J. Cell Set. ai, 415-422 (1976)
Printed in Great Britain
415
^CALCIUM LOCALIZATION IN ISLETS OF
LANGERHANS, A STUDY BY ELECTRONMICROSCOPIC AUTORADIOGRAPHY
S. L. HOWELL AND MARGARET TYHURST
School of Biological Sciences, University of Sussex,
Fainter, Brighton BNi 9QG, U.K.
SUMMARY
Attempts were made to localize the sites of uptake of "calcium in B cells of islets of Langerhans by electron-microscopic autoradiography. Despite the potential sources of error inherent
in the partial loss of radioactivity during fixation, and in the relatively high energy of emission
of this isotope, distribution of silver grains differed significantly from random in all experiments. Grains were concentrated over mitochondria and to a lesser extent over storage granules.
Incubation of islets in the presence of 10raMglucose and isobutylmethyl-xanthine before
fixation and autoradiography resulted in a small but not statistically significant reduction in
silver grains associated with the mitochondria. These results further indicate a dominant role
of mitochondria in the regulation of cytosolic calcium concentrations in pancreatic B cells.
INTRODUCTION
Calcium, and in particular changes in its intracellular concentration and binding
in B cells, may play an important role in the regulation of insulin secretion (Grodsky &
Bennett, 1966; Malaisse, 1973). We have therefore initiated a series of studies of
calcium distribution and accumulation by the pancreatic B cells, and this has so far
included metabolic studies of the regulation of ^calcium accumulation by homogenates and subcellular fractions (Howell & Montague, 1975), together with electronmicroscope X-ray microanalysis of the calcium content of various organelles in
frozen sections of unfixed tissue (Howell, Montague & Tyhurst, 1975). The results
suggest that mitochondria and storage granules contain high calcium concentrations,
while nuclei, endoplasmic reticulum and other cytoplasmic areas contained rather
lower concentrations. In metabolic studies of isolated subcellular fractions, mitochondria and microsomal fractions were shown to accumulate 4BCa in an ATPdependent reaction; mitochondrial uptake was diminished in the presence of cyclic
AMP (Howell & Montague, 1975).
We have now extended these investigations to include localization of "Ca in intact
B cells by electron-microscopic autoradiography.
416
S. L. Howell and M. Tyhurst
MATERIALS AND METHODS
Isolation of islets of Langerhans
Islets of Langerhans were isolated from rat pancreas by a collagenase digestion procedure
(Howell & Taylor, 1966) using a bicarbonate-buffered incubation medium (Gey & Gey, 1936)
containing 5-5 mM glucose and 2 mM CaCl, which was gassed with 95% O,, 5% CO, to
maintain a pH of 7-4.
Incubation with calcium-^
Autoradiography was performed on tissue which had previously been incubated with
"CaClj (50/iCi/ml; 28/iCi/mg; 31 fig Ca/ml, obtained from the Radiochemical Centre,
Amersham, Bucks.) in bicarbonate-buffered medium containing 55 or 20 mM glucose, or
20 mM glucose plus 0-5 mM isobutylmethyl-xanthine for 60 min. The islets were then rinsed
rapidly and fixed in 1 % glutaraldehyde (Polysciences Inc., Warrington, Pa.) in o-i M cacodylate buffer, pH 74, for 30 min, or in other fixatives as described under Results. Dehydration
through a series of concentrations of ethanol, clearing with toluene and embedding in an
epoxy resin were performed by a standard procedure.
In order to monitor loss of radioactivity during the dehydration and fixation, samples of
tissue were removed at each stage of the procedure, homogenized in toluene and the radioactivity present was determined in a liquid scintillation spectrometer (LS 233, Beckman
Instruments, Glenrothes, Scotland) using Instagel (Packard Instruments, Caversham, Berks.)
as scintillant.
Electron-microscopic autoradiography of Epon-embedded tissue was performed by a
procedure already described (Howell, Kostianovsky & Lacy, 1969; Howell & Whitfield, 1973)
using Ilford L4 emulsion. Exposure times varied between 4 and 6 weeks and development was
completed in Kodak D19 for 2 min. Sections were stained for 10 min in 50% ethanolic uranyl
acetate before examination and assessment of grain distribution was performed from electron
micrographs obtained at screen magnifications of 5000-10000.
