![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Clinical Microbiology Reviews, April 2002, p. 194-222, Vol. 15, No. 2
0893-8512/02/$04.00+0 DOI: 10.1128/CMR.15.2.194-222.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Channing Laboratory, Brigham and Women's Hospital,1,1 Harvard Medical School,2,2 Children's Hospital,3 Boston, MA 021153
SUMMARY INTRODUCTION Overview of CF and Bacterial Infection Historical Framework for the Study of Cystic Fibrosis CYSTIC FIBROSIS Clinical and Biochemical Aspects Diagnosis Uncovering the function of CFTR Biological function of CFTR after discovery of the gene Clinical Manifestations of Mutations in CFTR Genetic and Functional Aspects of Mutations in CFTR CFTR Mutations Epidemiology of CF and CFTR Mutations and Possible Advantages for Heterozygotes MICROBIOLOGIC ASPECTS OF CYSTIC FIBROSIS LUNG INFECTION Recovery and Distribution of Microbial Pathogens among CF Patients S. aureus, H. influenzae, and CF Role of Inflammation and P. aeruginosa Infection Early aspects of inflammation. Initiation and establishment of P. aeruginosa infection. Emergence of the mucoid phenotype. Progression of chronic infection. (i) Biofilms and quorum sensing. (ii) Ineffectiveness of the innate immune response to mucoid P. aeruginosa. (iii) Ineffectiveness of the acquired immune response during chronic P. aeruginosa infection. THERAPIES FOR CYSTIC FIBROSIS LUNG DISEASE Airway Clearance Chemotherapy Mechanisms of Antibiotic Resistance EMERGING PATHOGENS AFFECTING CYSTIC FIBROSIS PATIENTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES
| SUMMARY |
|---|
 Top NextReferences |
|---|
| INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
Lungs of CF patients are often colonized or infected in infancy and early childhood with organisms, such as Staphylococcus aureus and Haemophilus influenzae, that may damage the epithelial surfaces, leading to increased attachment of, and eventual replacement by, P. aeruginosa. However, adequate clinical studies to determine the role of these organisms in the pathogenesis of lung disease in CF patients have never been published. The recovery of these organisms from a bronchoalveolar lavage (BAL) fluid sample from the lung would be considered a frank infection in need of therapy. However, the role that S. aureus, nontypeable H. influenzae, and similar organisms isolated from oropharyngeal cultures play in the progression of CF patients to respiratory failure has not been determined. Rather, the pathogenic role of S. aureus and nontypeable H. influenzae in the development of lung disease in CF patients is inferred principally from clinical anecdote but is otherwise lacking any solid support from studies in the peer-reviewed literature.
Chronic infection with P. aeruginosa is the main proven perpetrator of lung function decline and ultimate mortality in CF patients. Chronic P. aeruginosa infection leads to epithelial surface damage and airway plugging, progressively impairing airway conductance, which results in a decline in pulmonary function. Intense inflammation characterized by neutrophil sequestration in the airways contributes to impaired clearance and plugging associated with the death of senescent cells. Airway damage also arises through neutrophil release of a variety of oxidants and enzymes.
CF has not always been a disease characterized by chronic pseudomonal sinopulmonary infection. Prior to 1946, the reported prevalence of CF pseudomonal infections was low (78). However, a variety of sources indicate that during the 1960s P. aeruginosa became the most prevalent organism in the airways of CF patients (229). The emergence of this pathogen coincided temporally with the introduction of regional centers that specialized in CF care. The adherence to standardized principles of multidisciplinary therapy by CF centers has been lauded as an important factor responsible for increasing the median survival from 14 years in 1969 to greater than 30 years currently in the United States (247). However, studies in Denmark pointed to CF centers as potential sites of increased risk for spread of P. aeruginosa (223, 225). Studies in the United States have corroborated these suspicions. In a study by Farrell et al. (90), the median pseudomonas-free period of the patients attending one center was more than five times that of those attending another CF center. The center with the earlier pseudomonal acquisition time was distinguished by an urban setting, admixing of young patients with older, P. aeruginosa-infected patients, and more opportunity for social interactions among the patients. Further studies in Denmark (95) demonstrated a decrease in pseudomonal colonization after institution in 1981 of cohort isolation (isolation of younger, uninfected patients from older patients more likely to carry infectious agents). Thus, the acquisition of P. aeruginosa by CF patients can be affected by different treatment settings.
