New and beneficial treatments that were introduced in the 1990s contributed to significant progress in helping physicians to understand and treat cystic fibrosis.

The 1990s were a decade of remarkable progress in the understanding and treatment of cystic fibrosis (CF), and that progress has contributed to ever-improving survival. CF is the most common lethal genetic disease in people of European extraction, affecting one in 4,000 infants born in the United States and Canada.¹ It is an autosomal recessive disorder; both parents of an affected infant must carry the gene, and each pregnancy carries a one-in-four chance of producing a child affected by CF. The carrier rate in Europeans and their descendants is about one in 28,1 with the carrier rate in other ethnic groups generally being lower.

The major manifestations of CF include intestinal malabsorption (which results in poor growth and abnormal stools) and chronic, recurrent lung infections that ultimately lead to early death in over 95% of those affected. Geneticists believe that the most common mutation causing CF, deltaF508, may have existed for over 50,000 years. It appears to have been dispersed through Europe in conjunction with several distinct migrations of early humans, which may account for the variable prevalence of deltaF508 in different regions.

The present understanding of the clinical manifestations and course of CF did not begin to evolve until the 1930s. The inheritance pattern was described in 1946 by Andersen and Hodges,² and the sweat test, which is still the most widely used means of diagnosing CF, was developed in the 1950s. Finally, in 1981, Knowles et al³ described an important and reproducible secondary effect of the gene: increased sodium absorption by epithelial cells. In 1983, Quinton and Bijman⁴ clearly demonstrated the characteristic cellular defect of CF, which is epithelial chloride impermeability. These two findings form the basis of current thought regarding the pathophysiology of CF.

When the report concerning chloride impermeability was published, efforts to find the gene were already well under way, terminating with a triumphant application of the techniques of reverse genetics. It is much easier to find the offending gene if a gene product is isolated first. In reverse genetics, the search takes place without any knowledge of the gene product. In 1985, a report was published linking the gene for CF and an enzyme, paroxonase, to chromosome 7. Within a few months of this discovery, three groups simultaneously reported in Nature the approximate location of the CF gene on the long arm of chromosome 7. This led to a frantic, highly competitive search that lasted until 1989, when the CF gene was isolated by Tsui and collaborators^5-7 in Toronto and Michigan. It was named the CF transmembrane regulator (CFTR) gene; it encodes a membrane-spanning glycoprotein some 1,480 amino-acid residues in length. The most common mutation, deltaF508, contains a deletion of three base pairs that results in the deletion of phenylalanine in position 508 of CFTR.

In 1990, it was thought that as few as five or six different mutations might cause CF. As of 2001, nearly 1,000 disease-causing mutations of the CF gene have been described around the world. These have been classified into a number of mechanism categories. Class I mutations do not produce full-length or functional protein. Class II mutations, which include deltaF508, are not processed correctly and cannot progress through the endoplasmic reticulum. deltaF508 folds incorrectly because of the deletion of a single phenylalanine at position 508 of the CFTR protein, rarely reaching the cell membrane. Instead, it is rapidly degraded before it leaves the endoplasmic reticulum (because its abnormal folding signals the cell to reject and degrade it). Class III mutations, after reaching the cell membrane, are defectively regulated. Class IV mutations are correctly inserted and regulated, but conductance is low. Examples of all of these exist, and mutations in each class have characteristics in common. Mutations in classes I and II are generally associated with pancreatic insufficiency. Classes III and IV are generally associated with pancreatic sufficiency and milder lung disease.

The CF genotype correlates closely with pancreatic function. There is also good correlation of the genotype and an important cause of male infertility: congenital bilateral absence of the vas deferens. In contrast, the genotype correlates only in broad terms with the severity of lung disease, the most lethal aspect of CF. It is known that the CFTR gene interacts with a number of other genes, including a group of newly described modifier genes that might partially explain the complex relationship between genotype and pulmonary phenotype.

