AMERICAN JOURNAL

OF

PHYSIOLOGY

Vol. 228, No. 4, April 1975.

Printed in U.S.A.

Kinetics of Na’ uptake transit by the pancreas

and transcellular

M. ROSSIER AND S. S. ROTHMAN Department of Physiologic, University of Calijornia,

San Francisco,

ROSSIER, M., AND S. S. ROTHMAN. Kinetics of Na’ uptake and tramcellular transit by the pancreas. Am. J. Physiol. 228(4) : 1199-1205. 1975. -nNa uptake into strips of rabbit pancreas was measured for up to 10 min. The uptake curve was characterized by the presence of two plateaus separated by an inflexion point; a first “plateau” or an approximation of steady-state uptake was observed between I and 3 min; between 3 and 4 min the slope of the uptake curve increased again, finally decreasing to a new and higher steadystate uptake between 4 and 6 min. The data suggest that the first part of the uptake curve (from 0 to 3 min) represents uptake into most if not all cells, and the second part (from 3 to 10 min) repsteady-state cellular uptake and of the resents the sum of “quasi” equilibration of the ductal compartment in series with the cells. In this model a substantial delay (2.5-3.25 min) elapses between the filling of cellular and ductal compartments which is apparently of intracellular origin, implying restricted Na+ diffusion within the cytoplasm and an intracellular Na+ gradient. If this model is correct, then the mean transit time for Na+ across the whole organ should be approximately 34 min and be primarily the result of transcellular transit. The mean transit time for Na+ across the whole organ in vitro was measured and found to be 3.5 min on the average. The step that accounts for most of this time appears to be the transepithelial transit of Na+. electrolyte cytoplasmic

secretion; restricted

ion transport; membranes; rabbit diffusion; intracellular gradient

pancreas;

OF A MOLECULE from an external (bathing) medium (b) into most tissues (cells) (c), assuming essentially instantaneous equilibration of the extracellular space with the molecule being examined, can be most simply modeled as a two-compartment system, b + c, with tissue uptake of the molecule being described by a single exponential function: UPTAKE

Y = A(1

-

eakt)

-

ewkct) + A,-J(~ -

where c and d refer to cell and duct. This model reduces to a two-compartment

94143

A d (1 -

e-“dt) E 0 under the following conditions: I) when the species considered is not transported from c to d, i.e., not secreted, 2) when the ductal compartment is of negligible size, being of either little volume or containing the molecule of interest at low concentrations relative to the cell, and 3) when b and d are essentially the same compartment, i.e., equilibration of duct contents occurs directly and rapidly from the bath in parallel with uptake into cells. In the present study we measured the uptake of Na+ into pancreatic tissue, mainly pancreatic strips (14). In this preparation the subepithelial layer is relatively insubstantial, and the tissue is comprised mainly of secretory cells and ducts. Therefore, uptake might well satisfactorily fit a threecompartment model (b e c $ d). It is not likely that it would be reduced to or toward a two-compartment system where Ad(l - e-kdt) Z 0 because, relative to the conditions just discussed, 1) Na+ is secreted and in fact is the main cation in pancreatic juice, 2) ductal sodium concentration is approximately 5 times higher than intracellular concentration, and 3) in pancreatic strips (as opposed to slices) only a few cut ends of some interlobular ducts are exposed to the bath and, as a result, the duct contents are isolated from the bath by the cells and direct equilibration of the duct contents from the bath is negligible for the duration of the experiments (see METHODS and Fig. 51). Na+ uptake into these tissue strips did not conform to either the simple three-compartment or double-exponential model (equation 2) or a single-exponential model (equation I). The data indicate that a substantial delay exists between the filling of compartments c and d, tissue uptake therefore essentially being described as the sum of an exponential and a sigmoid function. The evidence that suggests this model and a discussion of its implications form the body of this paper.

(1)

where Y is tissue uptake of the molecule at time t, A is uptake at the steady state, and k is the rate constant. However, even a simple model for tissue uptake in exocrine secretory systems is more complicated because it involves ducts (d) as well as cells (c). If these two compartments (c and d) fill in series with each other, then uptake would be expressed as a three-compartment system, b ( c C d, in which tissue uptake Y is the sum of two exponential functions, one representing cellular and the other ductal equilibration: Y = A,(1

Caltyornia

ewkdt) solution

(2) when

METHODS

Tissue preparation. White male New Zealand rabbits (2-3 kg) were used. When strips or slices of tissue were prepared, the animals were sacrificed by air embolism. The tissue strips (20-80 mg wet wt (WW)) were separated from the whole pancreas and surrounding connective tissue by gentle forceps dissection and then preincubated for 30 min at 37°C in 25 ml of bicarbonate-buffered Krebs-Henseleit (KH) (9) solution containing 11 mM glucose and constantly gassed with 95 % 02-- 5 % COQ. The strips once separated have a tendency to coalesce. To prevent this it was necessary to spread them between two sides of a folded large-pore nylon mesh and then keep them in place by threading silk through

1199

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

M.

