| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
DIAGNOSTICS, TYPING AND IDENTIFICATION |
Department of Laboratory Medicine1 and Second Department of Internal Medicine2, Nagasaki University School of Medicine, Nagasaki 852-8501, Japan
Correspondence Y. Hirakata hirakata{at}net.nagasaki-u.ac.jp
Received 13 June 2002 Accepted 23 December 2002
A detection system for Legionella DNA in blood samples based on the PCR was developed and evaluated in A/J mice with experimentally induced Legionella pneumonia. Primers were designed to amplify a 106 bp DNA fragment of the 16S rRNA gene specific to Legionella species. The PCR system could detect clinically relevant Legionella species including Legionella pneumophila, Legionella micdadei, Legionella bozemanae, Legionella dumoffii, Legionella longbeachae, Legionella gormanii and Legionella jordanis. The sensitivity of the PCR system was 20 fg extracted DNA. In the mouse model, the blood PCR was compared with results obtained by PCR on bronchoalveolar lavage fluid (BALF) samples, cultures of blood and BALF and detection of Legionella urinary antigen. Blood PCR was positive until 8 days after infection, while BALF PCR became negative on day 4. These results indicate that PCR using blood samples may be a useful, convenient and non-invasive method for the diagnosis of Legionella pneumonia.
Legionella pneumophila is one of the leading causes of bacterial pneumonia, particularly in susceptible individuals or the immunocompromised (El-Solh et al., 2001; Marston et al., 1997; Sopena et al., 1999). Legionella pneumonia is not always easy to diagnose, since clinical and radiographic features are often indistinguishable from those of pneumonias caused by other pathogens. Although serological examination has been one of the methods traditionally used for the diagnosis of Legionella pneumonia, paired sera are usually required. Moreover, as many as 25 % of patients with Legionella pneumonia may fail to exhibit diagnostic antibody titres (Harrison & Taylor, 1988). Culture of sputum or other respiratory samples such as a transtracheal aspirate (TTA) is another traditional method for the detection of Legionella species, reported to be specific and considered the ‘gold standard’ test (Yu, 1995). However, the sensitivity of the culture method has been reported to be as low as 10–60 % (Breiman & Butler, 1998; Waterer et al., 2001). In addition, it is frequently difficult to obtain respiratory samples from patients with Legionella pneumonia, since most patients have a non-productive cough; physicians frequently need to obtain samples by invasive methods such as TTA or bronchoalveolar lavage (BAL). Furthermore, culture often fails to isolate the pathogens when patients have already been treated with antibiotics, even when such agents are not clinically potent.
Recently, enzyme immunoassay (EIA) for the detection of Legionella antigen in urine has been used for diagnosis of Legionella pneumonia, with a reported sensitivity of 63–77 % (Benson et al., 2000; Dominguez et al., 1998; Stout & Yu, 1997). PCR has also been used as a rapid diagnostic method, employing samples of BAL fluid (BALF) (Cloud et al., 2000; Jaulhac et al., 1998; Jonas et al., 1995; Matsiota-Bernard et al., 1994; Weir et al., 1998) or pleural effusion (Breiman & Butler, 1998; Hirakata et al., 1996; Lo Presti et al., 2000). However, PCR detection of L. pneumophila using non-invasive specimens is preferable, particularly by family physicians and non-pulmonologists.
The aim of the present study was to examine the efficacy of PCR on blood samples as a rapid diagnostic method for Legionella pneumonia and to compare the results with those of PCR on BALF, cultures of blood and BALF samples and EIA to detect urinary antigen in a mouse model of Legionella pneumonia.
METHODS
Bacteria.
L. pneumophila serogroup 1 ATCC 33152T (Philadelphia 1T) was used for the animal experiments and for assessment of the sensitivity of PCR detection. The bacterial strains used for evaluation of specificity of PCR detection are summarized in Table 1.
|
Bacterial inoculum.
L. pneumophila ATCC 33152T was stored until use at –80 °C in Müller–Hinton broth containing 30 % glycerol. A portion of the stock was cultured on buffered charcoal yeast extract (BCYE)-
agar (Oxoid) for 4 days at 37 °C. A single colony grown on BCYE- agar was inoculated into 5 ml buffered yeast extract (BYE) broth (Difco Laboratories) and incubated under shaking at 130 r.p.m. for 24 h at 37 °C. A portion of the culture was inoculated into 5 ml fresh BYE broth and incubated until the OD595 of the culture broth reached 0.5. The bacterial suspension was centrifuged at 3000 r.p.m. for 15 min at 4 °C and the pellet was washed twice and resuspended in physiological saline. The bacterial suspension was adjusted to 1 x 109 c.f.u. ml–1 by turbidimetry.
Animal model of Legionella pneumonia.
