Physiol Rev
82: 1–18, 2002; 10.1152/physrev.00022.2001.
Conservation of Digestive Enzymes
STEPHEN ROTHMAN, CHARLES LIEBOW, AND LOIS ISENMAN
University of California-San Francisco, San Francisco, California; State University of New York, Buffalo,
New York; and Brandeis University, Waltham, Massachusetts
1
2
3
3
4
4
4
5
7
8
8
9
11
12
12
13
13
14
14
14
15
15
16
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
I. Introduction
II. Foundation of the Traditional Viewpoint: the Single Pass
III. Membrane Permeability
A. Intestinal epithelium
B. Transpancreatic transport
IV. Circulation Measured In Situ
A. Circulation of intact digestive enzymes by tracer analysis
B. Magnitude of circulation
C. Effect on the plasma pool
V. Attempts at Corroboration
A. Pancreatic arteriovenous concentration difference
B. Intravenous injection of radiolabeled digestive enzyme
C. Fate of digestive enzymes in the intestines
VI. Rate of Synthesis of Digestive Enzymes Compared With Their Rate of Secretion
A. Basal rate of protein secretion
B. Enhancing synthesis
C. Ribosomes
D. Energy requirement
E. Summary
VII. Membrane Protein Transport
VIII. Forms of Circulating Proteins
IX. Conservation of Other Digestive Enzymes
X. Conclusions
Rothman, Stephen, Charles Liebow, and Lois Isenman. Conservation of Digestive Enzymes. Physiol Rev 82:
1–18, 2002; 10.1152/physrev.00022.2001.—The traditional understanding is that an entirely new complement of
digestive enzymes is secreted by the pancreas into the small intestines with each meal. This is thought to be
necessary because, like food itself, these enzymes are degraded during digestion. In this review we discuss
experiments that bring this point of view into question. They suggest that digestive enzymes can be absorbed into
blood, reaccumulated by the pancreas, and reutilized, instead of being reduced to their constituent amino acids in
the intestines. This is called an enteropancreatic circulation of digestive enzymes.
I. INTRODUCTION
In this review we summarize experiments whose implications were of great interest when they were first
reported. They provided unexpected evidence that the
conventional belief that every meal is digested by an
entirely new complement of digestive enzymes (Fig. 1) is
incorrect. The data suggested that instead of being completely degraded in the small bowel with the food they
digest, a large fraction of the digestive enzymes secreted
by the pancreas are absorbed and recycled in an enteropancreatic circulation (Fig. 1).
www.prv.org
Not long after the original findings were reported (20,
48), the idea and the evidence were variously affirmed and
questioned by additional experimentation (26, 37, 44, 74,
76 –78). Disagreement ensued, and the subject became
controversial (6, 42, 75, 79, 87–90, 94). This was not just
because of the surprising nature of the observations, but
also because they seemed to require that the proteins
cross several cellular membranes in transit, a possibility
viewed as unlikely if not impossible at the time. In addition, the exact chemical forms of the circulating proteins
were not known. Were they modified, and if they so, in
what way? And, finally, it turned out that the negative
0031-9333/02 $15.00 Copyright © 2002 the American Physiological Society
1
2
ROTHMAN, LIEBOW, AND ISENMAN
studies were not failed attempts to reproduce the original
observations, but different experiments altogether that
neither proved the circulation hypothesis false nor explained the affirmative evidence in other ways.
Given these uncertainties, research on the topic soon
ended, leaving the existence of digestive enzyme conservation an unsettled matter, seemingly neither adequately
confirmed nor adequately refuted. Indeed, the evidence
has never been the subject of retrospective review and
analysis. This is our intention here some 25 years after the
initial discoveries. What did the evidence show, what
didn’t it show, and what remains to be learned?
II. FOUNDATION OF THE TRADITIONAL
VIEWPOINT: THE SINGLE PASS
It may seem surprising, but the traditional understanding that the pancreas produces an entirely new complement of enzymes with each meal is not based on direct
experimental evidence that this is the case, or even on
evidence that it is capable of such replacement. Nor is it
based on proof that the digestive enzymes are completely
degraded in the small intestines subsequent to their secretion, thereby requiring their replacement. What is
known is that 95–99% or almost all digestive enzyme
Physiol Rev • VOL
activity disappears as the chyme passes down the small
bowel into the colon (8, 40, 51, 68). This has been assumed to be the consequence of enzyme degradation in
transit and it alone.
This view is grounded in two presumptions. First,
because mixtures of activated digestive enzymes, for example, from pancreatic juice or tissue homogenates, left
standing at room temperature lose their activity over
time, presumably due to auto- and heterocatalytic proteolytic degradation, the same events would occur at least as
actively in the intestinal lumen. It has been believed,
though not proven, that the pro- or inactive forms of
proteolytic enzymes such as trypsinogen and proelastase
are completely and essentially instantaneously activated
when they enter the intestines. This is thought to be due
to the presence of enterokinase, an enzyme secreted by
the intestinal epithelium that splits a particular lysine
containing peptide bond near the COOH terminus of
trypsinogen. This then produces the active form of this
enzyme, trypsin, which in turn activates the other proteolytic enzymes in what is referred to as the activation
cascade. Subsequently, the active proteolytic enzymes
digest each other as well as nonproteolytic digestive enzymes such as amylase and lipase, as the ingested food is
being degraded.
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
FIG. 1. A: traditional view of the life history of the digestive enzymes secreted by the exocrine pancreas. From this
perspective, these products (open circles) (Ep) are secreted into the duct system of the gland, and through it enter the
small intestines (Ei), where they digest food and are subsequently degraded into their substituent amino acids (AAi) by
auto- and heterocatalytic catalysis (e). The amino acid products are then absorbed across the intestinal epithelium into
blood (AAb). An equivalent amount of amino acid is accumulated by pancreatic tissue (AAp), and the cycle is completed
when these amino acids are incorporated into new proteins (Ep) that are secreted by the pancreas in the ensuing period.
According to this scheme, each meal is digested by a completely new complement of digestive enzymes synthesized
during the interdigestive period. B: an alternative model that is supported by results discussed in this review. From this
perspective, only a small fraction of the digestive enzymes secreted by the pancreas (Ep) into the intestines (Ei) is
degraded during a single cycle of digestion (the degradative cycle is shown by narrow arrows on the right). Most is
absorbed into blood (Eb), accumulated by the pancreas (Ep), and secreted during the next digestive cycle (shown by the
broad arrows on the left). In this scheme, the pancreas does not produce a new complement of digestive enzymes for
each meal but conserves and reutilizes these protein products.
CONSERVATION OF DIGESTIVE ENZYMES
Physiol Rev • VOL
digestive enzymes. As a consequence, if digestive enzymes could be absorbed from the intestines into the
bloodstream, they could be taken up by the cells of the
pancreas and be secreted into the intestines where they
would be reutilized.
If such a process existed for substantial quantities of
digestive enzymes, they might be recirculated in a manner
similar to bile salts. Bile salts are released into the gut
from the liver and gallbladder where they emulsify fat.
After fat absorption takes place, the unneeded bile salts
are absorbed across the ileum, reaccumulated by the
liver, conserved, and secreted again with the next cycle of
digestion. This is called the enterohepatic circulation. The
liver only manufactures a small fraction of the bile salt
pool de novo each day, and the possibility that digestive
enzymes, which are far more energetically costly to manufacture, indeed are among the most costly biomolecules
to synthesize, might be involved in a similar conservation
process was intriguing.
The capacity to conserve digestive enzymes was first
tested in the mid 1970s. Two types of experiment were
carried out. In the first, the permeability of the relevant
membranes (pancreatic and intestinal) to particular digestive enzymes was assessed in vitro. Did processes
exist that would allow for a circulation? Second, given
their existence, could circulation be demonstrated in intact animals?
III. MEMBRANE PERMEABILITY
A. Intestinal Epithelium
The ileum was found to be permeable to the two
digestive enzymes that were tested, chymotrypsinogen
and amylase, using everted sacs of intestine (48) as well
as pieces of intestine as membranes separating two halves
of an incubation (Ussing) chamber (21) to measure transport. After labeled chymotrypsinogen was added to the
medium bathing everted sacs of rabbit ileum (the mucosal
compartment), it appeared in the internal medium of the
sac (the serosal compartment) at a rate some 10 times
that seen for a control protein, albumin (48). Similarly,
amylase (55 kDa) appeared in the serosal medium of an
Ussing chamber soon after being added to its mucosal
compartment (21). In this case, the enzyme was transported at a rate some two to three orders of magnitude
greater than seen for the control molecule, the much
smaller polysaccharide inulin (⬃9 kDa). In this study,
enzyme activity, rather than radioactivity, was followed to
ensure that the molecule had not been degraded en route.
Amylase was transported intact across the epithelium in
both directions, but was preferentially absorbed by an
active transport process (Fig. 2) (21).
It was possible to estimate the absorptive capacity of
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
The second presumption is that the epithelium of the
small bowel is completely impervious to proteins. If this
were true, then the only way that the digestive enzymes
could escape the intestinal enclosure would be by their
movement aborally into the large bowel. Because almost
all activity is lost during this passage, some sort of inactivation, presumably due to proteolytic degradation,
would have to occur. Throughout the 20th century there
was confidence, bordering on certainty, that biomembranes were impermeable to protein molecules. Not only
was there no direct experimental evidence for this conclusion, but there was considerable evidence to the contrary for the epithelium of the small intestines. It appeared to be permeable to a variety of protein molecules,
including various digestive enzymes (1– 4, 9, 13, 22, 32–34,
41, 52, 53, 66, 99 –101). These results, however, were
judged to be artifactual, pathological, quantitatively trivial, or exceptions to a general rule of impermeability.
Whatever the evidence, the intestines of adult animals
were widely believed to be impermeable to proteins, including digestive enzymes.
This conclusion was reinforced by a third presumption. It was believed that if digestive enzymes, particularly
active proteolytic enzymes, entered the bloodstream,
massive systemic pathology would result, including the
uncontrolled “digestion” of the body’s organs. This would
of course be life threatening, and because digestion normally occurs in the absence of such a cataclysm, it was
safe to assume that digestive enzymes were not absorbed
into blood.
However, two unavoidable facts argued against this
conclusion. First, digestive enzymes, including the various active proteolytic forms, were normal, even appreciable, constituents of blood (5, 7, 10, 11, 15, 17, 18, 36, 38, 39,
43, 73, 93, 96, 97, 98, 104). Although their presence might
primarily reflect release directly from the pancreas into
the bloodstream, their intestinal absorption could not be
excluded. Indeed, it was suggested by the presence of
active proteolytic digestive enzyme. Second, blood
plasma contained a variety of enzyme inhibitors, particularly protease inhibitors, whose presence could, and indeed seemed designed to, prevent or mitigate pathological occurrences that might occur pursuant to the
absorption of active enzyme. In fact, the second most
abundant plasma protein after albumin was a trypsin
inhibitor, ␣1-antitrypsin.