1200 the mesh. In some experiments the tissue was sliced with a razor blade in a chamber designed to produce slices not greater than 0.2 mm thick. The slices were otherwise treated in exactly the same manner as the strips. 22Na uptake (into sfr$s or slices). After preincubation 22Na (Cambridge Nuclear Corporation) was added to the bath (2.8 &i/mmol), and the strips (or slices) were removed at different time intervals, blotted, and weighed. z2Na activity in the bath and uptake into the tissue were measured by counting the gamma spectrum emission between 0.125 and 0.625 MeV. Na+ uptake was then calculated assuming ideal tracer kinetics (1) (22Na bath/Na+ bath = 22Na tissue/Na+ tissue), which appears to hold (see Table 3) at least in that at the steady state chemically and isotopically determined intracellular [Na+] are roughly equivalent. [ 14C]S~c~ose uptake. In some experiments the extracellular space (ECS) was measured using [l*C]sucrose (ICN Isotope and Nuclear Division) in order to correct total 22Na uptake for uptake into the ECS. Five microcuries of [l*C]sucrose (200400 mCi/mmol) were added to the medium at the beginning of the preincubation period 30 min before adding 2z\a. * Prelir n’n1 a r y experiments indicated that this time was adequate for [14C]sucrose equilibration in the tissue. In these tissue was also removed at different time inexperiments, tervals to measure 22Na uptake. After removal of the tissue from the bathing medium, it was weighed, dried at 100°C resuspended in 4.0 ml 0.1 N for 24 h, weighed again, HNO3 at 4”C, and shaken for another 24 h to extract the sucrose. Bath aliquots (0.1 ml) were also acidified with 3.0 ml 0.1 N HNO 3. A 1-ml sample of the acid solution of both tissue and bath fluid was solubilized in 1.0 ml of Bio-Solv (BBS-3, Beckman Instrument Company). Eighteen milliliters of toluene containing a mixture of fluors (Beckman fluoralloy TLA) was added. Samples were counted by liquid scintillation spectroscopy for [14C]sucrose uptake and corrected for overlap with 22Na beta and gamma emissions. EC’S measurements. The measurement of ECS with sucrose is the sum of interstitial fluid (ISF) plus any portion of ductal space (DS) that might contain sucrose. This latter component is probably negligible with tissue strips because: I) the cross-sectional area available for retrograde diffusion up the duct system in the strips is very small, only a few cut ends of ducts being exposed. 2) Secretory flow would act to 3) Even ignoring the opposing oppose such equilibration. effects of secretory flow, during the course of the experiment, sucrose would diffuse only a short distance, 2.6 mm, up essentially the whole duct system from the cut ends of the interlobular ducts. (A similar diffusion of Na+ would be accounted for by the sucrose space measurement, since in 10 min Na+ would diffuse a shorter distance, i.e., 2.4 mm.) 4) Our measurement of sucrose space represents a small ECS (Table 4) relative to many other tissues, although it is roughly equal to reported values for the albumin space of rabbit pancreas (19). It does not seem likely that this value also includes a substantial ductal space. 5) The ECS increases when the strips are sliced, which is consistent with slicing exposing previously inaccessible ducts to the tracer (see below). For these reasons it seems most likely that total ECS is accounted for primarily, if not exclusively, by the ISF in tissue strips. In unwashed tissue the Y intercept (or “zero time” Na+

ROSSIER

AND

S. S. ROTHMAN

uptake) was a good numerical estimate of the amount of 22Na in the ISF as independently determined by [i4C]sucrose (Table 4), indicating the rapid equilibration of Na+ within this space. The Y intercept was calculated as a linear fit using the method of least squares from the first four points on the curve (18). Ductal s-ace estimates. Strips were sliced in an attempt to estimate DS as an increase in the measured ECS. The use of this technique assumes that the increase in ECS due to slicing the strips is only the result of an increase in the accessibility of the ductal compartment to the marker and, as a corollary, that the marker is not distributed throughout the duct system in the unsliced material as discussed above. The thickness of a slice not being greater than 0.2 mm, the equilibration of sucrose should be complete within the ductal space at the end of the preincubation period. Slicing the tissue therefore should increase the steady-state sucrose uptake only if previously unaccessible ducts are exposed, unless slicing produces sufficient damage to cells so that a substantial number equilibrate rapidly with sucrose. Total tissue Na+ content. Tissue was weighed, dried, and its ion content equilibrated in 0.1 N HNO3 as described under [14C]sucrose uptake. The Na+ content of samples was then measured by flame-emission techniques. Intracellular Na+ concentration, [Na+]rc (meq kg-l of intracellular water), was calculated as follows: l