Our understanding of the genetic basis of CF was advanced greatly by the work of Dorothy Andersen, who in 1938 published a detailed study of 49 CF patients (6). In this study, cases were categorized into three groups based on the patients' age at death. Group I consisted of patients who died before the age of 1 week, group II consisted of patients who died between 1 week and 6 months, and group III consisted of patients who died between 6 months and 14.5 years of age. The cause of death of patients in group I was intestinal obstruction, while patients in groups II and III usually died of "respiratory complications." However, a broad range of observations demonstrated that despite these classifications, similar pathological conditions (for example, pancreatic lesions and malnutrition) could be seen in all patients. These observations led to an understanding of CF as a single disease with diverse effects rather than a loose collection of related disease states. Further evidence of the pleiotropic manifestations of CF was presented by di Sant'Agnese et al., who demonstrated that the sweat of CF patients contains abnormally high concentrations of sodium, chloride, and potassium (77), and by Shwachman et al., who made the interesting observation that seven CF patients (two males and five females), who had reached adulthood and married had universally failed to produce offspring (297).
Citing histological data from autopsy samples, Norris and Tyson (209) and Baggenstoss et al. (12) postulated that the physiological nature of the CF defect was a malformation of the pancreatic ducts, leading to defective secretion by various epithelial glands. While it was appreciated that pancreatic function could be normal in some patients with CF (pancreatic sufficient), most patients who were recognized to have the disease presented with large greasy stools (steatorrhea) due to pancreatic function that was inadequate for proper absorption of nutrients (pancreatic insufficient). Consequently, CF was viewed primarily as a disease of the digestive tract. Thus, early studies were biased toward severe cases and focused on detailed histological descriptions of the anatomical defects of the pancreas during progression of CF. While minor discrepancies exist between studies, several consistent observations warrant comment. Acini (glands) were found to contain concretions (dehydrated secretions) of various sizes; also, the acinar cells exhibited various degrees of flattening, resulting in a vaguely squamous appearance (6). The degree of flattening of the acinar cells appeared to be directly related to the size of the concretions. Furthermore, acini were often surrounded by fibrous or adipose tissue and were also occasionally infiltrated by fibroblasts, lymphocytes, plasma cells, or phagocytes (6, 209). Walters proposed that these sequelae stemmed from hyperplasia (overproliferation) of ductile epithelial cells (332). Such hyperplasia was presumed to compress the local acinar tissue, resulting ultimately in atrophy of the acini and their replacement by fibrous or adipose tissue (332). The islets of Langerhans are usually reported to be normal in terms of their architecture but have been reported at times to be less frequent in the CF pancreas than in a normal pancreas. These changes are now appreciated to be due to autodigestion of pancreatic tissue from enzymes trapped in concretions.
More recently, it has been proposed by Freedman et al. (96) that dysfunction of the acinar tissue in CF may be due to an imbalance in the utilization of free fatty acids in the phospholipids of CF patients. These workers reported that the phospholipids of CFTR knockout mice contain a higher than normal proportion of arachidonic acid, at the expense of docosahexaenoic acid (DHA). Orally administering DHA to the knockout mice corrected the membrane defect and restored normal histology to the affected tissues, suggesting that physiologic defects in CF are not due to an inability of the CF intestine to absorb certain fatty acids but, rather, are due to a defect in fatty acid synthesis or utilization.
Bacteriological studies on the lungs of CF patients date to the turn of the century. In 1905, Landsteiner (169) reported that of 15 CF lung samples, 9 were culture positive. S. aureus was the predominant agent isolated, although Staphylococcus albus (i.e., coagulase-negative staphylococcus, probably Staphylococcus epidermidis) and Streptococcus haemolyticus (i.e., Streptococcus pyogenes) were also identified in the cultures (169). As treatment standards for CF patients improved over the years, the average mean survival of CF patients increased dramatically. Most notable among such advances was the refinement of nutritional regimens (4, 19, 290) and the advent of antibiotic chemotherapy (122, 193, 197, 228). While essentially all patients prior to the 1950s died by the age of 10 years, reports published in the 1950s (192) and 1960s (297) described a considerable proportion of CF patients surviving well beyond this age. By the 1990s, approximately one-third of CF patients were surviving to adulthood (93). This increased mean survival has had a dramatic impact on the nature of CF as an infectious disease, since the longer survival of CF patients has created opportunities for the establishment of infection by bacteria other than Staphylococcus. Two reports published in 1968 by Burns and by Burns and May demonstrated that the sera of CF patients contained antibodies to both P. aeruginosa and Klebsiella spp. (43, 44). Furthermore, the presence of serum antibodies to P. aeruginosa correlated perfectly with bacterial carriage, as assessed by sputum culture of the microorganism. Today, P. aeruginosa is the most prevalent pulmonary pathogen in CF patients.
