0021-972X/99/$03.00/0
The Journal of Clinical Endocrinology & Metabolism
Copyright © 1999 by The Endocrine Society
Vol. 84, No. 4
Printed in U.S.A.
Prolonged Exposure of Human b-Cells to High Glucose
Increases Their Release of Proinsulin during Acute
Stimulation with Glucose or Arginine*
KATLEEN HOSTENS†, ZHIDONG LING, CHRISTIAAN VAN SCHRAVENDIJK,
DANIEL PIPELEERS
AND
Diabetes Research Center, Vrije Universiteit Brussel, Brussels, Belgium
ABSTRACT
The disproportionate hyperproinsulinemia in type 2 diabetes has
been attributed to either a primary b-cell defect or a secondary dysregulation of b cells under sustained hyperglycemia. This study examines the effect of a 10- to 13-day exposure to 20 mmol/L glucose on
subsequent proinsulin and insulin release by human islets isolated
from nondiabetic donors. Compared to control preparations kept at 6
mmol/L glucose, the high glucose cultured b-cells released more proinsulin and less insulin during perifusion at 5, 10, or 20 mmol/L
glucose. The lower amounts of secreted insulin resulted from a
marked reduction in cellular insulin content (5-fold lower than in
controls). The higher amount of secreted proinsulin is attributed to
the sustained state of cellular activation that is known to occur after
prolonged exposure to high glucose levels. This activated state of the
b-cell population is also held responsible for its higher secretory responsiveness to 5 mmol/L arginine at a submaximal (5 mmol/L) glucose concentration (8-fold higher proinsulin levels than in the control
population). It results, together with the reduction in cellular insulin
content, in 7- to 10-fold higher proinsulin over insulin ratios in the
medium; at 5 mmol/L glucose, this extracellular ratio is similar to that
in the cells. These data add direct support to the view that a disproportionate hyperproinsulinemia can result from a sustained activation of human b-cells after prolonged exposure to elevated glucose
levels. (J Clin Endocrinol Metab 84: 1386 –1390, 1999)
P
examine whether this condition leads to a disproportionately
increased release of proinsulin by human b-cells.
ROINSULIN levels are increased in patients with type 2
diabetes (1– 4) as well as in subjects at risk for this
disease (5– 8). This may result from a delay in proinsulin
processing and/or from an accelerated release of immature
granules, which are known to contain higher proportions of
proinsulin (9 –11). The disproportionate elevation of circulating proinsulin has been related to the degree of impairment in the b-cell secretory capacity (12), thus supporting the
view that it reflects a primary b-cell defect (13, 14). There is
also evidence that this alteration is the consequence of an
increased secretory demand, like that occurring in persistent
hyperglycemia (9). In rat models of hyperglycemia, both islet
tissue and plasma were indeed characterized by increased
proportions of proinsulin over insulin (15–18). In hemipancreatectomized patients, hyperproinsulinemia was also
found to develop as a result of an increased b-cell demand
(19). Chronically elevated glucose levels do not necessarily
impair b-cell functions in the sense of reducing their activity.
Prolonged exposure of isolated rat or human b-cells to high
glucose concentrations was shown to induce a hyperactivated state, which is maintained after subsequent incubations at lower glucose levels (20, 21). In the present work, we
Received August 18, 1998. Revision received January 11, 1999. Accepted January 12, 1999.
Address all correspondence and requests for reprints to: Dr. D. Pipeleers, Diabetes Research Center, Vrije Universiteit Brussel, Laarbeeklaan
103, B-1090 Brussels, Belgium. E-mail: [email protected].
* This work was supported by grants from the Juvenile Diabetes
Foundation (JDF-DIRP 995004), the Belgian Fonds voor Wetenschappelijk Onderzoek (FWO G.0376.97), and the services of the Prime Minister (Interuniversity Attraction Pole P4/21).
† Research Assistant of the Belgian Fonds voor Wetenschappelijk
Onderzoek.
