|
|
| |||||||||||||||||||||||||||||||||||||||
Background In patients with cystic fibrosis, infection with Pseudomonas cepacia is associated with poor outcomes. However, the extent of person-to-person transmission and the source of P. cepacia infection after lung transplantation are not well defined. Using DNA-based typing systems, we sought to determine the genetic relatedness of P. cepacia infection at one cystic fibrosis center.
Methods We analyzed 65 P. cepacia isolates gathered over a period of eight years at a single cystic fibrosis center from 17 clinic patients and from 5 patients who underwent double-lung transplantation. The isolates were analyzed by ribotyping and chromosomal fingerprinting based on pulsed-field gel electrophoresis.
Results Analyses of serial isolates revealed that each clinic patient and transplant recipient harbored a different P. cepacia clone that was persistent. In the transplant recipients, the preoperative and postoperative isolates were identical. In the two patients with disseminated infection after lung transplantation, isolates from multiple sites were identical and indicated clonal expansion of the previous respiratory P. cepacia strain. Pulsed-field gel electrophoresis proved both more discriminative and more practical than ribotyping as a means of defining the genetic relatedness of the P. cepacia isolates.
Conclusions Our serial analyses in patients with cystic fibrosis at one center found distinct strains of P. cepacia persistently infecting each patient and no evidence of person-to-person transmission of this organism. P. cepacia infection after lung transplantation was due to the persistence of the strain present before transplantation.
Double-lung transplantation is a recognized treatment for end-stage pulmonary disease resulting from cystic fibrosis. Since the trachea and sinuses remain bacterial reservoirs postoperatively, it was once feared that persistent infection with multiply resistant organisms would be common after transplantation and immunosuppression7. However, centers treating lung-transplant recipients without P. cepacia infection have found no difference in infection rates between recipients with cystic fibrosis and those without it8,9. Thus, pulmonary infection, even with multiresistant microorganisms, has not become a criterion for exclusion from transplantation9,10. The consequences of P. cepacia infection threaten to change this situation dramatically.
In a Toronto series the post-transplantation acquisition of P. cepacia was strikingly associated with morbidity and mortality11,12. Such observations have led to the suggestion that excluding patients with P. cepacia from transplantation would improve survival rates overall and reduce transmission to other transplant recipients with cystic fibrosis12,13,14,15.
Before microbiologic criteria for the selection of candidates for transplantation can be developed, the degree of transmissibility, clonality, and persistence of P. cepacia must be better understood. Two molecular genetic assays can be used to address such epidemiologic questions: ribotyping, the determination of restriction-fragment-length polymorphisms (RFLPs) associated with the multicopy ribosomal RNA operon (rrn), and chromosomal fingerprinting, resolved by pulsed-field gel electrophoresis (PFGE).
Our study of P. cepacia had two purposes: to define the molecular epidemiology of P. cepacia infection at one cystic fibrosis center and to compare the epidemiologic discriminative power of the molecular genetic assays. This allowed us to determine whether lung-transplant recipients with P. cepacia infection acquired new strains or retained their pretransplantation clones despite exposure to clinic and hospital flora, antibiotic therapy, and immunosuppression. We also determined whether an infecting P. cepacia clone persisted in clinic patients, and whether clones were transmitted within the cystic fibrosis center. The results have direct implications for clinical management and demonstrate the necessity for specific molecular genetic approaches to identify whether a particular P. cepacia isolate has epidemic potential.
Methods
Patient Population and Bacterial Strains
P. cepacia strains were isolated from patients at the University of North Carolina Cystic Fibrosis Center from 1985 through 1993. Though placed in private rooms when hospitalized, infected patients were not otherwise separated from other clinic patients. Patients awaiting double-lung transplantation were not segregated on the basis of P. cepacia infection and were required to participate in a support group and pulmonary rehabilitation.
