The second factor is the localization and level of expression of the gene in organs, tissues, and cells. Knowledge of the pattern of expression helps define the role of the gene in the different structures. Although the gene sequence is present in all cells, significant levels of messenger RNA (mRNA) and protein are usually present only in those tissues and organs in which gene action is required. However, the relation between CFTR dysfunction and the pathophysiology of the disease is poorly understood. A fuller description requires not only details of the alterations in the gene and its protein but also information on the actual rates of CFTR-mediated secretion and their effect on the specific organs affected in the disease. In addition, each organ has intrinsic features that influence the generation of damage. We present a hypothesis that helps explain the manner in which these various elements combine to produce organ damage and lead to the final phenotype.

Several studies have shown that predictions about disease progression based on knowledge of the mutation or mutations present in patients are possible only in relation to pancreatic status [4-6]. This observation is consistent with the intrafamilial concordance of pancreatic sufficiency and insufficiency [7]. Two other gastrointestinal manifestations, meconium ileus and liver disease, show partial familial concordance [8, 9], but it is not clear why these are present in only some patients, even among those with the same genotype. Intestinal obstruction and meconium ileus have been detected in at least two thirds of the fetuses with cystic fibrosis, yet meconium ileus at birth appears in only 10% to 15% of newborns with cystic fibrosis [10]. Histopathologic changes involving biliary obstruction occur in most patients with cystic fibrosis, but liver disease does not develop in all of them. However, the definition of liver disease is not standardized, and the incidences in different medical centers are difficult to compare.

Lung disease is also not easily correlated with genotype [5, 6]. The lungs of neonates with cystic fibrosis do not show definite abnormalities, and the respiratory complications that cause the death of 95% of patients with cystic fibrosis occur progressively after the neonatal period. However, severe complications can appear early in life and lead to the death of some patients. In a study of 146 autopsies, bronchiectasis was found in all patients who survived for 6 months or longer [11], indicating that lung damage develops soon after birth.

Although a few males with cystic fibrosis have been reported to be fertile, nearly all are azoospermic because of absent or atrophied distal epididymis, vas deferens, or seminal vesicles, suggesting that in the male duct, sequelae of CFTR malfunction are present regardless of the genotype. Congenital bilateral absence of the vas deferens without pulmonary or gastrointestinal disease may represent the mild end of the cystic fibrosis clinical spectrum [12] and appears to be a phenotype conferred by a particular set of mutations. Further, because some patients with congenital bilateral absence of the vas deferens have only a single known mutation in CFTR, even heterozygosity at the CFTR locus can interact with other genetic or nongenetic factors to cause the disease [13]. Because patients with congenital bilateral absence of the vas deferens are only recognized in adult life because of their infertility, further studies are necessary to determine whether pathologic disorders are present at birth or during infancy. Male duct abnormalities, caused by progressive obstruction and obliteration, are perhaps the paradigmatic result of absent or altered CFTR.

Because differences in CFTR genotype alone cannot account for the wide spectrum of clinical manifestations of the disease, the influence of other genetic or environmental elements must be considered. For example, a missense mutation (R117H), observed in persons with one of three phenotypes (asymptomatic persons, patients with pancreatic-sufficient cystic fibrosis, and patients with congenital bilateral absence of the vas deferens), occurs on two chromosome backgrounds that lead to the three different phenotypes [14].

The expression of CFTR has been analyzed in both rodents and humans by mRNA in situ hybridization and by protein detection [15-23]. Constitutive levels of CFTR mRNA are first seen in gastrointestinal tissues in early development and are maintained throughout life. The expression is primarily restricted to less differentiated cells, including intestinal crypt cells, pancreatic duct and centeroacinar cells, bile duct epithelial cells, or specialized cells such as mucin secretory cells in gallbladder epithelia and the Brunner glands [16]. Most neonates with cystic fibrosis have pathologic changes in gastrointestinal structures, although with variable severity. These include meconium ileus and dilation of the Brunner glands and the crypts of Lieberkuhn, with considerable accumulation of secretions. Pancreatic ducts are plugged by inspissated secretions and feature mild intra- and interlobular fibrosis. Liver abnormalities include periportal fibrosis, excess mucous in bile ducts, and focal biliary cirrhosis [11].