It was the realization that digestive enzymes might be
absorbed into blood that gave rise to the idea of an
enteropancreatic circulation. It had been discovered that
some of these proteins could be transported from the
interstitial space of pancreatic tissue across the basolateral surface of its secretion cells (29, 30, 45, 47), and it
was appreciated that material accumulated by these cells
might subsequently be secreted into the gland’s duct system and enter the intestinal lumen along with endogenous
3
4
ROTHMAN, LIEBOW, AND ISENMAN
the intestines from these measurements. With the use of
conservative assumptions (that transport only occurs
across ileal epithelium and that the maximal rate possible
was being observed under the particular conditions of the
study), 50 – 65% of the amylase secreted by the rabbit
pancreas in response to maximal stimulation would be
absorbed over a comparable period of time (21). Thus, at
least as a first-order approximation, the potential existed
to transport sufficient quantities of amylase across the
intestinal epithelium to support a significant conservation
of this enzyme.
B. Transpancreatic Transport
Likewise, the potential for transport across the pancreatic epithelium, i.e., from the interstitial spaces of the
gland to its duct system, was assessed. An organ culture
preparation of rabbit pancreas was used for this purpose.
The rabbit pancreas is extremely thin and can be studied
whole in vitro without vascular perfusion (80, 86). The
whole organ is removed from the animal and placed in an
incubation chamber where it is bathed by a physiological
salt solution. The gland’s exocrine secretion can be collected pure and undiluted by cannulating the pancreatic
duct. As in the intestinal studies, amylase and chymotrypsinogen were the test molecules.
Physiol Rev • VOL
IV. CIRCULATION MEASURED IN SITU
A. Circulation of Intact Digestive Enzymes
by Tracer Analysis
A variety of studies had established that intestinal
and pancreatic epithelia were permeable to digestive enzymes and that this permeability was of sufficient magnitude to accommodate a conservation process. Although
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
FIG. 2. Amylase transport in the mucosal to serosal direction
(M3 S) across ileal epithelium in vitro. The appearance of amylase in
the serosal medium is shown as a function of time after its addition to
the mucosal bathing fluid, either in the presence or absence of oxygen
(95% O2-5% CO2 or 95% N2-5% CO2). The results show active amylase
absorption. [Modified from Goetze and Rothman (21).]
When radiolabeled forms of these enzymes were
added to the medium bathing the gland, providing access
to its interstitial fluid compartment, they appeared in
ductal fluid over time. Like the ileal epithelium, the pancreatic epithelial layer was permeable to these proteins.
Also, like the intestines, transport seemed to be selective.
For example, labeled chymotrypsinogen was secreted at
20 times the rate observed for the control protein albumin
(48). Again, the secretion of the exogenous material was
also measured as enzyme activity (31) to demonstrate that
the enzymes were transported across the gland intact and
functional.
The rate of transpancreatic transport could be substantial, greatly surpassing basal rates of endogenous secretion. For example, maximal transpancreatic amylase
secretion was some 15 times the rate of unstimulated
endogenous secretion, or ⬃20% of a maximal endogenous
response to pharmacological stimulants. For chymotrypsinogen, transpancreatic secretion was 7.5 times the unstimulated endogenous rate, or ⬃30% of a maximal response. This was in the absence of stimulants such as the
hormone cholecystokinin (CCK) or acetylcholine that
were found to greatly enhance transpancreatic transport
as well as endogenous secretion (28, 48).
Transport across the pancreas was discovered to be
the result of protein movement through the gland’s secretory epithelial or acinar cells, not extracellularly via paracellular shunts (31, 55–57, 92). For example, when chymotrypsinogen was added to the bathing medium several
hours before applying a maximal cholinergic stimulus, the
response elicited by the stimulant subsequently was more
than three times the maximal response seen without preincubation (31) (Fig. 3). This salutary effect was the result
of the prior uptake and storage of the exogenous enzyme
by the secretion cells. If endogenous amylase pools were
labeled with radioactive amino acids, the addition of unlabeled amylase to the medium bathing the gland not only
reduced the specific radioactivity of the secreted amylase
by some 90%, meaning that ⬃90% was of exogenous origin, but produced a large (72%) absolute reduction in the
secretion of endogenous (labeled) protein (31). This was
the result of the mixing of exogenous and endogenous
enzyme in the cell and their subsequent competition for
exit from it.
CONSERVATION OF DIGESTIVE ENZYMES
5
FIG. 3. The transpancreatic transport of chymotrypsinogen (ChTg)
under basal conditions (basal), in the presence of 1 mg/ml chymotrypsinogen A (⫹ChTg), after the administration of acetyl-␤-methylcholine
(1 mg/100 ml bath fluid) (⫹Mch), and after the administration of Mch
subsequent to a 2-h preincubation in 1 mg/ml chymotrypsinogen A
(ChTg ⫹ Mch). Values are given (as ␮mol substrate split/min by secretion) for the peak 10-min collection period, except for “basal.” [Modified
from Isenman and Rothman (31).]
the fact that the pertinent membranes were permeable
suggested that circulation would occur given the opportunity, it was important to demonstrate in situ that it did.
To test for circulation, labeled chymotrypsinogen was
instilled into the upper duodenum of anesthetized rabbits,
and secretory fluid was simultaneously collected from the
cannulated pancreatic duct to monitor for its appearance
(Fig. 4). Shortly after intestinal instillation, a small
amount of labeled protein (48) (Fig. 4) was recovered in
ductal fluid. Electrophoretic profiles established that the
label was still associated with chymotrypsinogen and had
not been reincorporated into the various endogenous proteins secreted by the gland.
B. Magnitude of Circulation
Although this confirmed the existence of an enteropancreatic circulation, a different approach was needed
to assess its magnitude and hence the possibility of conservation. To make such a determination, the secretion of
digestive enzymes by the pancreas of anesthetized rabbits
was followed in three different experimental situations
(Fig. 5) (20). In the first situation, secretion was collected
directly from the duct system of the gland via a catheter
and permanently removed from the animal. Equal
amounts of albumin were instilled into the intestines instead of the secreted enzymes. In the second experimental situation, after removal of a small aliquot for the
measurement of its protein content, the secreted fluid was
immediately returned to the intestinal lumen close to the
Physiol Rev • VOL
FIG. 4. The circulation of labeled chymotrypsinogen. Top: drawing
shows the path followed by labeled bovine chymotrypsinogen (solid
circles) after its instillation into the duodenum (1), through its absorption (2), pancreatic uptake (3), and subsequent secretion (4). Bottom:
graph shows the concentration of the labeled chymotrypsinogen as it
appeared in pancreatic secretion after its instillation into the small
intestines. Electrophoresis demonstrated that the label remained associated with chymotrypsinogen. [Modified from Liebow and Rothman
(48).]
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
site of its original secretion to follow its normal route. In
the third situation, the secreted fluid was collected and
injected into circulatory system via a large vein, rather
than being allowed to enter the intestinal lumen. All of
this was done in the continued presence of a maximal
cholinergic stimulus to drive protein secretion.
In the absence of enzyme circulation, there was no
reason to expect that the diversion of the secreted proteins from the animal would influence the subsequent rate
of protein secretion by the gland. That would be determined by the cholinergic agonist acting locally on the
secretion cells, and there was no basis for thinking that
the removal of already secreted protein from the body
would affect its presence. For the same reason, reintroducing the secreted enzymes into the intestines, or injecting them into the bloodstream, was not expected to affect
6
ROTHMAN, LIEBOW, AND ISENMAN
secretion. From the single pass perspective, whatever one
might do to the material after it was secreted, removing it
from the animal, letting it enter the duodenum, injecting it
into blood, or anything else that we might think of, would
not be expected to affect the ensuing rate of secretion.
On the other hand, if enzyme conservation occurred,
under the right circumstances protein secretion would be
altered by each of these manipulations. It would be reduced if sufficient circulating enzyme was removed from
the animal by its diversion. Whereas the reinstillation or
intravenous injection of the same material would ameliorate that reduction by reintroducing the diverted enzyme
Physiol Rev • VOL
FIG. 6. The relationship between the intestinal instillation or intravenous injection of digestive enzyme and the amount recirculated. Data
from both instilled and injected studies form a single function with a
slope of 0.92; that is, 92% of the instilled or injected material is recirculated. [Modified from Goetze and Rothman (20).]
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
FIG. 5. The bulk circulation of digestive enzyme. Top: drawing
shows three experimental conditions: 1) pancreatic secretion (open
circles) diverted from the animal via a catheter, 2) pancreatic secretion
returned to the duodenum after a sample is removed for measurement of
its protein contents, and 3) pancreatic secretion injected intraveneously.
Intestinal instillation and intravenous injection were carried out every 30
min. A cholinergic agonist (methylcholine chloride) and the peptide
hormone secretin were injected at half-hourly intervals for the duration
of the experiment, starting at time 0. Bottom: graph shows the amount
of protein in pancreatic secretion after the 3 different treatments. The
instillation or intravenous injection of the secreted protein (2 and 3)
enhanced subsequent protein secretion two- to threefold compared with
the diverted control (1). [Modified from Goetze and Rothman (20).]
and allowing it to be recycled. Furthermore, at steady
state, it would be immaterial whether the material had
been added back to the intestinal lumen or injected directly into blood. The effect would be roughly the same
either way, since only the location of reintroduction
would have differed.
As it turned out, all three conditions, diversion, intestinal instillation, and intravenous injection, altered secretion by the gland and in the particular ways predicted
by the circulation hypothesis. When the secreted material
was diverted from the gut, protein secretion by the gland
declined in a roughly exponential fashion (Fig. 5). Both
reinstillation into the intestines and intravenous injection
of the secreted material ameliorated this decline and to
approximately the same degree (Fig. 5).
Perhaps the most convincing observation was that
protein secretion by the pancreas could be restored when
pancreatic fluid was injected into blood. This made no
sense in terms of a single pass mechanism. Why would
digestive enzyme in blood increase the rate of protein
secretion? The fact that the effect of intestinal instillation
and intravenous injection were comparable seemed to
rule out the possibility that digestive enzymes in the intestines had stimulated secretion by means of some sort
of intestinal nervous or hormonal action.
The efficiency of circulation could be estimated by
comparing the rate of secretion with diversion to that
after readministration of the secreted material (either
intestinal instillation or vascular injection). When the
amount readministered was plotted against the increase
in pancreatic secretion, input was related to output by a
linear function with a slope of 0.92 (Fig. 6). That is, the
efficiency of circulation was 92%. In accordance with the
theory, the data fit the same function whether the material
was added to the intestines or injected into the bloodstream.
