Main

CF is a lethal genetic disease affecting the exocrine function of multiple organs lined by epithelial cells(1). The disease is characterized by obstructive lesions within ducts and disturbances of macromolecules and electrolyte secretion(2). In 1989, theCFTR gene responsible for CF was cloned(3–5). The protein product of the CFTR gene is a cAMP-regulated chloride channel situated on the apical membrane of a variety of epithelial cells(6). Mutations in theCFTR gene give rise to defective function or the absence of CFTR in epithelial cells(2) resulting in aberrant chloride transport and a reduction in electrolyte and water secretion at the affected epithelial surfaces.

The gastrointestinal tract of CF patients is affected in utero. Accumulation of thick, viscid intestinal secretions produce intestinal obstruction, or meconium ileus, in approximately 15% of affected neonates(7). The CF exocrine pancreas also develop pathologic changes in utero and in early infancy(8, 9). Approximately 85% of CF patients have clinical evidence of pancreatic insufficiency at diagnosis(10). The remaining patients, termed pancreatic-sufficient, have evidence of exocrine pancreatic dysfunction, but maintain adequate pancreatic reserve for normal digestion(10). The severity of pancreatic disease is determined by the type of CFTR gene mutation(11). Biophysical studies of pancreatic ducts showed that bicarbonate is secreted across the epithelial apical membrane on a chloride-bicarbonate exchanger(12). Hence, a reduction in CFTR chloride channel activity would result in decreased ductal bicarbonate secretion, which is responsible for the manifestations of disease in CF exocrine pancreas. Clinical data confirm that the pathogenesis of pancreatic disease in CF arises from impaired ductal fluid secretion(13, 14). Consequent to reduced ductal fluid, secretory proteins, in the form of digestive enzymes and other proteins of acinar origin, precipitate in the ductal lumen causing ductal obstruction(15, 16). Histopathologically, intralobular ducts are plugged with inspissated secretions which result in ductal dilatation proximally and atrophy of the acini. Damaged tissue then becomes replaced by fibrous tissue and fat(17).

Mouse models for CF have been developed by targeting the CFTR gene locus in embryonic stem cells(18–21). Mouse models in which CFTR function is completely abolished (Cftr-/-) show severe gastrointestinal manifestations remarkably similar to those seen in CF patients before effective treatment(21). Gastrointestinal manifestation in Cftr-/- mice are even more severe than in man; 25% of the Cftr-/- mice die at birth from complications closely resembling meconium ileus. Survivors remain relatively stable during suckling, but a dramatic increase in mortality occurs with introduction of solid food after weaning, and most die within 10 d of weaning from intestinal blockage and/or perforation. In contrast, the exocrine pancreas of Cftr-/- mice(20) showed minor focal histologic abnormalities comprised of dilatation and obstruction of several small pancreatic ducts in half of the 10 Cftr-/- animals examined. Because the majority of Cftr-/- mice die of intestinal obstruction shortly after being weaned to a solid diet(21), it has not been possible to fully evaluate the exocrine pancreas of the developing or adult Cftr-/- mouse.

We have successfully used a liquid diet to feed theCftr-/- mice at weaning which has dramatically improved long-term survival(40). This provided the opportunity to study the impact of CF on the developing rodent exocrine pancreas. We describe the developmental changes in the exocrine pancreas of the long-livedCftr-/- mice compared with littermate controls(Cftr+/+ and Cftr+/-) at specific ages from suckling to adulthood. A control group weaned to the liquid diet was included to assess the influence of diet on somatic growth, exocrine pancreatic development, and alterations in exocrine pancreatic pathology.