| CYSTIC FIBROSIS |
|---|
|
TopPreviousNextReferences |
|---|
F508 CFTR (which alone would cause abnormal sweat electrolyte levels) but who have a third mutation in one of their CFTR alleles, R553Q, may have normal sweat electrolyte levels (83), indicative of possible compensatory, second mutations in an allele that otherwise leads to elevated sweat chloride levels. For these reasons, diagnosis is confirmed by genetic analysis. While genetic screens are able to identify more than 90% of occurrences of the more than 1,000 known CFTR mutations, a negative screen does not ensure a normal CFTR genotype since the commercial screens that are currently available detect only the 70 most prevalent CFTR mutations. Diagnoses that remain unclear after sweat testing and genotyping may be confirmed by a test that directly measures CFTR function, such as nasal potential difference testing (a method for real-time measurement of transepithelial electrical potential resulting from ion transport through channels including CFTR [154]). Uncovering the function of CFTR The deciphering of the biological function of CFTR began in 1953 with the observation (77) that the sweat of CF patients contains abnormally high electrolytes levels. This observation has ultimately led to demonstrations by several researchers that CF patients have abnormalities in chloride conductance in and out of cells (243, 279). Normally, as the isotonic secretions travel from the acinus of the sweat gland to the surface of the skin, the epithelial cells lining the ducts act to reabsorb NaCl, resulting in hypotonic sweat. However, the sweat ducts of CF patients are impermeable to Cl-. Thus, the NaCl remains in the secretions, and the sweat is salty (Fig. 1). Later studies by Sato and Sato (267) showed that unlike normal glands, CF sweat glands fail to secrete fluid in response to ß-adrenergic agonists that stimulate cyclic AMP (cAMP) production, yet CF glands produced normal amounts of cAMP. Thus, the Cl- conductance defect was located downstream from adenylate cyclase, at the level of the chloride channel or regulator. Studies utilizing the patch-clamp technique, which enables observations of single ion-channel activity, suggested that the defect lay in the regulation of a chloride channel, called the secretory channel, which was studied by multiple investigators using a variety of epithelial tissues. This channel has the following properties: outward rectification (implying a preference to transport Cl- ions into rather than out of the cell), moderate conductance, and activation by cAMP and protein kinase A (PKA) and, under some conditions, protein kinase C (PKC). Outwardly rectifying chloride channels (ORCC) can be found in epithelial cells from CF patients, but these fail to respond to PKA and PKC. This observation prompted historical speculation that if the CF gene product did not encode the ORCC, it was perhaps a regulator of the ORCC (132, 175, 272).
|
|
|
|
Other features of CF are not clearly related to the role of CFTR as a coordinator of fluid secretion. For example, the link between CFTR defects and congenital bilateral absence of the vas deferens (CBAVD), which renders most males with defective CFTR infertile (55), is unclear. The role that CFTR plays in some tissues in which it is highly expressed, such as heart and kidneys, which have conspicuously normal phenotypes in CF patients (aside from the somewhat greater propensity of CF patients to develop kidney stones) is also not clear.
The CFTR protein (Fig. 4) is a member of the ATP binding cassette (ABC) family of transporters. Members of this protein family are found in mammals, insects, yeast, and bacteria and include the multidrug resistance efflux pump (MDR1) (257), the transporters associated with antigen processing (TAP1 and TAP2) (307), and the bacterial histidine permease (178, 206). The highly conserved motif that defines the ABC family of proteins includes a membrane-spanning domain, containing six membrane-spanning peptides, followed by a nucleotide binding domain (NBD), which is responsible for the ATP binding and hydrolysis that supplies energy to drive the opening and closing of the ion channel. Like many members of the ABC family, CFTR consists of a tandem repeat of this motif.
|
The NBDs of CFTR are responsible for the binding and hydrolysis of ATP and provide the energy necessary for channel activity (8). A recent study (104) reports that the two NBDs of CFTR function in a coordinated fashion to sequentially mediate the opening and closing of the ion channel pore. Thus, the N-terminal NBD (NBD-1) hydrolyzes one molecule of ATP to open the channel and then the C-terminal NBD (NBD-2) hydrolyzes a second molecule of ATP to close the channel. The NBDs of CFTR contain motifs that are common to the nucleotide binding folds of many ABC- and non-ABC family proteins, but they also contain motifs that are restricted to the ABC family or to a subset of ABC proteins. The most widely conserved ABC family motifs are the Walker A motif (at the amino-terminal end of each NBD) and the Walker B motif (at the carboxy-terminal end of each NBD), which function in the binding and coordinate interaction of ATP and Mg2+, respectively. The Walker A motif has the sequence GxxGxGK(S/T), where x is any amino acid. Mutations within the Walker A motif have a strong negative effect on channel activity (245). The Walker A motif is also referred to as the phosphate loop or P-loop, since its constituent amino acids make direct contacts with the 
-, ß-, and 
-phosphates of ATP. The Walker B motif has a less stringent sequence, which can be represented R(x)6-8
D, where x is any amino acid and is any hydrophobic amino acid. In addition to the two widely conserved Walker motifs, NBDs of the ABC family proteins contain a motif which is unique to this family—the signature motif or C motif. The C motif lies between the Walker A and B motifs, just upstream of the Walker B motif, and has the consensus sequence LSGGQ. The NBD of CFTR also contain a fourth conserved motif that appears in only a subset of ABC family proteins. This motif is the center region, which lies at the midpoint between the Walker A and B motifs. While mutations within the Walker A, Walker B, and signature motifs usually impair protein function, the importance of the center region is variable depending on the particular transporter protein. The yeast -factor transporter STE6, for example, is tolerant of mutation within its center region (18), while other ABC proteins (CFTR, for example) are quite sensitive to mutations in this motif. For example, the most common mutation, F508 CFTR, is a center-region mutation.
