OUP user menu

Chlamydia-like bacteria in respiratory samples of community-acquired pneumonia patients

Susanne Haider , Astrid Collingro , Julia Walochnik , Michael Wagner , Matthias Horn
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01099.x 198-202 First published online: 1 April 2008

Abstract

Chlamydia-like bacteria, obligate intracellular relatives of Chlamydia trachomatis and Chlamydophila pneumoniae, are widely distributed in nature. Using a two-step nested and semi-nested PCR approach targeting the 16S rRNA gene, we found DNA of Chlamydia-like bacteria in respiratory samples from patients with community-acquired pneumonia. Of 387 cases tested, four (1.03%) tested positive if only sequences showing less than 99.9% 16S rRNA gene sequence similarity to known Chlamydiae were considered. These included for the first time Protochlamydia amoebophila, Waddlia chondrophila, and ‘Candidatus Rhabdochlamydia porcellionis’-related sequences. This study extends previous findings suggesting an association of Chlamydia-like bacteria with respiratory disease, but a causal link between these microorganisms and respiratory tract infections has yet to be established.

Keywords
  • Parachlamydia
  • Protochlamydia
  • Simkania
  • Waddlia
  • Rhabdochlamydia
  • Acanthamoeba

Introduction

Pneumonia is one of the most frequent infections of humans and animals and the third most common cause of death due to infectious disease worldwide (Welte et al., 2004). Among bacterial pathogens, the leading causes of community-acquired pneumonia (CAP) are Streptococcus pneumoniae, Haemophilus influenzae, Mycoplasma pneumoniae, and Chlamydophila pneumoniae. Chlamydophila pneumoniae, in particular, has been estimated to be responsible for 2–43% of all cases of CAP (Wellinghausen et al., 2006). However, in about 50% of all cases, the causative agent of pneumonia remains unknown (reviewed in Bartlett et al., 1998). Recently, a number of novel bacteria have been identified that are moderately related to C. pneumoniae and that have been proposed to represent emerging pathogens (Greub & Raoult, 2002; Friedman et al., 2003). These obligate intracellular Chlamydia-like bacteria (also known as environmental Chlamydiae) have been grouped into the novel families Waddliaceae, Parachlamydiaceae, and Simkaniaceae (Everett et al., 1999; Rurangirwa et al., 1999). They are widely distributed in nature and show an extremely broad host range. They live as endosymbionts in free-living amoebae and are able to infect and thrive in insects, crustaceans, reptiles, fish, birds, marsupials, and mammals (reviewed in Corsaro & Greub, 2006). Additionally, a large number of unidentified Chlamydiae exist in various environmental and clinical samples, indicating that chlamydial diversity is still underestimated (Horn & Wagner, 2001; Corsaro et al., 2003). To date, Chlamydia-like bacteria have mostly been implicated in human respiratory disease, mainly based on serological and molecular data (reviewed in Friedman et al., 2003; Corsaro & Greub, 2006). For example, in one study, 2.6% of adult patients hospitalized with CAP (n=308) showed high or increasing IgA and/or IgG titers against Simkania negevensis, which was taken as evidence for an acute infection (Lieberman et al., 1997). In another study, IgM antibody titers ≥50 against Parachlamydia spp. have been observed in 3.7% of CAP patients (n=376), which was considered to represent past infections (Marrie et al., 2001). Two of these patients showed an even higher antibody titer (≥400), indicating an acute infection (Marrie et al., 2001). Additional evidence for a human pathogenic potential of Chlamydia-like bacteria might be the documented ability of some Parachlamydiaceae to infect and replicate within mammalian cells (Greub et al., 2003; Collingro et al., 2005a; Casson et al., 2006). However, the actual prevalence of Chlamydia-like bacteria in clinical specimens is difficult to assess, as they are not detected by conventional diagnostic procedures. In this study, we investigated respiratory samples from CAP patients for the presence of Chlamydia-like bacteria by a novel nested, highly sensitive PCR approach. This work was part of the German medical research network CAPNETZ, which aims to improve the current knowledge, diagnostic, and therapy of CAP (Welte et al., 2004).

Materials and methods

Clinical samples

Respiratory samples from 387 CAP patients recruited by CAPNETZ (Welte et al., 2004) were examined. The inclusion criteria were age ≥18, pulmonary infiltration visible in chest X-ray, and at least one of the following symptoms: cough, purulent sputum, and pathological sounds on auscultation (Wellinghausen et al., 2006). In total, 493 respiratory samples were analyzed, including primarily sputum (n=383), throat washings (n=101), and bronchoalveolar lavage (BAL, n=24).

