Main

CF, the most common, severe recessive disorder in Caucasian populations, is caused by mutations in the CF transmembrane regulator gene (CFTR)(1–3). Patients with CF have abnormalities of ion transport at many epithelial surfaces, including the gastrointestinal and respiratory tracts, sweat glands, and reproductive tissues(4). Because the protein product of CFTR is a Cl- channel(5), it is likely that the absence or deficiency of Cl- transport underlies the physiologic defects in CF patients. Additional abnormalities that have been observed in CF patients, including increased Na+ absorption in the respiratory tract(6) and the presence of excess mucus in various exocrine glands(7), are probably secondary consequences of the primary defect, although the mechanisms by which these are generated are not yet known. Similarly, the relationship between the primary defect and the clinical manifestations that lead to the morbidity and mortality of CF patients, including colonization by Pseudomonas, has not been completely unraveled(8).

The discovery of the defective gene may lead to specific treatments, including gene therapy(9), that correct the underlying biochemical problems. Development of such treatments, as well as understanding the relationship between CFTR dysfunction and pathology, should be facilitated by animal models. Although no natural models exist, several strains of mice defective in CFTR function have been developed by targeting the endogenousCftr locus. The mouse models have been produced by disruption ofCftr(10, 11) or by complete(12) or partial(13) duplication of one of the exons. In the case of two of the models(10, 11) the majority of the mice die during the first 3 wk after weaning. The cause of death appears to be intestinal obstruction, indicating that CFTR deficiency in mice leads to more severe gastrointestinal complications than are seen in human patients. The early mortality of these Cftr-/- mouse models precludes their use for long-term studies of therapy and for the analysis of the relationship between the biochemical defect and the development of pathology. In particular, it is not known whether these animals, if they survive long enough, will develop pulmonary disease similar to that observed with time in CF patients.

We report here the results of a study to prolong the lifespan of the CF mice developed by Snouwaert et al.(10) through the use of a specialized diet. We also report the analysis of the histopathology and function of the intestinal, respiratory and reproductive tracts in these long lived Cftr-/- mice. The histopathology in the pancreas is also described.

METHODS

Breeding and care of Cftr-/- mice. All experiments were carried out under protocols approved by the HSC Animal Care Committee. A breeding colony was established using heterozygous mice (Cftr+/-)(10) obtained from Dr. Beverly Koller, University of North Carolina. Tail clip samples of 14-d-old mice were processed for analysis of genotype as described(10). Mice identified as Cftr+/- were selected and placed in breeding pairs at 8 wk of age. All offspring from these breeding pairs were tail clipped and identified by ear notch at 14-16 d of age. Heterozygous(Cftr+/-) breeding pairs(10) were as productive as non transgenic mice of similar background. Animals that died early in the neonatal period (1-3 d) were not genotyped. Thereafter, newborn mice survived to weaning at 21 d. The average litter size at weaning was 8.3 mice with a mean weight of 9.5 g. In 10 consecutive litters studied,Cftr-/- mice were consistently the smallest in the litters with a mean weight of 6.9 g at weaning. None of the homozygousCftr-/- mice was within the weight range of their wild-type or heterozygous litter mates. Mice identified as Cftr-/- were weaned on to a special diet by 21 d of age.

The liquid mouse diet used (Liquidet F3017, Bio-Serv, Frenchtown, NJ) was prepared in sterile water according to the instructions of the manufacturer. Agway Rodent Chow (RMH #1000, Agway, Syracuse, NY) was used a standard solid chow. The composition of the liquid and solid diets was, respectively: protein(liquid: 14%; solid: 14%), fat (5%; 6%), fiber (1.5%; 4.5%), ash (5%; 8%), carbohydrate (69%; 63%). The protein component of the liquid diet was hydrolyzed casein. The caloric content of the liquid diet is about 1000 kcal/L, and an adult mouse will drink approximately 20 mL/d (20 kcal). Similarly, the caloric content of solid chow is 4 kcal/g, and an adult mouse will eat approximately 5 g/d (20 kcal). The liquid diet was fed to the animals in glass liquid mouse feeders (F3018, Bio-Serv, Frenchtown, NJ) suspended in the microisolator cages on stainless steel holders manufactured for this purpose. The feeders were sterilized by autoclaving and replaced daily. Fresh diet was also provided daily.

