Enteropancreatic circulation of digestive enzymes.

Intact digestive enzymes can be absorbed by the intestine and resecreted by the pancreas. The pancreas, therefore, appears to be able to recycle proteins much as the liver recycles bile salts, although the magnitude of this process remains uncertain.

Diffusion-like process can account for protein secretion by the pancreas.

When fluid secretion by the pancreas was mechanically blocked, amylase secretion into the duct ceased. When flow was reduced in a graded fashion by the application of a back pressure, amylase output was reduced proportionately and amylase concentration in secretion was maintained constant. Thus, the secretion of digestive enzyme from the cell into the duct appears to be dependent upon the concentration of enzyme in the duct system. This behavior is most simply explained by diffusion-like (concentration dependent, bidirectional) fluxes of digestive enzyme across the plasma membrane. A unidirectional process, such as exocytosis, whose rate should be unaffected by fluid flow, cannot readily explain these results.

Selective pancreatic enzyme secretion due to a new peptide called chymodenin.

Chymodenin, a peptide newly isolated from porcine duodenum, elicited a rapid threefold increase in secretion of chymotrypsinogen while increasing secrelion of total protein by only, about 40 percent; lipase secretion was unaffected. The pancreas appears able, under chymodenin stimulation, to rapidly alter the transport of an individual enzyme as opposed to producing en masse secretion of enzyme.

The endocrine secretion of human insulin and growth hormone by exocrine glands of the gastrointestinal tract.

The exocrine pancreas, liver, and submandibular glands of the rat were used to express and secrete two exogenous, human protein hormones (growth hormone and insulin) into blood at physiological concentrations. Transfection, expression, and secretion were achieved by the in vivo retrograde injection of plasmid DNA into the secretory ducts of these glands. Pancreatic acinar cells secreted physiological concentrations of growth hormone into the circulation, and its secretion was enhanced by cholinergic stimulation. A human insulin gene was engineered to allow normal processing of insulin in non-beta cells. With this gene, the secretion of human insulin by the exocrine pancreas normalized elevated blood glucose levels in diabetic rats. These in vivo observations demonstrate the utility of retrograde ductal administration of naked DNA into exocrine organs as a novel method for the regulated systemic delivery of protein-based pharmaceuticals.

Protein flux across the membrane of single secretion granules.

We have applied, for the first time to our knowledge, X-ray microscopy to measure the mass of protein contained in single sub-cellular membrane-bound structures and to make high resolution, time-resolved observations on them. Using this method we have been able to follow the flux of protein out of secretion (zymogen) granules isolated from the acinar cells of the exocrine pancreas. The results provide direct visual and quantitative confirmation of the hypothesis that the membrane enclosing this object is permeable to its various contained proteins, although the mechanism remains unknown.

A trans-membrane pore can account for protein movement across zymogen granule membranes.

We have reported that the membrane of zymogen granules, secretion vesicles from the exocrine pancreas, is permeable to its contained proteins by measuring both the loss and accumulation of protein in response to mass action forces [1-3] However, the mechanism of transport has remained unknown. Here we consider evidence that this transport occurs through trans-membrane pores. Using freeze-fracture electron microscopic methods, Cabana et al. [4] have reported the presence of a 15 nm intramembrane particle in zymogen granule membrane which contains a 5 nm (+/- 0.1 nm, S.D.) diameter lucent center. In this article, we propose that this structure is a pore through which proteins can be transported, and test this hypothesis by comparing the predicted phenomenological permeability coefficient for transport by diffusion via this structure, to that calculated from protein flux measurements on granules using an X-ray microscope. The predicted and experimental values were essentially identical and hence support the hypothesis that this structure could be a protein transporting channel.

Transpancreatic transport of digestive enzyme.