In addition to the two protein domains already described (membrane-spanning domain and NBD), CFTR contains a regulatory (or R) domain, which modulates the channel activity of CFTR (49, 318). Only a small subset of ABC family proteins contain an R domain. Other examples of R domain-containing ABC proteins are the yeast YCF1 protein, which confers cadmium resistance to yeast (317), and the MRP1 multidrug resistance gene (145). A model of R domain function has been proposed by Ma et al. (180), who suggests that when dephosphorylated, the R domain interacts with the N-terminal NBD (NBD-1), thus blocking the ATP binding site on the NBD. With ATP unable to bind, channel opening cannot occur. According to the same model, when the R domain becomes phosphorylated, it either dissociates from the NBD or interacts with the NBD in a different way, such that the ATP binding site of NBD-1 is available. Further evidence for the important role of the R domain in channel activity was obtained through the analysis of R-domain point mutants, which demonstrated either reduced channel activity or complete loss of activity (219). Deletions in the R domain result in regulatory defects (180). The physical interaction of the R domain with NBD-1, as well as the phosphorylation dependence of this interaction, has recently been demonstrated (205), supporting the model of Ma et al. However, the actual mechanism by which the ion channel activity of CFTR is regulated is probably more complex. For example, data from other groups has shown that phosphorylation in the R domain affects channel activity differently depending on which serine residue(s) of the R domain is phosphorylated (341, 345).
|
F508. An important advance in the identification of CFTR mutant alleles was the refinement of molecular genetics approaches. Early attempts to track the distribution of CF alleles relied on restriction fragment length polymorphism linkage analysis, which was fraught with error due to the loose linkage of common marker haplotypes with the CFTR gene. The advent of techniques such as single-strand conformational polymorphism allowed the screening of large numbers of individuals for mutant alleles. This, together with a more rigorous study of the epidemiology of the disease, allowed the identification of many novel CFTR alleles, some with subtle clinical manifestations. In addition to mutations resulting in changes in amino acids comprising the CFTR protein, other mutations affecting transcript production, such as the so-called 5T allele, can impact CFTR levels. In the 5T allele there is a variation in the length of the polypyrimidine tract in the splice acceptor site at the 3' end of intron 8. While the normal splice acceptor site at this location has nine tandem thymidines, shortening of the polypyrimidine tract first to seven and then to five thymidines, results in inefficient splicing of the CFTR transcript. Although this mutation was originally characterized as the causative factor of CBAVD (58), the 5T allele has more recently been shown to manifest itself as the cause of mild but clinically detectable pulmonary dysfunction (55). Moreover, this mutation can cause pancreatic insufficiency, although the latter sequela appears to be uncommon (146). The mild and highly variable phenotype associated with the 5T allele has complicated its characterization. In a study conducted in 1997, Kerem et al. reported that of a cohort of approximately 150 subjects, those with typical or atypical CF had an approximately threefold-higher incidence of the 5T allele than did individuals without CF (147). The subjects in this study who carried the 5T allele presented with a wide range of symptoms including asthma, bronchitis, bronchiectasis, meconium ileus, and pancreatic insufficiency. Even the sweat chloride levels of 5T allele carriers exhibited a wide range from normal to elevated, confounding the use of sweat chloride levels as the pathognomonic feature of CF. The 5T allele probably results in a more severe phenotype only when it is present in a compound heterozygous state with another mutant allele (T. Bienvenu, J. Lepercq, J. P. Allard, D. Hubert, C. Francoval, C. Beldjord, and J. C. Kaplan, Letter, Ann. Genet. 41:63-64, 1998).
The F508 CFTR mutation is by far the most common mutant allele, accounting for some 70% of all mutant CFTR alleles. This mutation is a deletion of CTT, containing the third nucleotide of the ATC codon for isoleucine at position 507 and the first two TT nucleotides of the TTT codon for phenylalanine at position 508 within the first NBD. The wild-type ATC codon becomes ATT, which also codes for isoleucine, and the normal coding sequence of a GGT codon for glycine at position 509 remains intact. Cellular localization experiments and CFTR glycosylation experiments have demonstrated that in many cell lines the F508 mutant CFTR protein does not traffic to the Golgi network (335), a requirement for membrane expression. More recently, it has been shown that the F508 CFTR protein is mostly retained in the endoplasmic reticulum but slowly leaks to the endoplasmic reticulum-Golgi intermediate compartment (107). This incompletely processed F508 CFTR protein is eventually degraded intracellularly. Ward et al. showed that ubiquitin-F508 CFTR conjugates accumulate in cells when proteosomes are inhibited (333), suggesting that degradation of F508 mutant CFTR protein proceeds at least partially through a ubiquinated intermediate. Consistent with these results is the observation that F508 CFTR is found in extremely small quantities at the apical plasma membrane of cultured cells (335). The low level of membrane expression of CFTR in F508-homozygous cells results in low or unmeasurable chloride ion conductance (338). This reduced ion conductance is probably responsible, at least in part, for the abnormal composition of epithelial secretions in CF patients. Processing defects similar to those described for F508 CFTR have been noted in mutant CFTR alleles containing missense mutations in the third cytoplasmic loop (in the second membrane-spanning domain). These missense mutants migrate faster than wild-type CFTR in polyacrylamide gels, demonstrate sensitivity to the enzyme endoglycosidase H (suggesting incomplete processing of N-linked carbohydrates), and have chloride channel activity suggesting alterations in their "open probability" (287).