Mice were housed in a barrier unit in sterile microisolators with corncob bedding, changed daily, and provided with sterile water in addition to the liquid diet. Mice were kept in a 12-h light-dark cycle, and the room was ventilated at 20 air changes/h with high efficiency particulate filtered air.Cftr-/- mice were not separated by sex, and efforts were made to introduce weanling animals into cages with older females so that they could be introduced to the liquid feeders promptly.

Assessment of growth and survival. Groups ofCftr-/- mice on liquid diet were weighed weekly from weaning to 8 wk of age, with the weights of males and females tabulated separately. Twenty animals were measured from each group of males and females for each age group. These weights were compared with identical measurements on control mice(both wild type and Cftr+/-) fed either mouse chow or liquid diet, and housed in identical conditions within the same room.

To determine whether survival was altered by the liquid diet when compared with commercial mouse chow, 23 Cftr-/- mice were weaned to mouse chow at 21 d of age. Animals were killed if found moribund, but most were found dead. Autopsies were performed on all animals. Survival was compared with that of 26 Cftr-/- mice born at similar times and placed on the liquid diet no later than at 21 d of age. No mice had other unrelated congenital anomalies. Survival was recorded to 60 d.

Histopathologic studies. Groups of mice, both male and female, were killed and prepared for histologic evaluation at weaning (21 d), 40-60 d, and 80-100 d of age. Some animals that had become ill and were electively killed were also included in some of these studies. The mice were anesthetized with an intraperitoneal injection of 3.6% sterile chloral hydrate, 0.1 mg/10 g of body weight. Once anesthetized, the chests were opened and a 25-gauge needle introduced first into the right and then the left ventricles. A slow infusion of 10% buffered formalin (total volume 2 mL/ventricle) was introduced into each ventricle, and the vena cava was opened to allow release of the formalin. The lumen of the gut was then perfused with 15 mL of buffered formalin introduced into the lumen of the duodenum. Infusion pressures were maintained at a level to just distend the serosa as the lumen filled. The rectum was opened to allow drainage of the perfusate. The trachea was then opened, and a 22-gauge short angiocatheter was introduced and tied securely into the lumen. The lung and heart were removed en bloc from the chest and suspended in a tube of 10% neutral buffered formalin. The angiocatheter was attached to a 3-mL plastic syringe with no plunger. Buffered formalin was introduced into the syringe so that the lung slowly filled. The syringe was kept topped up with formalin so that a fixation pressure of 3-5 cm was maintained. After 3 d, the lungs were removed from the device, and sections were cut for histologic analysis.

The gall bladder, pancreas, duodenum, ileum, colon, liver, lung, nose, and gonads were taken for histologic analysis. To ensure uniformity of gut sampling, the samples were taken at the following locations: duodenum, 1 cm distal to the pylorus; ileum, just proximal to the ileocecal sphincter; and colon, just distal to the ileocecal sphincter. Two cross-sections and two longitudinal sections were made for each region of gut. The tissues were rinsed, dehydrated through 100% ethanol, and embedded in paraffin. Serial paraffin sections, 4-5 μm thick, were mounted onto glass slides and deparaffinized. Sections for examination of basic morphology were stained with Harris' hematoxylin for 10 min, washed multiple times with water, counterstained with eosin, and rinsed. Adjacent sections for examination of glycoprotein content were stained with 1% Alcian blue in 0.1 N acetic acid, pH 2.5-3.0, for 10 min, rinsed with water, and counterstained with 1% aqueous neutral red for 5 min. Stained sections were dehydrated, and coverslips were resin-mounted for photomicroscopy. All sections were photographed on a Reichert-Jung Polyvar photomicroscope by bright-field illumination using a 40× objective lens. Using histologic landmarks, adjacent sections were matched and photographed to determine both the morphology and glycoprotein content of the same region.

Ussing chamber short-circuit measurements. Isc measurements were carried out using isolated segments of intestinal epithelia. After cervical dislocation, tissues were removed from mice, cut longitudinally, rinsed with buffer (defined below), and subsequently mounted unstripped of underlying connective tissue across Ussing chamber oblong apertures (0.22 cm2) and supported by micro-pins. The potential difference across the tissue was short-circuited using a voltage-current clamp(Physiologic Instruments, San Diego, CA). To determine transepithelial resistance, transepithelial voltage was periodically stepped for 0.5 s from 0 mV to a preset voltage, and resistance (R) was determined from Ohms law (r = dV/dI). Experiments were carried out in medium with the following composition (mM): 118.5 NaCl, 3.5 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 K2HPO4, 25 NaHCO3, 10 glucose (pH 7.4). During experiments, the medium was gassed with 5% CO2/95% O2 at 37°C. To raise intracellular concentrations of cAMP, a mixture containing isobutyl-methylxanthine and 8-4-chlorophenylthio-cAMP, and forskolin was introduced into the solutions, bathing both sides of the tissue, to generate final concentrations of 100, 100, and 10 μM, respectively. Statistical comparisons among groups used an unpaired t test, whereas analyses of temporal, agonist, or drug-induced changes used pairedt tests.

