Main

Patients with CF may develop a variety of gastrointestinal disorders(1, 2) that are generally considered secondary to pancreatic insufficiency. However, recent studies have demonstrated that CFTR is widely expressed in the gut, suggesting that in CF this organ could be primarily involved(3). The importance of the intestinal involvement has been recently underlined in the study of animal models of CF. In CFTR(-/-) mice, severe alterations were found in the crypts of Lieberkühn. Another prominent feature is goblet cell hyperplasia(4). Despite the data of molecular biology demonstrating CFTR expression, in patients with CF, previous histologic studies described that the structure of the intestinal mucosa is relatively conserved. The height of the villi is described as normal, and alterations have been reported only in the crypts that appear dilated and containing dense material(58). A study by SEM confirmed these findings, showing a mucosal surface with preserved architecture and covered by dense mucous membranes(9). Data of TEM are very scarce and, in a review of the literature, we have found only a single study on the intestinal mucosa of patients with CF(10). This study did not describe alterations of any of the intestinal cell types(10), and it has been widely cited in the following years, representing strong support for the hypothesis that the intestinal involvement in CF is not primary but secondary to the pancreatic deficiency. However, the described normality of the small bowel in CF is surprising considering that intestinal cells express CFTR at the apical membrane and that the small bowel shows important biochemical and metabolic alterations in patients with CF(1113). Thus a discrepancy seems to exist between biochemical and ultrastructural data. However, in the cited study(10), TEM features were obtained in osmic acidfixed specimens, and this method is presently no longer accepted for the scarce preservation of ultrastructural details and for the large number of induced artifacts. The ultrastructural normality of the intestinal mucosa in patients with CF, has been not controlled by more recent methods and, therefore, the discussion about a primary or secondary involvement of the small bowel in patients with CF is presently performed in the absence of an adequate morphologic knowledge of this structure. In our opinion, considering the high prevalence and the social cost of this disease, it could be important to exclude the existence of an ultrastructurally detectable enteropathy in patients with CF. Therefore, we have performed an ultrastructural study in a group of patients with CF and gastrointestinal symptoms. The purpose of the present study was to evaluate by ultrastructural techniques whether the intestinal mucosa is effectively normal as previously described or if ultrastructural lesions are present.

METHODS

The study group consisted of nine patients aged between 4 mo and 12 y (mean age, 5 y, 4 mo). In all cases the diagnostic criteria of CF were fulfilled. For genetical evaluation of CF, the 31 mutations most frequent in Italian populations were studied. All the patients had gastrointestinal symptoms(vomiting, abdominal pain, constipation) and/or growth failure. In none of the patients was there reported a history of meconium ileus. Some clinical data of the patients are reported in Table 1. Biopsies at the ligament of Treitz were obtained with a pediatric Watson capsule before starting any therapy for CF. For TEM, the tissues were fixed with glutaraldehyde 2% in 0.1 M phosphate buffer, postfixed in 1% osmium tetroxide in the same buffer for 1 h, dehydrated in graded ethanols, embedded in Epon-Araldite, and sectioned in a Ultracut E ultramicrotome. Ultrathin sections were stained with lead citrate and uranyl acetate and observed under an EM 10 electron microscope (Zeiss). About 10 crypts were studied in each patient.

Table 1 Clinical features of the patients with CF

For SEM, the tissues were fixed with glutaraldehyde 2% in 0.1 M phosphate buffer, postfixed in 1% osmium tetroxide in the same buffer for 1 h, dehydrated in graded ethanols, critical point dried (CPD 030, Balzers), fixed to stubs with colloidal silver, sputtered with gold (MED 010, Balzers), and examined with a scanning electron microscope (DSM 690, Zeiss).

An evaluation of goblet cells opening diameter was performed by SEM in six patients with CF. Two patients in which the intestinal surface was almost totally covered with mucus were excluded by morphometric evaluations. In each patient, four villi were selected at low magnification and in each of them five goblet cells were selected in the mid-villous zone, therefore 20 goblet cells were evaluated in each patient. The mean diameter [(largest diameter + smallest diameter)/2] of goblet cell opening was calculated at an enlargement of ×5000. Morphometrical evaluations were also performed in a control group randomly selected from our files, composed of seven children with failure to thrive or short stature, but not suffering from CF and with a well preserved structure of the intestinal mucosa. All control subjects were studied with the same protocol as the patients with CF.