DNA purification and PCR assays

DNA extraction was performed at the CAPNETZ central service unit using the QIAamp DNA-Blood Mini Kit (Qiagen, Hilden, Germany) according to the instructions of the manufacturer (Wellinghausen et al., 2006). For PCR screening we developed and used a new, highly sensitive nested PCR approach targeting the 16S rRNA gene. Previous studies have shown that PCR screening for Chlamydia-like bacteria is extremely susceptible to contaminations (Corsaro & Greub, 2006), possibly due to the ubiquitous occurrence of Chlamydia-like bacteria and their amoeba hosts. The assay used in this study tried to minimize this risk by targeting a rather large DNA fragment in the first step of the nested PCR, which should thus be less prone to contamination than PCR targeting shorter and much more stable DNA fragments. This first step of the nested PCR assay amplified the near full-length 16S rRNA gene (c. 1530 bp) of most known Chlamydia-like bacteria, including the Parachlamydiaceae, Waddliaceae, and Simkaniaceae. The forward primer PCf excluding the Chlamydiaceae (Horn & Wagner, 2001) was used in combination with the Chlamydiales-specific reverse primer 16S2 (Pudjiatmoko et al., 1997) at an annealing temperature of 56 °C (Table 1). Subsequently, a 1 : 100 dilution of the PCR product was used as template for the second step of the nested PCR assay, which employed the Chlamydiales-specific primer set 16SigF2/16SigR2 (modified from Everett et al., 1999), amplifying c. 290-bp fragments, at an annealing temperature of 61 °C. The sensitivity of this nested PCR assay was assessed with Protochlamydia amoebophila DNA. The lowest limit of detection was 10 fg DNA, corresponding to c. four bacterial cells (data not shown) and demonstrating the high sensitivity of this assay. Samples that tested positive in this first, nested PCR assay were subsequently further analyzed by an additional, semi-nested PCR using the PCR product from the first step of the nested PCR assay (which was not sufficient to be sequenced directly), but primers 16SigF2 and 16S2 in the second step (Table 1). In contrast to the first nested PCR assay, this semi-nested PCR generated a 16S rRNA gene fragment of sufficient length (c. 1510 bp) for a detailed phylogenetic analysis. Positive controls (P. amoebophila UWE25 DNA; Collingro et al., 2005b) and negative controls (no DNA added) were included in all PCR reactions. Amplification products were purified by the QIAquick PCR Purification Kit (Qiagen, Vienna, Austria) and sequenced directly.

View this table:
1

PCR primers used in the 16S rRNA gene-targeted PCR assays

Sequence analysis

The software package arb (Ludwig et al., 2004) was used to check for chimeric sequences and to perform phylogenetic analysis. Nucleotide sequences were deposited at GenBank/EMBL/DDBJ under accession numbers EU090706 to EU090709.

Results and discussion

To study the incidence of Chlamydia-like bacteria in CAP patients, 493 respiratory samples from 387 patients were examined using a novel combination of a nested and a semi-nested PCR assay. Of these, 33 samples (6.69%) tested positive in the first, nested PCR assay, but only 15 samples (3.04%) tested positive for the presence of Chlamydia-like organisms in both the nested and the semi-nested PCR assay. The difference between the results from those two tests might indicate false positives in the first nested PCR assay (due to short, degraded DNA fragments amplified in the second step of the nested PCR assay, but not targeted by the second step of the semi-nested PCR assay; Table 1). Ambiguous samples were thus excluded from further analysis, although the negative controls included in all PCR assays remained negative. Subsequent comparative sequence analysis of the amplificates from the semi-nested PCR demonstrated that one amplificate was unspecific showing no significant database hit, while 10 sequences were identical or almost identical (99.9–100% sequence similarity) to each other and to the 16S rRNA gene of Parachlamydia acanthamoebae Berg 17 (n=7) or P. amoebophila UWE25 (n=3).

On the one hand, the high 16S rRNA sequence similarity to known Parachlamydiaceae might not be very surprising as also all C. pneumoniae strains recovered from humans are virtually indistinguishable based on their 16S rRNA sequences (99–100% similarity; Pettersson et al., 1997). On the other, however, this could also indicate that these sequences represent PCR contaminations from organisms also handled in our laboratories, despite all possible care taken. To minimize the risk of analyzing false-positive data, we thus preferred to exclude all sequences from further analysis that shared ≥99.9% 16S rRNA sequence similarity with known sequences. Phylogenetic sequence analysis of the remaining four sequences allowed us to assign them to the chlamydial families Parachlamydiaceae (n=1), and Waddliaceae (n=1), and to ‘Candidatus Rhabdochlamydia porcellionis’ (n=2), respectively.