Measurements of pulmonary function. Pulmonary function measurements were made on 15 Cftr-/- (age 175 ± 40 d) and seven control mice (age 213 ± 58 d), and full curves were obtained on six Cftr-/- and four control mice. Although the two sets of mice were matched for age, the Cftr-/- mice were significantly smaller than the controls (mean weight 23.7 versus 33.5 g, p < 0.005). Therefore, the measurements were normalized per body weight. The mechanical properties of the lungs were measured using the passive flow volume technique(14) modified for mice. The snout of the mouse was pressed gently into a silicon rubber sealed face mask. The tidal volume was then measured with a high sensitive low dead space pneumotachograph (series 8431B, Hans Rudolph, Inc., Kansas City, MO; Fleisch model comparison 0000) with the flow signal integrated for volume. Mask pressure was also recorded. The end inspiration was identified by computer, and the airway was occluded with a solenoid valve to induce the Hering Breuer reflex. The occlusion was released after 100 ms, and the subsequent passive flow volume curve was recorded. The compliance and resistance of the respiratory system was calculated from the curve.

Studies of breeding performance. Breeding pairs ofCftr-/- mice or of Cftr-/- and heterozygous mice of both sexes were established at 8 wk of age. Liquid diets were maintained for all mice, even if one of the pair was heterozygous. The fertility of the heterozygous animals was established by successful matings with alternate heterozygous partners. For experiments to test the fertility ofCftr-/- males on solid diets, animals were selected at 8-12 wk for large body size and general good health. Only mice with no recorded history of gastrointestinal illness and that had attained a body weight of 20 g were used. These animals were each placed into a cage with two to four heterogyzous females, 6-8 wk of age, to determine whether sufficient numbers of Cftr-/- offspring could be produced before the death of the male to make it a viable breeding undertaking. The mice were provided with liquid and solid food for 5 d to permit the males to adapt to the solid diet.

RESULTS

Survival and Growth of Cftr-/- Mice

As had been previously reported(10),Cftr-/- mice died within 31 d (10 d after weaning) when maintained on solid food (Fig. 1), and the gastrointestinal tracts were invariably obstructed, usually at the ileocecal sphincter or in the colon (not shown). CF patients commonly suffer from subacute bowel obstruction due to distal intestinal obstruction syndrome(15), which is effectively treated with a balanced fluid and electrolyte lavage solution. We therefore reasoned that a liquid nutrient diet might effectively maintain an unobstructed intestine inCftr-/- mice and prolong their survival.Cftr-/- mice placed on this diet at weaning had close to 90% survival at 60 d (Fig. 1). With few exceptions, liquid-fed control mice (both wild type and Cftr+/- litter mates) gained weight at the same rate as those on solid diet (Fig. 2). Cftr-/- mice remained consistently smaller than both controls, with weights reduced by 18-34% (p < 0.0001)(Fig. 2).

Figure 1
figure 1

Survival of Cftr-/- mice placed on liquid or solid diets at weaning. All 23 mice weaned to solid chow died within 31 d; 23 of 26 mice weaned to liquids were still alive at 60 d. Two deaths occurred before 26 d and one at 32 d. These latter animals were runted and had intestinal obstruction. One animal had been ill at the time of weaning, did not thrive on the liquid diet, and had evidence of inflammatory bowel disease.

Full size image
Figure 2
figure 2

Weights of Cftr-/- mice fed liquid diet compared with control mice fed liquid or solid diet. Each group contained 20 animals. Cftr-/- mice were smaller than both groups of controls at all ages (**p < 0.0001). Control mice fed solid and liquid diets were within the same weight range except for females at 4 and 5 wk and males at 6 and 8 wk, when the liquid-fed animals were smaller than those on solid chow (*0.005 < p < 0.02).

