Development of a Serum-Based Taqman Real-Time PCR Assay for Diagnosis of Invasive Aspergillosis

Abstract

We compared an Aspergillus fumigatus-specific Taqman real-time PCR assay for the diagnosis of invasive aspergillosis (IA) with the enzyme-linked immunosorbent assay Platelia Aspergillus method. Patients with proven or probable IA had positive results with at least one of the two methods. The PCR method seems to be more specific, and a combination of the two should improve the diagnosis of IA.

Due to the absence of a highly sensitive and specific “gold standard” method for the early diagnosis of invasive aspergillosis (IA) (5, 6), we developed an Aspergillus fumigatus-specific Taqman real-time PCR assay. We retrospectively evaluated it on serum samples from 41 patients with suspected or confirmed IA and compared its performance with that of the enzyme-linked immunosorbent assay (ELISA) Platelia Aspergillus method, which detects Aspergillus galactomannan antigens.

Primers and probe were designed by comparison of the sequences of rRNA genes (GenBank database) from A. fumigatus (accession no. U28460), other Aspergillus spp. (A. flavus, A. niger, A. nidulans, A. terreus, and A. versicolor), other filamentous fungi (Paecilomyces spp., Penicillium spp., Fusarium spp., and Alternaria spp.), yeasts (Candida spp. and Trichosporon spp.), and bacteria (Staphylococcus spp., Streptococcus spp., Enterobacteriaceae, Pseudomonas spp., and Actinomyces spp.). The target, a 67-bp DNA fragment specific to the multicopy gene encoding the 28S rRNA of A. fumigatus, was chosen with the Primer Express, version 1.0, program for Taqman technology. The primers were 28S-466 (5′-CTC GGA ATG TAT CAC CTC TCG G-3′) and 28S-533 (5′-TCC TCG GTC CAG GCA GG-3′), and the Taqman probe was 28S-490 (5′-6-carboxyfluorescein-TGT CTT ATA GCC GAG GGT GCA ATG CG-3′-6-carboxy tetramethylrhodamine).

For the PCR assay, the QIAmp DNA minikit (Qiagen) was used to extract the DNA from A. fumigatus IHEM 6117 (Scientific Institute of Public Health-Louis Pasteur, Brussels, Belgium), other fungi and bacteria studied, and all serum samples. Purified DNA was eluted in 50 μl of elution buffer and stored at −20°C until use. The 50-μl PCR mixture contained 1× Taqman universal PCR Master Mix, 500 nM each primer, 100 nM probe, and 15 μl of DNA. DNA was amplified in a Thermocycler/ABI Prism 7700 sequence detector (Applied Biosystems) by incubation at 95°C for 10 min followed by 45 amplification cycles (15 s of denaturation at 95°C and 1 min of hybridization and elongation at 60°C). The fluorescence of the PCR products was detected by the same apparatus. A calibration curve was constructed for each assay to determine the amount of DNA present in serum samples (concentration versus number of copies). For this, the A. fumigatus target was ligated into pUC18 and used to transform a competent Escherichia coli strain. PCR inhibitors were systematically sought by amplification of the β-globin gene.

We tested 207 serum samples from 41 immunocompromised patients (21 adults and 20 children) hospitalized in the Hematology, Reanimation, or Infectiology department of Hôpital Necker-Enfants Malades, Paris, France (n = 39), Cochin Hospital, Paris, France (n = 1), or Amiens Hospital (n = 1) between January 1999 and July 2001. Patients were considered to have proven, probable, or possible IA according to the revised European Organization for Research and Treatment of Cancer/Mycoses Study Group classification (2). At least two serum samples (range, 2 to 14 samples) were obtained from each patient over a maximum of 2 months before and/or after suspected infection (day 0 [D0]). Samples were collected in dry tubes and stored at −20°C. We also tested 39 serum samples from 29 control patients: 19 from 19 healthy subjects (11 adults and 8 children), 8 from 4 children with cystic fibrosis and A. fumigatus bronchial colonization, and 12 collected 15 and 30 days after bone marrow transplantation from 6 adults without suspected IA.

Aspergillus antigens were detected with the ELISA Platelia Aspergillus kit (Bio-Rad). Results were considered positive when they were equal to or higher than the cutoff value (1 ng of galactomannan/ml of serum). A. fumigatus DNA was detected with our specific PCR method. All samples were tested in duplicate. The sensitivity (related to the mean minimum number of cycles necessary to detect fungal DNA) was 34.3 cycles (1 copy). The reproducibility was good, with a linear function and a correlation coefficient of 0.983 for four DNA concentrations (1 to 10³ copies) amplified on five different days.