Countering these observations is that of Kalin et al. (142), who used a panel of monoclonal and polyclonal antibodies to CFTR to localize the expression of the F508 CFTR protein in tissues from CF patients. They found a tissue-specific variation in membrane expression of F508 CFTR ranging from none to levels commensurate with those of wild-type CFTR. Notably, they reported that the expression levels of the F508 CFTR protein equaled that of wild-type CFTR expression in intestinal and respiratory tract tissue sections, the two tissues primarily affected in CF patients. Thus, membrane expression of F508 CFTR may be normal, implicating aberrant protein function as the cardinal feature leading to disease. Smith et al. (303) showed that there was no chloride conductance activity in airway tissues from F508 CF patients, consistent with a loss of CFTR function in these patients. On the other hand, Engelhardt et al. (87) reported that the submucosal glands are the site in the respiratory tract where CFTR is prominently expressed, and they could not detect any CFTR protein in patients with the F508 CFTR allele. At this point the debate over whether the major problem with the F508 CFTR protein is in its expression or function is not resolved.
Interestingly, recombinant expression of F508 CFTR in either insect cells or frog oocytes resulted in levels of mutant CFTR protein in the plasma membrane similar to wild-type levels (71). This observation led Denning et al. to conclude that the F508 mutation encodes a temperature-sensitive protein (xenopus oocytes and insect Sf9 cells are routinely cultured at room temperature). They also identified a second defect in F508 CFTR: F508-homozygous cells grown extensively at reduced temperatures did not recover a level of ion conductance commensurate with the increase observed in normally processed CFTR protein (71). Thus, in these studies, the F508 CFTR exhibited a defect in chloride ion conductance. While wild-type CFTR protein has a probability of 0.34 of being in the open state, F508 CFTR protein has an open probability of only 0.13. Therefore, the F508 mutation affects CFTR by at least two distinct mechanisms, reducing the levels of protein reaching the plasma membrane and diminishing the ion channel activity of the CFTR protein that does reach the cell surface.
F508 allele of CFTR is extremely common in certain populations; it is estimated to be carried at a frequency of 2 to 5% in Caucasians of European descent. Although the F508 allele accounts for 70% of all mutant CFTR alleles, there is considerable regional variation, from as low as 27% of CF alleles in Turkey (131) to nearly 87% of CF alleles in Denmark (283). Data obtained in a multicenter epidemiological study of CF show that as of 1998, approximately 50% of genotyped CF patients were homozygous for the F508 mutation while an additional 25% were compound heterozygotes in whom one CFTR allele contained the F508 deletion (50). Cystic fibrosis occurs in males and females with approximately equal frequencies, although clinically male patients have slightly better health than do female patients (70). Although the prevalence of the CF mutation was originally theorized to be due to genetic drift (347) and was somewhat later theorized to result from a single genetic event (88), we now know that in addition to the genetic event that produced the predominant mutant CFTR allele, F508, the more than 1,000 other mutant alleles have arisen as independent events (351). The high frequency of the F508 allele of CFTR in specific populations suggests that selective pressure has been operative, perhaps due to heterozygote advantage. Studies aimed at discovering such an advantage have focused primarily on examination of physiological functions that are usually impaired in clinical CF: pulmonary function, intestinal absorption, and reproductive function. Results obtained from such studies have so far failed to uncover a consistent interpretation favoring the heterozygote advantage theory. Examination of pulmonary function among CFTR wild-type carriers and F508 CFTR heterozygotes has, in one study, suggested that the latter are protected against asthma (276). However, these results have been refuted by others (195).
Early suggestions of increased fertility secondary to decreased fetal loss, along with a greater proportion of male births among CF heterozygotes, were not confirmed when results were analyzed after ascertaining the true parentage of affected offspring used to identify obligate carriers (139). Others have reported no overall increased fecundity in F508 CFTR heterozygotes but have found that smoking was a potential modifier, with nonsmoking heterozygous parents showing increased family size and smoking heterozygous parents having decreased family size (63).
