# Increase in zymogen granule volume accounts for increase in volume density during prenatal development of pancreas.

код для вставкиСкачатьTHE ANATOMICAL RECORD 207:487-501(1983) Increase in Zymogen Granule Volume Accounts for Increase in Volume Density During Prenatal Development of Pancreas THOMAS H. ERMAK AND S.S. ROTHMAN Laboratory of Cellular Dynamics, Department of Physiology, School of Medicine, University of California, San Francisco, CA 94143 ABSTRACT The sudden increase in volume density of zymogen granules in acinar cells of the fetal rat pancreas was examined with particular attention to the respective roles of granule size and number in this event. Volume density increased some twelvefold, from about 3%of cytoplasmic volume a t 17 days to about 45%at 20 days, following a sigmoidal pattern in which the greatest rate of increase occurred during day 18. This increase in volume density was primarily the result of a n increase in granule volume. Zymogen granule diameter increased from 0.55 pm at 17 days to 1.20 pm at 20 days, a n order of magnitude increase in average granule volume. The total number of granules in the tissue increased in proportion to the increase in organ weight (cell number and size), but changes in the number of granules per unit cytoplasmic volume were minor (+ 40%) in comparison to the increase in volume density. The distribution of granule diameter was roughly normal and unimodal a t each time interval, and the increase in average diameter over time was marked by a n increase in the upper limit of the size distribution and a n increased percentage of large granules. The size of condensing vacuoles also increased during this period, and their distributions were roughly coextensive with those seen for zymogen granules a t the same time. The potential origins of changes in granule size are discussed, as well as the important effect that size has on the number of granules observed in “two-dimensional” tissue sections viewed in the electron microscope. If size is not considered in our estimates, then we underestimate the numerical density in cells with small granules compared to those with large granules. The results indicate the central role of granule size, as opposed to number, in determining granule volume density in the embryonic pancreas. The zymogen granule content of the pancreatic acinar cell can potentially be changed by varying either the size or number of granules, or both. Until recently, however, changes in the volume density of secretion granules, i.e., the percentage of cytoplasmic or total cell volume occupied by granules, have been thought to occur primarily as a result of changes in the number of granules (Palade, 1959, 1975; Jamieson and Palade, 1977). This is because granules have been considered to be of constant size once formed and because digestive enzymes have been thought to be secreted by the loss of granules by fusion of granule membrane with the apical cell membrane (Palade, 1959).Within this context, changes in granule volume density have been conceived of in terms of changes 0 1983 ALAN R. LISS, INC. in number as a result of the formation of new granules or loss of old granules through exocytosis. However, recent studies indicate that a major portion of the alterations in granule volume density seen during active secretion can be due to changes in the size of granules (Ermak and Rothman, 1980, 19811, and, moreover, that these changes involve fluctuations in the size of individual, formed zymogen granules. These size changes are significant in light of physiological evidence that pancreatic proteins may not be secreted by exocytosis but by a diffusion-based mechanism involving transport of enzymes across the zymogen granule membrane (Rothman, 1975, 1980). Received August 2, 1982; accepted May 25, 1983. 488 T.H. ERMAK AND S.S. ROTHMAN In this article, we examine the respective roles of granule size and number as volume density increases for the first time during prenatal development of the rat pancreas. This period is unique in that the rate of enzyme accumulation far surpasses the rate of enzyme secretion (Doyle and Jamieson, 1978; Larose and Morisset, 1977; Werlin and Grand, 1979). During this time, the concentration of digestive enzymes in the tissue increases over a thousandfold (Rutter et al., 1968; Sanders and Rutter, 19741, and zymogen granules accumulate within the tissue and increase greatly in size (Parsa et al., 1969; Pictet et al., 1972). This situation produces the largest zymogen granules normally found in the gland, such that granules in the newborn rat average about six times the volume of granules seen in fasted adult animals (Ermak and Rothman, 1980). The volume density, size distribution, and number of zymogen granules in the tissue were determined by stereology and quantitative electron microscopy over a 4-day period of rapid granule accumulation (17-20 days of gestation). This time interval immediately follows the first appearance of granules in cells (Pictet et al., 1972)when enzyme specific activity increases most rapidly (Sanders and Rutter, 1974). During this period, granule size increased to higher mean values by a shift in a roughly normal distribution, and average granule volume increased by about one order of magnitude. Granule number increased in proportion to the increase in tissue mass, but despite the dramatic visual appearance of increasing granule number in electron micrographs, increases in granule size accounted for virtually the entire measured increase in volume density. bryos were chosen at 17 days, four a t 18 days, and three at 19 and 20 days. More animals were sampled at early time intervals because some cells at these times lacked zymogen granules altogether. The whole pancreas of embryos from other litters was extirpated and then weighed on a n electrobalance to measure organ mass. Electron Microscopy Tissue samples were fixed for 1.5 hr in 1.5% glutaraldehyde and 1%formaldehyde in 0.08 M cacodylate buffer (pH 7.2), postfixed for 1.5 h r in 1% OsO4 in the same buffer, stained en bloc with uranyl acetate, embedded in English araldite, and viewed in a JEM-100B electron microscope as previously described (Ermak and Rothman, 1981). Stereology Volume density was determined by point counting analysis according to Weibel(1969) using a transparent overlay with 1-cm spacings. The volume density (VV) of zymogen granules and condensing vacuoles was determined in relation to cytoplasmic volume (total cell volume minus nuclear volume), i.e., VV =VzgNcyt.The volume density of the cytoplasm was determined in relation to total The method of samcell volume, V, tNcell. pling was standiardized by using only acinar cells sectioned through the cell apex and nucleus (a sagittal section). Such a procedure is commonly used to examine changes of state for polarized epithelial cells where random sampling cannot be performed without a n extremely large sample size (Ermak and Rothman, 1980, 1981; Geuze and Kramer, 1974; Helander, 1978). Sections were photographed at X4,OOO and printed at x 10,000. In order to account for acinar cells without granules at 17 and 18 days, samples were MATERIALS AND METHODS further standardized by selecting only those Animal Procedures sagittally sectioned cells in the upper left Fetal rats were obtained from mated fe- corner of grid squares (Weibel, 1969). Two male Sprague-Dawley (Simonsen) rats on 17, tissue blocks per animal (13-18 electron mi18, 19, and 20 days of embryonic develop- crographs per animal) were examined at 17 ment. Day 0 was defined as the day a vaginal and 18 days, and one block per animal (10 plug was detected. Females were sacrificed electron micrographs per animal) a t 19 and by spinal section (after light etherization). 20 days. Some electron micrographs containThe body cavity and uterus were opened and ing two cells were included. individual embryos placed in a petri dish The number of granules per cell profile (N,) containing Earle’s balanced salt solution. was determined by counting the total numPancreatic rudiments from randomly chosen ber of granules in the plane of section embryos were dissected with the aid of a through a single ceII. The numericaI profile binocular microscope, and a portion of the density (NA; number per 100 test points or pancreas adjacent to the stomach and spleen approximately 100 pm2 of surface area was fixed for electron microscopy. Five em- counted at X10,OOO) was determined by di- 489 ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS viding the number of granules per cell profile by the number of test points over the cytoplasm times 100. The numerical density, or number of granules per unit volume of cytoplasm (Nv), was determined from NA by two independent methods. First: where NA and Vv are as defined above, 1.38 is a shape constant for spheres (&) , and K is a proportionality constant that depends upon the size distribution of the granules (Weibel, 1963; Weibel and Gomez, 1962).The larger the standard deviation of the mean size, expressed as a percent, the larger K is. For each time interval during the prenatal period, K was determined from the size distribution using the equation 812 K=(g) 21, where D1 is the arithmetic mean diameter (D) and D3 is the cube root of the third moment about the origin of the distribution. D3 was estimated from each size distribution by the following equation: D3 = + (d213 + . . . n + (dn)3)111’ where dl through d, equal the individual granule diameters in the distribution and n equals the total number of granules in the population (Williams, 1977). The same value for K (K = 1.1)was used for all time inter2% from vals, because K differed by only population to population, and because the standard deviation was a constant proportion of the mean diameter at all times prenatally (Fig. 9). Second: * N v =NA 3 , sectioned near the equator, i.e., the granule membrane was nearly vertical to the plane of section. Thus, granules sectioned through the poles (or