Statistical comparison of the data were performed with the t test for unpaired samples

RESULTS

At light microscopy, in patients with CF, the intestinal mucosa was usually normal, and only three patients showed slightly shortened villi. These patients did not differ significantly from the others with regard to symptoms, genotype, or age. In all cases, the crypts in the lamina propria appeared normal or showed a slightly enlarged lumen.

In the patients with CF, the most characteristic feature was the presence of an abundant mucus layer covering the intestinal surface. The areas devoid of mucus showed that the mucosal architecture was usually preserved; however, in the three patients with histologic lesions, the mucosa appeared cerebriform for the presence of short villi with intervillar anastomoses. At higher enlargement, in these patients, the surface of the villi was irregular for the presence of numerous scissurae. In these patients, some alterations of the absorbing cells were detectable, but in the other patients the absorbing cells displayed a rather normal surface. In the other patients, in the villi, the absorbing cells were generally well preserved with normal appearing microvilli.

In all the patients, the goblet cells showed unusual characteristic features of the apical membrane. In some cells, the apical portion protruded from the surface with a large sack containing mucus droplets (Figs. 13). Similar mucus-containing sacks were often found at the surface of the villi, detached from goblet cells and stuck at the glycocalix of absorbing cells (Figs. 4 and 5). In other cells, the apical membrane showed irregular blebs (Figs. 6 and 7). These alterations of the apical membrane were readily visible both in SEM and in TEM specimens. Morphometric evaluations demonstrated goblet cell opening significantly wider in patients with CF than in the control group (mean diameter ± SD, 4.4 ± 0.8μm versus 2.2 ± 0.2 μm; p < 0.001). The apical membrane of the goblet cells showed atypical features (mucus containing sacks or blebs) in 70-80% of the cells examined.

Figure 1
figure 1

The SEM appearance of a representative goblet cell(arrowhead) in a normal intestinal mucosa (original magnification,×4700).

Figure 4
figure 2

With SEM, mucus-containing sacks (asterisks) are visible at the surface of the absorbing cells, detached from goblet cells(original magnification, ×3000).

Figure 5
figure 3

With TEM, a mucus-containing sack (asterisk) is visible at the surface of the absorbing cells, detached from goblet cells. The SEM and TEM features are consistent (original magnification,×5000).

Figure 6
figure 4

With SEM, in this goblet cell (asterisk) the apical membrane shows characteristic blebs (original magnification,×8750).

Figure 7
figure 5

With TEM, in this goblet cell (asterisk) the apical membrane shows characteristic blebs (arrows) (original magnification, ×4700).

An evident ultrastructural lesion was also found in the upper portion of the crypts in all the patients studied. In this portion, which can be studied only by TEM, the lumen was filled with dense material, and the basal lamina was thickened. In about the 60% of the crypt examined, the cells showed degenerative features ranging from accumulation of lysosomes in the apical portion of the cell to cytoplasmic swelling and vacuolization (Fig. 8). These lesions were evident both in secretory and in immature absorbing cells. These aspects were in evident contrast in comparison with the well preserved ultrastructure of the cells in the deepest portion of the crypt or of the absorbing cells at the surface of the villi (Fig. 9). In the control group, no ultrastructurally detectable lesions were found.

Figure 8
figure 6

With TEM, an evident cellular vacuolization(asterisks) is visible in the upper portion of a crypt. g, gland lumen (original magnification, ×5000).

Figure 9
figure 7

At TEM, the well preserved ultrastructure of the absorbing cells (e) at the surface of the villi is visible (original magnification, ×4000).

DISCUSSION

In CF, the existence of a small bowel mucosal dysfunction has been suggested by studies that demonstrate secretory-absorptive defects apparently unexplained by exocrine pancreatic insufficiency or that persist after adequate pancreatic replacement therapy(14, 15). In patients with CF, mucin has an increased sedimentation coefficient, a higher content of carbohydrates having a longer chain length, an increased proportion of fucose and galactose(16), or abnormal fucosylation(17). The three-dimensional conformation of CF mucins leads to increased antigenicity(18). A dysfunction of the chloride transporter resulting in a decreased chloride secretion has been demonstrated(19, 20), also with utilization of secretagogues and bacterial toxins(21, 22). Therefore, in accordance with clinical observations, recent biochemical data show that the primary defects in CF concern both chloride transport and mucus secretion(15). However, there is no general consensus about the morphologic counterpart of such functional and biochemical abnormalities. At light microscopy, a normal architecture of the villi in the small bowel is described as the rule in CF(514). However, occasional abnormalities have been described in small bowel biopsy specimens. There may be some minor abnormalities such as excess of mucus at the surface, and the mucosal crypts may have increased numbers of goblet cells and contain dense secretions within the lumen(23). In addition, unexplained abnormalities have been described in small bowel biopsy specimens including minimal shortening of villi, increased cellularity, and edema of the lamina propria. In 8 out of 17 studied patients, Hill et al.(24) found a thin mucosa with reduced villous height, alterations described as consistent with cow's milk sensitive enteropathy.