Materials and Methods
Preparation of human islet cells
Pancreata were obtained from organ donors at European hospitals
affiliated with b Cell Transplant (Brussels, Belgium) and Eurotransplant
(Leiden, The Netherlands) (22). After collagenase digestion, the tissue
suspensions were gently dispersed and submitted to gradient centrifugation (21). The fractions enriched in islet cell clumps were isolated and
cultured in serum-free medium as described previously (21). Preparations for this study were precultured for 2–3 days before distribution
over two dishes containing Ham’s F-10 medium with 1% BSA, 2 mmol/L
glutamine, and either 6 or 20 mmol/L glucose. They were then further
cultured for 10 –13 days, with medium replacements every other day. At
the end of this culture period, the two preparations were collected from
the dishes and washed before samples were taken for immunocytochemistry and for DNA (21), proinsulin, and insulin assays; the rest of
the material was used for perifusion. The cellular composition of the test
fractions and their total number of b-cells were determined as previously
described (21). At the time of perifusion, the preparations contained 76 6
2% endocrine cells (57 6 2% b-cells and 21 6 2% a-cells); dead cells
represented less than 6%. They are called islet cell preparations instead
of islets because the isolation and culture procedures resulted in a
progressive dispersing of the initial islet structures, a step that we consider useful for enrichment of living endocrine cells.
Perifusion of human islet cells
A multiple microchamber module (Endotronics, Inc., MN) with
build-in pump and thermostat was used for perifusion of the human islet
cells (23). Cultured islet preparations were loaded on preformed Bio-Gel
P2 columns (Bio-Rad Laboratories, Inc., Richmond, CA) and perifused
with Ham’s F-10 medium supplemented with 0.5% (wt/vol) BSA (fraction V, RIA grade, Sigma Chemical Co., St. Louis, MO), 2 mmol/L
glutamine, and 2 mmol/L CaCl2 (final concentration) and equilibrated
with 95% O2-5% CO2 (23). During the first 20 min, the medium contained
2.5 mmol/L glucose. The cells were then exposed to 10-min pulses of
1386
PROINSULIN RELEASE FROM HUMAN b-CELLS
increasing glucose concentration in the presence or absence of 5 mmol/L
arginine, each pulse alternating with a 10-min phase at 2.5 mmol/L
glucose (Fig. 1). The flow rate was 1 mL/min; samples were collected
over 1 min and assayed for immunoreactive insulin and proinsulin.
Insulin and proinsulin assays
The human insulin RIA was carried out as described previously (23).
Human proinsulin displays a 25% cross-reactivity in this assay. As the
amount of proinsulin measured in the samples represents maximally
12% of the corresponding insulin levels, it can be concluded that this
cross-reactivity causes maximally a 3% error in the quantification of
insulin and a 1% error in the ratio of the measured proinsulin over
insulin values. The proinsulin RIA is based on a two-step nonequilibrium procedure with negligible cross-reactivity with human insulin and
C peptide (24). We used polyclonal goat antihuman proinsulin from
Linco Research, Inc. (St. Charles, MO) and human proinsulin standard
donated by F. Sodoyez-Goffaux (University of Liege, Liege, Belgium).
Assay samples were incubated with the antibody for 18 h at 20 C before
[125I]human proinsulin (also provided by Dr. F. Sodoyez-Goffaux) was
added, and the incubation was continued for 24 h. Bound and free
proinsulin were separated by centrifugation after incubation (20 min, 20
C) with horse antisheep-coated Sepharose (Pharmacia Decanting Suspension 2, Pharmacia Biotech, Uppsala, Sweden), and pellets were
washed with phosphate-buffered saline containing 0.25% BSA and
counted. Standard curves were calculated on-line by RIA-Calc software
Pharmacia, Wallac, Finland), and values were accepted if they showed
two linear dilutions. The sensitivity of the assay was 5 pmol/L (blank
2 3 sd), and the interassay coefficients of variation were, respectively,
FIG. 1. Effect of prolonged culture at
high glucose (20 mmol/L) on insulin (upper panel) and proinsulin (lower panel)
release from human b-cells. Human islets cells were cultured at 20 (f) or 6 (E)
mmol/L glucose before perifusion at different glucose concentrations, in the absence or presence of arginine (5 mmol/
L). The rate of hormone release is
expressed as femtomoles of insulin or
proinsulin per min/103 b-cells. Data
represent the mean 6 SEM of six independent experiments.