Full size image

To determine whether long-lived Cftr-/- mice remained susceptible to obstruction, nine mice on the liquid diet until at least 125 d were placed on the solid diet. Two animals died within 4 d of gastrointestinal complications, and four other animals survived between 27 and 40 d and then died of either bowel obstruction or gastrointestinal bleeding. The remaining three were still alive after 100 d on solid food.

Histopathology of Cftr-/- Mice on the Liquid Diet

Intestinal tract. Observations were made on four mice per test group. No morphologic differences were observed at weaning (21 d) between the intestinal tract of healthy Cftr-/- mice(Fig. 3,A-D) and control mice (Figs. 4,A-D). The mucosae of healthy, liquid-fed 60-70-d-oldCftr-/- mice demonstrated substantial growth, with ileal crypts becoming deeper than those of 21-d-old mice (Fig. 5A). Unexpectedly, stored mucin granules in goblet cells stained less intensely (i.e. paler blue) than at the earlier age, and secreted mucin was present in the external milieu (Fig. 5B). Colonic crypts were also deeper, the lower crypt goblet cells were cavitated, and many epithelial cells appeared cuboidal rather than columnar(Fig. 5C), suggesting accelerated mucin secretion(16). The crypt lumina were filled with mucin(Fig. 5D), and its decreased intensity suggested that inCftr-/- mice both stored and secreted mucins are chemically different from those at 21 d.

Figure 3
figure 3

Light microscopy of the ileum and colon of Cftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 21-d liquid-fed Cftr-/- female.(A) Ileum. The epithelium contains columnar absorptive cells and goblet cells. Paneth cells (arrow) are visible at the base of the crypts. (B) Ileum. Alcian blue-stained mucin is visible in goblet cells, and a little secreted mucin is present in the crypt lumina or intervillus spaces. (C) Colon. Goblet cells are numerous in the crypts, but few are present on the surface. (D) Colon. Goblet cells contain uniformly densely stained mucin granules. Secreted mucin is visible streaming from the epithelium and lining the crypt lumina (arrows). No abnormalities are noted in any of the sections.

Full size image
Figure 4
figure 4

Light microscopy of the ileum and colon of Cftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 21-d liquid-fed wild-type female. (A) Ileum. The mucosal thickness is the same as that in Figure 3A.(B) Ileum. Alcian blue-stained mucin is visible in goblet cells; little is present in the lumen. (C) Colon. Goblet cells, although present, are less robust than those in Figure 3C.(D) Colon. Mucin in goblet cells and accumulated at the base of the crypts stains dark blue.

Full size image
Figure 5
figure 5

Light microscopy of the ileum and colon of Cftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 60-d liquid-fed Cftr-/- male.(A) Ileum. Mucosal thickness has increased due to the presence of deeper crypts. (B) Ileum. Goblet cells are apparent on the villi and in the crypts, but stored mucin stains fainter than that in wild-type mouse(Fig. 6B). Crypt lumina are not enlarged, but intervillus spaces contain streaming mucin. (C) Colon. The crypt lumina are dilated and the epithelium somewhat flattened. Some goblet cells within the crypt are cavitated (arrowheads), suggesting accelerated mucin secretion. (D) Colon. Goblet cells are present in the epithelium, but the mucin granules stain less intensely than those of the wild-type mouse. Secreted mucin fills the crypt lumina.

Full size image

Maintenance of two control (wild-type) siblings on the same liquid diet for 60 d did not produce the same trends (Figs. 6,A-D). Specifically, there was a reduction, rather than an increase, in the mucosal thickness in all sections of the intestine, no accumulation of secreted mucin in the lumen, and no suggestion of a hypersecretory state. Villi in the ileum were also less numerous than in the Cftr-/- mice, and the crypts were shallower (Fig. 6A versus 5A). Although reduced in number, goblet cells of the liquid-fed control mice stained normally (i.e. blue), both on the villi and in the crypts(Fig. 6B). The colonic mucosa also demonstrated slight atrophy (Fig. 6C), but the mucin stained normally(Fig. 6D). Thus, the changes observed in theCftr-/- mice could not be attributed to the liquid diet.