Sixty-five P. cepacia isolates associated with cystic fibrosis were characterized, including 24 from five patients with cystic fibrosis who underwent transplantation; four of these patients were referred from other centers16. Forty-one isolates were from 17 clinic patients with cystic fibrosis who were receiving routine care during the same period. Independently isolated American Type Culture Collection (ATCC) strains of P. cepacia were also included: four from environmental sources and four from patients without cystic fibrosis.
DNA Hybridization
Chromosomal DNA was prepared as previously reported17,18,19. EcoRI RFLPs of multicopy ribosomal RNA operons were analyzed with an entire rrnB operon probe of Escherichia coli19,20.
PFGE
Chromosomal DNA plugs21 were incubated with SpeI (Boehringer). Restriction fragments were separated by PFGE with a CHEF Mapper (Bio-Rad) through 1 percent agarose gel (Bio-Rad) at a field strength of 6 V per centimeter and with initial and final pulse times of 1.2 and 54 seconds, respectively. Fragment sizes were determined, and a bar-code-format translation of chromosomal-fingerprint profiles was made with an Apple OneScanner with a Macintosh Quadra 950 cpu running Gene Construction Kit (Texto).
Statistical Analysis
PFGE chromosomal fingerprints were compared with use of the criteria of Prevost et al.22 On the basis of ribotyping criteria for P. cepacia,23 ribotypes were considered equivalent when a comparison of hybridizing fragments revealed a difference of three or fewer bands between two patterns. Subribotypes were considered identical when the patterns of two isolates differed by no more than one band. A quantitative pairwise comparison of RFLP patterns was carried out with the Dice coefficient of similarity (D), calculated as D = 2nxy/n1 + n2, where n is the total number of DNA fragments from strain X, n2 is the total number from strain Y, and nxy is the number of identical fragments24,25. A D value for two PFGE RFLPs of 
0.90 represents closely related strains, whereas unrelated strains have a value of 
0.60. Remarkably, intervening values have not been observed, and values between 0.50 and 0.60 are rare.
Results
Comparative Characterization of Control P. cepacia Isolates According to PFGE Profile and Ribotype
SpeI chromosomal RFLPs were characterized for ATCC P. cepacia isolates from diverse sources to provide a basis for comparison with strains associated with cystic fibrosis (Figure 1). Each isolate possessed a distinct profile of 11 to 19 fragments, with a mean [±SD] D of 0.32 ±0.10 on pairwise comparison. The variability was similar when analyzed with AseI (data not shown).
|
|
Figure 3 shows PFGE-resolved chromosomal fingerprints of isolates obtained before and after double-lung transplantation from one patient with transient post-transplantation P. cepacia colonization (Patient 4) and three patients with isolates cultured from various sites after transplantation (Patients 1, 2, and 3). The RFLP pattern within each patient's set of serial isolates remained unchanged over the interval of 1 to 13 months spanning preoperative evaluation, transplantation, and postoperative care. Furthermore, each such chromosomal-fingerprint profile was significantly different from other strains obtained from transplant recipients (D = 0.28 ±0.10).
|
PFGE Characterization of Isolates from Clinic Patients
PFGE-resolved chromosomal RFLPs of serial isolates from eight clinic patients plus prototypic RFLPs of four transplant-associated clonal lineages (Figure 3) are displayed in bar-code format in Figure 4. These results demonstrate that the strain (or strains in the case of one patient) from each patient monitored over an eight-year period was distinct from those of all other clinic patients (D = 0.20 ±0.09; range, 0.06 to 0.36). Single isolates from nine additional clinic patients were also disparate (data not shown). This level of chromosomal RFLP diversity among strains from 17 clinic patients exceeds but is not significantly different from that for independent ATCC isolates (0.20>P>0.10). RFLP comparison of transplantation-associated strains (Figure 4, lanes 2, 3, 4, and 5) with strains from clinic patients (Figure 4, lanes 6 through 36) revealed that each of the former was distinct from the latter (D = 0.17 ±0.09), confirming that these were epidemiologically distinct strains.