CNa 1IC

[total =

Naf + DS -.-.-Na+)] tissue Na+ - (ISF --~~ ___ 100 - (%ISF + %DS)

-

X--- 100-

Whole pancreas in vitro After cannulating the duct in situ, the pancreas was removed, mounted in a chamber (14, 15), and allowed to equilibrate for 1 h in 100 ml KH solution. 22Na (3 pCi/mmol) was then added to the bath, and its rate of appearance in juice was measured by touching the end of a polyethylene cannula to filter paper every 10 or 15 s (depending on the rate of secretion) for lo-20 min and by subsequently collecting 5- or 1 0-min samples of secretion for an additional 40-50 min. The delay in the appearance of the isotope in secretion due to a cannula “dead space” was calculated from the volume of the cannula and the flow rate and was substracted from the time scale in Fig. 4. Curvefitting. All values from the three series of experiments presented in RESULTS (see Uptake curve) were pooled after substracting their corresponding Y intercept (Fig. 1). The pooled weighted (as a function of the variance at individual points) data were analyzed using a BMD X 85 nonlinear least-squares program (3) on a IBM 360/75 computer. RESULTS

Uptake curve. Na+ uptake into pancreatic strips increased with time to reach steady state in approximately 4-6 min (Figs. 1 and 2). In addition, this curve appears to plateau between the 1st and 3rd min. That this observed form (including a “plateau”) (Figs. 1 and 2) reflects the true form of the function and is not a statistical sampling abnormality is supported by four related observations. First, two series of experiments performed 4 yr apart by two different investigators in the laboratory, with slightly different techniques, not only showed basically the same Na+ uptake pattern with plateaus, but also a striking similarity in absolute values

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

P.4NCREATIC

SODIUM

UPTAKE

1201 TABLE

Q r 5

E

1. Steady-state Na+ uptake into the jirst pool Time, min

1.5 2 2.5 3 4 /I

I

* Steady-state independently. tent.

-+ / @ 0

I 3

I 2

I

I 4

I 6 Time

I 8

I IO

(Minutes)

1. **Na uptake (peq* 100 mg - WW-l) vs. time in pancreatic Pooled data from 3 series of experiments are shown (see RESULTS, uptake curve). l % & SEM corrected for uptake into ISF by subtractFrom time-O, n = 8, 6, 18, 14, 13, 15, ing Y intercept (see METHODS). 20,19,8,6,5, 11,17,17,13. FIG.

strips.

im .-4)

6

5 sl-

I t

0;

0

I

2

3

4 Time

6

1

J

8

IO

(Minutes)

FIG. 2. **Na uptake (peq* 100 mg WW-l) vs. time in pancreatic strips, a comparison between 2 different series of experiments. l K done by SSR in 1969. From time 0, n = 8,6,6, 8,6,8, 7,6,2, 4, 4, 4, 4. l Z & SEM, done by MR in 1973. From time 0, n = 12, 6, 7, 7, 13, 13, 6, 6, 5, 7, 12, 13, 7. (For details see RESULTS, uptake curve).

when the two series of experiments were equated for ISF (Fig. 2) (The major difference in technique between the 1969 and the 1973 experiments was that in the earlier experiments tissue was washed twice for 5 s in ice-cold KH solution and then blotted, whereas in the more recent series the tissue was only blotted. (Y intercept = 0.45 peg* 100 mg WW-l in the former and 2.07 peq* 100 mg WW-l in the latter experiments.)) Second, in a third series of experiments in which 22Na and [14C]sucrose uptake were measured simultaneously on the same piece of tissue in order to examine the possibility that the plateau was only apparent, viz., the result of some unspecified time-dependent variation in ISF, still no additional Na+ uptake was observed between 1.5 and 3 min (Table 1). Third, in slices prepared to estimate DS as described in METHODS, Na+ uptake measured at selected points (2 and 8 min) approached steady-state distribution more rapidly than in strips (Fig. 5, I and 11) (at 2 min, strips were at 52 & 9 % (SE, n = 5) of steady state, (0.01 > P > slices at 88 A 9 % (SE, n = 6) of steady-state 0.005, one-tailed t test)). That is, simply altering the geometry of the system (by slicing) led to the disappearance of the plateau. Finally, we have reported (14) technique-