CONSERVATION OF DIGESTIVE ENZYMES
C. Effect on the Plasma Pool
Although these experiments provided convincing evidence for bulk circulation, they did not actually follow
the movement of the secreted products through the relevant compartments, most importantly into and through
blood. This was subsequently shown in awake rats. As in
the rabbit studies just discussed, rats were given a maximal cholinergic stimulus for an extended period of time
(3 h), after which the distribution of amylase in various
organs and tissues of the body was assessed and compared with that observed before stimulation (61). Amylase distribution was appraised by measuring the enzyme’s activity, rather than by means of its physical
separation or by radioactive labeling, to ensure that the
distribution of active enzyme was being surveyed.
The cholinergic stimulant was known to produce a
significant reduction in the amylase content of the rat
pancreas and to increase amylase secretion into the intestines by at least an order of magnitude over unstimulated rates (81). Under these conditions, the decrease in
the enzyme contents of the gland is roughly equal to the
amount it secretes. If recirculation occurred efficiently,
then reducing the tissue content of digestive enzyme
should add approximately the same amount of protein to
the circulating pools and should be measurable there.
After an overnight fast (again, to maximize tissue
content and minimize food in the small bowel), over 99%
of the amylase activity in the rat was found in three
compartments: the pancreas with 92.1%, the intestines
(mucosa plus contents) with 5.9%, followed by the plasma
(plus interstitial fluid) with 1.3% (61). How did things
change after a 3-h course of stimulation? According to the
single pass hypothesis, the amylase lost from the pancreas should be found exclusively in the intestines, primarily in the small intestines, minus whatever degradation had taken place.
Remarkably, after 3 h of continuous highly augmented secretion into the small intestines, no significant
elevation in amylase activity was seen at this location
(Fig. 7). That copious amounts of the enzyme had entered
the small bowel could be confirmed by large increases in
amylase activity seen in the cecum (two orders of magnitude above controls levels) and colon (61). Why wasn’t
there a comparable, or even a greater increase, in the
small intestines? This could be explained in three ways.
The secreted amylase might have simply passed into the
FIG. 7. Change in the percentage of the
whole body distribution of amylase activity
attributable to the pancreas, small intestine
(intestines), and blood plus interstitial fluid
(blood) after 3 h of cholinergic stimulation
(depletion, stipled bars) and a subsequent
2-h period of recovery (solid bars). Note that
at 3 h (depletion) amylase lost from the pancreas is in great part found in blood, not the
small intestines. Similarly, the recovery of
pancreatic amylase content is accompanied
by an equivalent loss of amylase from blood.
[Modified from Miyasaka and Rothman
(61).]
Physiol Rev • VOL
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
Circulation seemed to be a process of substantial
magnitude for the particular conditions of study. We underline this phrase to stress the importance of appreciating the circumstances in which circulation was examined.
The results demonstrated, and only demonstrated, that
when secretion is driven by maximal cholinergic stimulation for some 5 h in animals that were previously fasted
overnight, circulation was substantial. A period of maximal, prolonged stimulation was chosen to ensure depletion of tissue enzyme contents, and the animals were
fasted overnight to ensure the presence of substantial
enzyme stores in the pancreas initially and to provide a
more or less food-free intestinal tract.
One would not expect to see a similar effect in any
and all circumstances, notably in the absence of substantial and prolonged stimulation of secretion to mobilize
large amounts of product for circulation, or in the absence of an appropriate fast to maximize pancreatic enzyme contents and provide an empty small bowel. In
other conditions, for example during unstimulated secretion or after short periods of stimulation, the amount
circulating might be small and difficult to measure. And if
the animals were not fasted, the presence of food in the
intestines might delay circulation in an uncharacterized,
meal-dependent manner as food substrate competed with
the absorptive process for the enzymes.
7
8
ROTHMAN, LIEBOW, AND ISENMAN
Physiol Rev • VOL
reciprocal fashion (61) (Fig. 7). That is, the rate at which
amylase disappeared from blood was correlated with its
reappearance in the pancreas. It was not distinct from it,
as we might expect if we were dealing with two unrelated
events, for example, if the disappearance of amylase from
blood was the consequence of its breakdown in the liver
or its excretion via the kidney, while its appearance in the
pancreas was due to the synthesis of new amylase.
V. ATTEMPTS AT CORROBORATION
Taken together, the observations just outlined provided a compelling case for the conservation and circulation of digestive enzymes. It was not long before the
phenomenon was corroborated (24, 37). However, almost
as quickly there were reports that questioned its existence
(44, 74, 78). As noted at the outset, these negative observations were not failed attempts to reproduce the findings
that we have just discussed, such as the effect of the
diversion, reinstillation, and intravenous injection of secreted protein on pancreatic function, but entirely different experiments that focused in particular on one prediction of the circulation hypothesis. For circulation to
occur, enzyme in blood must be taken up by and transported across pancreatic tissue into the gland’s duct system. Two subsidiary hypotheses, or what were thought to
be subsidiary hypotheses, of this understanding were
tested.
A. Pancreatic Arteriovenous Concentration
Difference
The first hypothesis was that if uptake occurs, digestive enzymes in plasma should be extracted during the
passage of blood through the pancreatic capillary bed. As
a consequence, they should be present at lower concentrations in venous blood leaving the gland than in arterial
blood entering it. When such measurements were made,
venous blood had a higher, not a lower, amylase concentration (44). This meant that the enzyme was being released from the pancreas into blood, not withdrawn from
it, as predicted. From this observation it was concluded
that digestive enzymes are not taken up by the pancreas
from blood and as such cannot cross the gland. Of course,
if they cannot cross the gland, then their circulation
through the pancreas is not possible.
However, what was observed only applied to the
particular conditions that were examined. One could not
generalize, as was done, from it to all states and circumstances. Given that digestive enzymes equilibrate across
the basolateral membrane of the acinar cell (29, 30, 45), it
could not be concluded that there would always be a net
movement of amylase from tissue to blood whatever the
physiological state. This turned out to be particularly
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
large bowel from the small intestines, as the elevated
levels there might suggest. Alternatively, it might have
been degraded or otherwise inactivated in the small intestines. Finally, it may have been absorbed.
As it happened, it was clear that the missing amylase
had not simply passed through. Although enzyme levels in
the cecum and colon were greatly elevated, they still only
represented a tiny fraction of the activity found at the
same time in the small bowel even without any elevation
(⬍1%) (61). Nor did it seem likely that degradation accounted for the missing amylase activity. Although the
rate of amylase degradation in the small intestines is
unknown, it must be far lower than its rate of secretion.
How else could the enzyme be of use in digestion? Whatever its rate, if we assume that degradation is increased in
proportion to the increase in amylase secreted, we would
nonetheless expect to see a substantial, indeed a proportional increase in amylase activity. And, of course, however much degradation may have occurred, how do we
explain the fact that increases at least proportionately as
great as those seen in the large bowel and cecum were not
observed in the small intestines?
This left the third possibility. Amylase was absorbed.
This could be assessed simply by looking for the missing
amylase in blood or in tissues that might be accessed
through the bloodstream. When plasma levels were measured, much of the missing amylase activity was found.
Plasma and interstitial fluid amylase now represented
⬃13% of the total amylase contents of the body, an order
of magnitude increase above the ⬃1% seen before stimulation. It was now the second largest amylase pool after
the pancreas, and more than double the intestines (61).
Most importantly, the increase accounted for most of the
amylase that had been secreted by the pancreas (Fig. 7).
Because it was known that amylase is also secreted
directly from the pancreas into blood (28, 60), the elevated blood levels might have been the consequence of
this process, not the absorption of enzyme. But this made
little sense, since in this case we would have to conclude
that almost all of the amylase secreted by the gland after
cholinergic stimulation had been directed into the bloodstream, not the gut, and this was observably false (28, 81).
The occurrence of circulation was further validated
by what happened when the stimulant was withdrawn.
Two hours after its withdrawal, the percentage of amylase
in the pancreas had recovered to values close to those
found in fasted rats (61) and once again contained ⬃90%
of the total in the body. This rapid replenishment was
consistent with microscopic studies in which recovery in
the volume density (the volume of the cell occupied) of
zymogen granules (the structures in which the digestive
enzymes are stored) occurred over roughly the same period of time for the same conditions in the same species
(81). Whatever the rate of recovery, as pancreatic levels of
amylase increased, its level in plasma fell in a roughly
CONSERVATION OF DIGESTIVE ENZYMES
B. Intravenous Injection of Radiolabeled
Digestive Enzyme
The second prediction was that if radiolabeled digestive enzymes were injected into the systemic circulation
they would be accumulated by the pancreas and subsequently found in its exocrine secretion. Two groups (74,
76 –78) tested this hypothesis, and both reported that little
if any of the injected material was secreted by the gland
for the times and circumstances they examined. Like the
workers who measured arteriovenous concentration difference, they concluded that transport of digestive enzymes across the pancreas does not occur and hence that
their circulation through this gland is not possible.
The supposition seemed to be that if the circulation
hypothesis was correct, then most, or at least a large
percentage, of the injected material would appear forthwith in exocrine secretion as the gland rapidly cleared the
plasma of the labeled protein, much like the kidney clears
it of glucose. One group thought that the recovery of 3– 4%
of the injected material in pancreatic secretion was too
small to be considered meaningful.
The use of radiolabeled enzymes in these studies was
predicated on the understanding that the radioactive material “traced” the movement of the bulk of unlabeled
digestive enzyme; that is, if little labeled enzyme crossed
the pancreatic epithelium, then the same could be said of
unlabeled digestive enzymes of endogenous origin. This is
called tracer kinetics, and for it to apply two things must
be known. First, the labeled material must have equilibrated with the unlabeled form of the substance in all
relevant compartments, in this case in the various extracellular pools of the body. Second, the ratio of the labeled
to the unlabeled material, referred to as specific radioactivity (SRA), in each of these compartments must be
Physiol Rev • VOL
known. Only then can we calculate how much unlabeled
material is moved for a given amount of radioactivity.
Such knowledge was lacking in these studies and consequently it could not be assumed, as it was, that the radiolabeled material traced the movement of endogenous substances. The fact that only small, seemingly negligible
quantities of labeled protein appeared in pancreatic secretion said nothing about the movement of unlabeled
protein.
On entering the systemic circulation, either as a result of intravenous injection or absorption, marker molecules like the labeled digestive enzymes that were used,
first equilibrate with their endogenous forms in plasma,
and, where they exit the capillary beds, in the various
interstitial fluid compartments of the body. That is to say
the labeled digestive enzymes would not be “attracted” to
the pancreas as if it were some sort of magnet but would
be dispersed broadly throughout the body. Only after this
broad distribution takes place can there be any sizable
clearance by the pancreas.