|
Ribotype Characterization of Isolates from Transplant Recipients and Clinic Patients
Figure 5 shows the ribotypes of strains from four lung-transplant recipients and six clinic patients. Two sets of serial isolates from lung-transplant recipients (Patients 2 and 4) and four sets of serial isolates from clinic patients (Patients 22, 23, 25, and 26) had stable rrn RFLPs. Alterations in one or three bands indicate variation within ribotypes23 in sets from two transplant recipients (Patients 1 and 3) and one clinic patient (Patient 21). Among the isolates from Patient 1, two variants of a ribotype were recognized in the pretransplantation strains; after transplantation, one variant persisted and a third appeared as well. In Patient 3, one subribotype was identified before transplantation and two subribotypes were identified afterward (Figure 5, lanes 29 and 30 vs. lanes 31 and 32).
|
Though the ribotypes of strain sets from Patients 22, 24, and 26 were distinct, the ribotypes of the strain sets from Patients 23 and 25 (Figure 5) had very similar rrn RFLP profiles, again differing by only 3 bands of a total of 17 per pair, indicating that a single ribotype was shared by strains from these two clinic patients. These two variants had a D of 0.82, as compared with a mean D of 0.37 ±0.20 for all ribotype comparisons within the set of strains from clinic patients. The PFGE chromosomal RFLPs of these ribotype-identical strain sets were distinct (D = 0.18). Thus, PFGE-based chromosomal fingerprinting was better able than ribotyping to discriminate among strains within a species. In contrast, the ability of ribotyping to differentiate between species is shown in Figure 5: a Xanthomonas maltophilia strain isolated transiently can be seen to differ from the P. cepacia strains because it lacks the species' signature conserved pair of bands.
Isolates from Disseminated Infection in Transplant Recipients
P. cepacia bacteremia developed in one transplant recipient two months postoperatively, and a second had P. cepacia isolated from venous blood drawn from an access port. Blood samples from the latter patient (Patient 4), who survived (Figure 3 and Figure 5), and isolates from multiple sites from Patient 1, who died, yielded ribotype designations and PFGE RFLPs identical to each patient's pretransplantation isolates, indicative of clonal expansion of the original respiratory strain. The patients who died with P. cepacia infection survived 2.5 months (Patient 1), 17 months (Patient 2), and 30 months (Patient 3) after transplantation (Figure 3 and Figure 5). Chronic P. cepacia pulmonary infection with the strain present before transplantation persisted throughout this period. Two patients were free of infection 10 and 30 months after transplantation16.
Discussion
Our DNA-based analyses demonstrate that over an eight-year span, 17 patients at one cystic fibrosis center had discrete strains of P. cepacia. Analysis of serial isolates further revealed that all but one clinic patient persistently harbored a discrete clone, and of five lung-transplant recipients, those with postoperative P. cepacia remained infected by the strain present before transplantation.
As characterized by chromosomal RFLPs, no identical or closely related strains were isolated among the 22 patients with cystic fibrosis, indicating a total absence of patient-to-patient transmission or acquisition from a common source. Thus, there is a striking disparity between the diversity of P. cepacia strains identified at this cystic fibrosis center and earlier reports of center-specific, predominant strains presumably transmitted between patients or acquired from a common source3,5. The latter reports led to the incorrect inference that cystic fibrosis centers generally harbor one or more highly transmissible P. cepacia strains. Our findings cannot be explained by any stringent infection-control practices, nor by novel, nontransmissible P. cepacia lineages unique to this cystic fibrosis center, since four of the five patients with lung transplants were infected at other centers before transplantation. Rather, our results suggest that these patients were infected by distantly related strains of low transmissibility. Ironically, the hypothesis that transmissibility varies markedly among strains of P. cepacia is strongly supported by both our results and prior evidence of epidemic P. cepacia strains at other centers3,4,5.
Our results may help clarify the controversy regarding the appropriateness of transplantation for patients with cystic fibrosis who are infected with P. cepacia. The wisdom of this procedure has been questioned because of the potential for transmission, the high morbidity and mortality associated with P. cepacia infection after lung transplantation,11,12,15 and the possibility that when immunosuppressed, these patients will acquire new, more virulent strains. Our results show no evidence of transmission or of the acquisition of new, more virulent strains. Rather, they directly demonstrate that clonality characterizes P. cepacia infection after lung transplantation. Despite transplantation, and in one case a second transplantation, persistent infection rather than reinfection contributed to postoperative morbidity (Figure 3).