Prenatal respiratory tissues differ qualitatively and quantitatively from postnatal lung samples. Despite significant CFTR expression during fetal life [16-18], no pathologic changes are seen at birth in most neonates with cystic fibrosis. The earliest respiratory change described is hyperplasia of submucosal glands. However, expression in submucosal glands is not seen in fetuses and can only be detected after birth, when the glands are differentiated into mucous and serous glands [19]. In adult respiratory tissues, serous submucosal glands are the main sites of expression [20], specifically in the membrane of serous cell secretory granules [21], although CFTR is also detected in some nonciliated cells at every level of the distal lung [22].

In male reproductive tissues, even though all the structures of the excurrent ducts are derived from the mesonephric duct, the head of the epididymis is the only segment that is consistently present in patients with cystic fibrosis. Interestingly, of all male ducts, this portion of the epididymis has the highest and most consistent levels of expression throughout life [23], suggesting that organ damage occurs downstream from the region with the highest CFTR function. Female reproductive tissues (including the cervix, endometrium, and fallopian tubes) and salivary and sweat gland ducts also have measurable levels of CFTR expression at various stages of development. These structures, however, do not contain significant abnormalities at birth, and the disorders seen later in life are more related to organ dysfunction than to altered structure. Overall, the temporal level of CFTR expression and the pathologic findings in neonates with cystic fibrosis are not completely correlated.

A rational theory of pathogenesis must reconcile the different functional and anatomical abnormalities in patients with the levels of CFTR expression [24]. We hypothesize that in addition to the effects of the genotype and the rate of CFTR-mediated secretion present in each epithelia, nongenetic factors, such as the properties of the affected organs, contribute to variable damage that leads to the different phenotypes (Figure 1). The structure of the ducts, especially their size and complexity, and the nature of the secretions within them must influence the degree of abnormality (Table 1). Thus, the absence or dysfunction of CFTR can lead to different functional consequences in the organs in which the gene is expressed. As a result of the diminished secretion of electrolytes and, by extension, fluid, caused by the absence of CFTR, the concentration of macromolecules in the lumen of the ducts increases significantly. Biochemical abnormalities in the macromolecules in the secretions of patients with cystic fibrosis have also been reported [25]. The proteins precipitate, form plugs, slow ductal flow, and lead to blockage. One could therefore predict that in tissues that express CFTR and have high protein or mucous loads and low flow rates, the absence of CFTR function leads to obstruction and abnormality (Table 1). On the other hand, organs that express CFTR and have a low protein or mucous load and high flow rate will sustain no significant damage Table 1 [24].

View larger version (33K): [in this window] [in a new window]   Figure 1. Flow chart describing the three factors that lead to organ damage in cystic fibrosis: the genotype, the secretion of cystic fibrosis transmembrane regulator (CFTR), and the intrinsic properties of the affected organs. Several additional factors influence the final phenotype. In the case of pancreatic disease, a clear correlation has been shown between genotype and phenotype, with a specific set of mutations (for example, R117H) leading to pancreatic sufficiency and the rest, including DF508, to pancreatic insufficiency. Male duct disease is the paradigmatic finding in male patients with cystic fibrosis, suggesting that male duct abnormalities are present regardless of the type of mutation. Intestine and liver disease are probably influenced by other genetic and environmental factors, as evidenced by the lower prevalence of meconium ileus in newborns compared with fetuses. The lung, protected by unknown factors during intrauterine development, is strongly influenced by respiratory pathogens and the host inflammatory response after birth.