Limitations of autoradiography with i5Ca
The problem of the poorer resolution obtainable in autoradiographs of 45Ca (ji emission
£'m»i= °'2S MeV) compared with that obtained with 3H (Emlix = 0018 MeV) must also
be considered. Although a detailed analysis of the resolution obtainable with 45Ca has not been
published, Salpeter & Salpeter (1971) have discussed the problem in some detail, using 14C
(Ew*i = 0-156 MeV) as a model. They found that the half distance (within which 50 % of the
developed grains from a line source fall) increased from 180 nm for *H to 280 nm for U C, an
increase in energy of 9-fold, and predicted that isotopes of even higher energies, such as "S or
w
Ca, could be used with little further loss of resolution, since scattering of electrons by the
emulsion would be reduced to a minimum. Values for point sources would be expected to be
1-7 times the line source values. Localization of activity within B cells should be possible since
the smallest organelles considered (the storage granules) have a diameter of about 290 nm
including the enveloping sac.
RESULTS
Preliminary experiments showed that the use of glutaraldehyde resulted in an
overall loss of 42-48 % of the 46Ca originally present in the tissue before fixation,
while osmium tetroxide either as a primary fixative or as postfixation after glutaraldehyde caused significant additional losses (Table 1). This is consistent with previous
findings in studies of strontium accumulation by mitochondria in other tissues by
Greenawalt & Carafoli (1966). Osmium fixation was therefore not employed in the
present studies, and in these conditions the procedures after glutaraldehyde fixation
Calcium localization in islets
Table 1. Loss of
45
4*7
Ca during fixation of islets of Langerhans
% radioactivity
retained in tissue
Unfixed islets
0-5 % glutaraldehyde
1 % glutaraldehyde
3 % glutaraldehyde
1 % glutaraldehyde-formaldehyde
2%OsO 4
3 % glutaraldehyde + 2 % OsO4
1 % glutaraldehyde, followed by dehydration in
ethanol, clearing in toluene
100
55
58
52
55
31
42
54
Results shown are the means from triplicate observations of the percentage retention of
radioactivity in batches of 20 islets which were fixed in the conditions shown.
Table 2. Distribution of silver grains over B cells after incubation of
islets of Langerhans in various conditions
Distribution
predicted
from
morphometry
(random
distribution)
Nucleus
Nucleus/cytoplasm
Total
Rough endoplasmic
reticulum
Endoplasmic reticulum/
cytoplasm
Total
Mitochondria
Mitochondria/cytoplasm
Total
13
3
16
20 mM
20 mM glucose
+ isobutylmethylxan thine
glucose
(0-5 mM)
12
2
11
11
14
14
5 "5 I'M
glucose
3
16
18
6
23
21
2
7
9
Storage granules
Granules/cytoplasm
Total
1
12
13
Cytoplasm
P value of x* for difference
from random distribution
40
—
4
IS
3
27
9
29
37
38
10
4
16
17
20
9
22
2
IS
17
7
<o-o5
<oos
Differences in grain distribution between B cells incubated in the 3 different conditions were
not statistically significant.
418
S. L. Howell and M. Tyhurst
through to embedding could be accomplished without further loss of radioactivity
(Table i).
The analysis of grain distribution was made by listing those organelles which fell
within an arbitrary area denned by a radius of 400 nm from the centre of each
grain. The grains were assigned as lying over mitochondria, granules, nuclei, roughsurfaced endoplasmic reticulum, ground cytoplasm or over combinations of any two
of these. Results are shown only for the major organelles overlapped with ground
cytoplasm; other combinations were either seen only infrequently (e.g. granules/rough
endoplasmic reticulum, mitochondria/rough endoplasmic reticulum) and were
Table 3. Association of silver grains with plasma membrane in B cells
incubated with 5 or 20 mm glucose or with 20 mu glucose + isobutyl-methyl-xanthine
Grains per 100 fim membrane
5 mM glucose
20 mM glucose
20 mM glucose+ 0-5 mM isobutyl-methyl-xanthine
16
13
14-5
A total length of 300 ftm of plasma membrane was assessed in each case.