Figure 6
figure 6

Light microscopy of the ileum and colon of Cftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 66-d liquid-fed wild-type male. (A) Ileum. Villi are well shaped, but the crypts are shallower than those of theCftr-/- mouse. (B) Ileum. Crypts and intervillus spaces are clear of secreted mucin. (C) Colon. Goblet cells in the crypts show no signs of cavitation. (D) Colon. Mucin stains purplish-blue in the goblet cells. Secreted mucin is absent in the crypt lumina. Crypts are shallower than those of the Cftr-/- mouse.

Full size image

After 80-90 d, the ileal mucosae of six liquid-fed Cftr-/- mice examined demonstrated even more overt crypt dilation than at 60-70 d(Fig. 7A). A granular eosinophilic material was also present in the crypt lumina; its staining characteristics resembled granules in the Paneth cells at the crypt base. Stored and secreted mucin showed even weaker (paler blue) staining than at earlier ages (Fig. 7B). Colonic crypts were very deep and grossly dilated, but no granular eosinophilic material was present in the lumina (Fig. 7C); this is consistent with the reported lack of Paneth cells in mouse colon(17, 18). Goblet cells and crypt lumina contained abnormal mixtures of of pink and blue mucin granules, and vacuolated cells were absent (Fig. 7D). Liquid-fed wild-type mice, by 80-90 d, demonstrated progressive thinning of the intestinal mucosae and no crypt dilation or mucin accumulation. In the ileum, villi were sparse and crypts were shallow (Fig. 8A). Few goblet cells were present, but their stored mucin stained normally (Fig. 8B). The colonic mucosae were also thin (Fig. 8C) and contained fewer goblet cells and no vacuolated cells (Fig. 8D).

Figure 7
figure 7

Light microscopy of the ileum and colon ofCftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 86-d liquid diet-fed Cftr-/- male.(A) Ileum. The crypts show dilatation, epithelial flattening, and accumulation of granular eosinophilic material in the lumina. (B) Ileum. Few goblet cells with faintly stained mucin are present in the crypts, but pale blue material has accumulated in the distended lumina. (C) Colon. Deep crypts are dilated and contain both robust (i.e. intact) goblet cells and cavitated goblet cells (arrowheads). No eosinophilic granular material is noted. (D) Colon. Crypt goblet cells contain heterogeneously (pink and blue) stained mucin granules and poorly stained mucin is streaming out from the crypt cells into the lumen.

Full size image
Figure 8
figure 8

Light microscopy of the ileum and colon ofCftr-/- and wild-type mice maintained on a liquid diet. For all figures, the bar equals 50 μm; samples A and C are stained with hematoxylin and eosin and samples B and D are stained with Alcian blue/neutral red.

(A-D) A 91-d liquid-fed wild-type male. (A) Ileum. The mucosa is comprised of shallow crypts and few villi. (B) Ileum. Goblet cells containing mucin are still present on the villi and in the crypts. The lumen is clear of any secreted mucin. (C) Colon. The mucosa demonstrates slight atrophy, and fewer goblet cells are present on the epithelial surface or in the upper crypt regions. (D) Colon. Darkly(i.e. normally) staining mucin is stored within the goblet cells and little secreted mucin is observed.

Full size image

Six chow-fed wild-type mice were compared with liquid-fed controls andCftr-/- mice. All the chow-fed mice had more numerous villi in their ileal mucosae. Their crypts (Fig. 9A) were deeper than those of wild-type mice on the liquid diet, but not as deep as those ofCftr-/- mice, and contained no eosinophilic granules. Goblet cell mucin stained blue (i.e. normally), and crypts and intervillous spaces were relatively clear of secreted mucin (not shown). The colonic mucosa was well developed and contained numerous goblet cells and vacuolated cells(Fig. 9B) which were absent in the colons of both the liquid-fed Cftr-/- (Figs. 5C and7C) and wild-type mice (Figs. 6C and8C).

Figure 9
figure 9

(A-B) Light microscopy of the ileum and colon of a 65-d chow-fed wild-type male, stained with hematoxylin and eosin. For both figures, the bar equals 50 μm. (A) Ileum. Numerous villi are present and the crypt lumina show no dilation or accumulation of eosinophilic granules. Paneth cells are present in the crypt bases. (B) Colon. The colonic crypts contain both goblet cells (arrows) and darker staining vacuolated cells (arrowheads). The latter were not seen in colons of any of the liquid-fed animals.