Male, 8-week-old, specific-pathogen-free A/J mice (Japan SLC Inc.) were used for all experiments. The mice were housed in a pathogen-free environment and received sterile food and water. These mice were inoculated intratracheally with L. pneumophila as described previously (Yanagihara et al., 1997). Briefly, mice were anaesthetized with pentobarbital and then 0.1 ml bacterial suspension (approx. 105 c.f.u. per mouse) was inoculated through the mouth into the trachea using the outer sheath of an intravenous needle (Top Co.). Body weight was measured on the days indicated after infection. The experimental protocol was approved by the Animal Care and Use Committee of Nagasaki University.
Sample collection.
Urine samples were collected from mice housed in metabolic cages (Tecniplast). For sampling blood and BALF, five mice were selected at random on the days indicated and anaesthetized. An incision was made through the skin of the chest to expose the trachea, lungs and heart. Cardiac blood samples (about 800 µl per mouse) were obtained using a 26-gauge needle and 1 ml syringe containing a small amount of EDTA. A portion of blood was used for leukocyte counts. BALF was also obtained from these mice. For this purpose, the trachea was intubated with an intravenous catheter (3 Fr. Atom Co.), which was connected to a 1-ml syringe. The lungs were washed three times with 0.6 ml cold Ca2+- and Mg2+-free PBS (Gibco).
Recovery of L. pneumophila from BALF and blood samples.
BALF and blood samples were cultured on BCYE- agar containing vancomycin, polymyxin B and amphotericin B (WYO agar; Eiken) and 5 % sheep-blood agar (Nissui Pharmaceutical) for 7 days.
DNA extraction and PCR.
DNA was extracted from blood samples with a Nucleo Spin blood kit (Clontech) and from BALF with a Nucleo Spin tissue kit (Clontech). Bacterial DNA was extracted with a DNeasy tissue kit (Qiagen). A 106-bp region of the Legionella 16S rRNA-encoding gene was amplified using the 20mer primers LSP-GCG-S (sense, 5'-GCGGCTACCTGGCCTAATAC-3'; designed in this study) and Cp3.2 (antisense, 5'-CCAACAGTAAGTTGACATCG-3'; Jonas et al., 1995). PCR included initial denaturation at 94 °C for 4 min followed by 35 cycles of annealing at 56 °C for 1 min, extension at 72 °C for 1 min and denaturation at 94 °C for 1 min and a final step of extension for 10 min at 72 °C. For analysis of the PCR product, 2 % agarose gel electrophoresis was performed with 10 µl of the reaction solution and the DNA fragment was confirmed using ethidium bromide staining.
Detection of Legionella antigen in urine.
Legionella antigen in urine was detected with Legionella urinary antigen EIAs from Binax and Biotest AG.
RESULTS AND DISCUSSION
PCR sensitivity and specificity
We first confirmed that the primers used in our study could detect as little as 20 fg purified L. pneumophila DNA prepared in Tris/EDTA (Fig. 1). In another control study, it was possible to detect as little as 200 fg L. pneumophila DNA in 200 µl blood obtained from healthy volunteers. PCR assays using whole-blood samples have been considered difficult because of the presence of inhibitors in blood. PCR detection of Legionella in the serum or buffy coat has been reported previously (Aebischer et al., 1999; Lindsay et al., 1994; Murdoch & Chambers, 2000; Murdoch et al., 1996, 1999), but not in whole-blood samples. Recently, commercially available kits for DNA extraction have facilitated PCR on whole-blood samples. The PCR product was detected in L. pneumophila serogroups 1–6, Legionella longbeachae serogroups 1 and 2 and five other Legionella species tested. On the other hand, PCR products were not detected for seven other bacterial strains unrelated to Legionella (Table 1). PCR amplification using primers for the mip gene has been widely used, since these primers can detect most Legionella species, with the exception of Legionella geestiana (Ratcliff et al., 1998, 2001). For our PCR assay, we chose primers that amplify 106 bp of the 16S rRNA gene, which may have certain advantages compared with methods reported previously. Firstly, amplification of the 16S rRNA gene may be more sensitive than that of the mip gene (Engleberg et al., 1989; Iwamoto et al., 1994) because multiple copies of the 16S rRNA gene exist in bacteria. Furthermore, amplification of 5S rRNA failed to detect several clinically relevant Legionella species, such as Legionella jordanis (Brieland et al., 1994). Our primers could detect most of the relevant species of Legionella, including L. jordanis, with high specificity.
