| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Diabetes Research Center and the Department of Pharmaceutical and Biochemical Analysis (B.V.D.B), Vrije Universiteit Brussel, Brussels, Belgium
Address all correspondence and requests for reprints to: Prof. D. Pipeleers, Department of Metabolism and Endocrinology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium. E-mail: dpip{at}mebo.vub.ac.be
Abstract
Isolated human islets were examined for the rates of conversion and release of newly formed (pro)insulin-like peptides. The rate of proinsulin (PI) conversion was 2-fold slower in human ß-cells (t1/2 = 50 min) than in rat ß-cells (t1/2 = 25 min). During the first hour following labeling of newly synthesized proteins, PI represented the main newly formed hormonal peptide in the medium; its release was stimulated 2-fold over the basal level by 20 mmol/L glucose. During the second hour, newly synthesized hormone was mainly released as insulin, with 10- to 20-fold higher rates at 20 mmol/L glucose. Prolonged preculture of the islets at 20 mmol/L glucose did not delay PI conversion, but markedly increased the release of newly formed PI, des31,32-PI, and insulin at both low and high glucose levels. Our data demonstrate that 1) the release of PI provides an extracellular index for the hormone biosynthetic activity of human ß-cells; 2) an acute rise in glucose exerts a stronger amplification of the release of converted hormone than in that of nonconverted hormone; and 3) prolonged exposure to high glucose levels results in an elevated basal release of converted and nonconverted PI; this elevation is not associated with a delay in PI conversion, but is attributed to the hyperactivated state of the human ß-cell population, which was recently found to be responsible for an elevation in basal rates of hormone synthesis. These in vitro observations on human ß-cells provide a possible explanation for the altered circulating (pro)insulin levels measured in nondiabetic and noninsulin-dependent diabetic subjects.
PROINSULIN (PI) conversion requires the activity of two endoproteases, PC1 (PC3) and PC2, and one exopeptidase, carboxypeptidase H (1, 2, 3). These enzymes are evolutionary conserved, but their regulation and mode of action differ among species (4, 5) and experimental models (6, 7, 8, 9, 10). Although a mutation in carboxypeptidase H was recently observed in hyperproinsulinemic fa/fa mice (11), it is still unknown whether an imbalance between prohormone synthesis and enzyme activities contributes to the increased plasma levels of PI and des31,32-PI that are often observed in noninsulin-dependent diabetes (12). The possibility should also be considered that the elevated glucose levels in diabetes influence the biosynthetic and enzyme activities in islet ß-cells (13). In the present study, we have investigated the effects of acutely and chronically elevated glucose on the rate of PI conversion and the release of PI-like products in human ß-cell preparations.
Materials and Methods
Islet cells
Human islets were isolated within the framework of a collaborative program on islet cell transplantation in diabetes (14). Isolated human islets were cultured for 3–7 days in Ham’s F-10 (Life Technologies, Paisley, Scotland) containing 0.5% BSA (Boehringer Mannheim, Mannheim, Germany) at 6.7 or 20 mmol/L glucose (15). After culture, preparations consisted of 70% endocrine cells (15). Rat islet isolation and ß-cell purification were carried out as previously described (16); ß-cells were cultured for 20 h in Ham’s F-10 medium before labeling (17).
Pulse-chase experiments
Human islets (1–5 x 105 cells/tube) or rat ß-cells (1–3 x 105 ß-cells/tube) were labeled for 30 min at 10 mmol/L glucose (known to activate all ß-cells) with 16.7 Ci/mmol [3H]tyrosine (SA, 48 Ci/mmol; Amersham, Aylesbury, UK), washed in cold Earle’s HEPES buffer with 1 mmol/L unlabeled tyrosine, resuspended (3–6 x 105 rat ß-cells or 0.3–2 x 106 human islet cells/mL) in chase medium (Ham’s F-10, 0.5% BSA, 2 mmol/L glutamine, 2 mmol/L Ca2+, 50 µmol/L isobutylmethylxanthine with 2.5 or 20 mmol/L glucose, and 1 mmol/L unlabeled tyrosine, gassed with 5% CO2-95% O2), and incubated in a shaking water bath at 37 C. After a 60- or 120-min chase period, tubes were centrifuged, supernatants were stored at -20 C for later analysis, and pellets were extracted in 500 µL 2 mol/L acetic acid with 0.25% BSA. One sample was extracted after the 30-min labeling period to determine the composition at chase time zero.