granules with “halos”) were excluded. This method is particularly convenient when comparing different physiological states and has given results similar t o those obtained by other techniques (for a discussion of the method, see Ermak and Rothman, 1980, 1981). Some granules at 19 and 20 days were elliptical or had angular outLines; they were measured by taking the average value of the major and minor axes [(a b)/2)]. For 19 and 20 days, the same electron micrographs were used for measuring granule diameter and volume density (~10,000).For 17 and 18 days, granule diameter was measured on different micrographs taken at higher magnification, because their diameter was smaller. At 17 days, micrographs were taken at x 10,000 and magnified to ~25,000,and at 18 days they were taken at ~ 7 , 0 0 0 and magnified to x 17,500. Fifteen electron micrographs were examined from each animal at these two time intervals, and granules from adjacent cells were measured in order to increase the sample size of granules. The precise magnification of each set of micrographs was determined with the aid of a carbon calibration grating. + Analysis of Data Data were collected separately from each animal and mean values combined to give an average value for the time interval the standard error (where N = the number of animals). For zymogen granule size, data from all animals in a time period were combined to give a mean value for the whole population f the standard deviation of the mean. The size distributions shown in Figures 8 and 12 include data from all animals. When necessary, the ratio of experimental variances was determined comparing intravs interanimal variance (F test; Snedecor and Cochran, 1980).In general, the observed variance for N = the number of cells (or electron micrographs) was not different from the observed variance for N = the number of animals. However, all standard errors are given as N = the number of animals. + where is the mean diameter of the population (DeHoff and Rhines, 1968). These two methods have given similar results under other circumstances (for a comparison of RESULTS these methods, see Mayhew, 1972) and do Organ Growth and Cell Structure here as well. At 17 days of gestation, the embryonic panThe diameter of zymogen granules was determined (to the nearest 0.5 mm on electron creas was transparent, pink, and weighed micrographs) by measuring only granules about 1.0 mg. During the following 3 days, T.H. ERMAK AND S.S. ROTHMAN 490 contained numerous zymogen granules, and by 20 days acinar cells were packed full of granules (Fig. 51, giving a visual appearance similar to that seen in the newborn (Ermak and Rothman, 1980). 16 14 -z IS 2 W 3 12 10 8 z S LZ 6 0 4 2 , / I(7) u 18 20 17 19 TI M E (days postcoitum 1 Fig. 1. Increase in weight of the embryonic pancreas from 17 to 20 days of gestation. Data are mean values f SEM. The number of animals included for each point is given in parentheses. the tissue became opaque and white in color, and organ weight increased exponentially, roughly doubling every day (Fig. 1).By 20 days, the pancreas was a rich white, similar to that seen in the newborn, and weighed a n average of 15 mg. At all time intervals, the gland was composed of well-formed acini, although a t 17 days and, t o a lesser extent, 18 days, sections of many cells and several acini lacked zymogen granules altogether (Fig. 2). At 17 days, about 35%of the acinar cells had no granules at all, and by 18 days that value had decreased to about 13%(Table 1). Zymogen granule content increased on each day. At 17 days, several small granules were seen adjacent to the apical cell membrane (Fig. 3). A few granules were occasionally observed near the Golgi complex or in the basolateral region of the cells. By 18 days of gestation, granules occupied a larger proportion of the cell apex (Fig. 4) and, in addition, were observed more frequently in the basolateral region. By 19 days, all cells surveyed Morphometric Analysis The size of the pancreatic acinar cell increased over twofold during the late prenatal period as estimated from the number of test points per cell (Table 1; Figs. 2-5). This was primarily due to a n increase in the nonnuclear cell volume (see cytoplasmic volume density, Table 11, and the increase in volume density of granules was responsible for the majority of that increase (about 68%). The percentage of cytoplasm occupied by zymogen granules increased by more than 12 times in a sigmoidal pattern during this period (Fig. 6). At 17 days, zymogen granules occupied about 3% of the cytoplasmic volume. By 18 days, volume density had increased slightly to 7%. Between 18 and 20 days of gestation, volume density increased exponentially to about 24% at 19 days and about 45% a t 20 days, approximately the same value seen a t birth (Ermak and Rothman, 1980). Condensing vacuole volume density likewise increased during this period from less than 1% at 17 days to about 5% by 20 days (Fig. 6), slightly higher than that seen at birth (Ermak and Rothman, 1980). Zymogen Granule Size The increase in volume density of zymogen granules was primarily due to a n increase in granule size rather than granule number. The mean volume (calculated from mean diameter) increased ten-fold during the 4-day prenatal period (Table 2; 0.09 vs 0.90 pm3). This volume change corresponded to a n in- Figs. 2-5. Increase in zymogen granule content during prenatal development of the rat pancreas. The size and number of granules per cell profile increases on each day until the entire cytoplasm is filled with granules. Compare cell size in each figure; it roughly doubles over the 4-day interval. The cells shown do not necessarily depict the average morphometric values (see also Tables 1-31, x 8,000. Fig. 2. Day 17; cells lack zymogen granules altogether. Fig. 3. Day 17; granules are small and are seen primarily below the apical plasmalemma. Fig. 4. Day 18; granules fill a greater portion of the cytoplasm. ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS 491 492 T.H. ERMAK AND S.S. ROTHMAN Fig. 5. Day 20; granules have a wide range of sizes and are now seen in the basolateral portions of the cell. x 8,000. TABLE 1. Cytoplasmic and cellular features during prenatal development Time period (days postcoitum) 17 18 19 20 No. of Cells without granules cells sampled (%I 98 88 30 30 34.6 12.5 - - Relative cell volume' 0.7 0.8 1.1 1.7 (%I Increase in cytoplasm due to zymogen granules (%Y 61.0 f 0.7 62.8 k 0.9 79.7 1.3 85.2 f 0.9 33.8 65.2 67.7 Cytoplasmic volume density' - 'Total test points per cell (cytoplasm + nucleus) relative to 100 test points ( = total test points per cell at 3 weeks postnatally; Ermak and Rothman, 1980). 'Percentage of total cell volume (cytoplasm + nucleus) SEM where N = number of animals. 'Relative to 17 days. 493 ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS crease in granule diameter from an average of 0.55 pm at 17 days to 1.20 pm at 20 days of gestation. The greatest change in size, as with volume density, occurred between 18 and 19 days with over a four-fold increase in mean volume being observed, or a doubling in the original volume every 6 hr. On 17 and 19 days there was approximately a 50%daily increase in mean volume. The increase in mean volume accounted for virtually all of the increase in volume density over this 4-day period, and the increases in volume that occurred from day to day were accompanied by corresponding increases in volume densit-y (Fig. 7). A plot of the two parameters, V and Vvzg, formed a linear function that intercepted the y axis near the origin and whose slope was such that doubling the volume of granules roughly accounted for a doubling of volume density. The fact that granule volume is related to volume density in a linear fashion indicates that the relationship between average diameter and volume density is described by an exponential function that intercepts the axes at or near the origin and not by a I linear function that intercepts the diameterNB axis at a substantial distance from the origin (Fig. 7). 50 40 -G? 30 CI) z w n W 5 20 -I 0 > 10 (F) 0-y I I I 17 18 19 20 21 TIME (days postcoitum) Fig. 6. Zymogen granule (ZG) and condensing vacuole (CV) volume density during prenatal development of the rat pancreas. Volume density is expressed as percentage cytoplasmic volume. Data are mean values SEM for N = 5 (17 days), N = 4 (18 days), and N = 3 animals (19 and 20 days). Condensing vacuole SEM Q 0.8 units for all points. Newborn WB) added for comparison, from Ermak and Rothman (1980). + Zymogen Granule Number Since increases in the volume of zymogen granules accounted for the majority of the increase in volume density, and volume density is the product of mean volume and the number of granules per unit volume, an increase in granule number could have contributed little to the increase in volume density. When the number of granules per unit cyto- TABLE 2. Mean diameter and volume of zymogen granules during prenatal development Time period (days postcoitum) 17 18 19 20 Newborn3 Magnitude of volume increase Mean diameter (pm)' Calculated mean volume (pm3) 0.55 f 0.20 N = 480 0.65 k 0.20 N = 595 1.05 f 0.30 N = 688 1.20 k 0.35 N = 1098 1.50 f 0.40 0.09 - 0.14 1.5 x 0.61 7 X 0.90 10 x 1.75 20 x 'Arithmetic mean to nearest 0.05 pm differ from interanimal variance. 