Use of Shell-Vial Cell Culture Assay for Isolation of Bacteria from Clinical Specimens: 13 Years of Experience
http://www.100md.com
微生物临床杂志 2005年第10期
Unite des Rickettsies, CNRS UMR 6020, IFR 48, Faculte de Medecine, Universite de la Mediterranee, 27 Boulevard Jean Moulin, 13385 Marseille Cedex 05, France
ABSTRACT
The shell-vial culture assay is performed routinely in our laboratory. Recently we revisited our experience of using the shell-vial culture assay for the isolation of microorganisms from various clinical samples. Over a 13-year period, we have isolated 580 bacterial strains (5%) from 11,083 clinical samples tested. Over the same period, 285 isolates of rickettsiae, bartonellae, or Coxiella burnetii were cultured from a total of 7,102 samples tested. These isolates include 55 Rickettsia sp. isolates, 95 Coxiella burnetii isolates, and 135 Bartonella sp. isolates. Based on our experience with the growth of fastidious microorganisms, we have used a centrifugation shell-vial technique called JNSP, for "je ne sais pas" ("I don't know [what I am growing]") for the isolation of other microorganisms. A total of 173 isolates were cultured from the 3,861 clinical samples tested using the JNSP method. Of these, 40 isolates had not been grown before on usual axenic medium. These include 2 Staphylococcus aureus isolates, 7 isolates of Streptococcus sp. and related genera, 6 Mycobacterium sp. isolates, 1 Nocardia asteroides isolate, 1 Actinomyces sp. isolate, 1 Brucella melitensis isolate, 2 Francisella tularensis isolates, 1 Mycoplasma pneumoniae isolate, and 1 Legionella pneumophila isolate. Using this protocol, we have also cultured intracellular bacteria such as Chlamydia trachomatis and we have performed the first culture and establishment of Trophyrema whipplei. Applied in our laboratory, the shell-vial culture generally exhibits a low rate of success. However, in some cases, this technique allowed microbial diagnosis when classical agar procedure and PCR were negative.
INTRODUCTION
The spectrum of infectious diseases is wide and changing. The reliable diagnosis of infectious disease remains a difficult problem even for specialized laboratories. Isolation of new microorganisms will allow the description of clinical features of new diseases and the characterization of new pathogens, enabling genetic descriptions, physiological analyses, improvement of diagnostic tools, and antibiotic susceptibility testing for bacteria. Therefore, the isolation of infecting bacteria is not only a mean of diagnosis but also a basis for the evaluation of much needed improved diagnostic assays and a route to enhance understanding of the diversity and the epidemiology of infections (21). The successful isolation of fastidious microorganisms is often based on cell culture (21). The success of this technique is based on two critical points. First, the ratio of microorganisms to cells should be as high as possible (21). Second, centrifugation enhances the adhesion of intracellular microorganisms to the cells (52). In 1976, a centrifugation-cell microculture system, called the shell-vial assay, was first used for the diagnosis of viral disease due to cytomegalovirus and allowed early antigen detection (22). This diagnostic tool is becoming one of the most commonly used techniques in clinical virology laboratories since it reduces turnaround time (19).
For a long time, Rickettsia spp., which are strict intracellular gram-negative bacteria, were isolated using embryonated egg and/or animal inoculations (44). As the National Reference Center for Rickettsiosis in France, our laboratory started to adapt the shell-vial assay for bacteria culture in 1986 (32). This technique, due to the highly infectious nature of such bacteria, is routinely performed in a biosafety level 3 laboratory and has allowed us to isolate first Rickettsia sp. (32) and second Coxiella burnetii (45) from various specimens. For several years, our laboratory has developed the facilities and experience for cultivating fastidious microorganisms, using this versatile cell culture system. This has allowed us to receive a large panel of samples from France and abroad, and, notably, this approach has allowed the isolation and the establishment of the first strain of the bacillus of Whipple's disease, Tropheryma whipplei, from a cardiac valve (39). From September 1991 to August 2003, our laboratory attempted to routinely achieve isolation of microorganisms from various clinical samples, arthropods, and animals by the shell-vial assay. Using this technique, we were able to isolate not only rickettsial species (28, 29, 32, 51) and Coxiella burnetii (45) but also fastidious bacteria, such as Francisella tularensis (13), Legionella pneumophila (27), Brucella melitensis (48), Chlamydia trachomatis (35), and T. whipplei (39). In this study, we report our complete 13-year experience of shell-vial culture assay.
MATERIALS AND METHODS
Specimen collection. From September 1991 to August 2003, our laboratory received 18,124 samples for culture from France and abroad with presumptive diagnoses of infectious diseases due to fastidious microorganisms. We have collected and managed the samples according to the patient's clinical presentation. If necessary, we further contact the sender for additional or more suitable samples to obtain relevant clinical and epidemiological data. Samples from hospitals outside of Marseille were frozen at –80°C before transport in dry ice. Samples, transport means, and diagnostic techniques are summarized in Fig. 1.
Axenic culture. For the samples from Marseille, all specimens were divided into three equal parts: one for conventional axenic culture, one for shell-vial culture, and one for PCR assays. Inoculation procedures for axenic culture were then performed in our local bacteriological laboratory. Thus, both axenic and cell cultures were processed in parallel. All the methods for attempted isolation of pathogens, including those for Brucella sp., Francisella sp., Legionella sp., and Bartonella sp., are summarized in Table 1. For the samples received from outside Marseille, we divided them into two parts: one for shell-vial culture and one for PCR. The axenic culture was performed in the local bacterial laboratory. Results were retrospectively sent to us. Thus, for these cases, axenic and cell cultures could not be processed in parallel.
Cell culture. Five different shell-vial protocols have been developed in our laboratory: one for the specific isolation of rickettsiae, one for the specific isolation of Bartonella sp., one for the specific isolation of Coxiella burnetii, one for the specific isolation of Trophyrema whipplei, and one for the unspecific research of other strictly or facultative intracellular bacteria, called the "JNSP" protocol. The abbreviation "JNSP" is derived from the French sentence "Je ne sais pas," which means "I do not know" and stands for "I do not know what I am growing." Cell type, inoculation, length of incubation, and culture revelation differed according to the suspected pathogen. Most often, human embryonic lung (HEL) fibroblasts are used because they have the advantage that once a monolayer is established, contact inhibition prevents further division and the cell can be used for prolonged incubation. All the procedures are summarized in Table 2. Briefly, culture was performed using the centrifugation-shell-vial technique (Sterilin-Felthan-England, 3.7 ml) using 12-mm round coverslips seeded with 1 ml of medium containing 50,000 cells and incubated in a 5% CO2 incubator at 37°C for 3 days to obtain a confluent monolayer (32).
Inoculation procedure. All manipulations were done within a laminar flow hood except the centrifugation. One milliliter of supernatant from each sample was mixed with 1 ml of RPMI 1640 medium (Gibco BRL life Technologies, Cergy Pontoise, France). The suspension was used to inoculate three shell vials with a centrifugation step of 700 g/min for 1 h in Eagle's minimal essential medium (MEM; Gibco) at 22°C. The inoculum was removed, and the shell vials were washed twice with sterile phosphate-buffered saline (PBS). The cells were then incubated in MEM with fetal calf serum (FCS; Sigma, St. Louis, MO) and L-glutamine at 37°C under a 5% CO2 atmosphere. All cell lines and culture reagents were checked weekly for bacterial contamination. The culture medium was removed at different times according to the samples and the bacteria as described in Table 1.
Detection and identification of bacteria. The detection procedure was performed at different days of incubation according to the specimens (Fig. 1). Detection of growing bacteria was assessed on coverslips directly inside the shell vial. We have used acridine orange, Gimenez (18), or immunofluorescence (32) staining. Gram, Giemsa, periodic acid-Schiff (PAS), and Ziehl-Neelsen staining could also be performed, depending on the suspected microorganism. For immunofluorescence staining and fixation with methanol, 100 μl of homemade rabbit polyclonal antiserum (anti-Coxiella burnetii, Bartonella sp., and Rickettsia sp.) or patient's serum diluted respectively to 1:200 and 1:100 in PBS with 3% nonfat dry milk was added. The slides were incubated at 37°C for 30 min. After three washes in PBS, 100 μl of a fluorescein-conjugated goat anti-rabbit immunoglobulin serum (Jackson Immunoresearch Laboratories, West Grove, Pa.) diluted 1:200 in PBS containing 0.2% Evans blue was incubated at 37°C for 30 min. After three washes with PBS, the coverslips were mounted (cells facedown) in phosphate-buffered glycerol medium (pH 8) and examined at a magnification of x400 with an epifluorescence microscope (Zeiss, Thornwood, NY). If immunofluorescence or other staining was positive, identification to the species level was performed by using the 16S rRNA gene (26) or PCR specific for Bartonella sp., Rickettsia sp., Coxiella burnetii, and Tropheryma whipplei as described elsewhere (11, 15, 20, 55). DNA was extracted from both the isolate and the fresh biopsy specimen, using the QIAmp tissue kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. Sequencing was performed using a commercially available sequencing kit (dRhodamine terminator cycle sequencing kit; PE Applied Biosystems, Warrington, England) according to the manufacturer's recommendations (26). Sequences obtained from the isolates were aligned with the sequences in the GenBank DNA sequence database with the program BLAST (version 2.0; National Center for Biotechnology Information).