METHODS

Cftr-/- mice: breeding and care. A breeding colony of heterozygous mice (Cftr+/-) was established using animals developed by Snouwaert et al.(21). This CF mouse model was produced by targeted disruption of the CFTR locus in embryonic stem cells. Tail-clipped samples of 14-d-old mice were genotyped(21). Mice identified as Cftr+/- were selected and placed in breeding pairs at 8 wk of age. All offspring were tail-clipped and genotyped at 10-16 d of age. Animals that died perinatally were not genotyped. Mice identified as controls (Cftr+/+ orCftr+/-) were weaned around 21 d to solid chow (Agway Rodent Chow, RMH #1000, Prolab, St. Marys, Ohio), or to a liquid diet(Liquidet, Bio-Serv, Frenchtown, NJ). Cftr-/- mice were weaned to the liquid diet exclusively. The dietary composition of the solid and liquid diets, respectively, was: protein 14 and 14%, fat 6 and 5%, fiber 4.5 and 1.5%, ash 8 and 5%, carbohydrate 63 and 69%. The protein in the liquid diet was hydrolyzed casein. The liquid diet was prepared in sterile water according to the instructions of the manufacturer and stored at 4°C. The diet was fed in glass liquid mouse feeders, suspended in the cages. The feeders were sterilized by autoclaving and replaced daily. A fresh diet was provided daily.

Mice were housed in sterile microisolators with corncob bedding, which was changed daily. In addition to the liquid diet, mice were provided with sterile water. Their room had a 12-h light-dark cycle and was ventilated at 20 air changes per h with high efficiency particulate filtered air.

Laboratory studies. Animals were fasted on the day of sacrifice, weighed, and killed by cervical dislocation. Pancreata were immediately isolated, trimmed of fatty tissues, weighed, and placed in 0.9% NaCl at 4°C. Each whole pancreas was homogenized in fresh 0.9% NaCl (1:50, wt/vol), using a Polytron tissue grinder (Brinkman Instruments, Toronto, Ontario, Canada). The final homogenate was aliquoted and kept at -70°C until assayed. A minimum of seven animals was assessed in each group at all ages. Pancreatic enzyme activities, DNA, and protein content were determined at suckling (10 and 18 d), after weaning (26 d), and at adulthood (46, 60, and 100 d). Pancreatic growth was assessed by gland weight, cell number (DNA content), cell size (protein content), and by comparison to somatic growth.

Protein content of the pancreas was measured by the method of Lowryet al.(22), using BSA as a standard. DNA was measured by fluorometric quantification as described by Downs and Wilfinger(23), using calf thymus DNA as the standard. Pancreatic enzyme assays were performed by a micromethod technique described by Laineet al.(24), using 96-well microtiter plates(Molecular Devices Corp., Menlo Park, CA). Amylase activity was determined using the Diagnostics Amylase reagent kit. The assay is based on the hydrolytic activity of amylase on p-nitrophenylglycosides, yieldingp-nitrophenol. Lipase activity was determined using the Lipase-SR reagent kit (Fisher Scientific, Toronto, Ontario, Canada), which is a turbidimetric method, employing triolein emulsion, in CaCl2, bile salts, and colipase. For both amylase and lipase measurements, pancreatic homogenates were diluted in PBS, 20 mM Na2HPO4, 140 mM NaCl, pH 7.4, containing 0.1% BSA and 0.5 mM phenylmethylsulfonyl fluoride. Ten microliters of the diluted samples were pipetted into wells, followed by 200μL of the appropriate substrate. Absorbance at 405 nm was measured by kinetic analysis, at intervals of 1 min for a period of 10 min, at room temperature, using a Thermomax microplate reader (Molecular Devices Corp.). Trypsin and chymotrypsin activities were determined by hydrolysis of the substrates Nα-benzoyl-L-arginine-p-nitroanilide and succinyl-Ala-Ala-Pro-Phe-p-nitroanilide, respectively, yieldingp-nitroanilide as the end product.Nα-Benzoyl-L-arginine-p-nitroanilide and succinyl-Ala-Ala-Pro-Phe-p-nitroanilide solutions (1 mM) were individually prepared by dilution in Tris buffer (100 mM Tris-Cl, 20 mM CaCl2, pH 7.9). Pancreatic homogenates were diluted in Tris buffer with the addition of 0.1% BSA (without phenylmethylsulfonyl fluoride) and incubated at 37°C with 5.5 U/mL porcine enterokinase (4.7 U/mg) for 1 h. Ten microliters of the mixture were added to wells, followed by 200 μL of the appropriate substrate. Absorbance was measured at 405 nm at intervals of 1 min for a period of 10 min at room temperature. Chemicals and reagents were purchased from Sigma Chemical Co., St. Louis, MO, except as noted.