Pathophysiology of CF
CF lung disease is synonymous with sticky mucus, and the connection between the CF gene and sticky mucus would seem straightforward. CFTR is a chloride channel and is clearly involved in controlling the volume and electrolyte composition of sweat. A similar dysfunction in the lung could explain many of the manifestations of CF via hypotheses of abnormal salt concentration and low volume of airway surface liquid. Unfortunately, most evidence to date suggests that airway surface liquid is isotonic and of normal volume, so that the connections between the gene defect, the known functions of CFTR, and the lung disease of CF are still unknown.

CFTR is expressed in the lung primarily in the serous cells of the airway submucosal glands. It is unlikely that these cells control the composition or volume of airway surface liquid; however, they probably control the composition, hydration, and quaternary structure of the mucins. Thus, high viscosity and other characteristics of sticky mucus might result from CFTR expression in the serous cells, but this has not yet been determined.

In spite of the lack of agreement on the basis of the chronic lung disease of CF, there is general agreement that the characteristic disorder in the airways involves three things: obstruction by thick, sticky mucus; chronic infection; and inflammation. There is, however, no agreement as to which of these is the starting point for this cycle of lung destruction. There is evidence to support any one of the three as the entry point. How and when the mucus becomes sticky are not yet well understood. Over the past few years, it has been learned that lung inflammation begins early in CF, and the products of unremitting inflammation account for much of the damage done there.

A number of new therapies are directed at controlling inflammation and decreasing the damage that it causes. CF research has led to the discovery of a new class of antibacterial molecules endogenous to the lung: the defensins. Exploitation of this discovery may result in more effective strategies for controlling lung infection. Prevention of lung infection with resistant bacteria is a goal of several lines of research, as is the development of an antipseudomonal vaccine. As a result of all this activity, survival and quality of life have improved and will continue to increase, in steps both large and small.

Gene therapy
When the CFTR gene was first isolated, some thought that gene therapy would follow quickly. After all, the airways provide an easy route to the target organ. Packaging the CFTR gene in an appropriate vector and then inhaling it directly down the airways seem straightforward enough, but early optimism soon faded as numerous unforeseen challenges arose.

Many viral and nonviral vectors have been evaluated in humans, with limited success. Although it is estimated that as little as 10% of normal CFTR function will suffice, achieving this level has been an elusive goal. Some of the failure to achieve induction of CFTR was blamed on the significant immunologic responses triggered by the viral vectors. Consequently, researchers turned to smaller, less provocative organisms. Most work currently focuses on adeno-associated virus (AAV), a very simple, single-stranded organism with only two genes. Although it seems less provocative to the immune system, AAV initially failed to induce expression of the normal gene. It turns out that AAV preferentially binds to receptors located on the basolateral (tissue) side of the epithelial cells that line the airways, resulting in much less efficient uptake when it is delivered via aerosol. In addition, when entering from the apical side, AAV is tagged for delivery to an intracellular trash can, preventing expression. By blocking the tagging process and redirecting the AAV to binding sites on the apical surface of the airway epithelium, the airway cells can be infected and the gene can be expressed for 150 days or more.

Effective packaging of a large gene like CFTR, some 750,000 base pairs long, presents additional challenges. Such large molecules are more likely than smaller molecules to be recognized by the cell as invaders. This might be overcome by dividing the gene into several parts and reassembling it after it enters the cell.

At this time, it is too early to know whether these new strategies will achieve greater expression and correction of the basic CF defect than past efforts did. Correcting the defect in the epithelial cells that line the lung will not work if the problem begins in the mucin-secreting cells of the submucosal glands. Researchers are looking at ways to target these submucosal-gland cells.

Improving CFTR function
If replacing a defective gene is not easily accomplished, improving the function of the inherited CFTR protein is a reasonable alternative strategy. Several years ago, it was shown that cooling CF cells or treating them with glycerol improved the folding of deltaF508 and increased its delivery to the apical cell membrane. Once in the cell membrane, deltaF508 functions well enough that CF-related lung disease might not progress, but neither cooling nor glycerol treatment is a practical clinical approach. Finding some practical means of improving deltaF508 delivery and function would affect the 90% of CF patients who have at least one deltaF508 gene and would, almost certainly, be a huge step forward.