High performance liquid chromatography (HPLC)
Islet cell extracts and media were analyzed by reverse phase
HPLC (18), using a Nucleosil 300–5C4 (5-µm
C4 column; 250 x 4 mm; Macherey Nagel, Düren,
Germany) connected to an LKB 2150 pump (Bromma, Sweden) and a 2152 HPLC
controller. Elution buffer consisted of acetonitrile (ACN) and 0.1%
trifluroacetic acid in Milli-Q water (Millipore Corp., Bedford, MA).
After equilibration at 28% ACN, elution (at 1 mL/min, 0.5 mL/fraction)
was started by increasing ACN over a linear gradient during 60 min to
29.6%. The column was washed with 80% ACN before the next run.
Collected fractions were mixed with 7.5 mL scintillation liquid (Ultima
Gold, Packard, Downers Grove, IL), and counted (Packard Tri-Carb).
Elution times were determined with human and rat insulin standards from
Novo Nordisk (Copenhagen, Denmark), human PI from Sigma Chemical Co.
(St. Louis, MO), human (h) des31,32-PI (lot A52-AWK-117),
des64,65-hPI (lot A52-AWK-100), split 32,33 hPI (lot
A52–9UL-141), and split 65,66 hPI (lot A52-AUK-94) from Eli Lilly
Research Laboratories (Indianapolis, IN). Purified rat ß-cell
extracts and media were separated with a similar ACN gradient, running
from 28.3–34% at 45 C (18). Elution times of the two rat PI forms
were determined by injection of ß-cell extract samples at the end of
the 30-min labeling period. Comparable profiles were found for samples
that were first immunoprecipitated with insulin antiserum and for those
that were not, allowing cell extracts and media to be analyzed without
pretreatment. In some human islet cell extracts the PI peak exhibited a
shoulder between standard positions of intact PI and split 65,66 PI
(asterisk in Fig. 1B
). This
shoulder could be immunoprecipitated by polyclonal anti-(pro)insulin
serum, by monoclonal antibody S2 to the intact AC junction of human PI,
but not by monoclonal antibody S53 to the BC junction (19, 20).
Therefore, it is considered a des31,32-PI-related
truncation product and was added to the des31,32-PI peak in
the analysis. Although degradation of newly formed PI and insulin was
not studied, HPLC analysis of the islet cell extracts was not
indicative of the presence of additional peaks that would correspond to
degraded hormone.
|
ß-Cell number in human islet samples was calculated from the DNA content and the percentage of insulin-positive cells in immunocytochemistry (16). Partly overlapping peaks in human samples were resolved by fitting a multipeak signal model to the data using the Levenberg-Marquardt algorithm for nonlinear regression on Matlab 4.0 (The Math Works, Natick, MA). The required fitting flexibility was obtained by exponentially modifying a tailing peak model, i.e. the Fraser-Suzuki function (21). Profiles from rat samples were analyzed for insulin and PI; in subsequent analysis, peaks of insulin I and II and of PI I and II were counted separately and/or in combination. For each time point, the cell extract and the corresponding medium were eluted; the radioactivity was counted per peak, then for all peaks together, and then expressed per 1000 ß-cells. The radioactivity per peak was finally calculated as a percentage of the total (pro)insulin-like radioactivity. The rate of PI conversion (t1/2) was calculated as the time required for the conversion of 50% of the newly formed PI. Data are expressed as individual values or as the mean ± SEM. Statistical significance was calculated by two-tailed paired Student’s t test.