'Relative to 17 days. 3From Ermak and Rothman (1980). SD, N = number of granules. For all samples, intraanimal variance did not 494 T.H. ERMAK AND S.S. ROTHMAN plasmic volume was actually measured, it was found to have increased by only about 40% (numerical density; Table 3) during a period in which volume density increased some twelvefold. This does not mean that the number of granules in the tissue was not greatly changed during this period, but rather that the increase that occurred was roughly in proportion to the increase in tissue mass (approximately a fifteenfold increase; Fig. 1). MEAN DIAMETER ( p m ) Size Distribution At each time interval, the distribution of granule size was roughly normal and unimodal (Fig. 8). As the mean diameter increased, the absolute range of granule size also increased due to an increase in the size of the largest granules (from 1.10-2.60 pm). For the distributions seen at 17, 18, and 20 days, the mean diameter was slightly smaller than the mode, and for the broader distribution seen at 20 days there were several size classes near the mean diameter (0.90-1.40 pm) each of which accounted for approximately the same proportion of the population. The smallest granules (0.20-0.70 pm) accounted for most of the population at 17 days (89%), but were only a minor component (about 8%)by 20 days. Both the mean diameter and standard deviation around the mean increased proportionately during the 4-day period, i.e., the two variables were related to each other in a linear fashion, and the line intercepted the axes near the origin (Fig. 9). Thus, the standard deviation of granule size remained a constant proportion of the mean size as size increased, i.e., about 30%. In other words, whether for the broad distribution of granule sizes seen at 20 days or the comparatively u .2 .4 .6 .8 1.0 1.2 MEAN VOLUME ( p m 3 1 Fig. 7. Calculatd mean granule volume @) or mean granule diameter (D) vs granule volume density during prenatal development. Data are absolute values for mean diameter, or derived, therefrom rather than rounded. Granule volume vs volume density is a linear function that regresses through the axes near the origin. Notice that the slope of this function is such that increases in granule volume lead to roughly proportional increases in volume density. Because volume is a cubic function of diameter. the points for diameter vs volume density were fitted to a power function 07 = 25x3). Both functions were significant at less than the 1%level and had correlation coefficients of greater than 0.998. TABLE 3. Zymogen granule number during prenatal development Time period (days 17 18 19 20 ' Numerical density4 No. per cell profile'.' 7.0 rf10.4 37.6 rf65.2 rf- 1.3 1.2 4.2 5.9 (Nv) Numerical K . NA~'' 15.1 i-3.0 19.5 1.9 43.4 k 4.9 45.0 ? 3.2 k SEM where N = number of animals. 'Per section of single cell. 3Number per unit test area = 100 test points = 100 pm' acinar cell cytoplasm. 4Numbcr per unit volume (100 pm3 acinar cell cytoplasm). 25.0 25.9 41.9 36.5 27.5 30.0 41.3 37.5 "p 0 - (D n - n - 0 Iu 0 PERCENT OF ZYMOGEN GRANULES n 2 m T.H. ERMAK AND S.S. ROTHMAN 496 narrow distributions seen at 17 or 18 days, the range of sizes remained a constant percentage of mean size. A similar relationship is also seen during the 3-week postnatal period as granule size decreases (Fig. 9) (Ermak and Rothman, 1980), although in this case the observed standard deviation of granule size is somewhat smaller (18-27%) and decreases slightly (as a percentage) as size decreases (viz., the line does not intercept the axes a t the origin). Condensing Vacuole Size Condensing vacuoles were only observed occasionally a t 17 and 18 days but became a salient feature of the acinar cells by 19 and 20 days of gestation (Fig. 10). They increased in diameter from a n average of 0.65 pm a t 17 days to 1.45 pm at 20 days, over a tenfold increase in average volume. During the first 2 days, condensing vacuoles increased in size more rapidly than zymogen granules, and thereafter their size remained roughly constant even though zymogen granules continued to increase in size (Fig. 11).At birth, condensing vacuoles were smaller than at 19 and 20 days and were also smaller than zymogen granules seen at the same time (Ermak and Rothman, 1980).Although condensing vacuoles prenatally had a larger mean I / diameter than zymogen granules for the same time periods, their distributions encompassed roughly the same range of sizes, i.e., condensing vacuoles had a n equally broad, if not a broader, range of sizes than zymogen granules (Fig. 12). DISCUSSION The pattern of development and population dynamics of the pancreatic zymogen granule from the onset of its appearance in the secretory cell at approximateIy 17 days of gestation to its full expression as the dominant cytologic feature of the cell by birth can be summarized as follows: 1)The volume density of zymogen granules increases by one order of magnitude from days 17 to 20 of gestation following a sigmoidal pattern with the greatest increase occurring during the 18th day. 2) The number of granules per unit cytoplasmic volume changes little (+ 40%) between days 17 and 20 despite the magnitudinal increase in volume density. 3) The volume of the average granule increases by one order of magnitude and accounts in great part for the observed increase in volume density. 