The intestinal mucosa is a very complex structure, composed of several cell types, and an important contribution to the knowledge of the phenotypic manifestation of CF in small bowel could be provided by electron microscopy. However, ultrastructural data are surprisingly scarce in the literature. In a study of 11 patients with CF, Freye et al.(10) revealed normal columnar and goblet cell structure and a coarse, fibrillar substance covered large areas of some biopsy specimens. In subsequent years, the results of this study have been repeatedly cited as a demonstration of the ultrastructural normality of small bowel mucosa in patients with CF(14, 15). They seemed to be confirmed by Poley(9) who evaluated with SEM the intestinal mucosa of three patients and found large areas of surface covered with sheets of mucus.

Our study is not in agreement with the previous description of an ultrastructurally normal intestinal mucosa in patients with CF(10), and we have found that an ultrastructurally detectable enteropathy exists. This enteropathy, which is difficult to detect at the level of light microscopy resolution, seems to be characterized by alterations located mainly in the crypt and goblet cells.

In our opinion, these lesion are not artifactual because they are found in specimens where the other cell types display a good preservation of the ultrastructural morphology. In addition, this ultrastructural lesion is quite different from those found in other pathologies of the small bowel, and in particular it is not consistent with cow's milk sensitive enteropathy. These findings were not previously described in patients with CF probably for the scarcity of material examined and for the methods used. The study of Freye et al.(10) was performed in osmic acid-fixed material that permits a poor preservation of ultrastructural details, and Poley(9) examined his patients only by SEM and therefore he did not study intestinal crypts. The pattern of mucosal lesion that we have found seems quite consistent with the finding obtained by methods of molecular biology. In digestive apparatus, the expression of CFTR is higher in the bowel than in the esophagus and stomach. In the small bowel, the CFTR is expressed mainly in the gland, where the lesion is more evident, whereas absorbing cells, which have only a small CFTR expression, have a rather normal ultrastructure. This expression of the CFTR seems to be very specific, and Trezise et al.(25) demonstrated that other members of the superfamily of A(TP) B(binding) C(assette) transporters such as the multidrug resistance P-glycoprotein have a complementary pattern of epithelial expression.

In our study, the two types of secretory cells displayed quite different ultrastructural patterns. In the gland, the cells showed vacuolization, swelling, and degeneration. This is probably due to a reduction of chloride secretion that is high in these cells. In the goblet cells, which secrete more complex macromolecules and in which CFTR-mediated chloride secretion is probably less important, the ultrastructural lesion is less severe and degenerative aspects are lacking. However, in mammals, the goblet cells display immunoreactivity for CFTR(26), and our study demonstrates that, in patients with CF there is an evident modification of the goblet cell phenotype, in particular, at its apical pole. This modification is probably due to a chemical-physical modification of the mucus, although a modified resistance of the apical membrane to the mucus transit cannot be excluded.

In conclusion, our study demonstrates that, in patients with CF, before starting a therapy, the intestinal mucosa displays characteristic ultrastructural lesions, mainly localized in sites where molecular studies have described the highest expression of CFTR. This type of lesion has never been described in other gastrointestinal diseases. However, the patients in our study group showed gastrointestinal symptoms or growth failure. Therefore, it is difficult to correlate directly the ultrastructural pattern with the genetic lesion and the possibility that the results may not be valid for CF patients receiving adequate pancreatic treatment exists. In addition, most patients with cystic fibrosis have low levels of essential fatty acids, a condition which is associated with ultrastructural lesions and a decrease in height of jejunal villi(27). Therefore, further studies must be done to verify that the intestinal lesion that we have described is directly related to a deficiency of CFTR in intestinal secretory cells or is a secondary consequence of a pancreatic lesion.

Figure 2
figure 8

With SEM, in patients with CF goblet cells often do not form pits, but a mucus-containing sack (asterisk) protrudes from the mucosal surface (original magnification, ×4700).

Figure 3
figure 9

The TEM appearance of the apical portion of a goblet cell in a patient with CF. It is evident the large cytoplasmic process(asterisk) that protrudes from the surface containing mucus droplets and displaying a cytoplasmic membrane (original magnification,×4700).