As a consequence, one would expect that only a
small percentage of the labeled material would end up in
the interstitial spaces of the pancreas initially and be
available for uptake and transport across the glandular
epithelium. We can estimate how much from the mass of
the pancreas in relation to the total soft tissue mass of the
body, assuming that the size of the interstitial space is a
relatively constant proportion of the fractional mass of all
soft tissues and that most capillary beds are equally permeable to these substances. Calculated in this way, the
interstitial space of the pancreas would contain ⬃0.15% of
the injected aliquot after its initial equilibration throughout the extracellular spaces of the animal. The actual
number might be somewhat smaller or larger depending
on the suitability of the assumptions used in making the
calculation, but even if it were an order of magnitude
larger, only a small percentage of the injected material
would be found in the interstitial spaces of the pancreas
initially. Given a value of 0.15%, finding 3– 4% of the injected label in pancreatic secretion is impressive.
In any event, as said, only after its broad dispersion
throughout the body can the labeled protein be effectively
redistributed to the pancreas and other tissues that might
accumulate it. If, for the sake of estimation, we assume
that this redistribution occurs by the same diffusive and
flow processes that led to the initial distribution, then it
would occur more slowly to the degree that the exogenous (labeled) material is diluted by its endogenous (unlabeled) forms in blood and the various interstitial spaces.
The greater the dilution, the longer redistribution would
take.
For the case at hand, we would expect a considerable
dilution because the concentration of digestive enzymes
in the relevant endogenous compartments is substantial,
in the tens to hundreds of nanograms per milliliter range
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
important since the researchers had infelicitously chosen
to look for net amylase extraction from blood under the
most unfavorable conditions possible. They examined
fasted dogs under general anesthesia. In this circumstance, the pancreas does not secrete significant quantities of digestive enzymes into the intestines. The amount
of enzyme secreted and made available for circulation
would be expected to be one to two orders of magnitude
less than during active secretion. Hence, given the existence of an ongoing release of digestive enzymes from the
pancreas into blood (28), it was not surprising that a net
uptake of enzyme by the pancreas was not seen. The
judgement that there would never be a net uptake of
amylase by pancreatic tissue because of what was observed in this particular circumstance was like concluding that because most passengers on the New York subway are heading from the Boroughs to Manhattan at 8 A.M.
that this would also be true at 5 P.M.
9
10
ROTHMAN, LIEBOW, AND ISENMAN
1. Tissue levels of radioactivity after injection of
[35S]methionine-labeled pancreatic exocrine proteins
into the blood circulation
TABLE
TCA-Insoluble Radioactivity,
dpm/100 mg tissue ⫻ 10⫺2
Change From
15 Min, %
Organ or Tissue
15 min
60 min
420 min
60 min
420 min
Kidney
Lung
Spleen
Liver
Skeletal muscle
Intestine
Pancreas
311
18
30
29
2
3
3
170
24
47
48
5
29
72
60
17
37
41
9
46
83
⫺56
⫹33
⫹57
⫹66
⫹250
⫹967
⫹2,400
⫺81
⫺6
⫹23
⫹41
⫹450
⫹1,533
⫹2,767
Times (in min) refer to time after injection. [Modified from Rohr
and Scheele (76).]
Physiol Rev • VOL
with a variety of other tissues. For example, the liver
contained about an order of magnitude more per unit
tissue mass. However, with the passage of time, the
amount in the pancreas increased greatly, indeed magnitudinally, while that in other tissues either decreased or
only increased modestly (Table 1). By 7 h after the injection, the pancreas had the highest concentration of any
tissue examined, and within an hour its contents had
increased some 24-fold and were higher than that found in
the liver. The second largest increase was for intestinal
tissue (Table 1), which presumably for the most part
comes from the pancreas. If we factor this in, at 7 h the
increase was 43-fold.
The authors thought that the presence of labeled
protein in the pancreas was due to the uptake of labeled
amino acids previously released from the injected protein
elsewhere in the body and its subsequent incorporation
into the endogenous protein pool of the gland. Their
evidence for this conclusion was that when they collected
secretion from the pancreatic duct of rats after the intravenous injection of a mixture of radiolabeled guinea pig
digestive enzymes, only radiolabeled rat protein was
found in secretion as estimated by two-dimensional gel
electrophoresis.
Although the amount of labeled protein in secretion
was small and the distinction between rat and guinea pig
spots in the electrophoretogram uncertain in some respects, at least some of the labeled protein appeared to
comigrate with unlabeled rat proteins, suggesting that
they might also be rat proteins. Two measurements were
needed to validate this interpretation. The first was a
control in which small amounts of labeled guinea pig
protein (similar to the amount of labeled protein found in
secretion) would be admixed with larger amounts of unlabeled rat digestive enzyme (the same amount placed on
the gel in the experiment) and subjected to electrophoretic separation. In this case, where small amounts of
labeled guinea pig proteins were known to be present,
were distinct guinea pig spots seen and were rat spots
devoid of label? The second measurement is to show that
all cellular label is associated with rat protein and that
guinea pig protein is not in the acinar cell. Neither measurement was provided then or subsequently, and as a
result the authors’ claim was not experimentally substantiated.
There was another issue. As in the studies of arteriovenous concentration difference, the researchers had
only studied the unstimulated glands of anesthetized animals. As said, under these conditions, the pancreas only
secretes small amounts of protein. Thus once again tissue
uptake as well as transpancreatic transport was being
sought in a setting in which we would expect both to be
minimal.
Finally and significantly, accepting their results as
presented, they neither explained away nor otherwise
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
(28). The amount of labeled protein moved per mole of
total digestive enzyme would be reduced in proportion to
this dilution. For example, given the same forces driving
movement, if the ratio of labeled to unlabeled protein
(SRA) was reduced by two orders of magnitude during an
initial period of equilibration of say 10 min, it would take
1,000 min, or ⬃16 h, for redistribution to occur. Dilution
by its unlabeled form would in this sense “trap” the labeled material in the animal’s extracellular spaces. Consequently, it is not justified to conclude, as was done, that
because little tracer surfaced in pancreatic secretion in
the short run, that more would not be found there in due
course.
Moreover, the appearance of labeled protein in pancreatic tissue does not guarantee its immediate secretion.
Exogenous digestive enzymes taken up by the acinar cell
mix with its endogenous stores (31) just as they do in the
extracellular spaces of the body and are diluted by them.
Except that the dilution in this case is far greater because
the pancreas contains these proteins at very high concentrations (mM). In any event, once again the labeled protein would be “trapped,” unable to leave the acinar cell in
any quantity due to its severe competitive disadvantage in
relation to the far more abundant unlabeled enzyme reserves. In fact, even the complete turnover of the cellular
enzyme pool would not guarantee its expulsion, because
it would remain, at all times, at a severe competitive
disadvantage for exit from the cell. Its release could only
be ensured if the gland was fully depleted of its protein
contents during active, stimulant-induced exocrine secretion.
Although it seemed unremarkable to these workers,
when they made observations over an extended period, a
substantial redistribution of the labeled material to the
pancreas was actually seen, just as predicted by the circulation hypothesis (Table 1) (76). Shortly after injection
of the radiolabeled enzyme (15 min), the pancreas contained a low concentration of labeled protein compared
CONSERVATION OF DIGESTIVE ENZYMES
accounted for what at the time was the already substantial published evidence, some of which we have discussed, that digestive enzymes are taken up by pancreatic
tissue from its interstitial space and transported across
the gland in rabbits. In a contemporaneous dialogue in
print about some of these issues (6, 42, 75, 79, 87–90, 94)
we argued that the “falsifying” studies were equivocal and
incomplete and subject to interpretations more benign to
the circulation hypothesis than had been allowed. We still
believe this to be true.
C. Fate of Digestive Enzymes in the Intestines
Physiol Rev • VOL
aspirates from different portions of the gut taken at different times. As expected, with the passage of time the
presence of the three enzymes in the upper small bowel,
or duodenum, first increased and then decreased, returning at variable rates over a 3- to 6-h period to values
similar to those seen before the meal. But almost everything else that was observed was not this straightforward.
As the chyme moved aborally, through the jejunum to
the ileum, only 1% of the lipase activity was still present,
but over 20% of the trypsin and some 75% of the amylase
made the passage without being either degraded or absorbed. The authors explained these large differences as
being due to variations in the resistance of the different
enzymes to degradation and additionally for amylase, to
the protective effect of its binding to the starch substrate,
although they presented no evidence to support either
suggestion. The observed differences might just as well
have reflected differences in the rate and location of the
absorption of the three enzymes. In any event, and significantly, it turned out that for all of the enzymes, what
passed into the lower small bowel was apparently intact
or almost intact enzyme, not substantially degraded products.
If one looks at the concentration of enzyme (amount
per unit volume of aspirate), although lipase concentration decreased by some 97% between the duodenum and
ileum as measured by activity, its immunoreactivity was
retained. That is, the enzyme appeared to be essentially
intact or at least sufficiently intact to react with its antibody. Trypsin, on the other hand, had become less immunoreactive but retained its enzymatic activity, suggesting
nondegradative structural change. Finally, amylase concentration increased.
The authors believed that they could exclude absorption from consideration because the disappearance of
immunoreactivity and enzymatic activity did not occur in
parallel for either lipase or trypsin. As said, they understood that for every absorbed molecule both its enzymatic
activity and immunoreactivity should disappear from the
gut in concert, and this was not seen. However, this
explanation holds only if one does not consider alterations to the molecules that might occur before their
absorption. For example, given that nondegradative
chemical modification takes place in the intestines, altered, but still intact, molecules might be absorbed as well
as or even in preference to the unaltered molecule.
In the final analysis, one could not conclude from this
study that the disappearance of enzyme activity was the
result of degradative processes alone, as was done. Not
only didn’t the data provide reason to choose between
degradative and absorptive explanations for the disappearance of enzyme activity, remarkably there was no
proof of enzyme degradation at all. In fact, the maintenance of enzyme activity (trypsin and amylase) or immunoreactivity (lipase) in the ileum seemed to argue against
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
As mentioned at the outset, the traditional understanding that the digestive enzymes are degraded in the
small bowel along with their substrates (the ingested food
and the other digestable contents of the small intestines)
was not the result of the direct measurement of degradation and its correlation to the loss of enzymatic activity. If
degradation is the fate of all digestive enzyme secreted
into the bowel subsequent to the ingestion of a meal, then
such measurements would show that over time the disappearance of enzymatic activity in various aboral locations was accompanied by the concomitant and quantitatively equivalent appearance of small peptides and amino
acids liberated by that degradation.