Members of the Parachlamydiaceae have been suggested previously to be associated with respiratory disease of humans (Corsaro & Greub, 2006). The Parachlamydiaceae sequence found in this study (CN823) shows 99.6% similarity to the 16S rRNA gene of P. amoebophila and represents the first P. amoebophila sequence from a human specimen.

Currently, the family Waddliaceae comprises the two species Waddlia chondrophila, isolated from an aborted bovine fetus (Rurangirwa et al., 1999), and Waddlia malaysiensis, isolated from urine samples of fruit bats (Chua et al., 2005). Waddlia-related sequences have not been amplified from human specimens in a systematic study before, but a recent report described a correlation between seropositivity against W. chondrophila and human fetal loss (Baud et al., 2007). The Waddlia-like sequence detected in this study (CN761) shared 98.0% and 99.7% nucleotide similarity with W. chondrophila strains WSU-85-1044 and 2032/99, respectively, and represents the first W. chondrophila-like sequence found in a human respiratory sample.

Two sequences (CN808 and CN554) were related to ‘Candidatus Rhabdochlamydia porcellionis’, a recently described symbiont of terrestrial isopods, forming a distinct lineage within the Chlamydiales, most closely related to the Simkaniaceae (Kostanjsek et al., 2004). 16S rRNA sequence similarity values to ‘Candidatus Rhabdochlamydia porcellionis’ were only 89.5% and 90.3%, respectively, and 89.1% and 89.4% to S. negevensis. The most similar known sequence was environmental Chlamydia clone P-11 (environmental Chlamydia lineage VI; (Horn & Wagner, 2001), sharing 90.4% and 93.5% sequence similarity, respectively. To date, only two distantly related ‘Candidatus Rhabdochlamydia porcellionis’-like sequences (c. 230 bp) have been detected earlier in human specimens; only one of these originated from a patient with an upper respiratory tract infection (Ossewaarde & Meijer, 1999).

Many Chlamydiae live as endosymbionts of free-living amoebae or are able to infect these protozoa (Essig et al., 1997; Kahane et al., 2001; Corsaro & Greub, 2006). Amoebae are therefore considered to play a key role in the adaptation of environmental bacteria to intracellular life within higher eukaryotes (Horn & Wagner, 2004). In addition, amoebae are well known as environmental reservoirs and vehicles of dispersal for bacterial pathogens such as Legionella pneumophila (Molmeret et al., 2005). Although Chlamydia-like bacteria have a broad host range and are able to infect phylogenetically different amoebae, the majority of them can use Acanthamoeba spp. as hosts. For this reason, all positive samples were also tested for the presence of amoeba DNA by performing an Acanthamoeba-specific PCR assay (Walochnik et al., 2004). However, all samples were negative for Acanthamoeba DNA, suggesting either the absence of Acanthamoeba spp. in these samples or their presence in a concentration below the detection limit (one amoeba trophozoite mL−1).

The four specimens that tested positive for Chlamydia-like DNA originated from four different patients, corresponding to 1.03% of all patients included in this study. Their mean age was 64.5 years (range 56–75) and no sex-related differences existed. Only two of the four patients were also positive for other known agents of CAP (Table 2), but all four patients tested negative by three different C. pneumoniae-specific PCR assays (Wellinghausen et al., 2006). From two patients, specimens from throat washings were also available, but these tested negative in all PCR assays, suggesting that throat washings are less suited for the detection of Chlamydia-like bacteria than BAL and sputum. All examined patients recovered from pneumonia after Clarithromycin or Moxifloxacin treatment for 8–19 days (Table 2).

View this table:
2

Chlamydia-like sequences detected in CAP patients

In conclusion, using a two-step nested/semi-nested PCR approach we detected DNA sequences from Chlamydia-like bacteria in respiratory samples of CAP in a highly sensitive and specific manner. This PCR assay allowed us to amplify a broad spectrum of Chlamydia-like bacteria and to characterize the recovered sequences phylogenetically. The presence of DNA from Chlamydia-like bacteria in 1.03% of CAP patients (n=387) adds to a number of recent studies suggesting a possible association of these microorganisms with human disease (Corsaro & Greub, 2006). However, taking into account Frederick's and Relman's revisions of Koch's postulates for sequence-based identification of microbial pathogens (Fredericks & Relman, 1996), current evidence is just a first step towards our understanding of a possible causal link between Chlamydia-like bacteria and disease in humans.

Acknowledgements

The authors gratefully acknowledge the help of Stephan Schmitz-Esser, and Elena R. Toenshoff during preliminary studies and Christian Baranyi for invaluable technical assistance. This work was funded by the German Federal Ministry of Education and Research grant 01KI0422/B1, Competence Network CAPNETZ to M.W., and by the Austrian Science Fund grant Y277-B03 to M.H.

Footnotes

  • Editor: Rob Delahay

References

View Abstract