ISF estimated with Cell weight includes

Na+ Uptake h 1 SEM (n = 6), j.4eq-100 mg “Cell Wt”-I*

1.97 2.06 1.94 1.58 [l*C]sucrose both cellular

+ & & ~fi

0.24 0.39 0.62 0.41 for each sample and ductal con-

independent alterations in the uptake curve previously under other environmental circumstances (e.g., loss of the plateau (metabolic inhibition) or changed form of the plateau (secretin stimulation)). Curvejitting. Curve fitting was employed to better demonstrate the existence of the plateau and the deviation of the actual curve from either a single exponential function (equation I) or the sum of two exponential functions with approximately 0, 0 intercepts (equation 2). The weighted data were fit to both these functions (for equation I, A = 4.35 rt .34 peq.100 mg WW+ (&SD) and k = 0.0059 A= .0012 s-? (&SD); and for equation 2, A, = 0.5426 rt .707, Ad = 4.0046 & .5171 PeqlOO mg WW-+ (MD), and k, = 0.0544 A .0115, kd = 0.0045 =f= .0017 s-l (*SD)), and while the data give a relatively satisfactory fit, the deviations of the actual data from these functions indicate that the “goodness of fit” alone is an inadequate criterion for determining the best compartmental model. For example, the first seven data points (O-90 s) from the observed uptake curve are systematically below the estimated equation I function. If the odds are even that the points would be either below or above the calculated function, the probability of this happening by chance is 1 in 125 trials. Furthermore, the next four experimental points (120-210 s) were all above the calculated equation 1 function. A similar comparison with equation 2 also shows a systematic maldistribution of the data points below and above the estimated function, although to a predictably lesser degree. The differences between the observed Na+ uptake curve and the theoretical curves obtained by fitting the same data to the functions of equations I and 2 are stressed in Fig. 3 in which the slopes of successive linear functions derived from the uptake curve are plotted for a series of overlapping intervals. The actual series of slopes (closed circles) does not decrease progressively to a single steady state (b = 0) as a similar plot for the data fit to equation I (open circles) or to equation 2 (triangles) requires. Rather the slope b has an inflection point, that is b decreases to a first minimum, then significantly increases and finally decreases, reaching a new minimum: the 90- to is about 2.5 240-s interval slope (b = 11.63 peq*~-l*lO-~) greater than the 45- to 180-s interval slope (b = 4.78 I.ceqs-l. 10e3) (0.01 > P > 0.005, one-tailed t test). Furthermore, the changes in the slope of the measured curve before and after the inflection point (see Fig. 3) are significantly different from the changes in slope of the fitted curves over this series of intervals: four of four points for equation I and three of four for equation 2 were significantly different (Table 2). Whole pancreas in vitro. The appearance of 22Na in secretion and its approach to steady-state specific activity was

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

1202

M.

ROSSIER

AND

S. S. ROTHMAN (3) 650 mg/ hr

8

/

) 550 mg/hr

1.6

3. Changes in slope of a sequence of linear regressions for overlapping series of time intervals derived from uptake curve. Changes in slope derived from actual uptake curve (a) and from curves obtained fitting weighted data of Fig. 1 to Eq. I (Y = A( 1 - e-“t)) (0) and to Eq. 2, (Y = A,( 1 - emkct) + A& 1 - eskdt) (A) are compared. Plot of slopes derived from actual data (a) shows inflection points at 45- to 180- and 90- to 240-s intervals : b corresponding to 90- to 240-s interval than b corresponding to (b = 11.63 peqs -lo 10-a) is 2.5 times greater 45. to 180-s interval (b = 4.78 peq*s+ 10-a) (one-tailed t test: 0.01 > P > 0.005). For 4 time intervals (including lo-90 to lo-90 to 60-210 s), points calculated from actual data (0) were significantly different from values obtained fitting data to Eq. I, and 3 of these intervals were significantly different from values obtained fitting data to Eq. 2. See Table 2 for probability values. FIG.

2. t Values comparing slopes derived from actual uptake curve ( l > and curves obtained by jtting the data to equation I (0 > and equation 2 (A) as shown in Fig. 3

TABLE

Eq 2 (0):Y

Int enal

t 4.52 -3.19 -6.85 -8.22 -4.32

045 15-120 30-150 45-180 60-210 l

One-tailed

= A (1 - 8’)

>p>* 0.01-0.005 0.025-0.00125 0.0025-0.0005 0.0025-0.0005 0.01-0.005

t 1.78 -1.84 -3.76 -6.16 -3.01

>p>* 0.1-0.05 0.1-0.05 0.01-0.005 0.0025-0.0005 0.025-0.0125

test.