1387
11%, 6%, and 5% for 20, 60, and 100 pg proinsulin/tube. The intraassay
coefficients of variation were, respectively, 3%, 4%, and 8% for 70, 35,
and 20 pg/tube. Statistical analysis was performed with StatView SE1
graphics for Macintosh (Abacus Concepts, Berkeley, CA). Results are
expressed as the mean 6 sem. Statistical significance of differences was
calculated by Wilcoxon rank sum test.
Results
Effect on cellular insulin and proinsulin content
Culture at 20 mmol/L glucose markedly reduced the cellular insulin content in each of the six preparations, but had
little effect on the cellular proinsulin content. The mean value
for insulin was 5-fold lower (0.9 6 0.2 pmol/103 b-cells; P ,
0.05) than that in the control condition with 6 mmol/L (4.3 6
1.3 pmol/103 b-cells), whereas that of proinsulin was only
25% lower (0.08 6 0.02 pmol/103 b-cells vs. 0.11 6 0.03 in
controls; P , 0.05). Consequently, the molar ratio of cellular
proinsulin over insulin was 4-fold higher in the 20 mmol/L
glucose-cultured cells (Table 1).
Effect on insulin and proinsulin release
At low glucose concentration (2.5 mmol/L), both preparations released comparable amounts of insulin (0.21 6 0.03
1388
JCE & M • 1999
Vol 84 • No 4
HOSTENS ET AL.
and 0.24 6 0.03 fmol/103 b-cells after culture at, respectively,
6 and 20 mmol/L; P . 0.05) and proinsulin (0.010 6 0.003 and
0.015 6 0.004 fmol/103 b-cells after culture at, respectively,
6 and 20 mmol/L; P . 0.05).
Rapid insulin secretory responses were measured after a
rise in glucose to 5, 10, or 20 mmol/L (Fig. 1). The insulin
release rate during these stimulations was 2- to 3-fold lower
(P , 0.05) after culture at 20 mmol/L glucose than in control
preparations (Table 2). This difference disappeared when
insulin release was expressed as a function of the corresponding insulin content (Table 3); it even reversed for the
5 mmol/L glucose stimulus, which caused, in relative terms,
a 2-fold higher insulin release from 20 mmol/L glucose cultured cells (P , 0.05; Table 3). Addition of arginine (5
mmol/L) to the glucose stimuli had little effect in control
preparations; only a small (20%) stimulation was seen with
the 10 mmol/L glucose pulse (Tables 2 and 3); on the other
hand, it induced 60% and 300% higher responses in b-cells
cultured in 20 mmol/L glucose that were stimulated by,
respectively, 10 and 5 mmol/L glucose (Tables 2 and 3).
During stimulation with 5 mmol/L glucose plus 5 mmol/L
arginine, the fractional release rate was 8-fold higher in 20
mmol/L glucose-cultured cells than in the control preparation (Fig. 2 and Table 3).
Proinsulin secretory responses to the glucose stimuli were
more pronounced in the 20 mmol/L glucose-cultured preparations in both absolute and relative values (Figs. 1 and 2,
and Tables 2 and 3). Addition of arginine did not influence
proinsulin release from control preparations, but increased
that from 20 mmol/L glucose-cultured cells that were pulsed
at 5 mmol/L glucose: the amount of proinsulin released was
now 8-fold higher than that in control cells (Table 2 and
Fig. 1).
TABLE 1. Effect of prolonged exposure to high glucose on
proinsulin/insulin ratio
Proinsulin/insulin ratio
Culture condition
Cells
Effluent
5 mmol/L glucose pulse
20 mmol/L glucose pulse
Control
20 mmol/L glucose
0.026 6 0.005
0.10 6 0.03a
0.018 6 0.009
0.005 6 0.002
0.12 6 0.05a
0.05 6 0.01a
The molar ratio is represented as the mean 6 SEM of six experiments. Statistical significance of differences with control was calculated by Wilcoxon rank sum test.
a
P , 0.05.
Ratio of proinsulin over insulin in effluent
The molar proinsulin over insulin ratio was markedly
higher in the effluent of b-cell preparations that were exposed to chronically elevated glucose levels. During perifusion at 5 or 20 mmol/L glucose, these ratios were, respectively, 7- and 10-fold higher in 20 mmol/L glucose-cultured
cells than in control preparations (Table 1). It was noticed that
the ratio in the 5 mmol/L glucose effluent was comparable
to that measured in the cells, whether they were cultured at
control or high glucose levels (Table 1).