When porcine alpha-amylase or bovine chymotrypsinogen A was added to the medium bathing the rabbit pancreas in short-term organ culture, the secretion of these enzymes collected via the duct system increased greatly. To determine if it was indeed the amylase added to the bath that was recovered in secretion, endogenous enzyme stores were prelabeled during a 4 h incubation with [3H]-leucine and the specific radioactivity of amylase in secretion followed. The addition of unlabeled exogenous amylase to the bathing medium reduced the specific radioactivity of secreted amylase by about 90% suggesting that the response was due to the transpancreatic transport of the added enzyme. This inhibition was maintained over time, and was a result, not only of the increased secretion of unlabeled enzyme, but also of a 72% steady-state inhibition in the secretion of endogenous (labeled) amylase. This latter observation indicates that the exogenous enzyme crosses the acinar cell and mixes with endogenous cellular stores. A cellular route is also suggested by the observation that the addition of amylase to the bath increased the amylase concentration in ductal fluid relative to that in the bath by about 20 times; it did not reduce it as would be expected if paracellular shunts were involved. In addition, a cellular pathway is suggested by the observation that a 2 h prior incubation in bovine chymotrypsinogen resulted in a greatly augmented chymotrypsinogen response to a maximal cholinergic stimulus. In all, the data support the prediction of the equilibrium theory of digestive enzyme secretion that enzyme secretion should be responsive to mass action, and the prediction of the enteropancreatic circulation hypothesis that a capacity exists for a substantial transpancreatic flux of digestive enzyme.

Equilibration of pancratic digestive enzymes across zymogen granule membranes.

Zymogen granules isolated from the rat pancreas released digestive enzyms in response to lowering the enzyme concentration of the medium in which the granules were suspended, and reabsorbed enzyme when the enzyme concentration of the medium was increased. The evidence suggests that the distribution of digestive enzyme between zymogen granule and medium is the result of an equilibrium process. This equilibrium is apparently expressed across the granule membrane since the disruption of the membrane barrier eliminated the delay in establishing new equilibrium between the bound and free enzyme. The release of enzyme was continuous (equilibration did not occur) when the suspending medium was continuously filtered, thereby maintaining a constant state of disequilibrium between intragranular enzyme and enzyme in the medium. The data suggest the following: (1) that digestive enzyme passes through the intact granule membrane bidirectionally, and (2) that there is a free cytoplasmic pool of digestive enzyme in intact cells which is in equilibrium with the enzyme content of the zymogen granule.

Amylase transport across ileal epithelium in vitro.

Amylase transport was measured across the rabbit ileum in vitro employing a modified Ussing chamber. Amylase was moved preferentially in the mucosal to serosal direction. Its rate of transfer was 2--3 orders of magnitude greater than that for inulin. Mucosal to serosal transport of exogenous amylase was completely inhibited in the absence of oxygen. There was also a constant release of endogenous amylase from intestinal tissue into both mucosal and serosal compartments in the absence of an exogenous source. An estimate of the rate of amylase absorption indicates that it may be of sufficient magnitude to account for the enteropancreatic circulation of amylase secreted by the pancreas during augmented secretion.

Protein concentration in the pancreatic zymogen granule.

We have estimated the concentration of protein in the zymogen granule, a major storage depot for secretory protein in the pancreatic acinar cell, in four different ways. Each of these approaches yielded roughly similar values. The protein concentration in the granule is approx. 135-270 mg protein/ml granule volume in rat and rabbit, as compared to an average value for the protein concentration of tissue of 135 mg/g tissue for rabbits and 183 mg/g tissue for rats. This is equivalent to an average molarity for the contained proteins of between 4-9 mM based on an estimated average molecular weight for the mixture of 30 000-40 000. An upper limit for the concentration of protein in these granules can reasonably be set at about double the overall concentration of protein in the non-granule portion of the cell.

Distribution of three hexose derivatives across the pancreatic epithelium: paracellular shunts or cellular passage?

It has been proposed that the pancreatic epithelium is permeable to three presumably passively distributed non-electrolytes, namely sucrose, inulin and mannitol, via paracellular shunts, and that the increased flux of sucrose and inulin seen during augmented digestive enzyme secretion is due to an increase in the permeability of these shunts. The present study considers this hypothesis by comparing the permeability of the epithelium to three different hexose derivatives, mannitol, inositol and 3-O-methyl-glucose, in both the unstimulated state and after the augmentation of protein secretion with a cholinergic drug. The epithelium was found to be more permeable to mannitol than to either inositol or 3-O-methyl-glucose. In the unstimulated state, the concentration of mannitol in ductal fluid at the steady state was approx. 54% of its concentration in the interstitium, as compared to 12% for inositol and 8% for 3-O-methyl-glucose. Cholinergic stimulation substantially increased the concentration of inositol and 3-O-methyl-glucose in secretion, but did not increase that of mannitol. The increase in the concentration of inositol occurred in the absence of an increase in its rate of transepithelial movement. Taken together, the results suggest that: (1) there is a substantial passage of mannitol through the cells of the epithelial layer, and (2) the increased concentration of inositol and 3-O-methyl-glucose in ductal fluid that occurs with stimulation is due to an increase in their efflux from secretory cells.