PCR methods designed for fungi and the targets used (often located within the 18S rRNA gene) vary considerably among authors. These methods have been used to detect the Aspergillus DNA in serum or total blood, but the two types of sample have not been compared for the same patients (3, 4, 7-11, 14, 15, 18). Due to the retrospective nature of our study, the tests were carried out on serum samples. To avoid cross-reactions with microorganisms other than the target (e.g., Penicillium spp.), we chose a target within the 28S rRNA gene, which is specific to A. fumigatus, the main agent responsible for IA. We selected a multicopy target to maximize sensitivity. With the Taqman technology, there is little risk of false positives due to contamination with previously amplified products. The specific DNA could be detected and quantified within 2 h, and the method can easily be standardized. Thus, this method is suitable for routine biological diagnosis.

The sensitivity and specificity of the ELISA for the diagnosis of probable and proven IA are about 90% when clinical or radiological symptoms are present but about 50% in their absence (12, 16). The frequency of false-positive results for at-risk pediatric populations is between 5.7 and 10% (13). Some studies compared this method and PCR. One study showed that nested PCR with a target within the 28S rRNA gene was more sensitive than the ELISA (18). Another found that the PCR method developed with a mitochondrial DNA target was less sensitive and should be used to confirm positive ELISA results (3).

In our study (Tables 1 and 2), all control serum samples were negative for the two tests. Of the 41 patients with IA, 26 were positive according to the PCR method: 20 had a mean of one copy, and 6 had between 1 and 10 copies. Twenty-six patients had proven or probable IA (14 adults and 12 children): 22 (7 of 7 proven, 15 of 19 probable), including all the adults, had a positive PCR test (84.6%), and 20 (5 of 7 proven, 15 of 19 probable) had a positive ELISA (75.2%). All gave a positive result with at least one of the two methods. When the two tests were positive for the same patient (5 proven, 11 probable), the PCR method was positive earlier than the ELISA for 3 patients with probable IA (patients 1, 13, and 16). In three (one proven [patient 7] and two probable [patients 10 and 16]) of the nine patients who had serological follow-up before developing fungal infection (three proven and six probable), the PCR test was positive before the radioclinical symptoms appeared (9 to 31 days before D0). For the two probable cases, the PCR test was positive before the ELISA. For four patients with probable IA, the PCR test was negative. Three were treated with antifungal agents on D0 and died from infectious respiratory distress syndrome. These results may be due to the lack of sensitivity of the method. Indeed, only very small amounts of A. fumigatus DNA circulate in the serum (between 100 fg/ml and 1 ng/ml). Similar levels were measured by Loeffler et al. in whole blood (11). Some authors observed that antifungal treatment rapidly results in negative PCR results (9). The detection of fungal DNA in the blood may be improved if DNA is extracted from a larger sample. It should also be possible to include a reverse transcription step.

TABLE 1.

Results of the PCR tests and ELISA for the three IA groups

IA group	No. (%) of:
	PCR test-positive patients/ no. tested	ELISA t-positive patients/ no. tested
Total	26/41 (63.4)	29/41 (71)
Proven	7/7 (100)	5/7 (71.4)
Adults	3/3 (100)	2/3 (66.7)
Children	4/4 (100)	3/4 (75)
Probable	15/19 (78.9)	15/19 (78.9)
Adults	11/11 (100)	7/11 (63.6)
Children	4/8 (50)	8/8 (100)
Possible	4/15 (26.7)	9/15 (60)
Adults	3/7 (42.9)	1/7 (14.3)
Children	1/8 (12.5)	8/8 (100)

TABLE 2.

Clinical, radiological, and biological features and test results for patients with proven, probable, and possible IA