Interpretation of cell culture. A shell-vial culture was defined as positive if a pathogenic microorganism was isolated. If nonpathogenic or opportunistic microorganisms such as coagulase-negative Staphylococcus sp., Micrococcus sp., fungi, or nonfermentative gram-negative bacilli were isolated, the same microorganism would have to be detected in parallel with axenic culture and/or PCR assays in order to conclude that it was a positive cell culture. If these microorganisms were not detected with another test, they were considered as cell culture contaminants.
Establishment of a strain. In order to establish the isolate, the supernatant of a positive shell vial was inoculated onto confluent layers of various cell types according to the isolate in a 25-cm2 culture flask.
Statistical analysis. To compare the data, we used the Fisher's test using Epi Info version 6.04a (Centers for Disease Control and Prevention, Atlanta, GA). P < 0.05 was considered statistically significant.
RESULTS
From August 1991 to September 2003, we received a total of 18,124 human samples and a total of 384 animal samples in our laboratory. The samples received were inoculated using one or more of the five protocols developed in our laboratory. Each of these adapted protocols was chosen according to the clinical information obtained from the clinician. More precisely, 10,699 human samples were inoculated in a cell culture using the shell-vial assay. Of these, 3,536 were inoculated using the Coxiella burnetii protocol, 2,619 using the Bartonella sp. protocol, 949 using the Rickettsia sp. protocol, 3,861 using the JNSP protocol, and 88 using the new Trophyrema whipplei protocol. Samples were obtained from 5,764 blood specimens, 984 lymph nodes, 722 cardiac valves, 567 skin biopsies, 431 cerebrospinal fluid (CSF) samples, 203 vitreous humor fluid samples, 282 placentas, 189 bone marrow specimens, 155 abscesses from diverse sites, 79 lung biopsies, 31 digestive biopsies, and 1,292 specimens from various other sites. The latter group included samples from bones, synovial and pericardic fluids, or milk. Three hundred eighty-four animal samples were inoculated into shell-vial assay: 36 samples were inoculated using the Coxiella burnetii protocol, 297 using the Bartonella sp. protocol, 32 using the Rickettsia sp. protocol, and 19 using the JNSP protocol. For these 13 years, the global results obtained with shell-vial culture are summarized in Fig. 2. Note that three times more samples were analyzed in 2002 than in 1992 (1,290 versus 421). In parallel, we observed that the number of positive cultures increased from 18 in 1992 to 117 in 2002. Results for the shell-vial culture using the JNSP protocol from January 1996 to September 2003 are shown in Fig. 3.
Coxiella burnetii shell-vial culture protocol. During these 13 years, among the 3,536 human samples (Table 3) inoculated using this protocol, Coxiella burnetii was detected in 91 samples (45) by immunofluorescence and Gimenez staining and definitive identification was systematically confirmed by specific PCR and sequencing. We have detected Coxiella burnetii in 47 of the 351 inoculated cardiac valves, 31 of the 2,338 inoculated blood samples, 1 of the 22 inoculated lymph nodes, and 8 of the 280 inoculated placentas. We detected Coxiella burnetii in two spleen biopsies and six vascular samples, but the total of the inoculated samples were not available. Among the 36 animal samples, Coxiella burnetii was detected in 4 samples (Table 2). These samples included one mouse spleen, two vaginal specimens from ewes, and one goat placenta specimen. A total of 95 strains of Coxiella burnetii were detected. We were able to establish 81 strains obtained from human samples and 4 strains obtained from animal samples, corresponding to 93% of the strains isolated in shell-vial culture.
Bartonella sp. shell-vial culture protocol. Among the 2,619 samples inoculated with this protocol (29), Bartonella sp. was detected in 88 human samples by immunofluorescence and Gimenez staining; the identification to the species level was systematically done by specific PCR and sequencing. More precisely, we have detected Bartonella sp. in 55 of 1,244 inoculated blood samples, 3 of the 101 inoculated cardiac valves, 3 of the 16 inoculated medullar aspirations, 2 of the 518 inoculated lymph nodes, and 2 of the 90 inoculated skin biopsies. Among the 297 animal samples inoculated, 47 Bartonella sp. isolates were detected (Table 3). These samples were obtained from cat's blood. A total of 135 strains of Bartonella sp. were detected. We were able to establish 113 strains (83%). Sixty-five established strains were obtained from human samples: 62 strains of Bartonella quintana and 3 strains of Bartonella henselae. All 47 strains obtained from animal samples were established: 27 Bartonella henselae strains, 19 Bartonella clarridgiae strains, and 1 Bartonella koehlerae strain.
Rickettsia sp. shell-vial culture protocol. Among the 949 samples inoculated with this protocol (51), Rickettsia sp. was detected in 52 human samples by immunofluorescence and Gimenez staining; definitive identification to the species level was systematically done by specific PCR and sequencing. More precisely, we detected Rickettsia sp. in 37 of the 337 inoculated skin biopsies and in 15 of the 510 inoculated blood samples. Among the 30 inoculated animal samples, we have detected Rickettsia sp. in 3 ticks and 2 lice. A total of 57 Rickettsia sp. isolates were detected, and 39 were established (70%) (Table 3). Thirty-six strains were obtained from human samples: 18 strains of Rickettsia conorii, 12 of Rickettsia africae, 4 of Rickettsia sibirica subsp. mongolitimonae (16, 40), 1 of Rickettsia prowazekii (5), and 1 of Rickettsia slovaca (6). We were able to establish two strains of R. slovaca and one of R. helvetica from ticks, but the two strains of R. prowazekii isolated from lice could not be established.
JNSP shell-vial culture protocol. In our laboratory, we have been performing the JNSP protocol almost routinely since 1996. Among the 3,861 samples inoculated with this protocol, we detected 175 bacterial species and 32 bacterial cell culture contaminants. These bacteria were detected by immunofluorescence and Gimenez staining; definitive identification to the species level was done by 16S rRNA PCR followed by sequencing. We identified 32 cell culture contaminants: 9 were coagulase-negative Staphylococcus, 1 was Alcaligenes faecalis, 6 were Micrococcus sp., 1 was Sphingomonas paucimobilis, 9 were Propionibacterium sp., and 6 were fungi. One hundred seventy-five microorganisms were obtained: 1 from the 56 inoculated lung biopsies, 2 from the 156 inoculated bone marrow, 2 from the 7 inoculated digestive biopsies, 5 from the 352 inoculated CSF samples, 5 from the 261 inoculated cardiac valves, 4 from the 198 inoculated aqueous humor fluid samples, 10 from the 1,599 inoculated blood samples, 10 from the 43 inoculated abscess specimens, 15 from the 140 inoculated skin biopsies, and 90 from the 432 inoculated lymph nodes. We also detected bacteria from 1 vascular aneurysm, 1 spleen biopsy, 2 liver biopsies, 6 bone biopsies, 6 pericardial biopsies and 15 synovial fluid samples, but the total of samples inoculated were not available for all of these samples. Among the 175 bacterial specimens detected at day 21 (date of cell culture reading) using the JNSP protocol, we have analyzed more precisely those results concerning Staphylococcus aureus, Streptococcus sp., and related genera and Mycobacterium sp. These results are summarized in Tables 4 and 5. Indeed, we have isolated 29 strains of Staphylococcus aureus by using shell-vial culture. Of these, 27 had been isolated using axenic medium, with the 2 remaining strains (7%) being isolated in cell culture alone. However, this difference was not statistically significant (P = 0.24). Both strains were secondarily established in axenic medium.
Twenty-six isolates of Streptococcus sp. and related genera were isolated using shell-vial culture, of which only 19 strains had been isolated using axenic medium. Of these 19 strains, 1 was Streptococcus F, 1 was Streptococcus G, 3 were Streptococcus A, 1 was Enterococcus faecium, 2 were Streptococcus bovis, 1 was Streptococcus pneumoniae, 1 was Gemella hemolysans, 1 was Streptococcus oralis, 1 was Streptococcus sanguis, 3 were Streptococcus mitis, and 4 were Streptococcus sp. Among the seven strains only isolated using shell-vial culture (24%), two Streptococcus B strains and one strain each of Gemella sanguinis, Streptococcus pneumoniae, and Streptococcus sp. were observed. This difference was statistically significant (P = 0.01). For five of the seven strains, attempts were made to establish in axenic medium (Table 4 and 5). We isolated a total of 56 Mycobacterium sp. strains using shell-vial culture. Fifty-two samples were cultured using both conventional techniques (8) and shell-vial assay. Forty-six of these 52 samples were positive for Mycobacterium sp. using conventional methods. Thus, 11.5% of the isolates of Mycobacterium sp. were only cultured in shell-vial assay. This difference was statistically significant (P = 0.01). Forty-nine Mycobacterium sp. isolates cultured in cell culture were also established in axenic medium. Only one isolate could not be established, and two others were not retested.
Fifty-five other isolates of microorganisms had also been cultured by the JNSP protocol in axenic medium. However, only four strains could be successfully established (Tables 4 and 5). We have isolated such fastidious microorganisms as Nocardia sp. and Actinomyces sp. on shell-vial culture, whereas axenic medium sheep blood agar medium was sterile. We isolated Francisella tularensis (13) from a skin biopsy and from a lymph node and Brucella melitensis (48) from a liver biopsy. These bacteria could only be cultured using cell culture, and isolation on standard chocolate agar failed. Legionella pneumophila (27) was isolated from a lung biopsy, whereas bacterial growth failed on buffered charcoal-yeast extract (BCYE) agar. We were able to isolate other strict intracellular bacteria such as Chlamydia trachomatis (35) from a lymph node and Trophyrema whipplei in a cardiac valve (Table 5). Before 2000, isolation and establishment of the T. whipplei strain from clinical samples were done using the JNSP protocol (10, 39). Since then, this protocol has been adapted for culturing T. whipplei.