Results
PI conversion
The rate of PI conversion, expressed as the half-time of
disappearance of the intact prohormone, was 2-fold slower in human
islets than in rat ß-cells (Table 1). A
similar difference was noticed when the rates of insulin appearance
were determined (data not shown). The rate of PI disappearance was not
influenced by the glucose concentration in the chase medium. At the end
of a 60-min chase at 2.5 mmol/L glucose, 60% of PI was converted. The
major conversion products were insulin and des31,32-PI;
only 6% appeared as other intermediates (Table 2). Comparable data were obtained at the
end of a 60-min chase at 20 mmol/L glucose (not shown).
|
|
). After a chase at 20 mmol/L
glucose, the same amount (87%) of PI was converted, but resulted in a
higher proportion of insulin (91% vs. 84% of converted
products) and a lower proportion of des31,32-PI (9%
vs. 15%) than for a 2.5 mmol/L glucose chase (Table 2). The
conversion rate was not decreased in islets cultured with 20 mmol/L
glucose (Table 1). Release of newly synthesized (pro)insulin
At the end of a 60-min chase at 2.5 mmol/L glucose, 2.7 ±
0.7% (n = 6) of newly formed hormone was discharged into the
medium, mainly (
80%) as PI (Table 3).
A higher proportion (10.7 ± 3.60%; n = 6) was recovered
after a 20 mmol/L glucose chase (Table 3); this increase occurred for
all peptides, but was most pronounced for insulin (12-fold elevation),
intermediate for des31,32-PI (6-fold), and lowest for PI
(3-fold; Table 3).
|
70%; Table 3). At 20
mmol/L glucose, 20.2 ± 4.8% (n = 5) of newly formed hormone
was released; stimulation was again highest for insulin (25-fold
elevation), intermediate for des31,32-PI
(5-fold), and lowest for PI (2-fold). Glucose stimulation thus reverts
the relative proportions of newly synthesized insulin and PI in the
medium: insulin now represents 65% of the newly formed peptides in the
medium, and PI comprises only 24%. As a result of these variations,
the medium ratio of newly formed insulin over PI increases with the
length of chase and with the glucose concentration in the chase medium
(Table 3).
In the three experiments in which human islets were also cultured at 20
mmol/L glucose, the release of newly synthesized hormone was markedly
increased during the 120-min chase period (Table 4). More than 20% of newly synthesized
hormone was released during this time period regardless of the presence
of a low (2.5 mmol/L) or a high (20 mmol/L) glucose concentration. This
release was thus exceptionally high at low glucose concentration (10-
to 40-fold higher than in 6.7 mmol/L glucose-cultured islet cells) and
was only marginally amplified at high glucose concentration (<2-fold
increased vs. 3- to 10-fold in 6.7 mmol/L glucose-cultured
islet cells; Table 4). With both low and high glucose chases, insulin
was the predominant form in the released newly synthesized hormone,
which led to a high medium ratio of insulin over PI (Table 4).
|
The present study demonstrates that the conversion of PI is 2-fold slower in human ß-cells than in rat ß-cells, at least under the selected in vitro conditions. In the rat, this rate is mainly derived from the more abundant PI-I, but human PI conversion is also slower than that of rat PI-II. This idea appeared when data were compared from previous studies on human and rat islet preparations (4, 5). It is unknown whether methodological differences, such as pancreas procurement and islet cell isolation, or differences in cellular composition of the tested rat and human preparations might influence their respective conversion rates. Although the prevailing glucose concentration did not influence the rate of conversion, we cannot exclude that other in vivo parameters, which are absent in vitro, maintain similar conversion rates in both species. We have not yet investigated whether part of the newly formed (pro)insulin products undergoes degradation under the present experimental conditions; if such a process did occur, it was not detected by our HPLC analysis.