Making such measurements admittedly presents significant problems, but an attempt was made by researchers to assess the situation in a simpler way. They compared observed levels of enzyme activity in the intestines
to the amount of the enzyme bound in a radioimmunoassay (40). As enzyme is degraded, the peptide fragments
produced by degradation would either be unable or become increasingly unable to bind to the antibody. The
parallel loss of enzymatic activity and immunoreactivity
would be consistent with the loss of activity by enzyme
degradation. Perhaps this can be appreciated with a
counter example. If enzymatic activity disappeared, but
not immunoreactivity, we would have to conclude that
some event other than degradation had produced the
reduction in enzyme activity.
The researchers who carried out this work understood that the parallel disappearance of enzymatic activity and immunoreactivity could be the consequence of
enzyme absorption as well as degradation. In this case, as
enzyme is absorbed both activity and immunoreactivity
would also disappear. Consequently, the authors concluded from their measurements that secreted digestive
enzyme was degraded, not absorbed, but careful examination of their data reveals something else.
Experiments were carried out on human subjects
after the ingestion of a meal comprised solely of starch.
Enzymatic activity and immunoreactivity of lipase and
trypsin, and the activity of amylase, were measured in
11
12
ROTHMAN, LIEBOW, AND ISENMAN
it. What was observed was exactly what would be expected if absorption had taken place: the disappearance
of enzyme from the bowel in the continued presence of
intact digestive enzyme within it.
VI. RATE OF SYNTHESIS OF DIGESTIVE
ENZYMES COMPARED WITH THEIR
RATE OF SECRETION
Physiol Rev • VOL
One way to estimate the rate of synthesis is from the
rate of secretion seen in the relative absence of exogenous nervous and hormonal stimuli, that is, “unstimulated” or “basal” secretion. Under these circumstances
tissue contents remain unchanged despite ongoing low
levels of secretion. As such, what is being secreted is
being quantitatively replaced. If replacement is due to
new synthesis, then the rate of secretion is equal to the
rate of synthesis.
If a significant fraction of new digestive enzyme does
not successfully negotiate the posttranslational phase of
production, that is, is degraded without ever being secreted, this method underestimates the true synthetic
rate. However, this is not the rate in which we are really
interested. Rather, what we want to know is how much
new protein is available to be secreted to replace previously released protein.
In the fasted anesthetized rat, secretion in the absence of exogenous stimuli ranges from ⬃0.25 to 1.0 mg
protein䡠g tissue⫺1䡠h⫺1 (81). Tissue levels of enzyme are
constant at rates of secretion in this range, and hence they
provide a good estimate of the “effective” rate of protein
synthesis, the amount of new protein available for secretion under these conditions. On the basis of these values,
between 6 and 24 mg protein are manufactured each day
per gram of tissue, or some 4 –35% of the amount secreted
(Table 2). With the assumption of the highest estimate for
synthetic rate and lowest for daily secretion, synthesis is
still only about one-third of that needed.
2. Estimates of the rate of protein synthesis
by the exocrine pancreas
TABLE
Estimated Range
Based On:
Basal secretion
Product of ribosome
number and time for
synthesis of peptide
chain
Oxygen consumption
Stimulant enhanced
Rate of secretion
Protein 䡠 g tissue⫺1 䡠 day⫺1
Tissue pool/day, %
6–24
9–34
8–17
5–10
7–29
11–24
7–14
10–41
70–140
100–200
Percent of tissue pool per day is based on 70 mg protein/g tissue
(that found in rat pancreas after an overnight fast). Basal secretion is
based on the equivalence of basal secretion to the effective rate of
synthesis. Product of ribosome number and time for synthesis of peptide
chain is based on 106 attached ribosomes per cell and a synthetic rate of
3 min/chain. Oxygen consumption compares oxygen consumption to the
cost of peptide bond synthesis. Stimulant enhanced refers to elevations
over basal rate (given in “basal secretion”) of 50% for 5 h (based on the
experimental literature). Rate of secretion is given for comparison. It is
based on 75% at the basal rate (0.25–1.0 mg 䡠 g⫺1 䡠 h⫺1) and 25% at 10 –20
mg 䡠 g⫺1 䡠 h⫺1 (estimated from data in the rat, Refs. 25, 61, 81).
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
Another way to approach the question of the conservation of digestive enzymes is to assess the capability of
the exocrine pancreas to manufacture sufficient quantities of these proteins to replace those it secretes (88). If
the gland is not capable of meeting this need, then a
conservation of digestive enzymes is not merely suggested, but required. Over time, shortfall would build on
shortfall, deficiency on deficiency, until the gland was
depleted of its enzyme contents and unable to respond to
stimuli.
The most common method of measuring protein
synthesis, following the rate of incorporation of radioactive amino acids into protein molecules, although
useful for comparing synthetic rate under different conditions is not satisfactory for quantitative measurements of the sort needed. There are, however, other
ways to assess the sufficiency of synthesis. Next we
consider three approaches, each dependent on different variables.
First we must settle on a value for protein secretion
to compare to estimates of synthetic capacity. As we have
already discussed in another context, the question is not
whether synthesis is capable of replacing what is secreted
under some circumstances, but under all circumstances,
particularly when rates of secretion are high and the gland
is partially or fully depleted of its enzyme contents, as is
frequently the case. The important question is whether
under these conditions the recovery of tissue digestive
enzyme content is due solely to new synthesis?
Using data from fasted anesthetized rats (25, 61, 81),
and extrapolating to daily rates of secretion, we estimate
the amount of protein secreted by the pancreas to be
between 70 and 140 mg protein䡠g tissue⫺1䡠day⫺1. We have
assumed that secretion occurs at low or unstimulated
rates for 75% of the day and at highly elevated rates for
25%. Under these conditions, tissue digestive enzyme
stores would turn over, in one way or another, once to
twice daily based on the observed digestive enzyme contents of 70 mg protein/g tissue in rats after an overnight
fast (25). This range is conservative, that is, biased toward
the possibility of synthetic sufficiency. Actual 24-h rates
of secretion in awake animals fed ad libitum would likely
be larger, not smaller.
A. Basal Rate of Protein Secretion
CONSERVATION OF DIGESTIVE ENZYMES
B. Enhancing Synthesis
Physiol Rev • VOL
long, or 15 h, only 5 mg of new protein would be produced. For recovery to be complete in this case, synthesis
would have to remain elevated for more than 4 days.
Obviously, this is far too long for animals like rats and
humans that feed daily or more frequently.
The replenishment of tissue contents normally occurs within a few hours of the cessation of active secretion. For the circumstances described above in rats, it
takes just 2 h to reach levels that are roughly equivalent to
those seen in fasted animals (25, 61). Even if it took 15 h,
or a full 24 h, synthesis could not account for restocking
at the increases in rate that have been observed.
C. Ribosomes
Another means of estimating the rate of synthesis is
to calculate the product of the time it takes to manufacture a single peptide chain and the number of ribosomes
actively engaged in the task, since proteins are manufactured in a one to one correspondence to ribosomes. It
takes several minutes to produce single peptide chains in
eukaryotic cells (49, 50, 67, 69, 71, 106). More particularly,
it takes ⬃3 min to manufacture a molecule of pancreatic
ribonuclease or amylase (67, 71). The number of ribosomes in the secretion (acinar) cells of the pancreas
engaged in digestive enzyme synthesis can be estimated
by counting those attached to the endoplasmic reticulum
in electron micrographs. This number is thought to be the
same, or almost the same, as the number of digestive
enzyme chains being manufactured based on the understanding that 1) digestive enzymes are produced exclusively on ribosomes attached to the membranes of the
endoplasmic reticulum (ER); 2) all of these ribosomes are
actively engaged in protein manufacture; and finally 3) the
great majority are busy manufacturing digestive enzymes,
not other proteins such as membrane proteins, proteins
destined for the ER itself or for organelles such as lysosomes. These discernments are widely believed to be true,
but whether they are or are not, by presuming them true,
we bias our estimate toward higher synthetic rates.
The ER is a striking feature of the pancreatic acinar
cell and occupies some 20% of the cell’s volume (16). It
appears to be comprised of numerous elongated, ribosome-studded membrane sacs that are stacked upon each
other. By making some simple assumptions about the
arrangement of ribosomes on the membranes and how
the ER sacs fill the volume of the cell they occupy, the
number of ribosomes present can be calculated. Specifically, and to further bias our estimate toward higher
synthetic rates, we assume that the membranes are arranged in orderly rows and that the ribosomes attached to
them are packed densely in a rectilinear array, separated
from each other by the diameter of a single ribosome. We
chose a 60-nm spacing between sacs based on what is
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
Couldn’t this shortfall be made up by increases in the
rate of synthesis after periods of active secretion? The
evidence does not support this possibility. There are numerous reports, in an abundant literature (for example,
Refs. 12, 23, 26, 35, 54, 62– 65, 70, 103), of elevations in
synthetic rate after the administration of stimulants of
secretion. However, the effect is modest, generally in the
range of 25–50%. Sometimes no enhancement is seen or
even a decrease, whereas on the other hand, the largest
increase ever reported is about a doubling (54).
Whatever value we might settle on, it would not be
nearly enough to meet the need. We should understand
that without boosting synthetic rate, even a gland that is
only slightly depleted of its enzyme contents during periods of active secretion would be unable to recover what
it lost whatever the rate of synthesis might be. When
secretion returns to low, “unstimulated” rates after a period of active secretion, protein manufacture once again
equals the rate of secretion, and consequently no accumulation of enzyme can occur.
As it happens, during periods of active secretion the
enzyme content of pancreatic tissue declines substantially, sometimes dramatically. For example, after 3 h of
highly augmented secretion in the awake rat, tissue content decreases by ⬃50% (25, 61). For synthesis to account
for the recovery under such circumstances, it must be
greatly elevated. The rate of recovery would be proportional to the increase in synthetic rate. For example, if
secretion was elevated by an order of magnitude in response to a meal, as commonly occurs, synthesis would
have to be increased by an order of magnitude for recovery to occur in the same time frame as loss from the
gland. If synthesis only increased by 50%, a relatively high
value, it would take 20 times as long to replace what was
lost as to lose it. More particularly, if secretion was magnitudinally elevated for 3 h and returned to the unstimulated state immediately thereafter, it would take 60 h, or
2.5 days, to replace that lost.