T. whipplei shell-vial protocol. Among the 88 human samples inoculated with this adapted shell-vial protocol, T. whipplei was detected in 11 human samples by immunofluorescence and Gimenez staining; definitive identification was systematically confirmed by specific PCR and sequencing. We detected T. whipplei in 4 of the 9 cardiac valves, in 2 of the 18 duodenal biopsies, in the 1 aqueous humor fluid sample, in 2 of the 17 blood samples, in 1 of the 5 CSF samples, and in 1 synovial fluid. We were able to established only three strains of T. whipplei: one from a cardiac valve sample, one from a duodenal biopsy sample, and one from a blood sample.
DISCUSSION
Since 1991, the number of samples received by our laboratory, as a diagnosis and research laboratory, for cell culture has increased. In the beginning, we were particularly interested in strict intracellular bacteria, and we have acquired good experience in this domain. Culture of strict intracellular bacteria is delicate work, and cell culture is an obligatory step for their isolation. Differences were sometimes observed between primary isolated strains and established strains. A successful isolation did not always lead to successful propagation of the isolate, as published in previous studies (28, 29). It is important to underline that the difference in the number of established strains of Rickettsia sp. (70%) in this study is statistically significant in comparison to the 44% reported in a previous study (P = 0.02) (28). However, this technique is still only suitable for clinical laboratories with cell culture and biohazard facilities.
Cell culture also seems to be an effective tool for the isolation of fastidious bacteria such as L. pneumophila, F. tularensis, B. melitensis, and B. henselae or B. quintana, as previously reported (13, 27, 29, 48). Herein, we have reported that positive cell culture has also been successfully obtained for other fastidious microorganisms such as Actinomyces sp., N. asteroides, and Mycobacterium sp. (14), whereas culture using adapted conventional medium was negative. Overall, the shell-vial culture assay was applied to a lot of samples with a very low rate of success (580/11,083 [5%]), but it is important to consider that, in some cases, the technique was also very helpful for isolation of nonfastidious microorganisms from clinical samples. For example, S. aureus is usually considered as a nonfastidious microorganism. However, some studies suggest that the possible intracellular location of S. aureus may be responsible for negative cultures on axenic medium (54). Besides, S. aureus small colony variants are characterized by their fastidious growth and the atypical morphology of theirs colonies on routine media, making recovery as well as correct identification difficult for microbiological laboratories (49). Indeed, their ability to interrupt electron transport and to form variant subpopulations affords S. aureus a number of survival advantages, including the ability of a subpopulation to persist in an intracellular situation within nonprofessional phagocytes (2, 50). Our data suggests that cell culture may be considered as an additional tool for helping to cultivate fastidious S. aureus. Surprisingly, a significant difference was observed for the isolation of Streptococcus sp. and related genera in cell culture in comparison to axenic medium. Streptococcus sp. isolates usually grow easily in axenic blood agar medium (38). However, invasive bacteria can survive in intracellular vesicles when phagosome-lysosome fusion does not occur and can multiply and eventually spread to adjacent cells (9). For example, Streptococcus A is resistant to phagocytosis (37), and a recent study has shown that this bacterium is able to trigger its own entry into nonphagocytic cells (36). Another study has demonstrated that some strains of Streptococcus B could survive in macrophages (7). Our data suggest that some streptococci may be more easily isolated by cell culture, but this needs to be confirmed.
Overall, our study indicates that shell-vial culture is rarely positive. However, in cases of persistent or recurrent infections with poor clinical response to standard antimicrobial therapy or in the absence of bacterial growth using conventional methods, this method (41) performed in a reference center could improve the accuracy of microbiological diagnosis and consequently the management of the patient. Thus, cell culture provides supplemental tools to elucidate the cause of microbial diseases when PCR and classical agar procedures are negative. Besides and interestingly, the attempts at establishment of one strain each of B. melitensis, S. pneumoniae, and Mycobacterium lentiflavum have only been possible on cell culture and have systematically failed in standard axenic media. These observations are quite difficult to explain. Apparently the cells are a necessary prerequisite since the organism is incapable of growing in the cell-free medium. Maybe there are some strains from of these bacteria with defects in some metabolic pathways compared with the majority of strains in the same species.
Recently, cell culture has been challenged by the development of genotype-based methods such as broad-range PCR (46). Indeed, cell culture is less sensitive than PCR. Besides hypothesizing the existence of "uncultivable" microorganisms, several authors have emphasized the critical role of molecular tools for the investigation of emerging infectious diseases and have required a reassessment of Koch's postulate (17). However, cell culture has irreplaceable advantages allowing the isolation and establishment of new or "uncultivable" bacterial pathogens. For example, using the cell culture in shell vial, we have destroyed the myth that considered T. whipplei to be "uncultivable" (39). Since then, the isolation and establishment of this bacterium have led to new perspectives for the management and treatment of Whipple disease (1, 12, 30, 31, 33, 34, 42). Also, to start with, T. whipplei was considered to be a hypothetical intracellular bacterium (25). The successful cell culture of the microorganism has allowed the complete sequencing of its genome (4, 43). Its analysis has shown multiple defects on several amino acid metabolic pathways. The supplementation of axenic culture medium with the missing amino acid has allowed the successful growth of T. whipplei on this medium (47). In our view, it is dangerous to neglect microorganism isolation, although as molecular diagnosis expands, the facilities for cell culture and isolation work may become more centralized to retain expertise and to provide the range and quality of service required (41). Our experience for Rickettsia sp. culture has allowed us to obtain isolates from either the tick or the patient and describe new rickettsial disease. Cell culture is the ultimate goal in rickettsial description. For example R. conorii was the only spotted fever group rickettsia reported in Africa. Using cell culture, the isolation of an R. africae strain from Amblyomma hebraeum ticks in Zimbabwe (23) and then from blood collected from a suspected patient using shell-vial culture allowed the description of the etiological agent of African tick bite fever (24). When rickettsial strains of unknown pathogenicity are first isolated in ticks, such as R. slovaca (3) and R. sibirica subsp. mongolitimonae (53), and second through culture of the rickettsial strains from patient isolates (6, 40), this can provide definitive evidence that these rickettsiae are human pathogens.
Finally, despite the fact that shell-vial culture is regularly performed in our laboratory, the number of isolated strains is small but precious for microbiological diagnosis. Culture remains a vital clue for the characterization of pathogens and the management of infectious diseases. The shell-vial method is invaluable in allowing the establishment of new clinical isolates for basic research, especially with regard to the obligate Rickettsia sp. and Coxiella burnetii. The use of genomic amplification in the clinical laboratory may be a useful technique, but as a supplemental and complementary method, and could never substitute for culture of the microorganism.
ACKNOWLEDGMENTS
We thank Kelly Johnston for reviewing the manuscript.
REFERENCES
Baisden, B. L., H. Lepidi, D. Raoult, P. Argani, J. H. Yardley, and J. S. Dumler. 2002. Diagnosis of Whipple disease by immunohistochemical analysis: a sensitive and specific method for the detection of Tropheryma whipplei (the Whipple bacillus) in paraffin-embedded tissue. Am. J. Clin. Pathol. 118:742-748.
Balwit, J. M., P. Van Langevelde, J. M. Vann, and R. A. Proctor. 1994. Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells. J. Infect. Dis. 170:1033-1037.
Beati, L., J. P. Finidori, and D. Raoult. 1993. First isolation of Rickettsia slovaca from Dermacentor marginatus in France. Am. J. Trop. Med. Hyg. 48:257-268.
Bentley, S. D., M. Maiwald, L. D. Murphy, M. J. Pallen, C. A. Yeats, L. G. Dover, H. T. Norbertczak, G. S. Besra, M. A. Quail, D. E. Harris, A. Von Herbay, A. Goble, S. Rutter, R. Squares, S. Squares, B. G. Barrell, J. Parkhill, and D. A. Relman. 2003. Sequencing and analysis of the genome of the Whipple's disease bacterium Tropheryma whipplei. Lancet 361:637-644.
Birg, M.-L., B. La Scola, V. Roux, P. Brouqui, and D. Raoult. 1999. Isolation of Rickettsia prowazekii from blood by shell vial cell culture. J. Clin. Microbiol. 37:3722-3724.
Cazorla, C., M. Enea, F. Lucht, and D. Raoult. 2003. First isolation of Rickettsia slovaca from a patient, France. Emerg. Infect. Dis. 9:135.
Cornacchione, P., L. Scaringi, K. Fettucciari, E. Rosati, R. Sabatini, G. Orefici, C. von Hunolstein, A. Modesti, A. Modica, F. Minelli, and P. Marconi. 1998. Group B streptococci persist inside macrophages. Immunology 93:86-95.
Drancourt, M., P. Carrieri, M.-J. Gevaudan, and D. Raoult. 2003. Blood agar and Mycobacterium tuberculosis: the end of a dogma. J. Clin. Microbiol. 41:1710-1711.
Falkow, S. 1991. Bacterial entry into eukaryotic cells. Cell 65:1099-1102.
Fenollar, F., M.-L. Birg, V. Gauduchon, and D. Raoult. 2003. Culture of Tropheryma whipplei from human samples: a 3-year experience (1999 to 2002). J. Clin. Microbiol. 41:3816-3822.
Fenollar, F., P. E. Fournier, D. Raoult, R. Gerolami, H. Lepidi, and C. Poyart. 2002. Quantitative detection of Tropheryma whipplei DNA by real-time PCR. J. Clin. Microbiol. 40:1119-1120. (Letter.)