Discussion
Chronic hyperglycemia can cause abnormalities in insulin
and proinsulin release (reviewed in Ref. 9). It is therefore
considered a pathogenic factor in the b-cell dysfunction of
type 2 diabetes (9, 25). The degree of fasting hyperglycemia
has been related to the disproportionate increase in proinsulin levels and, hence, to the rise in the proinsulin over
insulin ratio (12). It remains questionable, however, to what
extent this relationship expresses an effect of chronically
elevated glucose levels on the b-cells (10). The possibility
cannot be excluded that an increasing severity of the b-cell
defect leads to both a higher degree of hyperglycemia and a
further dissociation between proinsulin and insulin levels
(10). The present study demonstrates that prolonged exposure of normal human b-cells to high glucose concentrations
(20 mmol/L) brings them to a state in which they release
more proinsulin and less insulin upon glucose stimulation.
This leads to a 7- to 10-fold higher medium proinsulin over
insulin ratio than that for control cells that were cultured at
6 mmol/L glucose. These extracellular changes are comparable to those described in type 2 diabetes (1– 4). They reflect
in the first place the changes in cellular proinsulin and insulin
contents. The cellular insulin content was markedly decreased by culture at high glucose, probably as a result of an
excessive release that is not compensated by synthesis; this
leads to lower amounts of released insulin (20, 21). When
insulin release was expressed as a function of the cellular
insulin content, the b-cell secretory activity was not decreased, confirming the view that exposure to high glucose
does not reduce b-cell functions, at least not over 2-week
periods (20, 21). Previous work has indicated that this condition activates the majority of cells, increasing their mean
sensitivity to glucose (20, 21), a feature that has also been
noticed in type 2 diabetic patients.
TABLE 2. Secretory response after prolonged exposure to elevated glucose levels
Insulin release (fmol/103 b-cells)
Culture condition
Glucose pulse (mmol/L)
Arginine
Control
20 mmol/L
Proinsulin release (fmol/103 b-cells)
Control
20 mmol/L
5
2
1
11.1 6 3.4
11.7 6 3.2
5.1 6 2.7a
19.6 6 5.3b
0.1 6 0.03
0.1 6 0.02
0.4 6 01a
0.8 6 0.3a
10
2
1
44.3 6 13.2
54.0 6 13.4b
13.9 6 3.7a
22.3 6 5.1a,b
0.2 6 0.1
0.3 6 0.1
0.9 6 0.3
1.0 6 0.3
20
2
1
70.9 6 14.2
79.3 6 16.8
23.0 6 4.6a
29.2 6 6.1a
0.3 6 0.1
0.3 6 0.1
1.3 6 0.4
1.3 6 0.3a
Arginine was tested at 5 mmol/L. Release represents the amount of hormone secreted during a pulse. Statistical significance of differences
was calculated by Wilcoxon rank sum test.
a
P , 0.05 vs. control.
b
P , 0.05 vs. corresponding condition without arginine.
PROINSULIN RELEASE FROM HUMAN b-CELLS
1389
TABLE 3. Effect of glucose concentration during culture on the secretory response
Culture condition
Glucose pulse (mmol/L)
Insulin release (% content)
Proinsulin release (% content)
Arginine
Control
20 mmol/L
Control
20 mmol/L
5
2
1
0.3 6 0.1
0.3 6 0.1
0.6 6 0.2a
2.5 6 0.8a,b
0.1 6 0.03
0.1 6 0.03
0.4 6 0.1a
0.7 6 0.2a
10
2
1
1.2 6 0.3
1.4 6 0.3b
1.8 6 0.5
3.1 6 1.1b
0.2 6 0.1
0.3 6 0.1
0.9 6 0.2a
1.1 6 0.3a
20
2
1
2.1 6 0.5
2.3 6 0.6
2.9 6 0.7
4.2 6 1.4
0.4 6 0.2
0.3 6 0.2
1.4 6 0.3a
1.5 6 0.4a
Arginine was tested at 5 mmol/L. Release is calculated as described in Table 2 and expressed as a percentage of the cellular content. Statistical
significance of differences was calculated by Wilcoxon rank sum test.
a
P , 0.05 vs. control.
b
P , 0.05 vs. corresponding condition without arginine.