Even this is unrealistically optimistic, because it assumes that synthetic rate remains elevated for 60 h until
recovery is complete, and this is not seen. Synthetic enhancement occurs primarily in the presence of stimulants,
that is, during periods of tissue depletion, not recovery,
and in any event only lasts for a few hours. For the
example given above, the rat pancreas loses ⬃35 mg of its
digestive enzyme contents (50%) after 3 h of active secretion (25, 61). If the rate of synthesis were elevated by 50%,
and secretion returned to its unstimulated rate with removal of the stimulant, synthesis would only have provided about a milligram of new protein, or just a few
percent of the 35 mg needed for the recovery of tissue
contents. Even if the elevation in synthesis lasted well
beyond withdrawal of the stimulant, say five times as
13
14
ROTHMAN, LIEBOW, AND ISENMAN
D. Energy Requirement
A final way to judge synthetic rate is to ask whether
the pancreas can support the replacement of the digestive
enzyme pool from an energetic perspective? That is, does
it generate enough energy to replace what it loses by new
synthesis? As said, protein synthesis is among the most
costly biological reactions. The energy required to form
the peptide bond has been estimated to be ⬃3.5 kJ/g
protein (58, 59, 72, 95, 102). The pancreas’ available energy has been measured experimentally as oxygen consumption (14, 19). Metabolism is primarily oxidative in
this gland, and values for oxygen consumption reflect its
energy utilization relatively well. Thus, knowing the cost
of synthesis and the available energy, we can estimate
how much protein can be manufactured by the pancreas
per unit time, assuming, incorrectly of course, that all of
the energy is used for this purpose.
Oxygen consumption in the pancreas is ⬃0.01– 0.02
ml䡠g tissue⫺1䡠min⫺1 at standard temperature and pressure
in mammals (14, 19) or 14 –28 ml oxygen䡠g tissue⫺1䡠24
h⫺1. With the assumption that this in the main reflects
glucose oxidation, 1.2 J would be produced for every
milliliter of oxygen consumed, and as a consequence, the
pancreas would generate somewhere between 17 and 34
J䡠day⫺1䡠g tissue⫺1. This value is roughly in proportion to
Physiol Rev • VOL
the gland’s mass as a percentage of the total mass of the
animal (59), indicating that at least in the rat, the pancreas
is about as metabolically active as other tissues on average.
To replace the 70 mg/g tissue digestive enzyme pool
of the fasted rat once a day would require 238 J or almost
10 times the available energy. Thus putting the cell’s other
needs aside, ⬃10% of the digestive enzyme pool can be
replaced daily, or ⬃5–10 mg protein can be synthesized
per gram of tissue each day, or ⬃4 –15% of our estimate of
the daily rate of secretion (Table 2). With the assumption
of the highest estimate for synthesis and the lowest for
secretion, we would still have a shortfall of some sevenfold.
E. Summary
Although calculations such as these always present
uncertainties, and must to some degree be considered
speculative, it is hard to envision errors that would be of
sufficient magnitude to invalidate the conclusion that synthesis alone is not able to replace secreted protein under
circumstances in which active secretion occurs. The calculations are based on published and well-documented
experimental measurements, and moreover, to mitigate
error, wherever possible we have tried to choose values
that would maximize synthetic rate and minimize secretory capacity to bias our numbers toward synthetic sufficiency. The fact that three independent means of estimating the rate of synthesis, each depending on different
variables, yield quite similar results (5–10, 8 –16, and 6 –24
mg䡠g⫺1䡠day⫺1), suggests that the rate of synthesis in these
glands is within this range. If so, this is far less than any
reasonable estimate for the rate of secretion from actively
secreting glands. The conclusion that synthesis is unable
to replace what is secreted is reinforced by the fact that to
restock depleted glands by means of synthesis alone requires an enhancement, in either magnitude or duration,
of synthetic rate that is well beyond anything that has
been observed.
Whatever uncertainty there may be, these estimates
cannot be taken, either singly or together, as support for
the view that the gland is capable of meeting its need for
digestive enzymes solely by new synthesis. To the contrary, they provide support for the conservation of digestive enzymes and buttress the experimental evidence for
such a process presented above.
VII. MEMBRANE PROTEIN TRANSPORT
In section I we mentioned that some of the initial
hesitancy to accept the existence of a conservation of
digestive enzymes was due to the widely held belief that
biological membranes were impermeable to protein mol-
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
seen in electron microscopic images. Given this geometry, there are about a million attached ribosomes per cell.
If the stacking of ER membrane is not this regular or the
placement of ribosomes on their surface not this dense, or
both, as appears to be the case, then this is an overestimate, and hence overestimates synthetic rate, again biasing the result in favor of synthetic sufficiency.
With the use of 3 min as the time required to synthesize an average digestive enzyme chain, 20 protein molecules would be manufactured per hour per attached ribosome assuming that there is no significant delay between
the manufacture of one chain and the next on a given
ribosome (any delay would of course reduce the rate of
synthesis). Given a million attached ribosomes per cell,
⬃1.2 ⫻ 1016 protein molecules䡠h⫺1䡠g tissue⫺1 would be
manufactured. Applying Avogadro’s number and converting to mass (assuming an average molecular mass of 30
kDa for the digestive enzymes), 1 g of tissue is capable of
producing ⬃0.67 mg of digestive enzymes per hour. Given
the various assumptions made in making this calculation,
half this value, or 0.34 mg䡠g⫺1䡠h⫺1, is probably a more
realistic, but still a generous, estimate. This amounts to
8 –16 mg protein synthesized䡠day⫺1䡠g tissue⫺1, or about
6 –24% of the amount we estimate is secreted per day
(Table 2). Once again, there is a substantial shortfall.
Even if we use our highest estimate for synthesis and
lowest for secretion, the shortfall is fourfold.
CONSERVATION OF DIGESTIVE ENZYMES
VIII. FORMS OF CIRCULATING PROTEINS
Hesitancy to accept the existence of an enteropancreatic circulation was also due to the complete lack of
knowledge of the molecular forms of the circulating species. For example, were they proenzymes, active enzymes, inactivated enzymes, or all of these? And if molecules were modified, how and where did this happen?
This also posed a chicken and egg problem. Before it
made any sense to try to determine the forms of the
circulating species, one had to be convinced that circulation actually occurred. And if the fact that nothing was
known about the forms of the circulating species engendered skepticism about circulation itself, there seemed no
point in attempting to determine the nature of events that
did not occur.
Consequently, research along these lines was not
carried out, and all we have is speculation as fodder for
future work. We can start such speculation by asking
whether each and every proteolytic enzyme is activated
immediately after its secretion into the intestinal tract in
any and all physiological circumstances? As noted, there
is no evidence that this is the case, only widely held belief.
Knowing the circumstances of activation might be a first
step in helping us understand the relative roles of pro- and
activated enzymes.
Physiol Rev • VOL
Active enzymes might be temporarily inactivated or
otherwise transformed in the intestines before their circulation. In this regard, the pancreas secretes substantial
amounts of enzyme inhibitors, most notably pancreatic
trypsin inhibitor, along with its digestive enzymes. We
have no idea what, if anything, these inhibitors do in the
intestines. Certainly, trypsin inhibitor might inactivate
trypsin. But if so, to what end? Of course, there are many
others potential means of chemically altering the digestive enzymes, including, but not limited to, covalent modifications. For example, they might be phosphorylated,
sulfonated, chlorinated, or leader or activation peptides
reattached in some way. The experiments by Layer et al.
(40) discussed above provide proof that some sort of
nondegradative modification occurs in the intestinal lumen to whatever purpose.
Modification may also occur pursuant to transport
across the intestinal epithelium, or after the proteins enter the bloodstream. As discussed, protease inhibitors are
present at high concentrations in plasma, and we would
expect active proteolytic enzymes to bind to them,
thereby reducing the risk of a generalized proteolytic
degradation. Perhaps these inhibitors also serve as “carrier” molecules presenting the circulating enzymes to the
pancreas. Finally, when the enzymes return to the pancreas, they may be chemically modified before their resecretion. Although it is thought that each digestive enzyme has a particular established structure before its
secretion, and that whatever variations exist reflect the
presence of gene-based isoenzymes, work with mass spectrometry suggests a multiplicity of forms of amylase and
other digestive enzymes with only slightly different masses
(105). Moreover, and significantly, the forms in secretion are
not identical to those in the zymogen granule.
IX. CONSERVATION OF OTHER
DIGESTIVE ENZYMES
It is natural to ask whether other digestive enzymes,
like salivary amylase or gastric pepsin, might be conserved? After all, the same issues apply. Although there
has been very little research to explore these possibilities,
Miyasaka and Rothman, in studies discussed above (61),
found that when blood levels of pancreatic amylase were
greatly elevated subsequent to stimulation of pancreatic
enzyme secretion in rats, the amylase contents of both the
parotid and submandibular glands were also greatly elevated, with the parotid gland elevated by two orders of
magnitude. This remarkable uptake of pancreatic amylase
from blood suggests the possibility that subsequently it
might be secreted into the mouth, though measurements
of secretion were not made. In a related study, Fredette
and Rothman (unpublished data) found that the administration of isoproterenol to rabbits to stimulate the secre-
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
ecules. Because conservation required the transport of
digestive enzymes across both intestinal and pancreatic
epithelia, minimally across four separate membranes,
conservation seemed unattainable. Vesicle mechanisms
could have been invoked; however, the properties of
transport that were observed, for example, its equilibrating nature, suggested direct membrane permeation.
Since the first reports of a membrane permeability to
protein molecules in the pancreas about 30 years ago (46),
a large body of experimental evidence has been published
establishing the direct transport of many different protein
molecules across as many biological membranes. Beyond
providing evidence that such processes exist, these studies have made it clear that membrane protein transport is
ubiquitous and central to cellular life. Without these processes to sort proteins to their various assigned locations,
to the mitochondrion, the nucleus, the peroxisome, the
chloroplast, as well as to allow proteins to exit the cell,
cellular life as we know it would not exist. Movement
occurs variously by the passage of proteins through large
membrane channels, through the lipid barrier, or by
means of interactions with special membrane “carrier”
proteins. We will not review this evidence here; however,
we direct readers to older reviews that deal with membrane transport of proteins in the pancreas (82– 84, 91), a
relatively recent general review of the subject by the
authors (27), as well as the first series of monographs ever
published on this topic (85).
15
16
ROTHMAN, LIEBOW, AND ISENMAN
tion of amylase into the gland’s duct system also leads to
a large increase in plasma amylase levels. However, when
amylase secreted into the mouth was prevented from
traveling down the gastrointestinal tract by means of an
esophagostomy, plasma amylase levels fell by ⬃50%.
X. CONCLUSIONS
Address for reprint requests and other correspondence: S.
Rothman, Dept. of Physiology, Univ. of California-San Francisco, 3rd and Parnassus Ave., San Francisco, CA 94143– 0444
(E-mail: [email protected]).
REFERENCES
1. ALPERS DH AND ISSELBACHER KJ. Protein synthesis by rat intestinal
mucosa. The role of ribonuclease. J Biol Chem 242: 5617–5622,
1967.