Data analysis. Contents of pancreatic protein, DNA, amylase, lipase, trypsin, and chymotrypsin were calculated and expressed per total pancreas. Enzymatic activities were also expressed per mg of pancreatic protein. Measurements of growth and enzyme activity were presented as the mean± SD for each experimental group at six ages, two before weaning and four postweaning. Results were presented graphically. A t test was used for pairwise comparisons of the experimental groups at each age(solid-fed and liquid-fed controls, Cftr-/- animals and liquid-fed controls, and Cftr-/- and solid-fed controls). Before weaning (10 and 18 d) comparisons were made betweenCftr-/- animals and their suckling control littermates. Because multiple comparisons were made, a significance level of p = 0.005 was used to indicate differences between groups not likely due to chance. To assess the changes between age groups, each variable was modeled in a piecewise linear regression model with nodes at age 20 d (at weaning), and at age 60 d (adult). This regression model allowed for estimates of slope in the three age periods defined by the nodes and comparison of these slopes among the three experimental groups(25).

RESULTS

Somatic and Pancreatic Growth

Postnatal somatic growth is shown in Figure 1A. The solid-fed controls underwent rapid growth between 18 and 60 d, after which growth appeared to plateau. Somatic growth of the liquid-fed controls was similar to the solid-fed animals. Growth in the Cftr-/- animals was parallel to the two control groups, but body weight was reduced at all ages (p < 0.005).

Figure 1
figure 1

Body weight (A), pancreatic weight(B), and pancreatic to body weight ratio (C) of mice at various postnatal ages. Pancreatic growth in the controls andCftr-/- mice closely paralleled somatic growth. Liquid-fed(○-○) and solid-fed control mice (•-•) had similar pancreatic weights at all ages (except at 26 d). Body weights ofCftr-/- mice (□-□) were lower than both control groups before and after weaning. Pancreatic weight was significantly lower than the control littermates at most ages. The pancreatic to body weight ratio increased between 10 and 46 d in both control groups, then plateaued. The ratio for Cftr-/- mice showed a similar increase up to 26 d; deviation from controls was observed at 46 d. *p < 0.005, difference between Cftr-/- mice and liquid-fed (and suckling) controls. Δp < 0.005, difference betweenCftr-/- mice and solid-fed controls. Ωp < 0.005, difference between liquid-fed and solid-fed controls.

Full size image

Pancreatic growth of the solid-fed and liquid-fed controls paralleled somatic growth, increasing rapidly from 10 to 46 d, followed by a period of slower growth (Fig. 1B). Pancreatic weight did not differ in the control animals fed a liquid or a solid diet, except immediately after weaning (26 d) when pancreatic weight of the liquid-fed controls was lower than the solid-fed controls. At 10 d of age, pancreatic weight ofCftr-/- mice was not different from controls, but there was a slight deficit at 18 d. Cftr-/- mice had a slower rate of pancreatic growth up to 60 d and no further change after 60 d. In comparison with the control mice fed the identical diet, pancreatic weight of theCftr-/- mice was significantly reduced at both ages 46 and 60 d (p < 0.005).

Figure 1C shows the relationship between pancreatic and body weight (pancreatic/somatic weight ratio) in the three groups of animals. In both control groups there was a rapid increase in the ratio of pancreatic to body weight from 10 to 46 d; between 46 and 60 d the ratio dropped and then remained stable. Liquid-fed controls had a smaller pancreatic to body weight ratio at 26 d than did the solid-fed controls. In the Cftr-/- mice, pancreatic growth relative to body weight showed a similar increase up to 26 d, after which there was no increase. At 46 d, the pancreatic to somatic weight ratio of the Cftr-/- mice was lower than the solid-fed littermates.