|
Features of experimentally induced pneumonia
In previous studies (Brieland et al., 1994; Winn et al., 1982), animal models of Legionella pneumonia were established by incision of the trachea. In our mouse model of Legionella pneumonia, bacteria were inoculated intratracheally in a less invasive technique than those used previously. Therefore, the characteristics of infection in our mouse model resembled those of patients with Legionella pneumonia more closely than have other animal models. Acute pneumonia was confirmed by pathological examination of resected murine lungs stained with haematoxylin and eosin, as evident by cellular infiltration in alveoli and effusion. Body weight decreased gradually from 28.4 ± 1.0 g (mean ± SD) at baseline to 20.9 ± 1.5 g at day 4 after infection. However, body weight subsequently recovered gradually and was 24.4 ± 1.9 g on day 13. As expected, acute pneumonia was associated with leukocytosis (basal leukocyte count, 2250; day 5, 11 077 cells ml–1), but the count decreased to 3090 cells ml–1 on day 12 (Fig. 2). Our results were similar to those of Brieland et al. (1994) with regard to the severity of pneumonia. Their results in mice infected with Legionella showed that L. pneumophila grew exponentially in the lung during the 24–48 h post-inoculation and was gradually eliminated from the lungs during days 3–7 post-inoculation. Based on changes in body weight and leukocyte counts in our model, we concluded that pneumonia was most severe on days 3 to 5 after inoculation. In addition, the pathological changes in the lungs and inflammatory response were serious by 72 h after inoculation.
|
Evaluation of diagnostic methods in a mouse model
In the present study, we evaluated PCR on blood samples by comparing the results of this assay with the results of PCR on BALF samples, culture of blood and BALF samples and urinary antigen detection in a mouse model of Legionella pneumonia. Previous studies have reported the usefulness of PCR as a diagnostic test for Legionella pneumonia using respiratory specimens (Cloud et al., 2000; Hirakata et al., 1996; Jaulhac et al., 1998; Jonas et al., 1995; Lo Presti et al., 2000; Weir et al., 1998), with a sensitivity of 
90 % (Matsiota-Bernard et al., 1994). However, cough is non-productive in most patients with Legionella pneumonia, and clinical symptoms deteriorate rapidly and severely. Therefore, it is often difficult to collect respiratory specimens using invasive techniques such as BALF. Consequently, specimens that could be collected non-invasively might be preferable. L. pneumophila was isolated from cardiac blood samples obtained on days 1 and 2, but could not be detected on or after day 3. On the other hand, the pathogen was recovered from BALF on days 1, 2 and 3, but was not detected on or after day 4. PCR assays of BALF samples were positive only on days 1, 2 and 3 after inoculation, while blood PCR remained positive until day 8 after infection and became negative on day 9. Legionella antigen in urine was positive from days 1–12 (Fig. 2) and was detected continuously up to day 30 by both the Biotest and Binax EIA. Therefore, in clinical cases, in addition to urinary antigen detection, whole-blood PCR may allow the diagnosis of Legionella pneumonia during the first several days after onset of the disease.
We performed further experiments to examine the relationship between positive urinary antigen and illness (data not shown). In these studies, we first established an oral-administration Legionella model by intraoesophageal inoculation of 6 x 106 c.f.u. L. pneumophila to A/J mice. The leukocyte count did not increase and weight loss was not observed in this model. Urinary antigen became positive at 6 h after inoculation and remained positive for 2 weeks after onset of infection. Blood PCR was negative throughout the test period. The chance of detecting urinary antigen would be minimal, as healthy humans are not tested for Legionella infections. These results, however, suggest that a person who drinks contaminated water may have a positive urinary antigen test even in the absence of clinical features of the disease. Another advantage of PCR compared with urinary antigen detection is that PCR gives the possibility of identifying Legionella to the species level in future by using the multiplex PCR method (Clark et al., 1993) or sequencing (Ratcliff et al., 2001).
In conclusion, in the present study, we have demonstrated that Legionella DNA could be detected by PCR in whole blood in a mouse model, suggesting that whole-blood PCR is a useful method for the diagnosis of Legionella pneumonia. Before whole-blood PCR can be applied to human samples, its specificity must be confirmed by Southern blotting or sequencing. Several studies have reported the usefulness of urine PCR (Helbig et al., 1999; Maiwald et al., 1995; Socan et al., 2000) as a diagnostic method. Further prospective studies of a large number of patients are required to examine the sensitivity and specificity and clinical impact of whole-blood PCR and to compare the assay with PCR on BALF and urine samples.
ACKNOWLEDGEMENTS
ACKNOWLEDGEMENTS
We thank Dr Naoyuki Miyashita, Department of Microbiology, Kawasaki Medical School, for supplying C. pneumoniae strains.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
) and leukocyte count (
) in a mouse model of Legionella pneumonia. A/J mice were inoculated intratracheally with L. pneumophila (105 c.f.u. per mouse) and body weight and peripheral leukocyte numbers were measured on the days indicated after infection. Data represent means ± SD. WBC, Whole-blood count; p.i., post-inoculation. Moreover, blood and BALF samples were cultured and used for PCR and urine samples were examined for Legionella antigen; results are shown.