High-throughput screening is a technique that is being applied to the search for compounds that promote normal folding of deltaF508. Flavinoids and benzoquinoliziniums have shown some promise. Of about 500 compounds in these families already screened for activity, six have strong effects on CFTR activity in isolated cells. High-throughput screening allows thousands of such promising compounds to be tested each day, greatly enhancing the ability to search for effective new treatments. This has been referred to as the needle in the haystack approach; it becomes viable only when the number of compounds screened per day becomes quite large. There is a great deal to be done before clinicians will be able to take advantage of this current work.

In addition to increasing the amount of CFTR protein that reaches the apical membrane, it is also logical to improve the function of defective CFTR that reaches the membrane. A number of genes have been described that have products that interact with CFTR to regulate both the channel’s open time and the survival of CFTR in the cell membrane. These studies are very preliminary, and no clinical trials have resulted yet, but this is a promising avenue of inquiry.

Bypassing the defect
It is thought that the lack of clinically significant lung disease in animals with the CF gene might result from the presence of alternative chloride channels that allow the cells to bypass the CFTR defect. For example, of the half dozen and more mouse models created since the discovery of the CF gene, none have significant lung disease. Alternatively, the lack of submucosal glands in mice might explain this important difference. Bypassing the defect is an important strategy and is being investigated in at least six clinical trials. Some of the molecules under study have been available for some time; these include gentamicin, an antibiotic, and phenylbutyrate, a drug used to treat Hansen disease. The medical community can expect to hear more about this approach in the next few years.

Another alternative strategy is the blocking of the excess absorption of sodium that occurs in CF epithelium, which has an effect similar to increasing chloride secretion. Amiloride, a blocker of apical sodium channels, has shown some limited promise in this regard, and modifications of amiloride might produce a greater impact. Combining amiloride with chloride-channel activators might provide an additive benefit.

Mucolytics
The sticky mucus in CF lungs is, at least in part, the result of infection and inflammation, and mucus plugs are a nearly universal finding in CF lung disease. Therefore, doing something directly to affect the products of infection and inflammation in mucus might help to alleviate some of the airway obstruction.

The first product to reach the marketplace in the 1990s as a result of this strategy is recombinant human DNase (rhDNase), which breaks down extracellular DNA. The infected secretions in CF lungs contain as much as 10% DNA as a result of the breakdown of bacteria and neutrophils in chronically infected lungs. DNase alters the physical properties of CF mucus, both in vitro and in vivo, so that patients cough less and have less thick sputum. Currently, about 49% of CF patients in the United States receive rhDNase.8 Its impact on lung function is small but significant, on average, and there is significant variation in individual response. In a phase III trial,⁹ there was a greater than 10% difference after 1 month of use and only a 4.3% difference in forced expiratory volume in 1 second (FEV1) after one year. Using it twice a day provided no more benefit than once-daily dosing. Frequency of pulmonary exacerbations decreased by about 34% (a more impressive effect). Subsequent clinical experience with this drug confirms the early impressions. Some patients seem to respond well and others appear to see little clinical impact. There are no reliable predictors of who will respond, and some argue that the benefits to lung function are less important than the decrease in respiratory-tract infections. DNase produces no benefit in chronic bronchitis or non-CF bronchiectasis.

Recently, rhDNase was tested¹⁰ in young patients with mild lung disease to see what impact it would have. After 6 months, it produced only a 3% benefit in FEV1, but it decreased the frequency of intravenous antibiotic courses by 34%. Because the frequency of such courses in this population is low, however, the cost-benefit ratio of this treatment might not greatly favor its use.

Another compound studied that might decrease sputum viscosity is gelsolin, which breaks down polymeric filamentous actin (F-actin) (an important component of mucus). Polymeric F-actin actually makes up about 10% of CF sputum, but clinical studies¹¹ of gelsolin have, to date, been somewhat disappointing. Some early data¹² seem to show that gelsolin, and even DNase, might activate inflammatory mediators such as interleukin-8. If this is an important effect, it might explain the disappointing impact of these compounds on lung function.