Exocrine Pancreatic Development

Protein and DNA content. Total pancreatic protein content increased in parallel with pancreatic growth (Fig. 2A). There was protein accretion during suckling, which was followed by an increase in the rate of accretion up to 46 d in the liquid-fed and solid-fed controls. After 46 d, protein content increased slightly up to 100 d. Total pancreatic protein content was consistently lower in the liquid-fed control group, but the difference was significant only at 26 d (p < 0.005). Pancreatic protein content of the Cftr-/- mice did not differ from the control group at 10 d, but was reduced at 18 d. After weaning, protein content of the Cftr-/- mice was significantly lower(p < 0.005) than that of the solid-fed controls at all ages, and was less than that of the diet-matched controls, becoming significant at 100 d(p < 0.005).

Figure 2
figure 2

Postnatal changes in pancreatic protein (A) and DNA (B) content. Protein and DNA content of the liquid-fed and solid-fed control mice were similar at all ages, except at 26 d when protein content of the liquid-fed animals was lower. Pancreatic protein content of theCftr-/- mice was significantly reduced at suckling (18 d), and was lower in comparison with solid-fed control group at all postweaning ages. Pancreatic DNA content in Cftr-/- mice was less than solid-fed controls at 46 d. There were no significant differences at all other ages between the Cftr-/- mice and their control littermates. Symbols are the same as in Figure 1.

Full size image

Pancreatic DNA content of the control animals increased rapidly from 10 to 46 d; thereafter, DNA content was unchanged (Fig. 2B). There were no differences in pancreatic DNA content between the liquid-fed and solid-fed controls. Pancreatic DNA content of the 10-d-oldCftr-/- animals was the same as the control group; thereafter the rate of increase was significantly less than the control groups, but continued up to 60 d. Thus, there was no significant difference in DNA content in the three groups after 60 d.

Pancreatic enzyme activity. Total activities of individual enzymes are shown in Figure 3. During suckling, control mice showed little or no increase in pancreatic enzyme content. However, between weaning and 46 d, the solid-fed control mice showed a dramatic increase in the activity of individual enzymes. After 46 d, there was a lesser increase in enzyme activity, coinciding with reduced growth. The liquid-fed controls showed lesser increase in the activities of individual enzymes at all ages after weaning; the activities of individual enzymes were significantly reduced (p < 0.005) in comparison with age-matched solid-fed controls.

Figure 3
figure 3

Postnatal changes in total activities of individual enzymes (U/whole pancreas). In the solid-fed controls, little or no increase in enzyme activities occurred during suckling, but after weaning there was rapid increase up to 46 d. After weaning, liquid-fed controls showed reduced enzyme activities compared with the solid-fed controls. However,Cftr-/- mice exhibited reduced total lipase activities before weaning (18 d) and reduced amylase and lipase after weaning in comparison with both control groups. Trypsin and chymotrypsin in Cftr-/- mice were also reduced when compared with solid-fed controls. Symbols are the same as in Figure 1.

Full size image

The Cftr-/- mice showed lower total enzyme activities in comparison with both control groups. As in the controls, there was little or no increase in the activity of individual enzymes from 10 to 18 d, but lipase activities were significantly lower than those of the 18-d controls. Total activities of amylase and lipase were lower than those of both control groups at all ages after weaning (p < 0.005). Total activities of trypsin and chymotrypsin were also reduced in comparison with the solid-fed controls, but to a lesser extent when compared with the diet-matched controls.

Activities of individual enzymes, expressed per mg of protein, are shown inFigure 4. In the solid-fed control animals there was considerable variability in the postnatal changes of individual enzymes. Amylase activity showed no increase before weaning, but immediately after weaning, enzyme activity increased dramatically at 26 d, then leveled off. Lipase activity increased dramatically before weaning (10-18 d) and showed continuing maturation up to 46 d. The proteases, trypsin and chymotrypsin, showed a reduction in activity during suckling which was followed by a slight increase in activity in the 1st wk after weaning. Control animals weaned to the liquid diet showed little or no increase in amylase after weaning; amylase activities were significantly lower than those of the solid-fed controls at ages 26, 46, and 100 d (p < 0.005). Similarly, lipase activities were significantly lower than those of the solid-fed controls at 46 and 100 d(p < 0.005). Postweaning, the liquid-fed controls had consistently lower trypsin activity, but this difference was significant only at 26 d. Postnatal changes of chymotrypsin activity did not differ in the two control groups.