IA group	Patient no.a	Sex/age (yr)	Feature(s)b			Mycology		No. of samples tested on D0/total no. of samples tested	No. of positive samplese/time of first positive testf
			Clinical	Radiological		Site(s)c/result by EDd or histology	Culture
				Nonspecific	Specific (scanner)				ELISA	PCR
Proven	1 (D, H)	M/28	P		P	Lung/+	−	1/19	3/D22	5*/D2
	2 (D, H)	M/76	P/C	P	/	Lung/+	A. fumigatus	0/2	0/NA	1/D24
	3 (R)	M/52	P	P	/	Trachea/+	A. fumigatus	0/2	2/D6	1/D6
	4 (D, T)	F/4	P/C	P/C	−	Disseminated/+	/	5/5	1/D0	1/D0
	5 (D, T)	F/16	P	P	/	Heart, thorax/+	A. fumigatus	1/9	0/NA	3/D12
	6 (D, T)	M/2	P/C	P	C	Lung/+	A. fumigatus	2/3	1/D14	1/D14
	7 (H)	M/2	P/C	P	C	Lung, brain/+	A. fumigatus	3/5	1/D−9	3/D−9
Probable	8 (D, H)	F/46	P/C		C	−	−	0/2	2/D19	2/D19
	9 (H, B)	M/51	P	P	/	+	A. fumigatus	0/7	2/D7	3/D7
	10 (H)	M/72	P		P	−	A. fumigatus	2/5	−	4/D−29
	11 (D, H, B)	F/45	P		P	−	−	1/14	−	5*/D22
	12 (D, H)	M/59	P/C	P	−	/	/	1/3	3/D0	1*/D0
	13 (D, H, B)	F/38	P		P	−	A. fumigatus	0/11	10/D2	11*/D1
	14 (D, H)	F/63	P	P	/	−	A. flavus	0/7	−	5/D2
	15 (D, H, B)	M/36	P/C		P	−	−	1/8	−	2/D13
	16 (T)	M/48	P	P	/	+	A. fumigatus	3/4	1/D4	2/D−31
	17 (H)	M/58	P		P	+	A. fumigatus	0/4	3/D3	2*/D3
	18 (D, H)	F/67	P		P	−	−	4/7	5/D0	5*/D0
	19 (T)	M/6	P	P	/	−	−	1/6	4/D0	−
	20 (D, H, B)	M/4	P/C	P	C	−	−	1/6	6/D0	−
	21 (D, H)	M/<1	P	P	/	−	A. fumigatus	0/2	2/D15	2/D15
	22 (D, H, B)	F/2	P	P	/	/	/	2/8	7/D0	−
	23 (D, H, B)	M/7	P	P	/	−	−	1/2	2/D0	−
	24 (H, B)	M/18	P	P	/	−	−	1/6	3/D11	3/D11
	25 (D, H, B)	M/15	P	P	/	−	A. fumigatus	2/4	3/D0	2/D6
	26 (D, H)	M/4	P	P	/	/	/	3/5	3/D21	3*/D21
Possible	27 (D, H)	M/65	P	P	/	−	−	0/8	−	6/D20
	28 (D, H)	F/71	P	P	/	−	A. fumigatus	1/2	−	−
	29 (H)	M/68	P	P	/	−	−	0/5	−	2/D10
	30 (H)	F/52	P	P	/	/	/	1/3	−	−
	31 (D, H)	M/71	P	P	/	−	−	0/2	−	−
	32 (D, H, B)	M/29	P/C	P	−	/	/	5/6	1/D7	3/D−26
	33 (D, H)	M/77	P	P	/	−	A. fumigatus	1/2	−	−
	34 (T)	M/2	−	−	/	/	/	1/2	2/D0	−
	35 (D, T)	M/4	P	P	/	−	A. fumigatus	0/4	1/D39	1/D30
	36 (D, I)	F/16	−	−	/	−	−	0/3	3/D2	−
	37 (H, B)	F/2	P	−	/	−	−	2/4	4/D−22	−
	38 (D, H, B)	F/2	−	−	/	/	/	0/4	4/D20	−
	39 (D, H, B)	F/3	−	−	/	−	−	3/9	7/D0	−
	40 (H, B)	F/4	−	P	/	/	/	3/4	3/D0	−
	41 (T)	M/4	−	P	/	/	/	1/3	3/D−1	−

^aD, death; H, hematological disease; I, immunological disease; B, bone marrow transplantation; T, organ transplantation; R, human immunodeficiency virus infection.

^bP, pulmonary signs; C, cerebral signs; P/C, pulmonary and cerebral signs; /, not tested; −, absent.

^cWhere no site is listed, the site is pulmonary.

^dED, cytological direct microscopic examination.

^e*, PCR quantification of 1 to 10 copies.

^fNA, not applicable. Boldface indicates days before D0.

In the possible IA group, the pulmonary symptoms of two (patients 27 and 29) of the three PCR-positive adults improved following treatment with antifungal agents but relapsed when specific treatment was stopped. The four PCR-negative patients were never treated with antifungal agents during the observed period and did not die. The discrepancy between the PCR and ELISA results for children with possible IA suggests that the PCR test is more specific in this population. ELISA was suspected to have given false-positive results for 10 patients (3 probable and 7 possible) with an associated infection (candidiasis or systemic bacterial infection) that may have resulted in cross-reaction (17) or with mucositis or host-versus-graft disease resulting in epithelial lesions that favored the digestive translocation of ingested galactomannans (1).

Thus, our PCR assay and the ELISA method are complementary for the diagnosis of IA. All the patients with proven or probable aspergillosis had a positive result with at least one of the two methods. However, the PCR method seems to be more specific than the ELISA method for patients with possible IA. Given these promising results, it is necessary to evaluate this method in a prospective study.