Another major hypothesis posits that mutations that reduce CFTR production or activity confer protection against a potentially fatal infection, including suggestions that individuals heterozygous for CFTR mutations are resistant to influenza (293), tuberculosis (143), cholera (102), and typhoid fever (233). The proposals for influenza and tuberculosis resistance were speculative hypotheses lacking experimental data. Although Gabriel et al. (102) found evidence for decreased fluid secretion in the intestinal lumen of heterozygous transgenic CF mice challenged with cholera toxin, Cuthbert et al. (60) could not confirm this finding. Cholera is unlikely to have been the selective factor, since Vibrio cholerae did not enter Europe, the site of concentration of mutant CFTR alleles, until 1832 during the second pandemic. It has been proposed that in order for the F508 allele to reach its current level of occurrence, over two-thirds of Europeans of reproductive age would have had to die from cholera between 1832 and 1900, when public health measures bought the disease under control (20). Clearly, this disaster did not happen. However, other diarrheal toxins such as Escherichia coli labile toxin, could have been a factor in a heterozygous advantage for CF predicated on resistance to diarrheal disease.
Research evaluating the potential role of typhoid fever as selection for a heterozygous advantage in CF was based on the finding that the CFTR protein serves as a receptor for gastrointestinal epithelial cell internalization and submucosal translocation of Salmonella enterica serovar Typhi. Since typhoid fever is life-threatening and typically affects individuals aged 3 to 19 years (136), resistance to this disease would probably confer a reproductive advantage. Pier et al. (233) demonstrated that the efficiency with which serovar Typhi invades intestinal epithelium is directly related to the amount of available CFTR protein on the epithelial surface. Thus, the level of bacterial invasion into F508-heterozygous epithelium is 80% lower than the level of invasion observed with wild-type epithelium, probably owing to the lower level of CFTR expressed on F508-heterozygous epithelia (233). In an effort to apply this model on a population scale, we examined the correlation between the incidence of serovar Typhi infection in several European countries (46) and the prevalence of the F508 CFTR allele in the same geographic location one or two generations later (see reference 326 and references therein). The hypothesis driving this comparison was that outbreaks of serovar Typhi infection can serve as a selective pressure favoring maintenance of the F508 CFTR allele. The results of this comparison are shown in Fig. 5 and suggest that a positive correlation may, in fact, exist. Note that the F508 genotype data used in this correlation indicate the percentage of mutant CFTR alleles that are F508 and that the genotype data represent the allele frequencies approximately 23 years after the reported incidence of serovar Typhi infection. It must be stressed that this comparison reflects the selective pressure applied by only one factor (serovar Typhi infection) over a time frame (3 years) which is very brief in evolutionary terms. However, given the potential relationship between serovar Typhi infection and the occurrence of the F508 CFTR allele (233), the results of this comparison are provocative.
|
F508 deletion, several other mutations have been characterized which also affect CFTR in the vicinity of phenylalanine 508, suggesting that this region of the CFTR gene is a mutational hot spot. Examples of these mutations are I506, I507, I506V, and F508C. While deletions in this region of the protein (i.e., deletions at amino acid 506, 507, or 508) result in an extremely severe disease phenotype (204, 335), missense mutations, such as I506V and F508C, are benign (157). This is consistent with the view that the severe phenotype resulting from a deletion of phenylalanine 508 is the result of a perturbation in the amino acid spacing of the protein, probably leading to misfolding of the protein and improper trafficking and maturation. Identification of these other mutations near phenylalanine 508 was hindered by two factors. The first was the mild phenotype associated with missense mutations such as I506V. Second, analytic methods such as restriction fragment length polymorphism were not able to distinguish between in-frame deletions such as I506 and the far more common F508 allele. Therefore, the presence of these other in-frame deletions was not realized until more sensitive genetic techniques were developed. | MICROBIOLOGIC ASPECTS OF CYSTIC FIBROSIS LUNG INFECTION |
|---|
|
TopPreviousNextReferences |
|---|
Over the past decade, several studies have shown that there can be important differences related to the detection of CF pathogens in the lower airway when comparing results from deep throat cultures with those obtained using BAL fluid, particularly in young CF patients. Ramsey et al. (250) found, in nonexpectorating CF patients in optimal respiratory status, a high positive predictive value of oropharyngeal cultures for the presence of P. aeruginosa and S. aureus in the lower airway, which was determined by BAL cultures, but a poor negative predictive value; 46% of younger, nonexpectorating patients had P. aeruginosa in their BAL fluid but a negative throat culture. Similarly, 21% had Klebsiella spp. in their BAL fluid but not the oropharyngeal culture. Ironically, the same group later reported opposite findings (41): a high negative predictive value of the oropharyngeal cultures for presence of organisms in the lower airway, particularly if two sequential cultures were considered (85% for one culture, 97% for two), and a lower positive predictive value (69% for one, 83% for two). Similarly, for CF patients younger than 5 years, Rosenfeld et al. (261) reported a better (95%) negative predictive value of oropharyngeal cultures lacking detectable P. aeruginosa for ruling out the presence of this organism in the lower airway but a low (44%) positive predictive value for detecting its presence in BAL fluid. Armstrong et al. (11) obtained a similar result in CF infants diagnosed via a neonatal screening program, with oropharyngeal cultures having a high (97%) negative predictive value for CF pathogens (S. aureus, P. aeruginosa, and H. influenzae) but poor positive predictive value (41%). Thus, some data indicate that CF pathogens can be present in the lower airway, but not reliably detected by throat culture, while other data, mostly from patients younger than 5 years, suggest that positive throat cultures are not necessarily indicative of pathogens in the lungs. One important caveat is that BAL fluid samples are obtained from only a small portion of the lung, leaving the possibility that pathogens might be present in parts of the lung not sampled by lavage. Therefore, it seems that the positive findings of Ramsey et al. (247), which showed detection of pathogens in the BAL at a high rate, may be more convincing than the negative findings of the other groups. Newer techniques such as fluorescent in situ hybridization (127) analysis of clinical specimens from CF patients may improve on the specificity and sensitivity of the detection of CF pathogens.