Figure 4
figure 4

Postnatal changes in activities of individual enzymes(U/mg pancreatic protein). In solid-fed control mice, amylase and lipase activities increased rapidly between 18 and 26 d, and from 10 to 46 d, respectively. Liquid-fed controls showed reduced amylase and lipase activities after weaning, in comparison with the solid-fed controls. In both control groups, the proteases, trypsin and chymotrypsin showed a decline during suckling (10-18 d); little change occurred after weaning. Compared with diet-matched controls, there was a significant decrease in amylase activity in the Cftr-/- mice immediately after weaning; lipase activity was markedly reduced during suckling, and both enzymes remained low during adulthood. There were no significant differences in trypsin and chymotrypsin activities between the Cftr-/- mice and their age- and diet-matched controls. Symbols are the same as in Figure 1.

Full size image

In the Cftr-/- mice, there were marked differences in the postnatal response of individual enzymes. At 10 d of age, the enzyme content of amylase, lipase, trypsin, and chymotrypsin did not differ from the control group. Amylase activity in the Cftr-/- animals failed to show a rise in activity after weaning; in fact, amylase activity actually dropped between 18 and 26 d, which coincided with the most rapid increase in activity in the solid-fed control animals. Lipase activity in theCftr-/- animals showed markedly reduced activity between 18 and 46 d when compared with both control groups (p < 0.005). No increase in lipase activity occurred after 46 d. Amylase and lipase activity was significantly lower than in both control groups at all ages after weaning. Alteration of trypsin and chymotrypsin activities (per mg of protein) was less evident in the Cftr-/- mice. There were no differences in protease activities at any postnatal age between the Cftr-/- mice and the controls fed the same diet.

DISCUSSION

The objectives of our study were 2-fold. To assess the postnatal development of the exocrine pancreas of the “knockout” CF mice, and to evaluate for evidence of confounding pancreatic effects from the liquid diet. Postnatal pancreatic growth in the solid-fed control mice, which showed a rapid increase in pancreatic protein and DNA content, was in agreement with previous data in rats. Up to 6 wk of age, pancreatic weight increased more rapidly than body weight. Cell hyperplasia was most evident during suckling, and cell hypertrophy predominated immediately after weaning. Pancreatic growth was also measured by monitoring enzyme activities in the whole pancreas. Both amylase and lipase showed a slight increase of activities during the suckling period, becoming a steady increase with age, which correlated with proportionate changes in pancreatic weight gain. Pancreatic content of the proteases was not changed until after 18 d of age, then began to increase. Lipase activity, when expressed as units activity/mg of pancreatic protein, was low at 10 d of age, but increased progressively with advancing postnatal age (up to 46 d); whereas amylase activity was not modified until after 18 d postnatally, then was increased after weaning, and remained constant. Trypsin and chymotrypsin activities were initially high, dropped during suckling, and remained relatively unchanged thereafter. Individual enzymes exhibited variable maturity at birth, followed by nonparallel maturation during postnatal development was previously demonstrated by Marchaim and Kulka(26). These investigators suggested that the mechanisms for controlling biosynthetic up-regulation of individual enzymes were not necessarily the same.