FIG. 2. Effect of prolonged culture at
high glucose (20 mmol/L) on insulin (upper panel) and proinsulin (lower panel)
release from human b-cells. Human islets cells were cultured at 20 (f) or 6 (E)
mmol/L glucose before perifusion at different glucose concentrations in the absence or presence of arginine (5 mmol/
L). The rate of insulin or proinsulin
release is expressed as a percentage of
the cellular insulin or proinsulin content measured at start of the perifusion.
Data represent the mean 6 SEM of six
independent experiments.
The present data suggest that proinsulin release might
represent a better marker for the functional state of b-cells
with a history of sustained secretory demand, such as persistently high glucose levels. At all tested glucose concentrations, more proinsulin is released during stimulation than
from control cells. This increased release of the precursor
hormone is not caused by a delay in conversion (26). We
believe, rather, that it results from the sustained state of
cellular activation. That proinsulin can be used as marker for
this condition is in part attributable to the fact that high
glucose exposure only marginally decreases the cellular proinsulin content at least under the present conditions. We do
not know whether this will still be the case after the longer
exposure periods that characterize the in vivo situation.
Prior exposure to high glucose increased not only the b-cell
secretory responsiveness to a subsequent glucose stimulus,
but also that to the nonglucose secretagogue, arginine, when
supplemented at a submaximal glucose concentration. Priming for a subsequent arginine stimulus was detected in terms
of both insulin and proinsulin release rates. As for glucose
1390
HOSTENS ET AL.
stimuli, proinsulin again appeared to be a better marker, as
its extracellular concentration increased with priming. Thus,
when arginine was administered at 5 mmol/L glucose, proinsulin levels were 8-fold higher for high glucose-cultured
b-cells than for control cells.
In conclusion, after prolonged exposure to elevated glucose concentrations, human b-cells release more proinsulin
and less insulin when stimulated by glucose or arginine. The
markedly increased proinsulin over insulin ratio in the medium reflects the changes in cellular hormone content that
result from a persistent state of cellular activation. The high
glucose-exposed b-cells exhibit a higher responsiveness to
glucose and arginine. The proinsulin levels in the medium
appear to be a valid marker for measuring the functional
state of b-cells, in particular during or after conditions of
increased secretory demand.
Acknowledgments
The authors thank the personnel of the central unit of b-Cell Transplant for preparing the human islet cells, and Lutgart Heylen and Gabriel Schoonjans for technical assistance in the present work. They are
grateful to Dr. F. Sodoyez-Goffaux, P. Houssa, and M. Deberg (University of Liege, Liege, Belgium) for their help in setting up the proinsulin
assay.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
References
1. Ward WK, LaCava EC, Paquette TL, Beard JC, Wallum BJ, Porte Jr D. 1987
Disproportionate elevation of immunoreactive proinsulin in type 2 (non-insulin dependent) diabetes mellitus and in experimental insulin resistance.
Diabetologia. 30:698 –702.
2. Yoshioka N, Kuzuya T, Tanigychi M, Iwamoto Y. 1988 Serum proinsulin
levels at fasting and after oral glucose load in patients with type 2 (noninsulin-dependent) diabetes mellitus. Diabetologia. 31:355–336.
3. Temple RC, Carrington CA, Luzio SD, Owens DR, Schneider AE, Sobey WJ,
Hales CN. 1989 Insulin deficiency in non-insulin-dependent diabetes. Lancet.
1:293–295.
4. Saad MF, Kahn SE, Nelson RG, et al. 1990 Disproportionately elevated proinsulin in Pima Indians with non-insulin-dependent diabetes mellitus. J Clin
Endocrinol Metab. 70:1247–1253.
5. Haffner SM, Bowsher RR, Mykkänen L, et al. 1994 Proinsulin and specific
insulin concentration in high- and low-risk populations for NIDDM. Diabetes.
43:1490 –1493.
6. Roder ME, Vaag A, Hartling SG, Dinesen B, Lanng S, Beck-Nielsen H,
Binder C. 1995 Proinsulin immunoreactivity in identical twins discordant
for non-insulin-dependent diabetes mellitus. J Clin Endocrinol Metab. 80:
2359 –2363.
7. Kahn SE, Leonetti DL, Prigeon RL, Boycko EJ, Bergstrom RW, Fujimoto Y.
20.