2. ALPERS DH AND SOLIN M. Amylase in intestine found in supernatant
fraction of tissue after washing to remove surface material. Gastroenterology 58: 833– 842, 1970.
3. AMBRUS JL, LASSMAN HB, AND DEMARCHI EP. Absorption of exogenous and endogenous proteolytic enzymes. Clin Pharmacol Ther 8:
362–368, 1967.
4. AVAKIAN S. Further studies on the absorption of chymotrypsin. Clin
Pharmacol Ther 5: 712–715, 1964.
5. BASSO D, FARBIS C, MEANI A, DEL FAVERO G, PUNUCCI A, VIANELLO D,
PICCOLI A, AND NACCARATO R. Serum deoxyribonuclease and ribonuclease in pancreatic cancer and chronic pancreatitis. Tumori 71:
529 –532, 1985.
Physiol Rev • VOL
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
As we reexamined the evidence for a conservation of
digestive enzymes, we found it no less compelling than we
did 25 years ago. Likewise, we found the studies that
questioned its existence as incomplete as they seemed to
us all those years ago. Add to this the calculations of
synthetic sufficiency presented here for the first time, and
it seems that the conservation of digestive enzyme does
not merely occur but is a necessary agency.
In the most general terms, conservation has much to
recommend it. Cells and organisms do not have unlimited
amounts of energy to carry out their various life functions, and being able to do with less, being more efficient,
has crucial selective advantage and can be an important
determinant in the evolution of species. The traditional
single pass view of digestion in which a completely new
complement of digestive enzymes is manufactured for
each meal has the curious consequence of requiring the
organism to be particularly wasteful in its expenditure of
energy to manufacture these costly molecules to meet its
needs for sustenance, perhaps uniquely so among its metabolic and physiological processes, when just the opposite would seem desirable. As such, mechanisms that
avoid the need to manufacture an entirely new batch of
digestive enzymes for each meal, such as is the case with
bile salts, would seem to provide an important advantage
for an organism, particularly when food is scarce, as it
often is in nature.
6. BEYNON RJ AND KAY J. Enteropancreatic circulation of digestive
enzyme. Nature 260: 78 –79, 1976.
7. BORGSTROM A, KUKORA J, AND OHLSSON K. Studies on immunoreactive
pancreatic elastase 2 in human serum. Hoppe-Seyler’s Z Physiol
Chem 361: 633– 640, 1980.
8. BORGSTROM B, DAHLQVIST A, LUNDH G, AND SJOVALL J. Studies of
intestinal digestion and absorption in the human. J Clin Invest 36:
1521–1536, 1957.
9. BRAMBEL FWR. The passive immunity of the young mammal. Biol
Rev 33: 488 –531, 1958.
10. BUCHLER M, MALFERTHEINER P, SHADLICH H, NEVALAINEN TJ, FREISS H,
AND BEGER HG. Role of phospholipase A2 in human acute pancreatitis. Gastroenterology 97: 1521–1526, 1989.
11. CARRERE J, FIGARELLA C, GUY O, AND THOUVENOT JP. Human pancreatic chymotrypsinogen A: a non-competitive enzyme immunoassay,
and molecular forms in serum and amniotic fluid. Biochim Biophys
Acta 883: 46 –53, 1986.
12. COUTURE Y, MONGEAU R, DUNNIGAN J, AND J. Morrisset evidence that
protein synthesis can be increased in vitro following cholinergic
stimulation. Can J Physiol Pharmacol 50: 874 – 882, 1972.
13. DANFORTH E AND MOORE RD. Intestinal absorption of insulin in the
rat. Endocrinology 65: 118 –123, 1959.
14. DAVIES RE, HARPER AA, AND MACKAY IFS. A comparison of the
respiratory activity and histological changes in isolated pancreatic
tissue. Am J Physiol 157: 278 –282, 1949.
15. DUANE WC, FREIRICHS R, AND LEVITT MD. Distribution, turnover and
mechanism of renal excretion of amylase in the baboon. J Clin
Invest 50: 156 –165, 1971.
16. ERMAK TH AND ROTHMAN SS. Proliferation of RER cisternae in
postnatal rat pancreas. Am Zool 17: 932, 1977.
17. FRIDHANDLER L, BERK JE, MONTGOMERY KA, AND WONG D. Column
chromatographic studies of isoamylases in human serum, milk and
urine. Clin Chem 20: 547–552, 1974.
18. GEOKAS MC, LARGMAN C, BRODRICK JW, AND FASSETT M. Molecular
forms of immunoreactive pancreatic elastase in canine pancreatic
and peripheral blood. Am J Physiol Gastrointest Liver Physiol 238:
G238 –G246, 1980.
19. GERARD RW AND STILL EU. Studies on the physiology of secretin. V.
The effects on the respiration of the excised pancreas. Am J
Physiol 103: 232–234, 1933.
20. GOETZE H AND ROTHMAN SS. Enteropancreatic circulation of digestive enzyme as a conservative mechanism. Nature 257: 607– 609,
1975.
21. GOETZE H AND ROTHMAN SS. Amylase transport across ileal epithelium in vitro. Biochim Biophys Acta 512: 214 –220, 1978.
22. GOLDBERG DM, CAMPBELL R, AND ROY AD. Binding of human trypsin
and chymotrypsin by human intestinal mucosa. Biochim Biophys
Acta 167: 613– 619, 1968.
23. GRENDELL JH AND ROTHMAN SS. The recovery of the enzyme content
of the pancreas after maximal stimulation: synthesis or conservation? Gastroenterology 84: 1174, 1983.
24. HEINRICH HC, GABBE EE, BRUGGEMANN J, ICAGIC F, AND CLASSEN M.
Enteropancreatic circulation of trypsin in man. Klin Wochenscht
57: 1295–1297, 1979.
25. HO JJL AND ROTHMAN SS. Protein concentration in the pancreatic
zymogen granule. Biochim Biophys Acta 755: 457– 466, 1983.
26. HOKIN LE. The synthesis and secretion of amylase by pigeon pancreas in vitro. Biochem J 48: 320 –326, 1950.
27. ISENMAN LD, LIEBOW C, AND ROTHMAN SS. Protein transport across
membranes: a paradigm in transition. Biochim Biophys Acta 1241:
341–370, 1995.
28. ISENMAN LD, LIEBOW C, AND ROTHMAN S. The endocrine secretion of
mammalian digestive enzymes by exocrine glands. Am J Physiol
Endocrinol Metab 276: E223–E232, 1999.
29. ISENMAN LD AND ROTHMAN SS. The transport of protein through the
basolateral membrane of the pancreatic acinar cell. Physiologist
18: 259, 1975.
30. ISENMAN LD AND ROTHMAN SS. Transport of ␣-amylase across the
basolateral membrane of the pancreatic acinar cell. Proc Natl Acad
Sci USA 74: 4068 – 4072, 1977.
31. ISENMAN LD AND ROTHMAN SS. Transpancreatic transport of digestive
enzyme. Biochim Biophys Acta 585: 321–332, 1979.
CONSERVATION OF DIGESTIVE ENZYMES
Physiol Rev • VOL
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
oxygen consumption, and muscle protein turnover (Abstract). Proc
Nutr Soc 38: 33A, 1978.
MIYASAKA K AND ROTHMAN SS. The endocrine secretion of ␣-amylase
by the pancreas. Am J Physiol Gastrointest Liver Physiol 241:
G170 –G175, 1981.
MIYAKASA K AND ROTHMAN SS. Redistribution of amylase activity
accompanying its secretion by the pancreas. Proc Natl Acad Sci
USA 73: 5438 –5442, 1982.
MONGEAU R, COUTURE Y, DUNNIGAN J, AND MORRISET J. Early dissociation of protein synthesis and amylase secretion following hormonal stimulation of the pancreas. Can J Physiol Pharmacol 52:
198 –205, 1974.
MONGEAU R, DAGORN JC, AND MORRISET J. Further evidence that
protein synthesis can be decreased in vivo following hormonal
stimulation in the rat pancreas. Can J Physiol Pharmacol 54:
305–313, 1976.
MORISSET JA AND WEBSTER PD. Effects of atropine on pigeon pancreas. Am J Physiol 219: 1286 –1291, 1970.
MORISSET JA AND WEBSTER PD. In vitro and in vivo effects of pancreozymin, urecholine, and cyclic AMP on rat pancreas. Am J
Physiol 230: 202–208, 1971.
MORIYA H, MORIWAKI C, AKIMOTO S, YAMAGUCHI K, AND MASANORI I.
Studies on the passage of ␣-chymotrypsin across the intestines.
Chem Pharm Bull Tokyo 15: 1662–1668, 1967.
MORRIS AJ AND DICKMAN SR. Biosynthesis of mouse ribonuclease in
mouse pancreas. J Biol Chem 235: 1404 –1408, 1960.
PELOT D AND GROSSMAN MI. Distribution and fate of pancreatic
enzymes in small intestine of the rat. Am J Physiol 202: 285–288,
1962.
PETERS T JR. Cytoplasmic particles and serum albumin synthesis.
J Histochem Cytochem 7: 224 –234, 1959.
POORT C AND KRAMER MF. Effect of feeding on the protein synthesis
in mammalian pancreas. Gastroenterology 57: 689 – 696, 1969.
REDMAN CM, SIEKEVITZ P, AND PALADE G. Synthesis and transfer of
amylase in pigeon pancreatic microsomes. J Biol Chem 241: 1150 –
1158, 1966.
REEDS PJ, CADENHEAD A, FULLER MF, LOBLEY GE, AND MCDONALD JD.
Protein turnover in growing pigs. Effects of age and food intake.
Br J Nutr 43: 445– 455, 1980.
REIDELBERGER RD, O’ROURKE M, DURIE PR, GEOKAS MC, AND LARGMAN
C. Effects of food intake and cholecystokinin on plasma trypsinogen levels in dogs. Am J Physiol Gastrointest Liver Physiol 246:
G543–G549, 1984.
ROHR G, KERN H, AND SCHEELE G. Enteropancreatic circulation of
digestive enzymes does not exist in the rat. Nature 292: 470 – 472,
1981.
ROHR G, KERN H, AND SCHEELE G. Is their an enteropancreatic
circulation of digestive enzyme? Am J Physiol Gastrointest Liver
Physiol 244: G104 –G106, 1983.
ROHR G AND SCHEELE G. Fate of radioactive exocrine pancreatic
proteins injected into the blood circulation of the rat. Tissue uptake
and transepithelial excretion. Gastroenterology 85: 991–1002, 1983.
ROSENBLUM JL, NIESEN TE, RABB BK, AND ALPERS DH. Fate of circulating isoamylases in the rabbit. Am J Physiol Gastrointest Liver
Physiol 244: G254 –G260, 1983.