Although our efforts to prolong survival of Cftr-/- mice with the liquid diet was successful, we observed significant dietary effects on pancreatic development in the control animals. Somatic and pancreatic growth of chow-fed and liquid-fed controls did not differ, but total pancreatic protein content and activities of individual enzyme were reduced after introduction of the liquid diet at weaning. There are several plausible reasons for these observations. The intestinal phase of the pancreatic secretory response is mediated by neural and hormonal pathways which are activated by the presence of dietary products within the gastrointestinal tract. The dietary components of the liquid and solid diets were similar with respect to the proportion of protein, carbohydrate, and fat, but differed in the protein constituents, because liquid diet was in the form of hydrolyzed casein. In the liquid-fed control mice, the presence of predigested protein in the duodenum may have ineffectively induced release of endogenous secretagogues such as CCK and secretin, which in turn may attenuate or delay pancreatic development. In previous studies, intact protein fed to rats has been shown to stimulate CCK release, whereas hydrolyzed proteins and amino acids fail to induce CCK release(27). In addition to its role in stimulating pancreatic secretion and inducing hypertrophic growth of the pancreas, CCK plus other intestinal hormones are involved in regulating intestinal growth and maintaining mucosal integrity(28). Evers et al.(29) has reported that rats fed a liquid elemental diet developed atrophy of the intestinal mucosa, which was prevented by exogenous administration of bombesin or neurotensin. We observed, in the liquid-fed control mice, a progressive thinning of the intestinal epithelium(40). This suggests that the intestinal effects of the liquid diet may have affected release of CCK, which in turn induced secondary effects on the exocrine pancreas of both control andCftr-/- mice.

In the present study, the Cftr-/- mice showed significant reductions in body weight and pancreatic weight at various postnatal ages in comparison with both liquid-fed and solid-fed littermates. Because the reduction in growth of Cftr-/- mice became evident before weaning, use of a liquid diet cannot be the sole explanation for the observed differences. After weaning, the Cftr-/- mice exhibited somatic growth velocity that paralleled their liquid-fed control littermates, but actual body weight was reduced. Pancreatic growth and enzyme activities, notably of amylase and lipase, were significantly reduced throughout postnatal life. The reduced pancreatic weight was accompanied by reduced protein content, whereas DNA content was marginally affected. Although it is possible that the loss of CFTR function is responsible for the major alteration in pancreatic development, similar postnatal pancreatic effects to those described in this study have been observed in rats subjected to malnutrition(30–32). During suckling, total pancreatic lipase content in Cftr-/- mice was significantly reduced, as was lipase activity/mg of protein. After weaning, amylase activity/whole pancreas and mg of protein, was severely decreased, but there were lesser deficits in activities of trypsin and chymotrypsin(Figs. 3 and4). Both absolute and relative differences in pancreatic amylase and lipase activities were evident in theseCftr-/- animals, irrespective of being related to pancreatic protein content or whole pancreas. This finding is in agreement with previous malnutrition data in the rodents, where the various enzymes are affected to a varying extent, with lipase and amylase being most sensitive, and the proteases least affected. Together, these observations suggest that malnutrition may have caused developmental alterations that occurred within the acinar cells of Cftr-/- mice, where synthesis and/or storage of amylase and lipase were significantly reduced. It is important to stress that the Cftr-/- mice under investigation were not deliberately subjected to malnourishment, but despite provision of adequate nutrients in the liquid diet, reduced nutrient intake may have occurred as a result of persistent gastrointestinal disease and/or malabsorption.

In humans with CF, the characteristic pathologic lesions in various organs involve an obstructive element in which a duct or air passage is blocked by intraluminal macromolecules such as mucus and/or protein. Although recent studies have provided considerable information concerning the molecular defect in CF and yielded knowledge concerning the electrophysiologic function of CFTR, the precise mechanisms of disease pathogenesis remain unclear. Organ damage may result from combined effects of various factors, including the functional effect of the CFTR mutation, the organ dependence upon CFTR-mediated chloride secretion, the structural and physiologic characteristics of the organ, and the macromolecular content of the secretions.

Unlike the lung, which is normal at birth, the pancreas appears to be severely affected in utero, presumably because it is susceptible to flow-related damage within tortuous and elongated ducts, which contain a high concentration of proteinaceous secretions of acinar origin. The normal human pancreas when stained with an antibody to CFTR reveals the presence of CFTR only in the apical membrane of centroacinar and intralobular pancreatic ducts(33). Although there remains a possibility of an intracellular role, pancreatic acinar cells do not show the presence of CFTR within the apical membrane. In contrast to the observations in man, previous reports of the exocrine pancreas of the knockout CF mice show relatively few pathologic changes(21). It has been suggested that CFTR may play a lesser role in maintaining normal ductal function in the rodent. Gray et al.(34) demonstrated that pancreatic ductal cells of mice contain an alternate secretory pathway that is mediated by Ca2+-activated chloride channels. Ca2+-activated Cl- channels may be up-regulated in the Cftr-/- pancreas to compensate for the loss of CFTR function. In the rodent pancreas, Ca2+-activated fluid secretion might maintain sufficient duct lumen fluidity, which protects the Cftr-/- mice from developing gross pancreatic abnormalities(35).