followed in secretion from the whole pancreas in vitro. After subtracting the time delay due to cannula dead space, an absolute delay of 30-90 s elapsed before 22Na appeared in the juice. 22Na started to app ear in substantial amounts steady state in - 6-10 min. The after - 3 min, reaching mean transit time for Na+ movement across the whole pancreas can be determined by extrapolation of the cumulative steady-state output through the time axis. Figure 4 shows three different experiments with spontaneously different basal secretory rates. In nine preparations, the mean value of the mean transit time was 3 min 30 s with a relatively small 18 s standard error of the mean. Ductal space appears to be an insignificant element in the determination of transit time, because the mean transit time was not found to be

3.7

6 8 Time(minutes)

IO

12

14

FIG. 4. Cumulative 22Na output in juice from time 0 (corrected for bath specific activity) after addition of isotope to bath. Delay due to cannula dead space was subtracted from time axis and does not appear on this graph. Mean transit time is shown for 3 preparations spontaneously secreting at 85, 550, and 650 mg of fluid per hour. Time intercepts : 1.6 rnin (1) and 3.7 min (2, 3). Mean transit time : 3 min 30 ZIZ 18 s (A SEM, n = 9).

inversely correlated to the secretory rate over a wide range of flow from 85 to 450 mg/h. Conversely, transit time was directly correlated to secretory rate over this range (b = 0.79/min per 100 mg*h-l, r = 0.9029 and 0.01 > P > 0.001). DISCUSSION

Neither a two-compartment construct (equation I), which provides the simplest description of the uptake curve, nor a three-compartment model (equation Z), which gives a better phenomenological description by including the duct COIIIpartment, is able to satisfactorily account for the observations because of the presence of a plateau (between 1 and 3 min) in the uptake curve. The observed form of this curve can be accounted for by including a “delay” between the filling of compartments c and d in the model so that tissue uptake would be the sum of an exponential and a sigmoid function as shown in Fig. 51 where b +

c (

(delay)

(

d

The delay can also be modeled in an analogous a multicompartmental system in which

manner

as

In this case, uptake into the tissue would be the sum of n + 1 exponential functions, each function following the previous one with a delay approaching, but not equal to zero, leading to a progressively greater time intercept. Nature of compartments. In our discussion thus far, we have assumed that c is a cellular compartment and d is the ductal compartment and therefore that the delay is related to their sequential equilibration. In all likelihood the first assumption is true, that is, the first part of the uptake curve (from

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

PANCREATIC

SODIUM

1203

UPTAKE

3. Intracellular and ductal bra+ concentration as estimated from isotopic or chemical measurements

TABLE

Intracellular

Cell

Duct

[Na+]

CNa+l

and Ductal

Chemical

23 (4

33 (b*) 24 w

226

Cc*)

140 (et) Time

Time

sLi$%e

‘;I

100 0

5o

2

8

h

1 2

8

uptake patterns for =Na. b: bath compartment and = interstitial fluid, i.e., extracellular space without ductal space). c: cell cornpartmerit. d: duct compartment. In strips (Z), cut ends of ducts are relatively inaccessible to retrograde diffusion of isotope from bath (see METHODS and DISCUSSION), and filling of duct contents (d) would occur via cell compartment (c). In this case, tissue uptake can be modeled as a three-compartment system b it c FI d. As discussed in the text, a delay appears to occur between filling of c and d, so that tissue uptake becomes sum of an exponential and sigmoid function as shown in I. In tissue slices (ZZ) equilibration of duct compartment (d) occurs not only through cells (c), but also directly from bath (b) through cut ends of sliced ducts with a minimal delay. In this case tissue uptake can be modeled as a two-compartment system b F! c since d and b become a single course

5. Comparison

compartment

between

and

should

152 (d) .

(a) Derived from the mean value of the first plateau (Fig. 1). (b) Derived from the tissue sodium content : 61.0 f 1.8 meq. kg WW-’ (n = 7). (c) Derived from the mean value of the second plateau minus the mean value of the first plateau (Fig. 1). (d) From S. S. Rothman and F. P. Brooks (15). * Corrected for total ECS (minimal estimate) : B -I- C (Table 4) = 25%. t Corrected for total ECS (maximal estimate) : B + D (Table4) = 30%.

Time (Minutes)

Time h4inutes) FIG.