Isolates were typable by both ribotyping and PFGE-based chromosomal fingerprinting, and there was a high degree of reproducibility, indicated by our finding that RFLPs were stable over many generations for serial and disseminated isolates. Two previous reports found both genetic assays to be equally able to identify identical P. cepacia isolates from epidemic outbreaks, involving nosocomial transmission among patients without cystic fibrosis27 and transmission among patients with cystic fibrosis5. The characterization of such epidemic strains represents a challenge to reproducibility. However, for nonepidemic P. cepacia isolates, our results reveal PFGE to be more discriminative than ribotyping. PFGE-resolved RFLPs identified all independently isolated strains, but six isolates identified as distinct by PFGE were identified as belonging to three rather than six ribotypes by rrn RFLP analysis -- results similar to those of previous studies3,23. An alternative genetic assay involving rrn spacer polymorphisms amplified with the polymerase chain reaction and introduced recently to P. cepacia epidemiology28 has also proved less discriminative than PFGE-based epidemiology29.
The sensitivity, simplicity, and speed of PFGE-based analysis encourage its introduction as a routine method to resolve epidemiologic questions related to P. cepacia and other microorganisms. In contrast to ribotyping and virulence-factor gene probing, it provides a rapid, direct method applicable to all bacterial pathogens. With the development of automated fluorescence scanners and software for bar-code translation, storage, and interpretation of RFLP profiles determined by PFGE, the elucidation of precise epidemiologic relations should become routine clinical practice.
Our current phylogenetic studies focus on genetic changes accounting for the existence of P. cepacia lineages varying in transmissibility. Preliminary unpublished findings imply that at least one evolutionarily unique lineage exists that correlates with a highly transmissible epidemic phenotype. Molecular genetic characterization of this divergent lineage should yield rapid DNA-based assays for the differentiation of P. cepacia strains based on the degree of transmissibility. Such criteria would provide a rational basis for evaluating the necessity for the segregation of a P. cepacia-infected patient with cystic fibrosis. We are also exploring the molecular genetic mechanisms accounting for persistent infection, since it should be possible to exploit the very factors mediating clonal persistence to eliminate such infection.
Supported by research grants from the Cystic Fibrosis Foundation (to Drs. Steinbach and Goldstein) and by Biomedical Research Support from the National Institutes of Health through Boston City Hospital.
We are indebted to Drs. Robert Beall, Daniel Shapiro, and Alison Holmes for useful discussions; to Aram Chobanian and Howard Corwin for encouragement to initiate these studies; to Kathy Brown, Richard Kozak, and Ethan Johnson for assistance with clinical microbiologic techniques; and to Drs. Janet Geisselsoder, Hector Salinas, and David Schwartz (Bio-Rad Laboratories) for technical support.
Source Information
From the Department of Pediatrics (S.S.) and the Section of Molecular Genetics, Maxwell Finland Laboratory for Infectious Diseases (L.S., R.-Z.J., R.G.), Boston University School of Medicine and Boston City Hospital, Boston; and the Cystic Fibrosis Center (P.F.), Clinical Microbiology Laboratory (P.G.), and Division of Cardiothoracic Surgery (T.M.E.), University of North Carolina School of Medicine and Hospitals, Chapel Hill.
Address reprint requests to Dr. Goldstein at the Section of Molecular Genetics, Maxwell Finland Laboratory for Infectious Diseases, 774 Albany St., Boston, MA 02118.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Mahenthiralingam E., Campbell M., Speert D. P., Govan J. R.W., LiPuma J. J., Stull T. L., Steinbach S. F., Goldstein R.
Extract |
Full Text
N Engl J Med 1995;
332:819-821, Mar 23, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2010 Massachusetts Medical Society. All rights reserved. |
Previous