Fenollar, F., and D. Raoult. 2001. Molecular techniques in Whipple's disease. Expert Rev. Mol. Diagn. 1:299-309.
Fournier, P.-E., L. Bernabeu, B. Schubert, M. Mutillod, V. Roux, and D. Raoult. 1998. Isolation of Francisella tularensis by centrifugation of shell vial cell culture from an inoculation eschar. J. Clin. Microbiol. 36:2782-2783.
Fournier, P. E., M. Drancourt, H. Lepidi, M. J. Gevaudan, and D. Raoult. 2000. Isolation of mycobacteria from clinical samples using the centrifugation-shell vial technique. Eur. J. Clin. Microbiol. Infect. Dis. 19:69-70.
Fournier, P.-E., H. Fujita, N. Takada, and D. Raoult. 2002. Genetic identification of rickettsiae isolated from ticks in Japan. J. Clin. Microbiol. 40:2176-2181.
Fournier, P. E., H. Tissot-Dupont, H. Gallais, and D. Raoult. 2000. Rickettsia mongolotimonae: a rare pathogen in France. Emerg. Infect. Dis. 6:290-292.
Fredricks, D. N., and D. A. Relman. 1996. Sequence-based identification of microbial pathogens: a reconsideration of Koch's postulates. Clin. Microbiol. Rev. 9:18-33.
Gimenez, D. F. 1964. Staining rickettsiae in yolk-sac cultures. Stain Technol. 39:135-140.
Gleaves, C. A., T. F. Smith, E. A. Shuster, and G. R. Pearson. 1985. Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J. Clin. Microbiol. 21:217-221.
Hoover, T. A., M. H. Vodkin, and J. C. Williams. 1992. A Coxiella burnetii repeated DNA element resembling a bacterial insertion sequence. J. Bacteriol. 174:5540-5548.
Houpikian, P., and D. Raoult. 2002. Traditional and molecular techniques for the study of emerging bacterial diseases: one laboratory's perspective. Emerg. Infect. Dis. 8:122-131.
Hudson, J. B., V. Misra, and T. R. Mosmann. 1976. Cytomegalovirus infectivity: analysis of the phenomenon of centrifugal enhancement of infectivity. Virology 72:235-243.
Kelly, P. J., L. Beati, P. R. Mason, L. A. Matthewman, V. Roux, and D. Raoult. 1996. Rickettsia africae sp. nov., the etiological agent of African tick bite fever. Int. J. Syst. Bacteriol. 46:611-614.
Kelly, P. J., L. A. Matthewman, L. Beati, D. Raoult, P. Mason, M. Dreary, and R. Makombe. 1992. African tick-bite fever: a new spotted fever group rickettsiosis under an old name. Lancet 340:982-983.
La Scola, B., F. Fenollar, P. E. Fournier, M. Altwegg, M. N. Mallet, and D. Raoult. 2001. Description of Tropheryma whipplei gen.nov., sp.nov., the Whipple's disease bacillus. Int. J. Syst. Evol. Microbiol. 51:1471-1479.
La Scola, B., G. Michel, and D. Raoult. 1997. Use of amplification and sequencing of the 16S rRNA gene to diagnose Mycoplasma pneumoniae osteomyelitis in a patient with hypogammaglobulinemia. Clin. Infect. Dis. 24:1161-1163.
La Scola, B., G. Michel, and D. Raoult. 1999. Isolation of Legionella pneumophila by centrifugation of shell vial cell cultures from multiple liver and lung abscesses. J. Clin. Microbiol. 37:785-787.
La Scola, B., and D. Raoult. 1996. Diagnosis of Mediterranean spotted fever by cultivation of Rickettsia conorii from blood and skin samples using the centrifugation-shell vial technique and by detection of R. conorii in circulating endothelial cells: a 6-year follow-up. J. Clin. Microbiol. 34:2722-2727.
La Scola, B., and D. Raoult. 1999. Culture of Bartonella quintana and Bartonella henselae from human samples: a 5-year experience (1993 to 1998). J. Clin. Microbiol. 37:1899-1905.
Lepidi, H., F. Fenollar, R. Gerolami, J. L. Mege, M. F. Bonzi, M. Chappuis, J. Sahel, and D. Raoult. 2003. Whipple's disease: immunospecific and quantitative immunohistochemical study of intestinal biopsy specimens. Hum. Pathol. 34:589-596.
Maiwald, M., A. Von Herbay, D. N. Fredricks, C. C. Ouverney, J. C. Kosek, and D. A. Relman. 2003. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J. Infect. Dis. 188:801-808.
Marrero, M., and D. Raoult. 1989. Centrifugation-shell vial technique for rapid detection of Mediterranean spotted fever rickettsia in blood culture. Am. J. Trop. Med. Hyg. 40:197-199.
Marth, T., and D. Raoult. 2003. Whipple's disease. Lancet 361:239-246.
Masselot, F., A. Boulos, M. Maurin, J. M. Rolain, and D. Raoult. 2003. Molecular evaluation of antibiotic susceptibility: Tropheryma whipplei paradigm. Antimicrob. Agents Chemother. 47:1658-1664.
Maurin, M., and D. Raoult. 2000. Isolation in endothelial cell cultures of Chlamydia trachomatis LGV (serovar L2) from a lymph node of a patient with suspected cat scratch disease. J. Clin. Microbiol. 38:2062-2064.
Molinari, G., and G. S. Chhatwal. 1999. Streptococcal invasion. Curr. Opin. Microbiol. 2:56-61.
Molinari, G., M. Rohde, C. A. Guzman, and G. S. Chhatwal. 2000. Two distinct pathways for the invasion of Streptococcus pyogenes in non-phagocytic cells. Cell Microbiol. 2:145-154.
Petts, D. N. 1984. Colistin-oxolinic acid-blood agar: a new selective medium for streptococci. J. Clin. Microbiol. 19:4-7.
Raoult, D., M. L. Birg, B. La Scola, P. E. Fournier, M. Enea, H. Lepidi, V. Roux, J. C. Piette, F. Vandenesch, D. Vital Durand, and T. J. Marrie. 2000. Cultivation of the bacillus of Whipple's disease. N. Engl. J. Med. 342:620-625.
Raoult, D., P. Brouqui, and V. Roux. 1996. A new spotted-fever-group rickettsiosis. Lancet 348:412.
Raoult, D., P. E. Fournier, and M. Drancourt. 2004. What does the future hold for clinical microbiology Nat. Rev. Microbiol. 2:151-159.
Raoult, D., B. La Scola, P. Lecocq, H. Lepidi, and P. E. Fournier. 2001. Culture and immunological detection of Tropheryma whippleii from the duodenum of a patient with Whipple disease. JAMA 285:1039-1043.
Raoult, D., H. Ogata, S. Audic, C. Robert, K. Suhre, M. Drancourt, and J. M. Claverie. 2003. Tropheryma whipplei twist: a human pathogenic actinobacteria with a reduced genome. Genome Res. 13:1800-1809.
Raoult, D., and V. Roux. 1997. Rickettsioses as paradigms of new or emerging infectious diseases. Clin. Microbiol. Rev. 10:694-719.
Raoult, D., G. Vestris, and M. Enea. 1990. Isolation of 16 strains of Coxiella burnetii from patients by using a sensitive centrifugation cell culture system and establishment of strains in HEL cells. J. Clin. Microbiol. 28:2482-2484.
Relman, D. A. 1999. The search for unrecognized pathogens. Science 284:1308-1310.
Renesto, P., N. Crapoulet, H. Ogata, B. La Scola, G. Vestris, J. M. Claverie, and D. Raoult. 2003. Genome-based design of a cell-free culture medium for Tropheryma whipplei. Lancet 362:447-449.
Rovery, C., J. M. Rolain, D. Raoult, and P. Brouqui. 2003. Shell vial culture as a tool for isolation of Brucella melitensis in chronic hepatic abscess. J. Clin. Microbiol. 41:4460-4461.
Seifert, H., H. Wisplinghoff, P. Schnabel, and C. von Eiff. 2003. Small colony variants of Staphylococcus aureus and pacemaker-related infection. Emerg. Infect. Dis. 9:1316-1318.
Vaudaux, P., P. Francois, C. Bisognano, W. L. Kelley, D. P. Lew, J. Schrenzel, R. A. Proctor, P. J. McNamara, G. Peters, and C. Von Eiff. 2002. Increased expression of clumping factor and fibronectin-binding proteins by hemB mutants of Staphylococcus aureus expressing small colony variant phenotypes. Infect. Immun. 70:5428-5437.
Vestris, G., J. M. Rolain, P. E. Fournier, M. L. Birg, M. Enea, J. Y. Patrice, and D. Raoult. 2003. Seven years' experience of isolation of Rickettsia spp. from clinical specimens using the shell vial cell culture assay. Ann. N. Y. Acad. Sci. 990:371-374.
Weiss, E., and H. R. Dressler. 1960. Centrifugation of rickettsiae and viruses into cells and its effect on infection. Proc. Soc. Exp. Biol. Med. 103:691-695.
Yu, X., Y. Jin, M. Fan, G. Xu, Q. Liu, and D. Raoult. 1993. Genotypic and antigenic identification of two new strains of spotted fever group rickettsiae isolated from China. J. Clin. Microbiol. 31:83-88.
Zannier, A., M. Drancourt, J. P. Franceschi, J. M. Aubaniac, and D. Raoult. 1991. Interêt d'une technique de lyse cellulaire par choc thermique dans l'isolement des bacteries responsables d'infections osteo-articulaires. Pathol. Biol. 39:543-546.