Na+ Concns, meq.kg-1

Isotopic

compartmental

models ISF (ISF

approximate

a single

and

exponential

(Q. I). Uptake curves are shown only to demonstrate different kinetec patterns and do not have any quantitative meaning. Histogram presents experimental data for strips and slices equilibration followed predicted pattern: at 2 min in uptake was 52’% of the steady-state 8-min 2aNa uptake n = 5) while it was 88% in slices (II) (SEM: =!G 9’%, P > 0.005, one-tailed t test).

and shows that strips (I), 22Na (SEM: f 9%, n = 6) (0.01 >

0 to 3 .min) represents uptake mainly into an intracellular pool (b & c), since uptake into the ISF was accounted for independently., (There may be a very small component of uptake into the second pool during this time, depending on the slope of the sigmoid curve.) Whether this first pool includes all cells is another question. If uptake from 0 to 3 min was into all or most cells, then the intracellular Na+ concentration derived from the first .plateau should be the same as values obtained from chemical measurements for intracellular Na+ concentration, provided that the pool of exchangeable Na+ is approximately equal to the total. The intracellular concentration estimated from isotopic measurements was 23 meq/liter, whereas our estimated chemical measurement was 33 meq/liter (Table 3), which may indicate that either some intracellular Na+ is not rapidly exchanging or that a cell type(s) (at most one-third of the total) had not equilibrated with Na+ during this time or both. Because we are only able to estimate the actual DS size, the calculated differences in the two measurements may be more apparent than real, e.g., if instead of our average DS we use the upper quartile value of DS, the two measurements come to about the same value (23 and 24 meq/liter). The last part of the curve, from 3 to 10 min, appears to

be the sum of the apparent steady-state intracellular uptake (b F? c) and uptake into another pool, probably in series with the cells (c) as we have suggested and occurring with a substantial apparent delay. If we consider uptake into the second pool as reflecting parallel uptake inlo another pool instead, such as ductal filling occurring directly from the bath (b @ d), then the equilibration without substantial delay of the duct contents with 22Na via the cells (c $ d) would act as a shunt obliterating a delay due to b F! d equilibration. If a cellular parallel pool is considered, the delay would require that entrance into such pool be remarkably restricted relative to Na+, e.g., a plasma membrane almost impermeable to Na+. It appears therefore most likely that the second pool does fill in series with most or perhaps even all cells. Besides the obvious series compartment, the ducts, this second pool could be Another cell type in series or an intracellular compartment(s) in series with the first cellular pool. This would change little relative to the interpretation of the plateau, since the delay still exists and must result from restricted access (cytoplasmic delay) to the series compartment. If isotopic and chemical intracellular Na+ concentrations are actually the same and the first pool represents uptake into all cells, another substantial Na+-containing pool, cellular or subcellular, of course, would not be possible. While there still may be some indeterminacy regarding the nature of the series compartment, the following two observations support the assumption that it is the duct system. In the first place, if we assume that the second pool represents ductal equilibration in series with the cells (c & d), then the isotopic Na+ concentration derived from the second plateau uptake should be of the same order as the chemical ductal concentration as found in secretion. This was the case (Table 3). Second, if the duct is the second compartment, by slicing the tissue and producing rapid equilibration of the ductal compartment with the bath, the system should be reduced toward two compartments (equation 2 E equation I). This was found (see Fig. 511). Delay and its implications. While the delay implies similar events, regardless of the nature of the second or series compartment, the evidence just cited plus the implications of the mean transit time (see below) suggest that it is in fact the duct system, and for this reason we will interpret the data assuming a three-compartment model: bath, cell, and duct.

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

1204 TABLE

M. ROSSIER 4.

Interstitial

Ertracdll$aS;

Interstitial space Ductal

and ductal space estimates

Spaces

Group

(ISF) space*

(W

ured

* Ductal space = ECS in strips), determined

Means of Computation

% Total Tissue Vol

S?&

It

A B

Y intercept [14C]Sucrose

14.5 16.5

1.0 0.9

32 42

C D

All values Upper quartile

8.6 13.8

0.8 0.6

40 10

(as measured by steady-state

in slices) [l*C]sucrose

ISI? (as measuptake.