More problematic is the definition of the contribution to CF lung disease of S. aureus. This organism is usually cultured only from the nose of healthy individuals, not the throat or respiratory secretions, yet it is often considered to be among the first pathogenic organisms when isolated from the CF respiratory tract (6), usually by throat culture. Clearly the presence of S. aureus in the lower respiratory tract is representative of a pathologic situation, but the degree of pathology associated with its presence in the lungs has never been adequately assessed in CF patients. Ulrich et al. (328) located S. aureus in the lungs of three infected CF patients, and this microorganism, like P. aeruginosa, was found predominately in the mucus of obstructed airways. This finding is clearly indicative of a pathologic situation. What is not entirely clear is the proportion of CF patients with S. aureus in their lower airway causing frank disease. One likely reason why this has not been adequately determined had been the routine use of anti-staphylococcal antibiotics in this patient population, potentially preventing the progression of S. aureus infection to a highly pathologic state that could be readily identified clinically.
Nonetheless, attempts have been made to evaluate the efficacy of routine or intermittent use of antistaphylococcal antibiotics in CF patients. McCaffery et al. (191) identified 13 clinical trials of antistaphylococcal antibiotic trials in CF patients. These trials used 19 different antibiotics and a variety of clinical and laboratory outcomes and involved both intermittent and continuous administration of antibiotics. While sputum clearance of S. aureus was achieved in most studies, none documented a positive effect on pulmonary function or other clinical outcome. On the contrary, a recent study of 3,219 CF patients in the European Registry of Cystic Fibrosis demonstrated that continuous antistaphylococcal prophylaxis increases the rate at which patients' sputum cultures converted from P. aeruginosa negative to P. aeruginosa positive (251). The patients who were monitored in this 
3-year study were categorized both according to their age and according to whether they received continuous (200 or more days per year), intermittent (only during acute exacerbations), or no antistaphylococcal therapy. The results of this study showed that P. aeruginosa acquisition in the group receiving continuous antistaphylococcal therapy was significantly higher than in the those receiving no or only intermittent therapy. This difference was most significant in the 0- to-6-year age group and was not significant in the > 12-year age group. Importantly, monitoring of the lung function (forced expiratory volume in 1 s [FEV1], a measure of small airway colonization) and body mass index of the same patients showed no significant differences between the three antibiotic treatment groups during this same period, suggesting that differences observed in P. aeruginosa acquisition were not simply a function of the patients' general health (although Ratjen et al. do not completely disregard this possibility [251]). This report bolsters previous work that had been published only in abstract form (H. R. Stutman, Abstract, Pediatr. Pulmonol. Suppl. 13:542, 1994). Thus, while there is a consensus among clinicians about a beneficial effect from treatment of staphylococci associated with clearance of the organism from the sputum, there are no data indicating that this treatment leads to improved lung function or other clinical benefit. Indeed, several studies have shown that the presence of S. aureus and the absence of P. aeruginosa predicts long-term survival in CF patients after the age of 18 years (129, 130). In addition, the potential for increasing P. aeruginosa colonization as a consequence of suppression of S. aureus infection should not be overlooked.