Earlier studies of pancreatic development in rodents provide additional reasons for the relative lack of pancreatic pathology in the CF mouse. The rodent pancreas is immature at birth and undergoes significant maturational development postnatally. Although the acinar cells of the newborn rat are structurally developed and packaged with secretory proteins within zymogen granules, they are functionally immature. In the immediate postnatal period, isolated acinar cells are unresponsive to cholinergic and peptidergic (CCK) secretagogues(36). Thus, unlike the human newborn infant, there appears to be little protein secretion in the rodent, and its ductal system may be less important for maintaining fluid and electrolyte secretion in utero and in early postnatal life. As the rodent pancreas matures postnatally, there is a progressive increase in the intracellular concentration of soluble proteins, primarily in the form of zymogen granules, which occurs after proliferation of intracellular organelles involved in protein synthesis and transport. The immature rodent pancreas may be protected from the lack of ductal CFTR-mediated chloride secretion, but may become more susceptible to pathologic damage at a later age, coincident with maturation of the acini and their ability to secrete protein. In this respect, histologic data in longer lived Cftr-/- mice developed by Ratcliff's group(20) showed focal areas of blocked pancreatic ducts, suggesting the possibility of a CF-related pathologic defect due to aberrant ductal secretion of anions. A more recent morphologic examination by DeLisle(37), of the pancreas ofCftr-/- mice fed a liquid elemental diet (Peptamen), revealed acinar lumina that were dilated and filled with aggregated protein. Additionally, gp300, which in normal tissue is mainly localized to the zymogen granule membrane, was found to line the distended luminal membranes in CF tissue. Immunolocalization experiments showed that the aggregated material in the acinar lumina of the Cftr-/- pancreas contained amylase; in contrast, amylase in the control acini was largely localized to zymogen granules. DeLisle proposed that the requirement for coordinated acinar and duct cell secretion in regulating storage and solubilization of secretory protein may be related to gp300, which may play a role in the development of CF in the mouse pancreas. We have not yet systematically investigated the morphologic features of the developing pancreas in Cftr-/- mice.

The structural anatomy of the rodent hepatobiliary-pancreatic system also differs from that of man. In the rodent, hepatobiliary and pancreatic drainage are in sequence. The pancreas is broken into small lobules, each possessing a drainage system, with many paired ducts, which either enters the duodenum directly or passes into the bile duct, forming a common biliary-pancreatic duct. In contrast, in man, the two drainage systems are distinct, uniting distally at a common ampulla at the level of the duodenum. Further, the spontaneous (basal) rate of fluid secretion in the rodent is considerably higher than in humans, and their pattern of electrolyte secretion also differs(38). In the human, secretin normally stimulates a HCO3--rich fluid from the pancreatic ductal epithelium, whereas CCK evokes the secretion of enzymes from the acini. However, data obtained from studies of various rodents showed that there is marked electrolyte secretion in response to CCK which is not mediated via CFTR(39). It seems possible that high constitutive levels of anion secretion in rodent species, derived from both pancreatic acini and ductal cells, render the exocrine pancreas less susceptible to proteinaceous obstruction.

On the basis of our observations, pancreatic disease in theCftr-/- mice does not appear to be as severe as in humans with CF, probably as a result of species differences with regard to developmental physiology of the pancreas and its dependence on CFTR-mediated epithelial Cl- secretion. The results of this study cannot exclude the possibility that some or all of the pancreatic changes inCftr-/- mice were due to secondary effects of diet and malnutrition. However, postnatal somatic and pancreatic development of the CF animals was markedly different from that of liquid-fed control littermates. Furthermore, some of these changes were evident before the introduction of liquid diet at weaning. Future studies will more clearly distinguish the primary pathology of the exocrine pancreas attributable to CF from secondary factors.