In this light, the kinetics of Na+ uptake indicates that there is a delay between filling of the cellular and ductal compartments. It appears that only a substantial delay during which a quasi steady-state is achieved in the cell prior to any significant exit across the duct facing plasma membrane can produce the observed uptake pattern. The apparent delay has a duration of 2.5-3.25 min and is most likely to be of intracellular origin unless we assume that the transit time for Na+ across the ductal plasma membrane is of the order of minutes, which does not seem reasonable, particularly for a Na+-secreting system. (Transit times of the order of milliseconds, not minutes, would be likely even of a relatively Na+ impermeable cell because of the thickness of the cell membrane.) A substantial delay of intracellular origin is not consistent with the view that the most time-consuming step that determines the transit time of a molecule across a cell is its movement across the cell membrane and that the intracellular milieu can be considered a well-mixed compartment in which at least small molecules can diffuse as freely as in water. This view has been challenged as a general hypothesis by a few authors (12, 20). The presence of the delay, in conjunction with the existence of the first plateau, has the must be restricted within following implications : I) diffusion the cytoplasm. Calculations of the “diffusion coefficient” for Na+, assuming a path length of 10+&O pm (10 pm E duct indicate that cell thickness, 40 pm ‘1! acinar cell thickness), diffusion would be restricted by 2-4 orders of magnitude Na+ (Dz$ at 37°C = 1.7 X 10-b cm2 Dcarc = 0.14 - 2.2 X 1 O-8 cm2 s-l). This can be compared to the diffusion coefficient for Na+ in nerve fiber (6) (almost the same as in free solution) and skeletal muscle (5, 11) (reduced maximally by a factor of 2). 2) There must be incomplete mixing of Naf within the cytoplasm, i.e., most of the intracellular Na+ must be located toward the basal side of the cell. Such a possibility raises the question of how an intracellular Na+ gradient would be generated and why. Perhaps it may be related to an osmotic gradient, providing the driving force for flow in secretory systems much as intercellular osmotic gradients (2) have been suggested for absorptive systems. 3) Finally, intercellular shunt pathways are not likely to be a major route for Na+ movement across the pancreas. A tissue with large shunts would approximate a two-compartment model as shown for the slice preparation (Fig. 511). That is, if substantial equilibration occurred through an intercellular shunt (i.e., from ISF to duct), then the delay between filling of the first (cellular) and second (ductal) compartment would be obliterated by equilibration via the shunt flux. l

l

s-l;

AND

S. S. ROTHMAN

Mean transit time. An independent validation of the existence of an intracellular delay and as a derivative, the identity of the compartments (cell and duct) lies in experiments that measured the mean transit time for Na+ movement across the whole pancreas. Na+ movement across an epithelium often involves a complex series of epithelial layers so that there may be some question as to the main reason for a transit time of the order of minutes. In the rabbit pancress there are three potential sites within the tissue that can contribute to the delay: the subepithelial connective tissue layer, the ducts, and the secretory cells themselves. The connective tissue should not be a major barrier since it is relatively sparse in pancreas, and in any event 22Na equilibration within the ISF occurs very rapidly, [WJsucrose space coinciding well with the Y intercept, i.e., approximately zero time equilibration (Table 4). Movement down the ducts might represent a significant dead space, which could account for a transit time of the order of 3 min. However, for secretory rates from 85 to 450 mg/h, the mean transit time did not decrease with increased secretory rates, which indicates that the ducts apparently reach equilibrium at approximately the same time throughout much of their length over this broad range of secretory flow. Thus the most time-consuming process determining the rate of appearance of Na+ in secretion is in all likelihood Na+ movement across a single layer of epithelial cells separating the bath and the duct lumen It therefore seems likelv that the mean transit time across the whole organ and the dela Yed filling of presumably the duct compartment observed in the uptake curve are probably two independent reflections of the same phenomenon. If this is so, then the mean transit time should fall during the filling of the second or ductal pool on the uptake curve, i.e., between 3 and 4 min (Figs. 1 and 2). In fact, the mean transit time, 3.5 min, and the half-equilibration time of the second pool, 3.5 min, coincide almost exactly. This indicates that the second pool is in fact the ductal compartment and that the delay in the uptake curve does indeed reflect the transcellular transit of Na+. Cellular origin of delayed Na+ movement. We said that the kinetics of Naf uptake in strips imply restricted diffusion and incomplete mixing within the cytoplasm. A possible morphological basis for these two phenomena is found in the unusually large amount of rough endoplasrnic reticulum and high degree of intracellular inhomogeneity of the acinar cell (4). However, the delay might also be a more general event, common to other epithelia. This is suggested by the fact that the mean transit time for Na+ across the amphibian skin and bladder is 2-5 min (7) and 7-8 min across the small intestine (17), values of the same order as our present observation for the pancreas (3.5 min). For the intestine Schultz and Zalusky (17) suggested that the submucosal layer may be the dominant factor causing the delay, whereas Hoshiko and Ussing (7) suggested that the absorptive cell is the major barrier for amphibian skin and bladder. The latter investigators based their argument on the fact that the “build-up of the labeled flux follows a single exponential course, ” and furthermore, that frog skin and toad bladder with different connective tissue layer thickness have approximately the same mean transit time. Since in our case a delay of similar magnitude occurred apparently across a single layer of secretory cells, it may be inferred that the mean l