Zeaiter, Z., P. E. Fournier, H. Ogata, and D. Raoult. 2002. Phylogenetic classification of Bartonella species by comparing groEL sequences. Int. J. Syst. Evol. Microbiol. 52:165-171.(Frederique Gouriet, Flore)
ABSTRACT
The shell-vial culture assay is performed routinely in our laboratory. Recently we revisited our experience of using the shell-vial culture assay for the isolation of microorganisms from various clinical samples. Over a 13-year period, we have isolated 580 bacterial strains (5%) from 11,083 clinical samples tested. Over the same period, 285 isolates of rickettsiae, bartonellae, or Coxiella burnetii were cultured from a total of 7,102 samples tested. These isolates include 55 Rickettsia sp. isolates, 95 Coxiella burnetii isolates, and 135 Bartonella sp. isolates. Based on our experience with the growth of fastidious microorganisms, we have used a centrifugation shell-vial technique called JNSP, for "je ne sais pas" ("I don't know [what I am growing]") for the isolation of other microorganisms. A total of 173 isolates were cultured from the 3,861 clinical samples tested using the JNSP method. Of these, 40 isolates had not been grown before on usual axenic medium. These include 2 Staphylococcus aureus isolates, 7 isolates of Streptococcus sp. and related genera, 6 Mycobacterium sp. isolates, 1 Nocardia asteroides isolate, 1 Actinomyces sp. isolate, 1 Brucella melitensis isolate, 2 Francisella tularensis isolates, 1 Mycoplasma pneumoniae isolate, and 1 Legionella pneumophila isolate. Using this protocol, we have also cultured intracellular bacteria such as Chlamydia trachomatis and we have performed the first culture and establishment of Trophyrema whipplei. Applied in our laboratory, the shell-vial culture generally exhibits a low rate of success. However, in some cases, this technique allowed microbial diagnosis when classical agar procedure and PCR were negative.
INTRODUCTION
The spectrum of infectious diseases is wide and changing. The reliable diagnosis of infectious disease remains a difficult problem even for specialized laboratories. Isolation of new microorganisms will allow the description of clinical features of new diseases and the characterization of new pathogens, enabling genetic descriptions, physiological analyses, improvement of diagnostic tools, and antibiotic susceptibility testing for bacteria. Therefore, the isolation of infecting bacteria is not only a mean of diagnosis but also a basis for the evaluation of much needed improved diagnostic assays and a route to enhance understanding of the diversity and the epidemiology of infections (21). The successful isolation of fastidious microorganisms is often based on cell culture (21). The success of this technique is based on two critical points. First, the ratio of microorganisms to cells should be as high as possible (21). Second, centrifugation enhances the adhesion of intracellular microorganisms to the cells (52). In 1976, a centrifugation-cell microculture system, called the shell-vial assay, was first used for the diagnosis of viral disease due to cytomegalovirus and allowed early antigen detection (22). This diagnostic tool is becoming one of the most commonly used techniques in clinical virology laboratories since it reduces turnaround time (19).
For a long time, Rickettsia spp., which are strict intracellular gram-negative bacteria, were isolated using embryonated egg and/or animal inoculations (44). As the National Reference Center for Rickettsiosis in France, our laboratory started to adapt the shell-vial assay for bacteria culture in 1986 (32). This technique, due to the highly infectious nature of such bacteria, is routinely performed in a biosafety level 3 laboratory and has allowed us to isolate first Rickettsia sp. (32) and second Coxiella burnetii (45) from various specimens. For several years, our laboratory has developed the facilities and experience for cultivating fastidious microorganisms, using this versatile cell culture system. This has allowed us to receive a large panel of samples from France and abroad, and, notably, this approach has allowed the isolation and the establishment of the first strain of the bacillus of Whipple's disease, Tropheryma whipplei, from a cardiac valve (39). From September 1991 to August 2003, our laboratory attempted to routinely achieve isolation of microorganisms from various clinical samples, arthropods, and animals by the shell-vial assay. Using this technique, we were able to isolate not only rickettsial species (28, 29, 32, 51) and Coxiella burnetii (45) but also fastidious bacteria, such as Francisella tularensis (13), Legionella pneumophila (27), Brucella melitensis (48), Chlamydia trachomatis (35), and T. whipplei (39). In this study, we report our complete 13-year experience of shell-vial culture assay.
MATERIALS AND METHODS
Specimen collection. From September 1991 to August 2003, our laboratory received 18,124 samples for culture from France and abroad with presumptive diagnoses of infectious diseases due to fastidious microorganisms. We have collected and managed the samples according to the patient's clinical presentation. If necessary, we further contact the sender for additional or more suitable samples to obtain relevant clinical and epidemiological data. Samples from hospitals outside of Marseille were frozen at –80°C before transport in dry ice. Samples, transport means, and diagnostic techniques are summarized in Fig. 1.
Axenic culture. For the samples from Marseille, all specimens were divided into three equal parts: one for conventional axenic culture, one for shell-vial culture, and one for PCR assays. Inoculation procedures for axenic culture were then performed in our local bacteriological laboratory. Thus, both axenic and cell cultures were processed in parallel. All the methods for attempted isolation of pathogens, including those for Brucella sp., Francisella sp., Legionella sp., and Bartonella sp., are summarized in Table 1. For the samples received from outside Marseille, we divided them into two parts: one for shell-vial culture and one for PCR. The axenic culture was performed in the local bacterial laboratory. Results were retrospectively sent to us. Thus, for these cases, axenic and cell cultures could not be processed in parallel.
Cell culture. Five different shell-vial protocols have been developed in our laboratory: one for the specific isolation of rickettsiae, one for the specific isolation of Bartonella sp., one for the specific isolation of Coxiella burnetii, one for the specific isolation of Trophyrema whipplei, and one for the unspecific research of other strictly or facultative intracellular bacteria, called the "JNSP" protocol. The abbreviation "JNSP" is derived from the French sentence "Je ne sais pas," which means "I do not know" and stands for "I do not know what I am growing." Cell type, inoculation, length of incubation, and culture revelation differed according to the suspected pathogen. Most often, human embryonic lung (HEL) fibroblasts are used because they have the advantage that once a monolayer is established, contact inhibition prevents further division and the cell can be used for prolonged incubation. All the procedures are summarized in Table 2. Briefly, culture was performed using the centrifugation-shell-vial technique (Sterilin-Felthan-England, 3.7 ml) using 12-mm round coverslips seeded with 1 ml of medium containing 50,000 cells and incubated in a 5% CO2 incubator at 37°C for 3 days to obtain a confluent monolayer (32).
Inoculation procedure. All manipulations were done within a laminar flow hood except the centrifugation. One milliliter of supernatant from each sample was mixed with 1 ml of RPMI 1640 medium (Gibco BRL life Technologies, Cergy Pontoise, France). The suspension was used to inoculate three shell vials with a centrifugation step of 700 g/min for 1 h in Eagle's minimal essential medium (MEM; Gibco) at 22°C. The inoculum was removed, and the shell vials were washed twice with sterile phosphate-buffered saline (PBS). The cells were then incubated in MEM with fetal calf serum (FCS; Sigma, St. Louis, MO) and L-glutamine at 37°C under a 5% CO2 atmosphere. All cell lines and culture reagents were checked weekly for bacterial contamination. The culture medium was removed at different times according to the samples and the bacteria as described in Table 1.
Detection and identification of bacteria. The detection procedure was performed at different days of incubation according to the specimens (Fig. 1). Detection of growing bacteria was assessed on coverslips directly inside the shell vial. We have used acridine orange, Gimenez (18), or immunofluorescence (32) staining. Gram, Giemsa, periodic acid-Schiff (PAS), and Ziehl-Neelsen staining could also be performed, depending on the suspected microorganism. For immunofluorescence staining and fixation with methanol, 100 μl of homemade rabbit polyclonal antiserum (anti-Coxiella burnetii, Bartonella sp., and Rickettsia sp.) or patient's serum diluted respectively to 1:200 and 1:100 in PBS with 3% nonfat dry milk was added. The slides were incubated at 37°C for 30 min. After three washes in PBS, 100 μl of a fluorescein-conjugated goat anti-rabbit immunoglobulin serum (Jackson Immunoresearch Laboratories, West Grove, Pa.) diluted 1:200 in PBS containing 0.2% Evans blue was incubated at 37°C for 30 min. After three washes with PBS, the coverslips were mounted (cells facedown) in phosphate-buffered glycerol medium (pH 8) and examined at a magnification of x400 with an epifluorescence microscope (Zeiss, Thornwood, NY). If immunofluorescence or other staining was positive, identification to the species level was performed by using the 16S rRNA gene (26) or PCR specific for Bartonella sp., Rickettsia sp., Coxiella burnetii, and Tropheryma whipplei as described elsewhere (11, 15, 20, 55). DNA was extracted from both the isolate and the fresh biopsy specimen, using the QIAmp tissue kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. Sequencing was performed using a commercially available sequencing kit (dRhodamine terminator cycle sequencing kit; PE Applied Biosystems, Warrington, England) according to the manufacturer's recommendations (26). Sequences obtained from the isolates were aligned with the sequences in the GenBank DNA sequence database with the program BLAST (version 2.0; National Center for Biotechnology Information).
Interpretation of cell culture. A shell-vial culture was defined as positive if a pathogenic microorganism was isolated. If nonpathogenic or opportunistic microorganisms such as coagulase-negative Staphylococcus sp., Micrococcus sp., fungi, or nonfermentative gram-negative bacilli were isolated, the same microorganism would have to be detected in parallel with axenic culture and/or PCR assays in order to conclude that it was a positive cell culture. If these microorganisms were not detected with another test, they were considered as cell culture contaminants.