omitted from the analysis, or were never found at all (nucleus/granule). The distribution which was obtained for cells which were previously incubated with ^Ca in the
presence of low (non-stimulatory) glucose concentrations, of high glucose concentrations or of high glucose concentrations in the presence of 3-isobutyl-i-methylxanthine (a potent inhibitor of cyclic nucleotide phosphodiesterase which raises
intracellular cyclic AMP levels) are shown in Table 2. This table also shows results
of analyses of cells using a random grid of circles of 400 nm diameter. The distribution
of silver grains differed significantly from random (%2 analysis, P < 0-05) in each of
the 3 incubation conditions, but the differences between these were not significant
(Table 2). However, there was a tendency for a reduction in the concentrations of
grains over the mitochondria in islets which had been incubated with isobutylmethyl-xanthine, and for a correspondingly higher concentration of grains over the
rough surfaced endoplasmic reticulum in these conditions. A number of grains were
found to overlap the plasma membrane and these were analysed separately as possibly
associated with the plasma membrane, regardless of the other organelles which they
overlapped. Results of this analysis, which was made in terms of grains associated
with the plasma membrane per unit length of membrane for cells incubated in the
3 conditions, are shown in Table 3. There was no significant difference in the number
of grains per /tm of membrane in the 3 cases. Illustrative areas to show the types of
grain distribution which were obtained are shown in Figs. 1 and 2.
DISCUSSION
It is certainly possible to localize ^Ca within B cells by autoradiography, the
distribution of silver grains obtained being significantly different from random.
Because of the long track of the isotope (see Methods) it was frequently impossible
Calcium localization in islets
4*9
r
iS%
Figs, i, 2. Electron-microscopic autoradiographs of B cells of rat islets of Langerhans
which had been incubated with 46Ca in the conditions described in the text. Fig. i is
from an islet incubated with 20 mM glucose alone. Fig. 2 with 20 mM glucose plus
0-5 mM isobutyl-methyl-xanthine. Fixation, 1 % glutaraldehyde only; stain, uranyl
acetate in 50% ethanol. Fig. 1, x 20000; Fig. 2, x 25 000.
420
5. L. Howell and M. Tyhurst
Fig. 2. For legend see p. 419.
Calcium localization in islets
421
to assign grains to a single organelle but by using a circle of diameter equivalent to
the estimated half distance of the track, an assessment could be made of the likely
origin of the radioactivity in each case. The most common combination was organelleplus-surrounding ground cytoplasm and in these cases the possibility must obviously
be considered that the activity arose not from the organelle in question at all, but from
the cytoplasm. This seemed rather unlikely, since grains present over cytoplasm
alone were very few and were in fact below the level which would be expected from
random labelling. We have therefore assumed that ' organelle + cytoplasm' labelling
reflects spreading of the radioactivity from the organelle to the surrounding area and
not the reverse process, and have taken the sum of 'organelle' and 'organelle +
cytoplasm' grain distribution as representing the true accumulation of activity
within that organelle. This type of analysis of grain distribution has revealed an
association of 46Ca with storage granules which has not been observed in a preliminary
study using the direct overlap of grain with organelle analysis used previously.
The highest concentration of silver grains obtained after incubation of islets in any
of the conditions used was seen in the mitochondria, which showed an almost 4-fold
accumulation above the expected frequency, and the storage granules (1-4-fold).
This accumulation over granules and mitochondria is consistent with results previously obtained by X-ray microanalysis (Howell et al. 1975), which showed that the
highest cellular concentrations of total calcium were present in those organelles. The
extensive accumulation of labelled calcium by mitochondria in intact cells is also
consistent with previous metabolic studies using broken cell preparations (Howell
et al. 1975), which suggested that mitochondria play a predominant role in the
regulation of calcium accumulation by B cell organelles.
The physiological importance of the high calcium concentrations associated with
insulin storage granules and the exact location of calcium in or on the granules are
not clear. The resolution obtainable by microanalysis with the EMMA system does
not allow distinction between the presence of an ion on the granule membrane by
surface binding, as is certainly indicated from pyroantimonate precipitation patterns
(Herman, Sato & Hales, 1973; Schafer & Kloppel, 1974), and binding within the
crystalline granule matrix or in the space between the electron-opaque core and the
limiting membrane. The stable nature of the calcium pool associated with the
granules, and its apparently rather slow exchangeability with exogenous 45Ca during
short term incubations suggest a relatively firm association of calcium with some
constituent of the granule. At least a proportion of the calcium content of the granule
could be trapped or incorporated during the process of its formation and then
exchange slowly with the cytosolic calcium during its lifetime; this might account
for the rather low level of labelling of granules observed in these autoradiographic
studies and for at least part of the *6Ca efflux which is observed during active periods
of insulin secretion (Malaisse, 1973).