Full size image

About 10% of the Cftr-/- mice in our collection (presently> 150) are severely runted by 21 d, produce poorly formed, sticky stools, and die before 40 d of age from problems associated with malabsorption and obstruction. Animals with formed stools always fare much better, although a few have experienced temporary upper gastrointestinal hemorrhage and/or unexplained anorexia and weight loss around 50-60 d. Examination of three of the latter group revealed marked ileal and colonic epithelial exfoliation, crypt dilation with massive mucin secretion, and, in the ileum, accumulation of eosinophilic material. Numerous mononuclear cells and leukocytes were also present in the lamina propria, suggestive of an inflammatory reaction (not shown).

Pancreas and gall bladder. No abnormal histopathologic lesions were obvious under light microscopic examination in any animals up to 104 d of age.

Liver. Animals at 20 d or in the 60-80-d interval had no apparent lesions. One healthy, liquid-fed control mouse was killed at 91 d and had some liver cell vacuolation. One Cftr-/- female was healthy until 99 d, became ill at 100 d, and was killed and found to have marked hepatocyte vacuolation. One Cftr-/- male killed at 86 d was healthy when killed but had sparse focal granulomas, possibly secondary to the breakdown of the gut wall and portal vein infection, leading to the introduction of microorganisms into the liver. However, most killed animals in the 80-100-d range were healthy and had normal livers.

Nasal mucosa. The nasal mucosa in control animals ranging in age from 20 to 100 d were all histologically normal. Two out of threeCftr-/- mice in the 20-d age group were normal, but the third had nasolacrimal gland distention with a purulent exudate containing neutrophils and cellular debris. One Cftr-/- mouse at 60 d had submucosal glands distended by cell debris, neutrophils, and mucus. TwoCftr-/- animals of 80-100 d of age had pathologic signs, including one with moderate distention of the nasolacrimal ducts with cell debris and the other with nasolacrimal ducts distended with mucus and cell debris as well as some submucosal glands distended with mucus.

Reproductive tract. No visible lesions were seen at any age in the reproductive tract of male and female Cftr-/- mice. There were no signs of mucus accumulation or of obstruction or dilatation of the oviducts.

Intestinal Ion Transport of Liquid-Fed Cftr-/- Mice

Figure 10 summarizes transepithelial electrophysiologic data obtained from segments of intestines isolated from normal and Cftr-/- mice. After mounting in the Ussing chamber, the tissues were short-circuited (voltage clamped at 0 mV), and theIsc was recorded. After mounting of jejunal segments from normal mice, a basal Isc was present which spontaneously declined to lower levels (p < 0.0001; n = 15). After stabilization, the resting Isc was somewhat smaller in the jejunum isolated from animals fed the liquid diet (24.5 ± 2.1μA·cm-2; n = 5) compared with those from animals on the solid diet (40.8 ± 5.4 μA·cm-2; n = 9;p = 0.05). Both groups responded similarly (p = 0.10) when intracellular concentrations of cAMP were elevated. By analysis of pooled data, cAMP increased Isc from 35.0 ± 4.1μA·cm-2 to 57.4 ± 5.6 μA·cm-2(p < 0.0001). To confirm that the stimulated Isc was due to enhancement of Cl- secretion, the tissues were exposed to 100 μM bumetanide on the serosal side to inhibit the activity of the Na/K/2Cl cotransporter. Bumetanide inhibited the Isc by 25.9± 3.3 μA·cm-2 (n = 14), to a level which was not different from the Isc before cAMP stimulation(p = 0.27), indicating that the enhancement of theIsc was due to stimulation of Cl- secretion. InCftr-/- mice, the basal Isc was somewhat smaller in jejunum isolated from animals fed the liquid diet (19.4 ± 7.7 μA·cm-2; n = 7) compared with those from animals fed the solid diet (49.0 ± 9.8 μA·cm-2;n = 7) (p = 0.04), as observed in the normal mice. The basal Isc was not significantly different betweenCftr-/- and normal mice when they were compared according to their diet (p = 0.58 for liquid diet; p = 0.44 for solid diet), nor was it sensitive to bumetanide (not shown). These results suggest that the basal Isc is not mediated by CFTR and likely does not reflect Cl- secretion. Its ionic basis is unknown. Elevation of intracellular cAMP levels was without effect on Isc in jejunum from Cftr-/- mice fed the liquid diet (p = 0.27; n = 8) or from 20-d-old mice at weaning (p = 0.26;n = 6), in contrast to the response of jejunum from normal mice.