Thus, most of the data implicating a pathogenic role for S. aureus in development of CF lung disease comes from historical findings, reasonable speculation, and the fact that the presence of S. aureus in the lower airway is judged to be clinically important and in need of antibiotic therapy. It has been suggested (112) that early infection "primes" the CF airway for later infections by P. aeruginosa. Whether there is indeed a progression from S. aureus to P. aeruginosa infection is questioned by the study of Burns et al. (41), who found evidence by culture and serologic testing for a 97.5% P. aeruginosa infection rate in CF children by the age of 3 years. Certainly, before the advent of antibiotic therapy, S. aureus was regarded as the chief infectious agent in CF patients, although it was not clear if this situation was due to a primary defect in the innate immune system of the lungs or was secondary to another aspect of CF, such as malnutrition (339). However, since antibiotic therapy has been extensively used to treat S. aureus in CF patients, recent data indicating a primary role for S. aureus in pathogenesis of CF lung disease are lacking. The common use of antistaphylococcal therapy for CF patients in many parts of the world has raised questions about whether such prophylactic treatment enhances susceptibility to infection by other agents such as H. influenzae and P. aeruginosa (15, 105), as suggested in the clinical trial discussed above. Interestingly, chronic colonization of CF airways by P. aeruginosa is reduced in regions where antistaphylococcal therapy is administered strictly on an as needed basis, rather than prophylactically (111). Thus, at the moment it is clear that conclusions supporting a pathogenic role for S. aureus in CF lung disease come from clinical observations, principally the judgment that oropharyngeal cultures positive for S. aureus might also indicate its presence in the lower airway. Definitive studies showing a positive effect from treating S. aureus when isolated in oropharyngeal cultures obtained from CF patients' clinical specimens are lacking.
|
1-antitrypsin and secretory leukoprotease inhibitor, 20 of the 27 children, including 2 of the 4 infants, had active neutrophil elastase in the BAL fluid (24). Thus, a protease-antiprotease imbalance appears early in life for CF patients, potentially contributing to lung damage. A study by Konstan et al. (161) of BAL fluid from 18 CF patients with clinically mild disease again demonstrated abundant active neutrophil elastase, even in the presence of threefold elevated levels of 1-antitrypsin. More striking are the findings of Khan et al. (148), who studied BAL fluid samples from 16 infants with CF with a mean age younger than 6 months. Despite the young age of the patients and the absence in seven patients of pathogenic bacteria detectable in BAL fluid cultures (which may miss sampling the part of the lung where pathogens are present), the infants had increased numbers of neutrophils, as well as elevated levels of neutrophil elastase, 1-antiprotease, and the proinflammatory cytokine interleukin-8 (IL-8). Interestingly, Freedman et al. recently reported that the infiltration of neutrophils into the airways of transgenic CF knockout mice was substantially reduced when the mice were orally given the fatty acid DHA (96), which appears to be deficient in the phospholipids of CFTR-knockout mice (97).
The report of elevated levels of inflammatory mediators in CF lungs in which no pathogen could be detected challenged the traditional view that the CF lung is initially normal but becomes progressively damaged by acquired bacterial infection and resultant inflammation. Researchers have begun investigating the possibility that the increase in inflammatory mediators seen in the lungs of CF patients derives from an intrinsic property of the epithelium itself, perhaps an exaggerated inflammatory response to early pathogens such as S. aureus that are cleared or are undetectable by culture. Bonfield et al. (26) isolated bronchial epithelial cells from healthy control subjects and from patients with CF and measured the amounts of the secreted anti-inflammatory cytokine IL-10, as well as those of the proinflammatory cytokines IL-8 and IL-6. Cells from normal patients secreted IL-10 but no detectable IL-6 or IL-8, whereas cells from CF patients did not secrete IL-10 but produced both IL-6 and IL-8. DiMango et al. (76) had demonstrated earlier that multiple gene products of P. aeruginosa stimulated respiratory epithelial cells to secrete IL-8 and that for a given stimulus, CF cells produced four times the amount of IL-8 than that produced by a genetically complemented control cell line. This response is linked to greater amounts of nuclear (activated) NF-
B in CF respiratory epithelial cells compared to isogenic control cells that had been genetically complemented with episomal copies of normal CFTR (75). If it is true that lung disease in CF begins before any pathogen actually infects the lower airway, then the importance of early detection of CF through neonatal screening is underscored, since early detection allows early intervention. Current screening tests are based on the elevated levels of plasma trypsinogen found in most newborns with CF.
Initiation and establishment of P. aeruginosa infection. The ubiquity of P. aeruginosa (109, 115) in the environment probably underlies the high frequency of recovery of this pathogen from CF patients. The role of P. aeruginosa in human disease is usually opportunistic. Approximately 6 to 20% of CF patients carry P. aeruginosa in their gastrointestinal tracts asymptomatically (306) and without mounting a significant immune response to the organism. CF patients can acquire P. aeruginosa in their respiratory tracts at any time, with most studies indicating that 70 to 80% CF patients are infected by their teen years. As noted above (41), P. aeruginosa infection probably initially occurs within the first 3 years of life. After the onset of chronic infection, patients experience episodic exacerbations requiring antibiotic chemotherapy. Infection may result from social contacts or may be hospital acquired, but the diversity of P. aeruginosa clones isolated from CF patients suggests that most clinical isolates originate in the environment (41, 305).
The abnormal composition of the airway secretions of the CF lung is frequently cited as the host factor that predisposes CF patients to chronic co