Conversion of human PI results in accumulation of mainly des31,32-PI and insulin. The amount of newly formed des31,32-PI decreases with the time of processing, in particular at elevated (20 mmol/L) glucose levels, whereas the amount of newly formed insulin increases. The endopeptidase PC2 thus seems to mediate a glucose-dependent rate-limiting step in the formation of insulin; this finding might represent the functional consequence of the glucose-stimulated PC2 expression and biosynthesis that have been described in recent reports (13, 22, 23). A fraction of the newly formed PI and des31,32-PI is released within 60 min after formation of the hormone. This fraction does not increase during the second hour of chase. Release of PI is substantial (2–3%) at low glucose (2.5 mmol/L) and doubles (to 5–6%) at high glucose (20 mmol/L). That of des31,32-PI is low (0.5%) at low glucose (2.5 mmol/L) and increases 5-fold (to 2–3%) at high glucose (20 mmol/L). These data suggest that circulating PI and des31,32-PI levels will vary with the prevailing rates of both hormone synthesis and hormone release. The more marked stimulation of plasma des31,32-PI than of plasma PI observed during oral glucose tests thus seems to be the result not only of differences in respective plasma half-lives (24), but also of the stronger glucose stimulation on the release of newly formed des31,32-PI.
The fraction of newly formed hormone that is released as insulin increases with time and is greatly amplified by glucose (10- to 20-fold). As for rat ß-cells, this glucose-regulated release of newly synthesized insulin probably results from a glucose-induced recruitment of ß-cells into secretory activity (25). In the three experiments in which human ß-cells were also examined after chronic exposure to high glucose levels (20 mmol/L), this glucose-dependent amplification was no longer observed. In fact, this condition exhibited a high fractional release of newly formed insulin at low glucose levels (2.5 mmol/L); this finding is consistent with the observation that ß-cells exhibit a prolonged activation after culture at high glucose, even when subsequent glucose levels are low (15, 26). It is now shown that these ß-cells present a faster, rather than a slower, rate of PI conversion, which indicates that the rise in medium PI is not the result of a delay in its conversion but instead is an expression of the elevated biosynthetic activity of these cells. These data are also in agreement with the recent finding that PC2 (gene) expression is increased after chronically elevated glucose levels (13).
Our data support the idea that increased ratios of circulating PI over insulin, as observed in noninsulin-dependent diabetes, are not necessarily attributable to inadequate PI processing, but can result from the increased biosynthetic and secretory activities that occur under hyperglycemic conditions (27). Prolonged exposure to high glucose does not decrease the rate of PI conversion, but increases the release of newly formed PI-like peptides in the medium by more than 7-fold. This release is no longer dependent on the glucose concentration during the incubation, as might occur in hyperactivated cells (15); further studies are needed to identify the underlying secretory pathway as constitutive(-like) or regulated. It will also be necessary to analyze the release of preformed (pro)hormone in parallel. Only after this analysis will it become clear to what extent the presently described observations might be responsible for increased PI/insulin ratios in patients with or without noninsulin-dependent diabetes (28).
Acknowledgments
We thank the technical personnel at the Central Unit of ß Cell Transplant for providing the human islet preparations, and R. De Proft and A. De Loof for DNA measurements. Décio Eizirik is thanked for advice during preparation of the manuscript. PI conversion intermediate standards were kindly provided by Lilly Research Laboratories (Dr. H. Schmitt).
Footnotes
1 This work was supported by grants from the Flemish Community (GOA
92/97–1807), the Belgian Fonds voor Geneeskundig Wetenschappelijk
Onderzoek (F.G.W.O. 3.0132.91), and BIOMED (BMH 1-CT 92–0805). 
Received May 19, 1997.
Revised October 29, 1997.
Accepted December 15, 1997.
References
This article has been cited by other articles:
 |
|
 |
|
  K. Hostens, Z. Ling, C. Van Schravendijk, and D. Pipeleers Prolonged Exposure of Human {beta}-Cells to High Glucose Increases Their Release of Proinsulin during Acute Stimulation with Glucose or Arginine J. Clin. Endocrinol. Metab., April 1, 1999; 84(4): 1386 - 1390. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