4)The number of granules in the tissue increases in proportion to the increase in organ mass (increased cell size and cell num- / MEAN DIAMETER ( A m ) Fig. 9. Mean granule diameter vs standard deviation of the mean at each time interval during prenatal development 0 . Data are absolute values for mean diameter rather than rounded. The calculated linear regression was significant at less than the 1%level and had a correlation coeficient of 0.990. The relationship between size and variability seen during the 3-week postnatal period 0 is given for comparison (from Ermak and Roth- man, 1980). Prenatal points in ascending order are: 17, 18,19, and 20 days. Postnatal points in descending order are: newborn, 1, 0.5, 2, and 3 weeks. The relationships differ in two ways. First, zymogen granules are more variable in size prenatally, and second, the standard deviation prenatally remains a constant proportion of the mean diameter, whereas postnatally it becomes a decreasing proportion of the mean as diameter decreases. ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS 497 Fig. 10. Condensing vacuoles (CV) from acinar cells at 19 days of gestation have a wide range of sizes and electron densities. Because condensing vacuoles are more numerous in the supranuclear region, this section overestimates their volume density and number. x 6,500. ber). 5) Granule size conforms to a roughly normal distribution at all times. 6) The increase in mean diameter of granules is due to an increase in the upper limit of granule size and an increase in the proportion of large granules in the population. 7) As granule size increases so does the absolute variance around mean size, i.e., the range of granule size increases as average size increases. 8) This increase in variance is proportional to the increase in mean size and hence the p r o portional variance remains a constant percentage of mean diameter as size increases. 9) The volume of the average condensing vacuole increases by one order of magnitude. 10) Although the average size of condensing vacuoles is slightly larger than for zymogen granules, their size distributions are nevertheless roughly coextensive. 11) The maximum size of condensing vacuoles appears t o be reached by day 19, whereas zymogen granule size continues to increase until birth (at day 22). Granule Size and Volume Density The increase in volume density of enzymecontaining zymogen granules which fill the acinar cells as birth approaches might be expected to be attributable to changes in the concentration of granules within the developing cell. However, a dramatic increase is not seen, and the increase in volume density during this period is primarily due to increases in the volume of individual granules and not their numerical density. In other circumstances, the concentration of granules in the cytoplasm can even vary inversely with volume density. For example, during the first 3 weeks of postnatal life, the volume density of zymogen granules in the rat pancreas decreases by about 60%, while numerical density doubles (Ermak and Rothman, 1980).Conversely, in the parotid gland, volume density appears to increase as numerical density decreases during recovery after maximal stimulation (Cope and Wil- 498 T.H. ERMAK AND S.S. ROTHMAN 1.6 I.4 I.2 5 I.o E Y W 5 .8 z Q .6 9 n OZG w 5 0 cv .4 .2 I 17 I I 19 20 TIME (days postcoitum) I I 18 21 I NB Fig. 11. Increase in condensing vacuole size (0) during prenatal development of the rat pancreas. Increase in zymogen granule size ( 0 )given for comparison. Condensing vacuoles increase in size more rapidly than zymogen granules and then remain at a constant size as granule size continues to increase. Data are mean values & SEM for N = the number of animals. Newborn (NB) added from Ermak and Rothman (1980). liams, 1981). Thus, it is becoming evident that changes in the size of granules are often the major cause of large fluctuations in the volume density of granules in a cell. Granule Size and Number Despite the fact that numerical density did not change substantially during development, a visual examination of tissue sections under the electron microscope gives the unmistakable impression that a great increase has occurred in the concentration of zymogen granules. This appearance of greatly increasing numerical density is in great part delusory. Granules appear more numerous simply because cell size more than doubles, and the total number of granules in the cell and organ increases even though their concentration is not dramatically changed. In addition, larger granules give the appearance of greater numerical density, because they occupy a greater percentage of the cytoplasmic volume. Even so, if we measure the number of granules per unit surface area, we find that the numerical profile density is increased some threefold during this 4-day developmental period. The central reason for this disparity is related t o the fact that we examine twodimensional images and not the actual threedimensional space itself, and such images do not give us an accurate representation of the number of granules per unit volume. The number of objects seen in electron micrographs of sectioned tissue depends on numerous factors such as the size, shape, size distribution, and plane of section, as well as the actual