Establishment of a strain. In order to establish the isolate, the supernatant of a positive shell vial was inoculated onto confluent layers of various cell types according to the isolate in a 25-cm2 culture flask.
Statistical analysis. To compare the data, we used the Fisher's test using Epi Info version 6.04a (Centers for Disease Control and Prevention, Atlanta, GA). P < 0.05 was considered statistically significant.
RESULTS
From August 1991 to September 2003, we received a total of 18,124 human samples and a total of 384 animal samples in our laboratory. The samples received were inoculated using one or more of the five protocols developed in our laboratory. Each of these adapted protocols was chosen according to the clinical information obtained from the clinician. More precisely, 10,699 human samples were inoculated in a cell culture using the shell-vial assay. Of these, 3,536 were inoculated using the Coxiella burnetii protocol, 2,619 using the Bartonella sp. protocol, 949 using the Rickettsia sp. protocol, 3,861 using the JNSP protocol, and 88 using the new Trophyrema whipplei protocol. Samples were obtained from 5,764 blood specimens, 984 lymph nodes, 722 cardiac valves, 567 skin biopsies, 431 cerebrospinal fluid (CSF) samples, 203 vitreous humor fluid samples, 282 placentas, 189 bone marrow specimens, 155 abscesses from diverse sites, 79 lung biopsies, 31 digestive biopsies, and 1,292 specimens from various other sites. The latter group included samples from bones, synovial and pericardic fluids, or milk. Three hundred eighty-four animal samples were inoculated into shell-vial assay: 36 samples were inoculated using the Coxiella burnetii protocol, 297 using the Bartonella sp. protocol, 32 using the Rickettsia sp. protocol, and 19 using the JNSP protocol. For these 13 years, the global results obtained with shell-vial culture are summarized in Fig. 2. Note that three times more samples were analyzed in 2002 than in 1992 (1,290 versus 421). In parallel, we observed that the number of positive cultures increased from 18 in 1992 to 117 in 2002. Results for the shell-vial culture using the JNSP protocol from January 1996 to September 2003 are shown in Fig. 3.
Coxiella burnetii shell-vial culture protocol. During these 13 years, among the 3,536 human samples (Table 3) inoculated using this protocol, Coxiella burnetii was detected in 91 samples (45) by immunofluorescence and Gimenez staining and definitive identification was systematically confirmed by specific PCR and sequencing. We have detected Coxiella burnetii in 47 of the 351 inoculated cardiac valves, 31 of the 2,338 inoculated blood samples, 1 of the 22 inoculated lymph nodes, and 8 of the 280 inoculated placentas. We detected Coxiella burnetii in two spleen biopsies and six vascular samples, but the total of the inoculated samples were not available. Among the 36 animal samples, Coxiella burnetii was detected in 4 samples (Table 2). These samples included one mouse spleen, two vaginal specimens from ewes, and one goat placenta specimen. A total of 95 strains of Coxiella burnetii were detected. We were able to establish 81 strains obtained from human samples and 4 strains obtained from animal samples, corresponding to 93% of the strains isolated in shell-vial culture.
Bartonella sp. shell-vial culture protocol. Among the 2,619 samples inoculated with this protocol (29), Bartonella sp. was detected in 88 human samples by immunofluorescence and Gimenez staining; the identification to the species level was systematically done by specific PCR and sequencing. More precisely, we have detected Bartonella sp. in 55 of 1,244 inoculated blood samples, 3 of the 101 inoculated cardiac valves, 3 of the 16 inoculated medullar aspirations, 2 of the 518 inoculated lymph nodes, and 2 of the 90 inoculated skin biopsies. Among the 297 animal samples inoculated, 47 Bartonella sp. isolates were detected (Table 3). These samples were obtained from cat's blood. A total of 135 strains of Bartonella sp. were detected. We were able to establish 113 strains (83%). Sixty-five established strains were obtained from human samples: 62 strains of Bartonella quintana and 3 strains of Bartonella henselae. All 47 strains obtained from animal samples were established: 27 Bartonella henselae strains, 19 Bartonella clarridgiae strains, and 1 Bartonella koehlerae strain.
Rickettsia sp. shell-vial culture protocol. Among the 949 samples inoculated with this protocol (51), Rickettsia sp. was detected in 52 human samples by immunofluorescence and Gimenez staining; definitive identification to the species level was systematically done by specific PCR and sequencing. More precisely, we detected Rickettsia sp. in 37 of the 337 inoculated skin biopsies and in 15 of the 510 inoculated blood samples. Among the 30 inoculated animal samples, we have detected Rickettsia sp. in 3 ticks and 2 lice. A total of 57 Rickettsia sp. isolates were detected, and 39 were established (70%) (Table 3). Thirty-six strains were obtained from human samples: 18 strains of Rickettsia conorii, 12 of Rickettsia africae, 4 of Rickettsia sibirica subsp. mongolitimonae (16, 40), 1 of Rickettsia prowazekii (5), and 1 of Rickettsia slovaca (6). We were able to establish two strains of R. slovaca and one of R. helvetica from ticks, but the two strains of R. prowazekii isolated from lice could not be established.
JNSP shell-vial culture protocol. In our laboratory, we have been performing the JNSP protocol almost routinely since 1996. Among the 3,861 samples inoculated with this protocol, we detected 175 bacterial species and 32 bacterial cell culture contaminants. These bacteria were detected by immunofluorescence and Gimenez staining; definitive identification to the species level was done by 16S rRNA PCR followed by sequencing. We identified 32 cell culture contaminants: 9 were coagulase-negative Staphylococcus, 1 was Alcaligenes faecalis, 6 were Micrococcus sp., 1 was Sphingomonas paucimobilis, 9 were Propionibacterium sp., and 6 were fungi. One hundred seventy-five microorganisms were obtained: 1 from the 56 inoculated lung biopsies, 2 from the 156 inoculated bone marrow, 2 from the 7 inoculated digestive biopsies, 5 from the 352 inoculated CSF samples, 5 from the 261 inoculated cardiac valves, 4 from the 198 inoculated aqueous humor fluid samples, 10 from the 1,599 inoculated blood samples, 10 from the 43 inoculated abscess specimens, 15 from the 140 inoculated skin biopsies, and 90 from the 432 inoculated lymph nodes. We also detected bacteria from 1 vascular aneurysm, 1 spleen biopsy, 2 liver biopsies, 6 bone biopsies, 6 pericardial biopsies and 15 synovial fluid samples, but the total of samples inoculated were not available for all of these samples. Among the 175 bacterial specimens detected at day 21 (date of cell culture reading) using the JNSP protocol, we have analyzed more precisely those results concerning Staphylococcus aureus, Streptococcus sp., and related genera and Mycobacterium sp. These results are summarized in Tables 4 and 5. Indeed, we have isolated 29 strains of Staphylococcus aureus by using shell-vial culture. Of these, 27 had been isolated using axenic medium, with the 2 remaining strains (7%) being isolated in cell culture alone. However, this difference was not statistically significant (P = 0.24). Both strains were secondarily established in axenic medium.
Twenty-six isolates of Streptococcus sp. and related genera were isolated using shell-vial culture, of which only 19 strains had been isolated using axenic medium. Of these 19 strains, 1 was Streptococcus F, 1 was Streptococcus G, 3 were Streptococcus A, 1 was Enterococcus faecium, 2 were Streptococcus bovis, 1 was Streptococcus pneumoniae, 1 was Gemella hemolysans, 1 was Streptococcus oralis, 1 was Streptococcus sanguis, 3 were Streptococcus mitis, and 4 were Streptococcus sp. Among the seven strains only isolated using shell-vial culture (24%), two Streptococcus B strains and one strain each of Gemella sanguinis, Streptococcus pneumoniae, and Streptococcus sp. were observed. This difference was statistically significant (P = 0.01). For five of the seven strains, attempts were made to establish in axenic medium (Table 4 and 5). We isolated a total of 56 Mycobacterium sp. strains using shell-vial culture. Fifty-two samples were cultured using both conventional techniques (8) and shell-vial assay. Forty-six of these 52 samples were positive for Mycobacterium sp. using conventional methods. Thus, 11.5% of the isolates of Mycobacterium sp. were only cultured in shell-vial assay. This difference was statistically significant (P = 0.01). Forty-nine Mycobacterium sp. isolates cultured in cell culture were also established in axenic medium. Only one isolate could not be established, and two others were not retested.
Fifty-five other isolates of microorganisms had also been cultured by the JNSP protocol in axenic medium. However, only four strains could be successfully established (Tables 4 and 5). We have isolated such fastidious microorganisms as Nocardia sp. and Actinomyces sp. on shell-vial culture, whereas axenic medium sheep blood agar medium was sterile. We isolated Francisella tularensis (13) from a skin biopsy and from a lymph node and Brucella melitensis (48) from a liver biopsy. These bacteria could only be cultured using cell culture, and isolation on standard chocolate agar failed. Legionella pneumophila (27) was isolated from a lung biopsy, whereas bacterial growth failed on buffered charcoal-yeast extract (BCYE) agar. We were able to isolate other strict intracellular bacteria such as Chlamydia trachomatis (35) from a lymph node and Trophyrema whipplei in a cardiac valve (Table 5). Before 2000, isolation and establishment of the T. whipplei strain from clinical samples were done using the JNSP protocol (10, 39). Since then, this protocol has been adapted for culturing T. whipplei.