Comparison of grain distribution in cells which had been incubated in the presence
of high glucose concentrations or of high glucose concentrations with isobutyl-methylxanthine, in order to stimulate secretion of insulin disappointingly showed only
small differences in grain distribution. This may be because only the rather stable
422
S. L. Howell and M. Tyhurst
fraction of the calcium accumulated by mitochondria remains intact under these
conditions while the labile fraction, which might perhaps be expected to be lost
during fixation and dehydration, may be the one which is more readily affected by
cyclic AMP. These results do serve to show that B cell mitochondria can accumulate
calcium in situ as well as in the artificial conditions of isolated mitochondrial preparations used previously, while the marginal fall in association of calcium with mitochondria is at least consistent with previous observations of the effect of cyclic AMP
on calcium accumulation by isolated mitochondria in liver, heart and kidney (Borle,
1974) as well as in islets of Langerhans (Howell & Montague, 1975). Proof of whether
cyclic AMP might induce efflux of calcium from mitochondria in intact B cells may
have to await results of X-ray microanalyses of calcium distribution in frozen sections
of unfixed tissue which had been incubated in various conditions.
Financial assistance from the Medical Research Council, British Diabetic Association and
from Hoechst Pharmaceuticals is gratefully acknowledged. S.L.H. is a CIBA Fellow.
REFERENCES
BORLE, A. B. (1974). Cyclic AMP stimulation of calcium efflux from kidney, heart and liver
mitochondria, J. Membrane Biol. 16, 221-236.
GEY, G. O. & GEY, M. K. (1936). The maintenance of human normal cells and tumour cells
in continuous culture. Am. J. Cancer 27, 45-76.
GREENAWALT, J. W. & CARAFOLI, E. (1966). Electron microscopical studies on the active
accumulation of strontium by rat liver mitochondria. .7. Cell Biol. 29, 37-61.
GRODSKY, G. M. & BENNETT, L. L. (1966). Cation requirements for insulin release in isolated
perfused pancreas. Diabetes 15, 910—913.
HERMAN, L., SATO, T. & HALES, C. N. (1973). The electron microscopic localization of cations
to pancreatic islets of Langerhans and their possible role in insulin secretion. J. Ultrastruct.
Res. 42, 298-311.
HOWELL, S. L., KOSTIANOVSKY, M. K. & LACY, P. E. (1969). B granule formation in isolated
islets of Langerhans: a study by electron microscopic radioautography. J. Cell Biol. 42,
695-7O5HOWELL, S. L. & MONTAGUE, W. (1975). Regulation by nucleotides of calcium-45 uptake in
homogenates of rat islets of Langerhans. FEBS Letters, Amsterdam 52, 48-52.
HOWELL, S. L., MONTAGUE, W. & TYHURST, M. (1975). Calcium distribution in islets of
Langerhans. A study of calcium concentrations and of calcium accumulation in B cell
organelles. J. Cell Sci. 19, 395-409.
HOWELL, S. L. & TAYLOR, K. W. (1966). Effect of glucose concentration on incorporation of
[ s H]leucine into insulin in isolated rabbits islet of Langerhans. Biochim. biophys. Ada 130,
5I9-52L
HOWELL, S. L. & WHITFIELD, M. (1973). Synthesis and secretion of growth hormone by rat
anterior pituitary I. The intracellular pathway, its time course and energy requirements.
J. Cell Sci. 12, 1-22.
MALAISSE, W. J. (1973). Insulin secretion: multifactorial regulation for a single process of
release. Diabetologia 9, 167-173.
SALPETER, M. M. & SALPETER, E. E. (1971). Resolution in electron microscope autoradiography.
J. Cell Biol. 50, 324-332.
SCHAFER, H. J. & KLOPPEL, G. (1974). The significance of calcium in insulin secretion.
Virchows Arch. path. Anat. Histol. 362, 231-245.
(Received 20 October 1975)