Digestive end products release pancreatic enzymes from particulate cellular pools, particularly zymogen granules.

The effects of various amino acids and phosphorylated forms of glucose on the release of digestive enzymes from particulate cellular pools, particularly zymogen granules, were evaluated in rat pancreas. Whole tissue homogenates, as well as zymogen granules isolated either by differential centrifugation in 0.3 M sucrose or by preparation in buffered sucrose and subsequent centrifugation in a Percoll gradient, were studied. The basic amino acids L-arginine and L-lysine, sites of tryptic cleavage, caused the release of trypsinogen, but not chymotrypsinogen, whereas the aromatic amino acids L-phenylalanine and L-tryptophan, sites of chymotryptic cleavage, caused release of both trypsinogen and chymotrypsinogen. Neither led to the release of the starch-splitting enzyme amylase. All effects occurred within the range of normal plasma concentrations for these amino acids in the rat. Two amino acids, L-threonine and hydroxy-L-proline, that are not sites of cleavage by trypsin or chymotrypsin, and a nonmammalian amino acid, aminoadipic acid, did not lead to release of trypsinogen, chymotrypsinogen, or amylase. Two phosphorylated forms of glucose, glucose 1-phosphate and glucose 1,6-diphosphate, caused the release of amylase, but of neither trypsinogen nor chymotrypsinogen. Contrary to previous results, D-glucose was without effect, as was glucose 6-phosphate. We propose that certain digestive end products, by direct action on zymogen granules, cause the selective release of the enzymes involved in their evolution from polymeric substrates during digestion.

Comparison of stored and secreted rat pancreatic digestive enzymes by mass spectrometry: alpha-amylase.

As part of a continuing effort to better understand the mechanisms of protein secretion, we compared the mass of pancreatic digestive enzymes, in resting and stimulated states, both in secretion and in the zymogen granule to determine whether their secretion is accompanied by chemical modification. Mass spectra were obtained applying the electrospray method on samples separated by reverse-phase HPLC. We report here our results for alpha-amylase (1,4-alpha-D-glucan glucanohydrolase EC 3.2.1.1). The data illustrate structural differences between states and compartments for this enzyme. Multiple isozymes were identified from the mass spectra, varying roughly from 52 to 60 kDa. On the basis of mass comparisons, not all of the products seen in the zymogen granule were found in secretion, nor were all secreted isoforms in the granule. Stimulation of protein secretion with a cholinergic agonist, led to time-dependent changes in the number and masses of isoforms in secretion, leaving only one of five resolvable forms in the granule. Only one form, 55.5 kDa, was found in all samples, granule and secretion. In addition to these differences, microheterogeneities of 400 Da or less were observed. The data suggest the differential or non-parallel release of different amylase forms and their chemical modification during the secretion process. As such, release appears to involve a third, intermediate compartment, between zymogen granule to ductal space, such as the cytoplasm, in which chemical modification takes place.

The interior of a whole and unmodified biological object--the zymogen granule--viewed with a high-resolution X-ray microscope.

We report the ability of focused soft X-rays to visualize at spatial resolution well beyond that of the optical microscope (less than 100 nm) the interior of a small, whole biological object without fixation, staining, dehydration or sectioning. Quantitative estimation of its protein content with unique femtogram sensitivity is also reported. The present results represent a significant step towards the goals of natural imaging and chemical mapping of biological structures with soft X-rays.

Zymogen granule size in pancreas of nursing rats.

Dramatic depression in granule volume density and size was measured in acinar cells of postnatal rat pancreas following the initiation of feeding. Volume density decreased about threefold from 45% at birth to 16% 2 days thereafter. Mean granule diameter decreased from 1.50 micron to 0.80 micron, an 85% decrease in corresponding granule volume. At the same time, numerical density approximately doubled. At 2 days after birth, cells with smaller granules had lower volume densities, and differences in mean granule volume between cells accounted for most of the differences in volume density. Although the distribution of granule diameter in newborns was lognormal, the distribution at 2 days was heavily skewed to larger sizes. This was the result of skewed distributions within individual cells and not an artifact of sampling. The results corroborate the central role of granule volume in determining changes in the volume density of zymogen granules in the pancreas and suggest that zymogen granules can act as capacitors that can change size as a function of the enzyme contained within.