T. whipplei shell-vial protocol. Among the 88 human samples inoculated with this adapted shell-vial protocol, T. whipplei was detected in 11 human samples by immunofluorescence and Gimenez staining; definitive identification was systematically confirmed by specific PCR and sequencing. We detected T. whipplei in 4 of the 9 cardiac valves, in 2 of the 18 duodenal biopsies, in the 1 aqueous humor fluid sample, in 2 of the 17 blood samples, in 1 of the 5 CSF samples, and in 1 synovial fluid. We were able to established only three strains of T. whipplei: one from a cardiac valve sample, one from a duodenal biopsy sample, and one from a blood sample.
DISCUSSION
Since 1991, the number of samples received by our laboratory, as a diagnosis and research laboratory, for cell culture has increased. In the beginning, we were particularly interested in strict intracellular bacteria, and we have acquired good experience in this domain. Culture of strict intracellular bacteria is delicate work, and cell culture is an obligatory step for their isolation. Differences were sometimes observed between primary isolated strains and established strains. A successful isolation did not always lead to successful propagation of the isolate, as published in previous studies (28, 29). It is important to underline that the difference in the number of established strains of Rickettsia sp. (70%) in this study is statistically significant in comparison to the 44% reported in a previous study (P = 0.02) (28). However, this technique is still only suitable for clinical laboratories with cell culture and biohazard facilities.
Cell culture also seems to be an effective tool for the isolation of fastidious bacteria such as L. pneumophila, F. tularensis, B. melitensis, and B. henselae or B. quintana, as previously reported (13, 27, 29, 48). Herein, we have reported that positive cell culture has also been successfully obtained for other fastidious microorganisms such as Actinomyces sp., N. asteroides, and Mycobacterium sp. (14), whereas culture using adapted conventional medium was negative. Overall, the shell-vial culture assay was applied to a lot of samples with a very low rate of success (580/11,083 [5%]), but it is important to consider that, in some cases, the technique was also very helpful for isolation of nonfastidious microorganisms from clinical samples. For example, S. aureus is usually considered as a nonfastidious microorganism. However, some studies suggest that the possible intracellular location of S. aureus may be responsible for negative cultures on axenic medium (54). Besides, S. aureus small colony variants are characterized by their fastidious growth and the atypical morphology of theirs colonies on routine media, making recovery as well as correct identification difficult for microbiological laboratories (49). Indeed, their ability to interrupt electron transport and to form variant subpopulations affords S. aureus a number of survival advantages, including the ability of a subpopulation to persist in an intracellular situation within nonprofessional phagocytes (2, 50). Our data suggests that cell culture may be considered as an additional tool for helping to cultivate fastidious S. aureus. Surprisingly, a significant difference was observed for the isolation of Streptococcus sp. and related genera in cell culture in comparison to axenic medium. Streptococcus sp. isolates usually grow easily in axenic blood agar medium (38). However, invasive bacteria can survive in intracellular vesicles when phagosome-lysosome fusion does not occur and can multiply and eventually spread to adjacent cells (9). For example, Streptococcus A is resistant to phagocytosis (37), and a recent study has shown that this bacterium is able to trigger its own entry into nonphagocytic cells (36). Another study has demonstrated that some strains of Streptococcus B could survive in macrophages (7). Our data suggest that some streptococci may be more easily isolated by cell culture, but this needs to be confirmed.
Overall, our study indicates that shell-vial culture is rarely positive. However, in cases of persistent or recurrent infections with poor clinical response to standard antimicrobial therapy or in the absence of bacterial growth using conventional methods, this method (41) performed in a reference center could improve the accuracy of microbiological diagnosis and consequently the management of the patient. Thus, cell culture provides supplemental tools to elucidate the cause of microbial diseases when PCR and classical agar procedures are negative. Besides and interestingly, the attempts at establishment of one strain each of B. melitensis, S. pneumoniae, and Mycobacterium lentiflavum have only been possible on cell culture and have systematically failed in standard axenic media. These observations are quite difficult to explain. Apparently the cells are a necessary prerequisite since the organism is incapable of growing in the cell-free medium. Maybe there are some strains from of these bacteria with defects in some metabolic pathways compared with the majority of strains in the same species.
Recently, cell culture has been challenged by the development of genotype-based methods such as broad-range PCR (46). Indeed, cell culture is less sensitive than PCR. Besides hypothesizing the existence of "uncultivable" microorganisms, several authors have emphasized the critical role of molecular tools for the investigation of emerging infectious diseases and have required a reassessment of Koch's postulate (17). However, cell culture has irreplaceable advantages allowing the isolation and establishment of new or "uncultivable" bacterial pathogens. For example, using the cell culture in shell vial, we have destroyed the myth that considered T. whipplei to be "uncultivable" (39). Since then, the isolation and establishment of this bacterium have led to new perspectives for the management and treatment of Whipple disease (1, 12, 30, 31, 33, 34, 42). Also, to start with, T. whipplei was considered to be a hypothetical intracellular bacterium (25). The successful cell culture of the microorganism has allowed the complete sequencing of its genome (4, 43). Its analysis has shown multiple defects on several amino acid metabolic pathways. The supplementation of axenic culture medium with the missing amino acid has allowed the successful growth of T. whipplei on this medium (47). In our view, it is dangerous to neglect microorganism isolation, although as molecular diagnosis expands, the facilities for cell culture and isolation work may become more centralized to retain expertise and to provide the range and quality of service required (41). Our experience for Rickettsia sp. culture has allowed us to obtain isolates from either the tick or the patient and describe new rickettsial disease. Cell culture is the ultimate goal in rickettsial description. For example R. conorii was the only spotted fever group rickettsia reported in Africa. Using cell culture, the isolation of an R. africae strain from Amblyomma hebraeum ticks in Zimbabwe (23) and then from blood collected from a suspected patient using shell-vial culture allowed the description of the etiological agent of African tick bite fever (24). When rickettsial strains of unknown pathogenicity are first isolated in ticks, such as R. slovaca (3) and R. sibirica subsp. mongolitimonae (53), and second through culture of the rickettsial strains from patient isolates (6, 40), this can provide definitive evidence that these rickettsiae are human pathogens.
Finally, despite the fact that shell-vial culture is regularly performed in our laboratory, the number of isolated strains is small but precious for microbiological diagnosis. Culture remains a vital clue for the characterization of pathogens and the management of infectious diseases. The shell-vial method is invaluable in allowing the establishment of new clinical isolates for basic research, especially with regard to the obligate Rickettsia sp. and Coxiella burnetii. The use of genomic amplification in the clinical laboratory may be a useful technique, but as a supplemental and complementary method, and could never substitute for culture of the microorganism.
ACKNOWLEDGMENTS
We thank Kelly Johnston for reviewing the manuscript.
REFERENCES
Baisden, B. L., H. Lepidi, D. Raoult, P. Argani, J. H. Yardley, and J. S. Dumler. 2002. Diagnosis of Whipple disease by immunohistochemical analysis: a sensitive and specific method for the detection of Tropheryma whipplei (the Whipple bacillus) in paraffin-embedded tissue. Am. J. Clin. Pathol. 118:742-748.
Balwit, J. M., P. Van Langevelde, J. M. Vann, and R. A. Proctor. 1994. Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells. J. Infect. Dis. 170:1033-1037.
Beati, L., J. P. Finidori, and D. Raoult. 1993. First isolation of Rickettsia slovaca from Dermacentor marginatus in France. Am. J. Trop. Med. Hyg. 48:257-268.
Bentley, S. D., M. Maiwald, L. D. Murphy, M. J. Pallen, C. A. Yeats, L. G. Dover, H. T. Norbertczak, G. S. Besra, M. A. Quail, D. E. Harris, A. Von Herbay, A. Goble, S. Rutter, R. Squares, S. Squares, B. G. Barrell, J. Parkhill, and D. A. Relman. 2003. Sequencing and analysis of the genome of the Whipple's disease bacterium Tropheryma whipplei. Lancet 361:637-644.
Birg, M.-L., B. La Scola, V. Roux, P. Brouqui, and D. Raoult. 1999. Isolation of Rickettsia prowazekii from blood by shell vial cell culture. J. Clin. Microbiol. 37:3722-3724.
Cazorla, C., M. Enea, F. Lucht, and D. Raoult. 2003. First isolation of Rickettsia slovaca from a patient, France. Emerg. Infect. Dis. 9:135.
Cornacchione, P., L. Scaringi, K. Fettucciari, E. Rosati, R. Sabatini, G. Orefici, C. von Hunolstein, A. Modesti, A. Modica, F. Minelli, and P. Marconi. 1998. Group B streptococci persist inside macrophages. Immunology 93:86-95.
Drancourt, M., P. Carrieri, M.-J. Gevaudan, and D. Raoult. 2003. Blood agar and Mycobacterium tuberculosis: the end of a dogma. J. Clin. Microbiol. 41:1710-1711.
Falkow, S. 1991. Bacterial entry into eukaryotic cells. Cell 65:1099-1102.
Fenollar, F., M.-L. Birg, V. Gauduchon, and D. Raoult. 2003. Culture of Tropheryma whipplei from human samples: a 3-year experience (1999 to 2002). J. Clin. Microbiol. 41:3816-3822.
Fenollar, F., P. E. Fournier, D. Raoult, R. Gerolami, H. Lepidi, and C. Poyart. 2002. Quantitative detection of Tropheryma whipplei DNA by real-time PCR. J. Clin. Microbiol. 40:1119-1120. (Letter.)
Fenollar, F., and D. Raoult. 2001. Molecular techniques in Whipple's disease. Expert Rev. Mol. Diagn. 1:299-309.