Figure 10
figure 10

Ussing chamber Isc measurements obtained from isolated intestinal segments (colon, jejunum) from normalCFTR-expressing and Cftr-/- mice. Results from the jejunum are presented separately for animals maintained on liquid and solid diets (the Cftr-/- mice in the latter group were 20 d old), as well as the pooled data from both dietary regimes. Data for colon were pooled from both dietary regimes. Shown are the stable basalIsc before stimulation, the peak response to elevated cAMP, and the response to subsequent exposure to 100 μM bumetanide. The statistical significance of differences are presented in the text.

Full size image

Electrophysiologic parameters were also measured in colons isolated from normal and Cftr-/- mice. After a decline from initially higher levels (p = 0.001), the resting Isc was similar in colons isolated from animals fed the liquid diet (38.4 ± 6.4μA·cm-2; n = 12) compared with those from animals fed the solid diet (23.1 ± 4.0 μA·cm-2; n = 12; p = 0.06). By analysis of pooled data, cAMP increasedIsc from 30.8 ± 4.1 μA·cm-2 to 63.2± 12.4 μA·cm-2 (p = 0.001). Bumetanide (100μM) added to the serosal side inhibited the Isc by 34.0± 7.3 μA·cm-2 (n = 8), to a level which was slightly smaller than the Isc before cAMP stimulation(p = 0.024). These results indicate that the enhancement of theIsc was due to stimulation of Cl- secretion, and that a small component of the basal Isc was due to spontaneous Cl- secretion. In Cftr-/- mice, the basalIsc was 7.6 ± 1.8 μA·-2 (n= 7), a value significantly smaller than that observed in colons from the normal mice (p = 0.002). This difference suggests that the basalIsc is dependent upon expression of CFTR, but its bumetanide insensitivity suggests that it does not reflect Cl- secretion. Further experiments will be necessary to define its ionic basis and its association with CFTR expression. Elevation of intracellular cAMP levels was without effect on Isc in colons from Cftr-/- mice(p = 0.99; n = 8), in contrast to the response of colons from normal mice.

These data indicate that the cAMP enhancement of Isc in the normal mouse jejunum and colon was associated with the expression of CFTR, and support bumetanide evidence that the stimulated Isc in control mice largely reflected enhanced Cl- secretion.

Pulmonary Function of Liquid-Fed Cftr-/- Mice

No pathologic lesions were evident in the lungs of Cftr-/- mice when examined by light microscopy. Tidal volumes (per kg) and frequency of breathing were not significantly different for the two groups of animals(data not shown). The compliance (mL/cm H2O/kg) was 7.8 ± 3.5(SD) in the control group (n = 4) and 7.3 ± 3.6 in theCftr-/- group (n = 6), and were not significantly different. Similarly, there were no significant differences in the conductance(mL/s/cm H2O/kg) measurement (controls = 84 ± 15;Cftr-/- = 68 ± 19). Thus, there was no major impairment of the mechanical properties of the lung in theCftr-/- mice, other than their smaller size.

Fertility of Cftr-/- Mice on the Liquid Diet

When mated with either wild-type or heterozygous animals with proven fertility, no liquid-fed Cftr-/- males or females were fertile. On the other hand, three of five Cftr-/- males placed on a solid diet were successfully bred to heterozygous females, producing litters of comparable size to those of Cftr+/- males. The two males that failed to breed died within 12 d after the introduction of solid food.

DISCUSSION

We have demonstrated that the survival of Cftr-/- mice can be prolonged significantly by using a liquid diet. This result is consistent with the original observations on these mice that suggested that death after weaning occurred because of intestinal obstruction(10). Thus, reduction of solid components from the diet presumably decreases the opportunities for intestinal blockage and prolongs survival. The animals were still susceptible to obstruction, however, as demonstrated by the mortality that periodically occurred in older, liquid-fed mice or more dramatically when these were fed solid food. It was notable that in this latter case not all mice died as quickly as younger mice fed a solid diet, suggesting that other factors contribute to mortality. Because of the complex mixture of mouse strains in the original Cftr+/- mice(10), determining the role of other genetic loci must await pure bred strains of mice.

Adult Cftr-/- males raised on liquid diet and then selected for good health and large body size can survive for periods up to 6 wk on solid food. When introduced to a sufficient number of fertile females, these animals can produce a reasonable number of litters before succumbing to obstruction. These observations suggest that novel breeding strategies that capitalize on their apparently normal fertility (although shortened lifespan) can produce a larger number of Cftr-/- mice.

