guangzhou, China, 20 offering another tool for rapid diagnosis of T. marneffei infection.
A limitation of our study is that we only tested the assay on patient plasma samples. Whole blood or white blood cell buffy coat may be better diagnostic specimens as T. marneffei primarily replicates inside macrophages, which can be lost during plasma sample processing. Another limitation is that the Cp values of the PCR-positive specimens in our study are high, suggesting that the amount of pathogen DNA obtained for PCR is not high enough. Research for better ways to break down the fungal cell wall using both mechanical and chemical means during DNA extraction and for ways to increase the amount of input DNA used in PCR assays may advance the molecular diagnostics of fungal diseases.
Despite the limitations and the need for further research to improve the sensitivity, the clear advantages of this real-time PCR assay are its speed and high specificity. It takes 5-6 h to perform, which would probably result in a next-day diagnosis in most routine laboratories. This compares to an average of 5 days to detect a growth signal in the blood culture of patients using the highly efficient Bactec system (Table 2). Additionally, identification of T. marneffei requires a demonstration of temperature-dependent dimorphic growth, and subculture at 25°C generally takes another 2-3 days. This real-time PCR assay should not replace the need for conventional microbiology methods in diagnosing penicilliosis. However, in conjunction with culturing, it can be used as a rapid rule-in test that can make a significant difference in patient management by allowing antifungal therapy to begin sooner, particularly in patients without skin lesions, and has the potential to improve the outcomes of T. marneffei-infected patients.