Fournier, P.-E., L. Bernabeu, B. Schubert, M. Mutillod, V. Roux, and D. Raoult. 1998. Isolation of Francisella tularensis by centrifugation of shell vial cell culture from an inoculation eschar. J. Clin. Microbiol. 36:2782-2783.
Fournier, P. E., M. Drancourt, H. Lepidi, M. J. Gevaudan, and D. Raoult. 2000. Isolation of mycobacteria from clinical samples using the centrifugation-shell vial technique. Eur. J. Clin. Microbiol. Infect. Dis. 19:69-70.
Fournier, P.-E., H. Fujita, N. Takada, and D. Raoult. 2002. Genetic identification of rickettsiae isolated from ticks in Japan. J. Clin. Microbiol. 40:2176-2181.
Fournier, P. E., H. Tissot-Dupont, H. Gallais, and D. Raoult. 2000. Rickettsia mongolotimonae: a rare pathogen in France. Emerg. Infect. Dis. 6:290-292.
Fredricks, D. N., and D. A. Relman. 1996. Sequence-based identification of microbial pathogens: a reconsideration of Koch's postulates. Clin. Microbiol. Rev. 9:18-33.
Gimenez, D. F. 1964. Staining rickettsiae in yolk-sac cultures. Stain Technol. 39:135-140.
Gleaves, C. A., T. F. Smith, E. A. Shuster, and G. R. Pearson. 1985. Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J. Clin. Microbiol. 21:217-221.
Hoover, T. A., M. H. Vodkin, and J. C. Williams. 1992. A Coxiella burnetii repeated DNA element resembling a bacterial insertion sequence. J. Bacteriol. 174:5540-5548.
Houpikian, P., and D. Raoult. 2002. Traditional and molecular techniques for the study of emerging bacterial diseases: one laboratory's perspective. Emerg. Infect. Dis. 8:122-131.
Hudson, J. B., V. Misra, and T. R. Mosmann. 1976. Cytomegalovirus infectivity: analysis of the phenomenon of centrifugal enhancement of infectivity. Virology 72:235-243.
Kelly, P. J., L. Beati, P. R. Mason, L. A. Matthewman, V. Roux, and D. Raoult. 1996. Rickettsia africae sp. nov., the etiological agent of African tick bite fever. Int. J. Syst. Bacteriol. 46:611-614.
Kelly, P. J., L. A. Matthewman, L. Beati, D. Raoult, P. Mason, M. Dreary, and R. Makombe. 1992. African tick-bite fever: a new spotted fever group rickettsiosis under an old name. Lancet 340:982-983.
La Scola, B., F. Fenollar, P. E. Fournier, M. Altwegg, M. N. Mallet, and D. Raoult. 2001. Description of Tropheryma whipplei gen.nov., sp.nov., the Whipple's disease bacillus. Int. J. Syst. Evol. Microbiol. 51:1471-1479.
La Scola, B., G. Michel, and D. Raoult. 1997. Use of amplification and sequencing of the 16S rRNA gene to diagnose Mycoplasma pneumoniae osteomyelitis in a patient with hypogammaglobulinemia. Clin. Infect. Dis. 24:1161-1163.
La Scola, B., G. Michel, and D. Raoult. 1999. Isolation of Legionella pneumophila by centrifugation of shell vial cell cultures from multiple liver and lung abscesses. J. Clin. Microbiol. 37:785-787.
La Scola, B., and D. Raoult. 1996. Diagnosis of Mediterranean spotted fever by cultivation of Rickettsia conorii from blood and skin samples using the centrifugation-shell vial technique and by detection of R. conorii in circulating endothelial cells: a 6-year follow-up. J. Clin. Microbiol. 34:2722-2727.
La Scola, B., and D. Raoult. 1999. Culture of Bartonella quintana and Bartonella henselae from human samples: a 5-year experience (1993 to 1998). J. Clin. Microbiol. 37:1899-1905.
Lepidi, H., F. Fenollar, R. Gerolami, J. L. Mege, M. F. Bonzi, M. Chappuis, J. Sahel, and D. Raoult. 2003. Whipple's disease: immunospecific and quantitative immunohistochemical study of intestinal biopsy specimens. Hum. Pathol. 34:589-596.
Maiwald, M., A. Von Herbay, D. N. Fredricks, C. C. Ouverney, J. C. Kosek, and D. A. Relman. 2003. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J. Infect. Dis. 188:801-808.
Marrero, M., and D. Raoult. 1989. Centrifugation-shell vial technique for rapid detection of Mediterranean spotted fever rickettsia in blood culture. Am. J. Trop. Med. Hyg. 40:197-199.
Marth, T., and D. Raoult. 2003. Whipple's disease. Lancet 361:239-246.
Masselot, F., A. Boulos, M. Maurin, J. M. Rolain, and D. Raoult. 2003. Molecular evaluation of antibiotic susceptibility: Tropheryma whipplei paradigm. Antimicrob. Agents Chemother. 47:1658-1664.
Maurin, M., and D. Raoult. 2000. Isolation in endothelial cell cultures of Chlamydia trachomatis LGV (serovar L2) from a lymph node of a patient with suspected cat scratch disease. J. Clin. Microbiol. 38:2062-2064.
Molinari, G., and G. S. Chhatwal. 1999. Streptococcal invasion. Curr. Opin. Microbiol. 2:56-61.
Molinari, G., M. Rohde, C. A. Guzman, and G. S. Chhatwal. 2000. Two distinct pathways for the invasion of Streptococcus pyogenes in non-phagocytic cells. Cell Microbiol. 2:145-154.
Petts, D. N. 1984. Colistin-oxolinic acid-blood agar: a new selective medium for streptococci. J. Clin. Microbiol. 19:4-7.
Raoult, D., M. L. Birg, B. La Scola, P. E. Fournier, M. Enea, H. Lepidi, V. Roux, J. C. Piette, F. Vandenesch, D. Vital Durand, and T. J. Marrie. 2000. Cultivation of the bacillus of Whipple's disease. N. Engl. J. Med. 342:620-625.
Raoult, D., P. Brouqui, and V. Roux. 1996. A new spotted-fever-group rickettsiosis. Lancet 348:412.
Raoult, D., P. E. Fournier, and M. Drancourt. 2004. What does the future hold for clinical microbiology Nat. Rev. Microbiol. 2:151-159.
Raoult, D., B. La Scola, P. Lecocq, H. Lepidi, and P. E. Fournier. 2001. Culture and immunological detection of Tropheryma whippleii from the duodenum of a patient with Whipple disease. JAMA 285:1039-1043.
Raoult, D., H. Ogata, S. Audic, C. Robert, K. Suhre, M. Drancourt, and J. M. Claverie. 2003. Tropheryma whipplei twist: a human pathogenic actinobacteria with a reduced genome. Genome Res. 13:1800-1809.
Raoult, D., and V. Roux. 1997. Rickettsioses as paradigms of new or emerging infectious diseases. Clin. Microbiol. Rev. 10:694-719.
Raoult, D., G. Vestris, and M. Enea. 1990. Isolation of 16 strains of Coxiella burnetii from patients by using a sensitive centrifugation cell culture system and establishment of strains in HEL cells. J. Clin. Microbiol. 28:2482-2484.
Relman, D. A. 1999. The search for unrecognized pathogens. Science 284:1308-1310.
Renesto, P., N. Crapoulet, H. Ogata, B. La Scola, G. Vestris, J. M. Claverie, and D. Raoult. 2003. Genome-based design of a cell-free culture medium for Tropheryma whipplei. Lancet 362:447-449.
Rovery, C., J. M. Rolain, D. Raoult, and P. Brouqui. 2003. Shell vial culture as a tool for isolation of Brucella melitensis in chronic hepatic abscess. J. Clin. Microbiol. 41:4460-4461.
Seifert, H., H. Wisplinghoff, P. Schnabel, and C. von Eiff. 2003. Small colony variants of Staphylococcus aureus and pacemaker-related infection. Emerg. Infect. Dis. 9:1316-1318.
Vaudaux, P., P. Francois, C. Bisognano, W. L. Kelley, D. P. Lew, J. Schrenzel, R. A. Proctor, P. J. McNamara, G. Peters, and C. Von Eiff. 2002. Increased expression of clumping factor and fibronectin-binding proteins by hemB mutants of Staphylococcus aureus expressing small colony variant phenotypes. Infect. Immun. 70:5428-5437.
Vestris, G., J. M. Rolain, P. E. Fournier, M. L. Birg, M. Enea, J. Y. Patrice, and D. Raoult. 2003. Seven years' experience of isolation of Rickettsia spp. from clinical specimens using the shell vial cell culture assay. Ann. N. Y. Acad. Sci. 990:371-374.
Weiss, E., and H. R. Dressler. 1960. Centrifugation of rickettsiae and viruses into cells and its effect on infection. Proc. Soc. Exp. Biol. Med. 103:691-695.
Yu, X., Y. Jin, M. Fan, G. Xu, Q. Liu, and D. Raoult. 1993. Genotypic and antigenic identification of two new strains of spotted fever group rickettsiae isolated from China. J. Clin. Microbiol. 31:83-88.
Zannier, A., M. Drancourt, J. P. Franceschi, J. M. Aubaniac, and D. Raoult. 1991. Interêt d'une technique de lyse cellulaire par choc thermique dans l'isolement des bacteries responsables d'infections osteo-articulaires. Pathol. Biol. 39:543-546.
Zeaiter, Z., P. E. Fournier, H. Ogata, and D. Raoult. 2002. Phylogenetic classification of Bartonella species by comparing groEL sequences. Int. J. Syst. Evol. Microbiol. 52:165-171.(Frederique Gouriet, Flore)