No histologic abnormalities, specifically mucin hypersecretion or increased epithelial cell loss, were detected in the intestinal tracts ofCftr-/- mice at the time of weaning (21 d). Differences between Cftr-/- and control mice on the liquid diet developed as the animals aged, however, and were most pronounced in the distal ileum and colon, even without overt obstruction. In Cftr-/- mice, the ileal crypts were longer, suggesting epithelial hyperplasia(19) and expansion of the proliferative zone(20). Ileal crypts contained accumulated pale-staining mucin and eosinophilic material. The latter is similar in staining characteristics to granules found in Paneth cells(21), suggesting this may be their epithelial source. In the colon the crypt luminal accumulation of mucin was even more dramatic than in the ileum, suggesting a marked goblet cell hypersecretion. These changes were progressively greater as the animals aged. They could not be attributed to the liquid diet, because control mice receiving the same diet exhibited signs of atrophy, rather than crypt elongation, and no evidence of histochemical change or abnormal accumulation of mucin or other epithelial secretory product. As a result of the liquid diet both wild type and Cftr-/- groups showed a decrease in colonic vacuolated cells.

It is unlikely that mucin hypersecretion in healthyCftr-/- mice can be attributed to inflammatory mediators, because there was no widespread increase in defense cell populations or signs of inflammation in the overwhelming majority of these mice. A fewCftr-/- mice, although maintained on the liquid diet, still developed a clinical crisis, either obstruction or weight loss, accompanied by epithelial sloughing and accumulation of mononuclear cells and leukocytes in the lamina propria.

The pallor of Alcian blue (pH 2.5) staining of goblet cells inCftr-/- mice was a prominent abnormality and suggests that the mucins expressed by these animals contain an abnormal composition of anionic groups. Alcian blue has been used in the past to detect both sialic acid and sulfate groups in glycoconjugates, the selectivity of staining being a function of the pH of the stain. At or above pH 2.5, sialomucins and weakly acidic sulfomucins stain bright blue; strongly acidic sulfomucins are unstained(22). Biochemical studies have suggested that CF patients have an increased sulfate content in their intestinal(23) and respiratory(24) mucins. Additionally, CF glycoconjugates show either no change(23) or a decrease(25) in sialic acid content. Histochemical examination of respiratory mucin fractions has suggested that the increased sulfate content is associated with the most acidic mucin species(26). It is possible, therefore, that in Cftr-/- mice decreased sialylation and/or increased sulfation is responsible for the decreased affinity of mucin for Alcian blue. Future research, using more precise tools for measurement of mucin sulfation and sialylation, is required to define the differences in anion composition between intestinal mucins of wild type and Cftr-/- mice.

In agreement with previous observations(11–13, 27), elevation of intracellular levels of cAMP stimulated a current in the jejunum and colon of normal mice but not in that of Cftr-/- mice. The ability of theCftr-/- mice to live and grow on the liquid diet is therefore unrelated to expression of an alternative cAMP-activated Cl- secretory capacity in their intestines.

Diffuse expression of Cftr has been detected throughout the respiratory tract of both fetal and adult mice(28). Furthermore, CFTR expression in postnatal human lungs is largely restricted to serous submucosal glands(29, 30), and mice have few submucosal glands. These observations have raised the question as to whether Cftr-/- mice would develop lung disease similar to that of CF patients. No evidence of abnormal lung function or lung pathology was detected in 175-d-old mice, consistent with the above-mentioned observations.

The availability of long lived Cftr-/- mice now allows their use for in depth studies of CF pathology in this animal model. Because lack of CFTR function in the intestine leads to histologic damage, even if death is prevented, these animals can be used to determine the role of CFTR during the critical postnatal period during which much of rodent intestinal maturation takes place. As well, one can examine the relationship between CFTR function and the mucus hyperplasia that is so evident in these mice. Second, we do not observe abnormalities of lung function, consistent with absence of lung disease in these long lived Cftr-/- mice. It is possible, however, to use these mice to test hypotheses that CF lung disease is caused by an abnormal pulmonary response to infection, inflammation, or injury. Thus, the mice can be exposed to a variety of challenges or pathogens and the subsequent response studied. Third, the long-livedCftr-/- animals can be used to test therapies to overcome the gene defect. These could be based on gene delivery or, in the long-term, on pharmacologic approaches.