ROSENBLUM JL, RAAB BK, AND ALPERS DH. Hepatobiliary and pancreatic clearance of circulating pancreatic amylase. Am J Physiol
Gastrointest Liver Physiol 243: G21–G27, 1982.
ROSENBLUM JL, RAAB BK, AND ALPERS DH. Is their an enteropancreatic circulation of digestive enzyme? Am J Physiol Gastrointest
Liver Physiol 244: G103, 1983.
ROTHMAN SS. Exocrine secretion from the isolated rabbit pancreas.
Nature 204: 84 – 85, 1964.
ROTHMAN SS. Enzyme secretion in the absence of zymogen granules. Am J Physiol 228: 1828 –1834, 1975.
ROTHMAN SS. The passage of protein through membrane: old assumptions and new perspectives. Am J Physiol Gastrointest Liver
Physiol 238: G391–G402, 1980.
ROTHMAN SS. Protein Secretion: a Critical Analysis of the Vesicle
Model. New York: Wiley, 1985.
ROTHMAN SS. Nonvesicular Transport. New York: Wiley, 1985.
ROTHMAN SS. Membrane Protein Transport. Greenwich, CT: JAI
(Elsevier), 1995–96, vol. 1–3.
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
32. KABACOFF BL, WOHLMAN A, UMNEY M, AND AVAKIAN S. Absorption of
chymotrypsin from the intestinal tract. Nature 199: 815– 817, 1963.
33. KATAYAMA K AND FUJITA T. Studies on biotransformation of elastase.
I. Transport of 131I-labeled elastase across rat intestine in vitro.
Biochim Biophys Acta 288: 172–180, 1972.
34. KATAYAMA K AND FUJITA T. Studies on the biotransformation of
elastase. II. Intestinal absorption of 131I-labeled elastase in vivo.
Biochim Biophys Acta 288: 181–189, 1972.
35. KORC M, BAILEY AC, AND WILLIAMS JA. Regulation of protein synthesis in normal and diabetic rat pancreas by cholecystokinin. Am J
Physiol Gastrointest Liver Physiol 241: G116 –G121, 1981.
36. KOROTKO GF AND KURZANOV AN. Secretion of amylase and lipase in
composition of bile. Fiziol Zh SSSR Im I M Sechenova 64: 81– 89,
1978.
37. LAKE-BAKAAR G, RUBIO CE, MCKAVANAGH S, POTTER BJ, AND SUMMER125
FIELD JA. Metabolism of
I-labelled trypsin in man: evidence for
recirculation. Gut 21: 580 –586, 1980.
38. LARGMAN C, BRODRICK JW, GEOKAS MC, AND JOHNSON JH. Demonstration of human pancreatic anionic trypsinogen in normal serum by
radioimmunoassay. Biochim Biophys Acta 543: 450 – 454, 1978.
39. LARGMAN C, RAY SB, BRODRICK JW, AND GEOKAS MC. Clearance and
metabolism of circulating pancreatic proelastase in the rat. Am J
Physiol Gastrointest Liver Physiol 239: G504 –G510, 1980.
40. LAYER P, GO VLW, AND DIMAGNO EP. Fate of pancreatic enzymes
during small intestinal aboral transit in humans. Am J Physiol
Gastrointest Liver Physiol 251: G475–G480, 1986.
41. LEV R AND ORLIC D. Uptake of protein in swallowed amniotic fluid
by monkey fetal intestine in utero. Gastroenterology 65: 60 – 68,
1973.
42. LEVITT MD. Is there an enteropancreatic circulation of digestive
enzyme? Am J Physiol Gastrointest Liver Physiol 244: G103–G104,
1983.
43. LEVITT MD, ELLIS C, AND ENGEL RR. Isoelectric focusing studies on
human serum and tissue amylases. J Lab Clin Med 90: 141–152,
1977.
44. LEVITT MD, ELLIS CJ, MURPHY SM, AND SCHWARTZ ML. Study of the
possible enteropancreatic circulation of pancreatic amylase in the
dog. Am J Physiol Gastrointest Liver Physiol 241: G54 –G58, 1981.
45. LIEBOW C. Protein Transport in the Pancreas (PhD thesis). San
Francisco: Univ. of California, 1973.
46. LIEBOW C AND ROTHMAN SS. Membrane transport of protein. Nature
New Biol 241: 176 –178, 1972.
47. LIEBOW C AND ROTHMAN SS. Transport of bovine chymotrypsinogen
into rabbit pancreatic cells. Am J Physiol 226: 1077–1081, 1974.
48. LIEBOW C AND ROTHMAN SS. Enteropancreatic circulation of digestive enzymes. Science 189: 472– 474, 1975.
49. LOTFIELD RB. Speed of protein synthesis (Abstract). Federation
Proc 16: 82, 1957.
50. LOTFIELD RB. The biosynthesis of protein. Prog Biophys Biophys
Chem 8: 347–386, 1957.
51. LOW AG. The activity of pepsin, chymotrypsin, and trypsin during
24 h periods in the small intestine of growing pigs. Br J Nutr 48:
147–159, 1982.
52. MARTIN GJ, BOGNER RL, AND EDELMAN E. Further in vivo observations with radioactive trypsin. Am J Pharmacol 129: 386 –392, 1957.
53. MEGEL H, STRAUSS R, HO R, AND BEILER M. Detection of trypsin-like
activity in plasma of rats after oral administration of trypsin. Arch
Biochem Biophys 108: 193–199, 1964.
54. MELDOLESI J. Effect of caerulein on protein synthesis and secretion
in the guinea-pig pancreas. Br J Pharmacol 40: 720 –731, 1970.
55. MELESE T AND ROTHMAN SS. Distribution of three hexose derivatives
across the pancreatic epithelium: paracellular shunts or cellular
passage? Biochim Biophys Acta 763: 212–219, 1983.
56. MELESE T AND ROTHMAN SS. Increased phosphate efflux from the
acinar cell during protein secretion. Am J Physiol Cell Physiol 245:
C121–C124, 1983.
57. MELESE T AND ROTHMAN SS. The pancreatic epithelium is permeable
to sucrose and insulin across secretory cells. Proc Natl Acad Sci
USA 80: 4870 – 4874, 1983.
58. MILLWARD DJ AND GARLICK PJ. The energy cost of growth. Proc Nutr
Soc 35: 339 –349, 1976.
59. MILLWARD DJ, HOLLIDAY MA, BATES PC, DALAL S, COX M, AND HEARD
CRC. The relationship between dietary state, thyroid hormones,
17
18
ROTHMAN, LIEBOW, AND ISENMAN
Physiol Rev • VOL
98.
99.
100.
101.
102.
103.
104.
105.
106.
human pancreatic amylase including a comparison with human
parotid amylase. Biochim Biophys Acta 302: 345–361, 1973.
TEMLER RS AND FELBER JP. Radioimmunoassay of human plasma
trypsin. Biochim Biophys Acta 445: 720 –728, 1976.
URBAN E, ZINGERY AA, BRUNDRANT T, WESER E, AND ZIEGLER DM.
Permeability of adolescent rat intestine to pancreatic ribonuclease.
J Pediatr Gastroenterol Nutr 1: 267–272, 1982.
WALKER WA, ISSELBACHER KJ, AND BLOCK KJ. Intestinal uptake of
macromolecules: effect of oral immunization. Science 177: 608 –
610, 1972.
WARSHAW AL, WALKER WA, AND ISSELBACHER KJ. Protein uptake by
the intestine: evidence for absorption of intact macromolecules.
Gastroenterology 66: 987–992, 1974.
WEBSTER AJF, LOBLEY G, REEDS PJ, AND PULLAR JD. Protein mass,
protein synthesis and heat loss in the Zucker rat (Abstract). Proc
Nutr Soc 37: 21A, 1977.
WEBSTER PD, BLACK O, MAINZ D, AND SINGH M. Pancreatic acinar cell
metabolism and funtion. Gastroenterology 73: 1434 –1449, 1977.
WEICKMANN JL AND GLITZ DG. . Human ribonucleases. J Biol Chem
257: 8705– 8710, 1982.
YOUNG MK, TSENG HC, FANG H, LIANG W, AND ROTHMAN SS. Comparison of stored and secreted rat pancreatic digestive enzymes by
mass spectrometry. I. ␣-Amylase. Biochim Biophys Acta 1293:
63–71, 1996.
ZAMECNIK PC, KELLER EB, LITTLEFIELD JW, HOAGLAND MB, AND LOFTFIELD RB. Mechanism of incorporation of labeled amino acids into
proteins. J Cell Comp Physiol 47 Suppl 1: 81–101, 1956.
82 • JANUARY 2002 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.33.5 on July 4, 2017
86. ROTHMAN SS AND BROOKS FP. Electrolyte secretion from rabbit
pancreas in vitro. Am J Physiol 208: 1171–1176, 1965.
87. ROTHMAN SS, GOETZE H, AND LIEBOW C. Enteropancreatic circulation
of digestive enzyme. Nature 250: 78 –79, 1976.
88. ROTHMAN SS AND GRENDELL JH. Is there an enteropancreatic circulation of digestive enzyme? Am J Physiol Gastrointest Liver
Physiol 244: G101–G102, 1983.
89. ROTHMAN SS AND GRENDELL JH. The case of the disappearing digestive enzymes. Gastroenterology 85: 1438 –1442, 1983.
90. ROTHMAN SS AND GRENDELL JH. On the endocrine secretion of pancreatic digestive enzymes. Gastroenterology 87: 986 –987, 1983.
91. ROTHMAN SS AND LIEBOW C. The permeability of the zymogen granule membrane to digestive enzymes. Am J Physiol Gastrointest
Liver Physiol 248: G385–G392, 1985.
92. ROTHMAN SS AND MELESE T. Leaky cells of glandular epithelia. Int
Rev Cytol 112: 225–244, 1988.
93. RYAN J AND APPERT H. Circulatory turnover of pancreatic amylase.
Proc Soc Exp Biol Med 149: 921–925, 1975.
94. SCHEELE G AND ROHR G. Enteropancreatic circulation of digestive
enzymes. Gastroenterology 86: 778 –783, 1984.
95. SIMON O, MUNCHMEYER R, BERGNER H, ZEBROWSKA T, AND BURACZEWSKA L. Estimation of rate of protein synthesis by constant infusion
of labelled amino acids in pigs. Br J Nutr 40: 243–252, 1978.
96. SINGH H, PEPIN J, APPERT HE, PAIRENT FW, AND HOWARD JM. Amylase
and lipase secretion in the hepatic and intestinal lymph. Arch Surg
100: 253–256, 1970.
97. STIEFEL DJ AND KELLER DJ. Preparation and some properties of