number of objects themselves. For spheres such as the zymogen granule, size or diameter is particularly important, since small spheres are sampled in tissue sections less frequently than large ones (see method 2 [DeHoff and Rhines, 19681in Materials and Methods); hence fewer small granules are seen or counted in “two-dimensional” sections. To determine true numerical density, ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS 499 40 30 v) W J 20 17 d 0 3 0 s 10 c3 z v) z W 30 0 z t- 8 8 + z W 0 a W a 20d DIAMETER (ym) Fig. 12. Distribution of condensing vacuole diameters at 17, 19, and 20 days of prenatal development. Data are expressed as percentage of total population (N = 56 for 17 days, 178 for 19 days, and 151 for 20 days). Each size category represents a 0.2-pm interval. Distribution of zymogen granule diameters a t the same time (smooth curves) is given for comparison. we must correct for this sampling problem. Thus, at 17 days, our “two-dimensional” measurement of numerical density underestimates the true value per unit volume relative to 20 days when granules are more than twice their diameter. The effect of granule size on the number of granules counted in tissue sections must also be considered when conditions produce a decrease in granule size, rather than an increase. For example, decreases in the number of granules seen in a cell during augmented secretion are usually thought to reflect real decreases in granule number. However, changes in granule size also occur under these conditions (Ermak and Rothman, 19811, 500 T.H. ERMAK AND S.S. ROTHMAN one for increases in granule size and another for decreases that are observed under different circumstances (Ermak and Rothman, 1980, 1981). Finally, we did not observe profiles suggestive of the fusion of granules in electron micrographs from any of the samples studied. Increases in the size of new or forming granules appear to be a major cause of the observed increase in size. New granules are being added to the organ such that the number more than doubles every day. In addition, since larger and larger condensing vacuoles are seen during the developmental period (until day 19), larger zymogen granules are apparently being formed as a result. These new granules reflect the formation of an increasingly broad size range as time advances Causes of Changes in Granule Size and are not the summation of several narrow The increase in zymogen granule size that distributions of larger and larger size formed occurs during prenatal development may be at different times. Finally, the increase in zymogen granule the result of several phenomena, either singly or in combination: 1)fusion of formed size may be attributable to increases in the granules with each other; 2) the formation of size of formed granules. Decreases in the size larger granules with time (either accom- of individual formed granules have been obpanied by the discontinuance of the forma- served in the pancreas after feeding (Ermak tion of smaller granules, or in addition to and Rothman, 1981), and analogous inthem); and 3) an increase in the size of indi- creases might occur during prenatal development. The size distributions seen during vidual formed zymogen granules. Granule to granule fusion does not seem to development are unimodal, relatively norbe a dominant feature of the developing pan- mal functions. That is, there is no indication creas. This conclusion is based on the fact of separate distributions for older (smaller) that the size of condensing vacuoles in- granules and new granules. If indeed all creases in parallel with that of zymogen granules, old and new, fit a single distribugranules, indicating that new granules of tion of sizes, as appears to be the case, then progressively larger size are being formed of course, if the size of the average granule (that is, of course, if we assume that the size increases with time, then so must the size of pattern of condensing vacuoles indicates the the old (formed)granules. In conclusion, the dramatic increase in size pattern of newly forming granules). Other factors also argue against a major role granule volume density that occurs during a for fusion in the observed size changes. In brief period of prenatal development is priother circumstances, e.g., periods of fasting, marily due to an increase in the size of grangranule size remains constant, and, thus, we ules in the cell and not the number per unit would have to propose special circumstances cytoplasmic volume. Thus, size change can that permit fusion at one time and prevent it be the central determinant of changes in at another. In addition, the fusion of gran- granule volume density, and granule numules produces a decrease in surface-to-volume ber may vary much less than is usually ratio that, assuming the conservation of thought in producing such changes. Furthergranule membrane, would result in a sub- more, because the size of granules can stantial increase in granule volume (two change, the effect