An interesting new approach to thinning mucus in CF patients is the use of low dose macrolide antibiotics. This class includes erythromycin, clarithromycin, and azithromycin. It appears that macrolides interfere with the ability of pseudomonads to make alginate, a major component of the protective coating made by these bacteria. When pseudomonads reach a critical density, they generate alginate in large quantities, creating a biofilm that interferes with antibiotics and with the clearance of mucus from the airways. This effect of macrolides was first documented in a condition seen almost exclusively in Japan called diffuse panbronchiolitis. Although the cause of diffuse panbronchiolitis is unknown, the use of macrolides in these patients (who are, like CF patients, chronically infected by Pseudomonas aeruginosa) has cut long-term mortality from 90% to less than 30%. A multicenter trial of macrolides for CF patients will start in 2001.

Anti-inflammatory drugs
A report in Lancet¹³ in the mid 1980s suggested that every-other-day doses of steroids for CF patients provided significant benefit. A subsequent multicenter trial¹⁴ funded by the CF Foundation confirmed the presence of a benefit for lung function, but also revealed a number of serious long-term side effects. There were about 285 patients enrolled in the three arms of the study (placebo, prednisone at 1 mg/kg every other day, and prednisone at 2 mg/kg every other day). The high-dose arm was stopped early by a safety-monitoring committee because of a very high risk of developing CF-related diabetes and growth retardation. Even in the 1 mg/kg group, diabetes occurred in 12 of 95 subjects (13%) and growth retardation was seen in 24 of 95 subjects (25%). Even after the study (and the steroid use) had been over for 6 to 7 years, the boys who received steroids still lagged behind those in the placebo group in height by 13 percentile points.¹⁵

In spite of the problems encountered, this study sparked a great deal of interest in anti-inflammatory therapy for CF patients. The evidence for early inflammation in CF lungs also has mounted over the years, adding to the impetus for clinicians to consider some kind of anti-inflammatory therapy for their patients. As a result, about 38% of patients in the United States are using long-term inhaled steroids.¹⁶ Some of this use is justified by the presence of asthma, but some steroids are given in the hope that they will improve outcomes, even though several studies^17-19 of inhaled steroids in CF do not support this practice.

There are data to support the use of high-dose ibuprofen in CF patients. In a 4-year study²⁰ of a small number of patients, the rate of deterioration in lung function was more than eight times slower in young patients using ibuprofen than in the placebo group. In older patients, presumably representing more severe lung disease, there was little impact. There was no difference in hospitalization rates between groups. Another observation of the first study suggests that the benefit is maintained as long as the patients continue to use ibuprofen. The drug appears to be safe, but concerns about long-term safety and its lack of impact on the frequency of acute infections probably have interfered with widespread adoption of this therapy. It is currently used by only 6% of CF patients in the United States.⁸

At least four other anti-inflammatory drugs are in preclinical or phase I studies in CF patients. This is an important area of emerging therapy in CF, and we expect to see many changes over the next decade.

Fatty acid therapy
A very interesting potential treatment for CF focuses on fatty acids. Fatty-acid composition in cell membranes is important to several key cell functions. Fatty acids, to some extent, regulate inflammation, mucin secretion, membrane fluidity, and membrane trafficking. In addition, CF patients exhibit alterations in fatty-acid metabolism. A key abnormality might be a decrease in docosahexaenoic acid (DHA). Through a series of complex connections with other essential fatty acids, a defect in DHA metabolism might increase the production of arachidonic acid and, subsequently, inflammatory molecules.

In recent experiments²¹ on CF mice fed DHA, pancreatic and intestinal morphology approached that observed in normal mice. In addition, it markedly decreased neutrophil infiltration of the lungs of CF mice in response to aerosolized Pseudomonas endotoxin. Other fatty acids did not produce this benefit. In fact, some other fatty acids actually worsened the pancreatic disease. Clinical trials in humans will start soon, so clinicians will know, in a few years, whether DHA will do for people what it does for mice.

Dennis W. Nielson, MD, PhD, is director, division of pulmonary medicine, and director, cystic fibrosis center, The Children’s Medical Center, Dayton, Ohio; professor of pediatrics, Wright State University School of Medicine, Dayton; and chair, Ohio Cystic Fibrosis Consortium.