21.
22.
23.
24.
25.
26.
JCE & M • 1999
Vol 84 • No 4
1995 Proinsulin as a marker for the development of NIDDM in JapaneseAmerican men. Diabetes. 44:173–179.
Mykkänen L, Haffner SM, Kuusisto J, Pyörälä K, Hales CN, Laakso M. 1995
Serum proinsulin levels are disproportionately increased in elderly prediabetic
subjects. Diabetologia. 38:1176 –1182.
Leahy JL. 1996 Impaired b-cell function with chronic hyperglycemia: ‘overworked b-cell’ hypothesis. Diabetes Rev. 4:298 –319.
Kahn SE. 1996 Regulation of b-cell function in vivo. Diabetes Rev. 4:372–389.
Kahn SE, Halban PA. 1997 Release of incompletely processed proinsulin is the
cause of the disproportionate proinsulinemia of NIDDM. Diabetes.
46:1725–1732.
Roder ME, Porte Jr D, Schwartz RS, Kahn SE. 1998 Disproportionately elevated proinsulin levels reflect the degree of impaired b cell secretory capacity
in patients with noninsulin-dependent diabetes mellitus. J Clin Endocrinol
Metab. 83:604 – 608.
Porte Jr D, Kahn SE. 1989 Hyperproinsulinemia and amyloid in NIDDM.
Clues to etiology of islet b-cell dysfunction? Diabetes. 38:1333–1336.
Rachman J, Levy JC, Barrow BA, Manley SE, Turner RC. 1997 Relative
hyperproinsulinemia of NIDDM persists despite the reduction of hyperglycemia with insulin or sulfonylurea therapy. Diabetes. 46:1557–1562.
Leahy JL, Halban PA, Weir GC. 1991 Relative hypersecretion of proinsulin in
a rat model of NIDDM. Diabetes. 40:985–989.
Leahy JL. 1993 Increased proinsulin/insulin ratio in pancreas extracts of hyperglycemic rats. Diabetes. 42:22–27.
Alarcon C, Leahy JC, Schuppin GT, Rhodes CJ. 1995 Increased secretory
demand rather than a defect in the proinsulin conversion mechanism causes
hyperproinsulinemia in a glucose-infusion rat model of non-insulin-dependent diabetes mellitus. J Clin Invest. 95:1032–1039.
Gadot M, Ariav Y, Cerasi E, Kaiser N, Gross DJ. 1995 Hyperproinsulinemia
in the diabetic Psammomys obesus is a result of increased secretory demand on
the b-cell. Endocrinology. 136:4218 – 4223.
Seaquist EZ, Kahn SE, Clark PM, Hales CN, Porte D, Robertson RP. 1996
Hyperproinsulinemia is associated with increased b cell demand after
hemipancreatectomy in humans. J Clin Invest. 97:455– 460.
Ling Z, Kiekens R, Mahler T, et al. 1996 Effects of chronically elevated glucose
levels on the functional properties of rat pancreatic b cells. Diabetes. 45:
1774 –1782.
Ling Z, Pipeleers D. 1996 Prolonged exposure of human b cells to elevated
glucose levels results in sustained cellular activation leading to a loss of glucose
regulation. J Clin Invest. 98:2805–2812.
Keymeulen B, Ling Z, Gorus FK, et al. 1998 Implantation of standardized
b-cell grafts in a liver segment of IDDM patients: graft and recipient characteristics in two cases of insulin-independence under maintenance immunosuppression for prior kidney graft. Diabetologia. 41:452– 459.
Van Schravendijk CFH, Kiekens R, Pipeleers DG. 1992 Pancreatic b
cell heterogeneity in glucose-induced insulin secretion. J Biol Chem.
267:21344 –21348.
Bowsher RR, Wolny JD, Frank BH. 1992 A rapid and sensitive radioimmunoassay for the measurement of proinsulin in human serum. Diabetes.
41:1084 –1090.
Poitout V, Robertson RP. 1996 An integrated view of b cell dysfunction in type
II diabetes. Annu Rev Med. 47:69 – 83.
Zambre Y, Ling Z, Hou Z, et al. 1998 Effect of glucose on production and
release of proinsulin conversion products by cultured human islets. J Clin
Endocrinol Metab. 83:1234 –1238.