of size on the apparent granules of equal diameter will increase number of granules in cells must be acabout 40% in volume). The multiple fusion of counted for. granules necessary to produce the observed ACKNOWLEDGMENTS increase in granule diameter would result in We gratefully acknowledge the excellent a sizable decrease in granule electron density which was not observed. Furthermore, we technical assistance of Ms. Sara Nelson. This would have to propose two mechanisms, viz., study was supported by research grants and, therefore, the decreased number seen in electron microscopic sections is at least in part only an apparent decrease due to the decreased frequency with which we sample small granules. This is not to say that changes in granule number do not occur, but when they appear to, the number of granules must by counted and corrected for sampling probabilities in order to properly draw the conclusion that number has indeed changed. Assuming that changes in numerical density occur based on the visual examination of a two-dimensional image without considering this issue may lead to an erroneous conclusion about the existence and extent of changes in the number of granules contained within cells during secretion. 501 ZYMOGEN GRANULE SIZE IN PRENATAL RAT PANCREAS AM25664 and AM16990 from the National Institutes of Health and by a research grant from the Academic senate of the university of California. LITERATURE CITED Cope, G.H., and M.A. Williams (1981) Secretion granule formation in the rabbit parotid gland after isoprenaline-induced secretion: Stereological reconstructions of granule populations. Anat. Rec., 199:377-387. DeHoff, R.T., and F.N. Rhines (1968) Quantitative Microscopy. McGraw Hill, New York. Doyle, C.M., and J.D. Jamieson (1978) Development of secretagogue response in rat pancreatic acinar cells. Dev. Biol., 65:ll-27. Ermak, T.H., and S.S.Rothman (1980)Large decrease in zymogen granule size in the postnatal rat pancreas. J. Ultrastruct. Res., 70:242-256. Ermak, T.H., and S.S. Rothman (1981) Zymogen granules of pancreas decrease in size in response to feeding. Cell Tissue Res., 214:51-66. Geuze, J.J., and Kramer, M.F. (1974) Function of coated membranes and multi-vesicular bodies during membrane regulation in stimulated exocrine pancreas cells. Cell Tissue Res., 156:l-20. Helander, H.F. (1978) Quantitative ultrastructural studies on rat gastric zymogen cells under different physiological and experimental conditions. Cell Tissue Res., 189.287-303. Jamieson, J.D., and G.E. Palade (1977) Production of secretory proteins in animal cells. In: International Cell Biology. B.R. Brinkley and K.R. Porter, eds. Rockefeller University Press, New York, pp. 308-317. Larose, L., and J. Morisset (1977) Acinar cell responsiveness to urecholine in the rat pancreas during fetal and early postnatal growth. Gastroenterology, 73:530-533. Mayhew, T.M. (1972) A comparison of several methods for stereological determination of the numbers of organelles per unit volume of cytoplasm. J. Microsc., 96:37-44. Palade, G.E. (1959) Functional changes in the structure of cell components. In: Subcellular Particles. T. Hay. ashi, ed. Ronald Press co., New Yo& PP. 64-83. Palade, G.E. (1975) Protein secretion by the pancreas. Science, 189:347-358. Parsa. I.. W.H. Marsh. and P.J. Fitznerald (1969) Pancreas acinar cell differentiation. I.-Morpholo&c and enzymatic comparisons of embryonic rat pancreas and pancreatic anlage grown in organ culture. Am. J. Pathol., 57:457-487. Pictet, R.L., W.R. Clark, R.H. Williams, and W.J. Rutter (1972) An ultrastructural analysis of the developing embryonic pancreas. Dev. Biol., 29:436-467. Rothman, S.S. (1975) Protein transport by the pancreas. Science, 190:747-753. Rothman, S.S. (1980) The passage of proteins through membranes - old assumptions and new perspectives. Am. J. Physiol., 238:G391-G402. Rutter, W.J., J.D. Kemp, W.S. Bradshaw, W.R. Clark, R.A. Ronzio, and T.G. Sanders (1968) Regulation of specific protein synthesis in cytodifferentiation. J. Cell Physiol., 72(SuppL 1):l-18. Sanders, T.G., and W.J. Rutter (1974) The developmental regulation of amylolytic and proteolytic enzymes in the embryonic rat pancreas J. Biol. Chem., 249:35003509. Snedecor, G.W., and W.G. Cochran (1980) Statistical Methods. Iowa State University Press, Ames, Iowa. Weibel, E.R. (1963) Principles and methods for the moruhometric studv of the lung and other organs. Lab. invest., 12: 131-i55. Weibel, E.R. (1969)Stereological principles for morphometry in electron microscopic cytology. Int. Rev. Cytol., 26:235-302. Weibel, E.R., and G.M. Gomez (1962) A principle for counting tissue structures on random sections. J. Appl. Physiol., 17:343-348. Werlin, S.L., and R.J. Grand (1979) Development of secretory mechanism in rat pancreas. Am. J. Physiol., 236:E446-E450. Williams, M.A. (1977) Quantitative methods in biology. In: Practical Methods in Electron Microscopy. A.M. Glauert, ed. North-Holland, Amsterdam, Vol. VI. pp. 65-66. - L

1/--страниц