References
1. Hamos A, FitzSimmons SC, Macek M, et al. J Pediatr. 1997;132:255-259.
2. Andersen DH, Hodges RG. Celiac syndrome. Genetics of cystic fibrosis of the pancreas with a consideration of the etiology. Am J Dis Child. 1946;72:62-80.
3. Knowles M, Gatzy J, Boucher RC. Increased bioelectric potential difference across respiratory epithelia in cystic fibrosis. N Engl J Med. 1981;305:1489-1495.
4. Quinton PM, Bijman J. Higher bioelectric potentials due to decreased chloride absorption in the sweat glands of patients with cystic fibrosis. N Engl J Med. 1983;308:1185-1189.
5. Kerem BS, Rommens JM, Buchanan JA, et al. Identification of the cystic fibrosis gene: genetic analysis. Science. 1989;245:1073-1080.
6. Riordan JR, Rommens JM, Kerem BS, et al. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science. 1989;245:1066-1072.
7. Rommens JM, Ianuzzi MC, Kerem BS, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science. 1989; 245:1059-1065.
8. Cystic Fibrosis Foundation. Patient Registry—1999. Bethesda, Md: Annual Data Report; 1999.
9. Fuchs HJ, Borowitz DS, Christiansen DH, et al. Effect of aerosolized recombinant DNase on exacerbation of respiratory symptoms and on pulmonary function in patients with cystic fibrosis. N Engl J Med. 1994;331:637-642.
10. Kostan MW, Tiddens HA, Quan JM, Sy J, McKenzie S, Wohl ME. A randomized, placebo-controlled trial of two years’ treatment with dornase alpha (Pulmozyme) in cystic fibrosis patients aged 6-10 years with early lung disease. Pediatr Pulmonol. 2000;20:S442.
11. Tullis DE, Corey M, Dabora R, Scaramucci J, Magilavy D. A double blind, placebo-controlled dose-escalating, tolerability study of inhaled recombinant gelsolin in cystic fibrosis [abstract]. Pediatr Pulmonol. 1998;15:S265.
12. Perks B, Shute JK. DNA and actin bind and inhibit interleukin-8 function in cystic fibrosis sputa. In vitro effects of mucolytics. Am J Resp Crit Care Med. 2000;162:1767-1772.
13. Auerbach HS, Williams M, Kirkpatrick JA, Colten HR. Alternate-day prednisone reduces morbidity and improves pulmonary function in cystic fibrosis. Lancet. 1985;2:686-688.
14. Eigen H, Rosenstein BJ, FitzSimmons S, Schidlow DV. A multicenter study of alternate-day prednisone therapy in patients with cystic fibrosis. Cystic Fibrosis Foundation Prednisone Trial Group. J Pediatr. 1995;126:515-523.
15. Lai HC, FitzSimmons SC, Allen DB, et al. Risk of persistent growth impairment after alternate-day prednisone treatment in children with cystic fibrosis. N Engl J Med. 2000;342:851-851.
16. Study Center Report. Epidemiologic Study of Cystic Fibrosis. Cary, NC: Clintrials Inc; 1999.
17. Van Haren EH, Lammers JW, Festen J, et al. The effects of the inhaled corticosteroid budesonide on lung function and bronchial hyperresponsiveness in adult patients with cystic fibrosis. Res Med. 1995;89:209-214.
18. Balfour-Lynn IM, Klein NJ, Dinwiddie R. Randomized controlled trial of inhaled corticosteroids (fluticasone proprionate) in cystic fibrosis. Arch Dis Child. 1997;77:124-130.
19. Bisgaard H, Pedersen SS, Nielsen KG, et al. Controlled trial of inhaled budesonide in patients with cystic fibrosis and chronic bronchopulmonary Pseudomonas aeruginosa infection. Am J Crit Care Med. 1997;156:1190-1196.
20. Konstan MW, Byard PJ, Hoppel CL, Davis PB. Effect of high-dose ibuprofen in patients with cystic fibrosis. N Engl J Med. 1995;332:848-854.
21. Freedman SD, Shea JC, Blanco PG, Alvarez JG. Fatty acids in cystic fibrosis. Curr Opin Pulmonol Med. 2000;6:530-532.