Organs more vulnerable to luminal concentration defects include the pancreas, which has a high protein content in the ducts that are related to its exocrine function. A similar situation occurs in the male ducts, in which various proteins, including albumin, ₂-macroglobulin, transferrin, and androgen-binding protein, have been found in the luminal fluid of the mammalian epididymis [26]. The protein content of the body and tail is greater than that in the head, and this finding may explain why the body and the tail of the epididymis are affected instead of the head, even though the latter has higher levels of CFTR expression. In the prenatal intestine, a slow luminal flow and a moderate protein load (meconium) are present early in the second trimester. The increased viscosity of meconium in fetuses with cystic fibrosis is caused by the reduction in water content and by an increase in albumin and mucoprotein levels. An analogous situation occurs in the biliary tract of the liver, a complex ductal system in which moderate protein load (bile) is present early in the second trimester. The pathologic manifestations in these organs seem to be directly related to their inability to maintain the luminal hydration of ductal macromolecules.

The lung is a special case. One of the unique functions of the prenatal lung (which is liquid filled), in comparison to the postnatal lung (which is air filled), is the constant secretion of liquid with low protein content (100 times less than that in plasma). In most epithelial cells, a maturation-dependent switch from a secretory to an absorptive function occurs during the last trimester. After birth, the movement of macromolecules at an air-fluid interface through narrow passages leads to secretions of mucous glycoprotein. This process, together with the abnormal electrolyte transport (chloride impermeability and sodium hyperabsorption) in the respiratory epithelia of patients with cystic fibrosis, the colonization with respiratory pathogens, and the host inflammatory response [27], is responsible for the pathogenesis of the disease in the lungs. Thus, the final respiratory phenotype reflects the CFTR genotype, the organ damage that results from CFTR malfunction, and the above-mentioned environmental factors.

On the other hand, in some organs with marked CFTR expression, the abnormality at birth is more directly related to the functional consequences of absence or dysfunction of CFTR rather than to organ damage. The uterus and fallopian tubes have a low protein load in their lumens and a relatively high flow rate. Abnormal secretion levels have been noted in the cervix of neonates with cystic fibrosis, but definite uterine and fallopian tube damage has not been reported. However, functional consequence in adult women with cystic fibrosis are reflected in the patients' reduced fertility. The sweat glands are another organ in which functional consequences can be directly observed because an abnormal sweat test is the paradigmatic biochemical abnormality in the disease. Expression of CFTR in the ducts of sweat glands has been observed at both the mRNA and protein level [15, 16]. A similar situation occurs in salivary glands, which show significant ductal expression of CFTR [15, 16] but have no clinically important problems. As in the case of female reproductive organs, the protein load in sweat and salivary glands is lower than that in the other organs; the flow rate is also high. These facts may explain the absence of definite pathologic disorders in these structures and are shown in Figure 1.

Although our theory reconciles the levels of CFTR expression and pathologic disorders, some aspects need further investigation. These include an understanding of the causes of the differential prenatal and postnatal pathologic disorders. For example, what mechanisms protect the fetal lung from CFTR absence or dysfunction? In addition to the possible role of other chloride channels, the maternal environment in which the lung develops must be considered. The prenatal liquid-filled lung is almost impermeable to macromolecules and is also protected from infection. Postnatal lung pathogenesis is usually associated with infection and inflammation secondary to dense and thick mucous in an air-filled airway [25]. These differences between prenatal and postnatal lungs, in addition to other environmental factors, may explain why the lungs of patients with cystic fibrosis are protected from development of abnormalities during the prenatal period. We also do not know what factors spontaneously resolve the intestinal obstruction that is present in two thirds of the fetuses with cystic fibrosis but in less than 15% of newborns. In utero ingestion of placental fluid at late stages may clear the obstruction.

The answers to these and other similar questions may confirm (or disprove) the hypothesis, provide further insights into the pathogenesis of the disease, and help prevent organ damage in cystic fibrosis. To be successful, novel therapies derived from the discovery of CFTR [28, 29] must be based on a comprehensive understanding of the basis and source of the complex symptoms observed in patients. For example, therapeutic strategies that imitate the mechanisms protecting the lung in utero may be helpful in preventing the severe morbid conditions that develop soon after birth. Another reasonable approach may include the modification of the nature of the secretions in the affected organs by developing pharmacologic alternatives to reduce protein load or to augment flow.