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

PANCREATIC

SODIUM

UPTAKE

transit time of these other systems may also in good part reflect a delay of transcellular origin across the epithelial barrier rather than diffusional delays through the connective tissue, although they may also occur in some tissues. If this were true, restricted intracellular Na+ movement would have to be the result of some common property of the cytoplasm of these cells, such as an intracellular protein lattice, a polyelectrolyte gel (8, 10, 13), whereby Na+ movement across the cells would be restricted by multiple binding interactions along the route. The fact that intracellular Na+ activity may &ughly be equivalent to its concentration does not preclude such binding, since ions bound by polyelectrolytes are genera lly in an electrochemically active state (8, 10, 13). Whether or not the delay is uniquely the result of the complex structure of the pancreatic acinar cell or indi-

1205 cates an ordered intracellular remains to be determined.

milieu

(12, 20) for many

cells

The authors thank Dr. V. Licko for his assistance with computer techniques. This study was supported by a research grant from the Academic Senate of the University of California. M. Rossier was supported in part by the American Swiss Foundation for Scientific Exchange and by the Fonds National Suisse. A preliminary report of this work has appeared (Federation Proc. 33: 1252, 1974). Address reprint requests to: Professor S. S. Rothman, Dept. of Physiology, University of California School of Medicine, HSE740, San Francisco, Calif. 94 143. Received

for publication

10 April

1974.

REFERENCES 1. CIJRRAN, P. F., A. E. TAYLOR, AND A. K. SOLOMON. Tracer diffusion and unidirectional fluxes. Biophys. J. 7 : 879-90 1, 1967. 2. DIAMOND, J. M., AND J. M. TORMEY. Role of long extracellular channels in fluid transport across epithelia. Nature 210: 817-820, 1966. 3. DIXON, Mr. G. BMD Biomedical Computer Programs X-Series Supplement. Berkeley : Univ. of California Press, 1969. 4. FAWCETT, D. Mr. An Atlas of Fine Structure. The Cell. Philadelphia: Saunders, 1966, p. 147. 5. HARRIS, E. J. Ionophoresis along frog muscle. J. Physiol., London 124 : 248-253, 1954. 6. HODGKIN, A. L., AND R. D. KEYNES. Experiments on the injection of substances into squid giant axons by means of a microsyringe. J. Physioi., London 13 I : 592-6 16, 1956. 7. HOSHIKO, T., AND H. H. USSINC. The kinetics of Na24 flux across amphibian skin and bladder. Acta Physiol. Stand. 49 : 74-81, 1960. 8. KATCHALSKY, A. Polyelectrolyte gels. Progr. Biophys. Biophys. Chem. 4: l-59, 1954. 9. KREBS, H. A., AND K. HENSELEIT. Untersuchungen iiber die Harnstoffbildung im Tierkiirper. 2. Physiol. Chem. 210: 33-36, 1932. 10. KURELLA, G. A. Polyelectrolytic properties of protoplasm and the character of resting potentials. Symp. Membrane Transport Metab., Prague, 1960, p. 54.

11. KUSHMERICK, M. J., AND R. J. PODOLSKY. Ionic mobility in muscle cells. Science 166 : 1297-1298, 1969. 12. LING, G. N. A Physical Theory of the Living State: the Association-lnduction Hypothesis. New York : Blaisdell, 1962. 13. RICE, S. A., AND M. NAGASAWA. Polyelectrolyte Solutions. New York : Academic, 1961, p. 391. 14. ROTHMAN, S. S. The secretion of water and electrolytes by the pancreas. In: The Exocrine Pancreas, edited by I. T. Beck and D. G. Sinclair. London : Churchill, 197 1, p. 47. 15. ROTHMAN, S. S., AND F. P. BROOKS. Electrolyte secretion from rabbit pancreas in vitro. Am. J. Physiol. 208 : 117 l- 1176, 1965. 16. ROTHMAN, S. S., AND F. P. BROOKS. Pancreatic secretion in vitro in “Cl--free”, “CO2-free”, and low Na+ environment. Am. J. Physiol. 209 : 790-796, 1965. 17. SCHULTZ, S. G., AND R. ZALUSKY. Ion transport in isolated rabbit ileum. J. Gen. Physiol. 47 : 567-587, 1964. 18. SNEDECOR, G. W., AND W. G. COCHRAN. Statistical Methods. Ames, Iowa: Iowa State Univ. Press, 1956, p. 125. 19. SWANSON, C. H., AND A. K. SOLOMON. A micropuncture investigation of the whole tissue mechanism of electrolyte secretion by the in vitro rabbit pancreas. J. Gen. Physiol. 62 : 407-429, 1973. 20. TROSHIN, A. S. Sorption properties of protoplasm and their role in cell permeability. Symp. Membrane Transport Metab., Prague, 1960, p. 45.

Downloaded from www.physiology.org/journal/ajplegacy by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.