Differences in glucose recognition by individual rat pancreatic B cells are associated with intercellular differences in glucose induced biosynthetic activity

Differences in glucose recognition by individual

rat pancreatic B cells are associated with

intercellular differences in glucose-induced

biosynthetic activity.

R Kiekens, … , M Van De Winkel, D Pipeleers

J Clin Invest.

1992;

89(1)

:117-125.

https://doi.org/10.1172/JCI115551

.

In vitro incubated rat islet B cells differ in their individual rates of protein synthesis. The

number of cells in biosynthetic activity increases with the glucose concentration. Flow

cytometric monitoring of the cellular redox states indicated that islet B cells differ in their

individual metabolic responsiveness to glucose. A shift from basal to increased NAD(P)H

fluorescence occurred for 18% of the cells at 1 mM glucose, for 43% at 5 mM, and for 70% at

20 mM. The functional significance of this metabolic heterogeneity was assessed by

comparing protein synthesis in metabolically responsive and unresponsive subpopulations,

shortly after their separation by autofluorescence-activated cell sorting. The

glucose-sensitive subpopulation exhibited four- to fivefold higher rates of insulin synthesis during

60-min incubations at 2.5-10 mM glucose. Its higher biosynthetic activity was mainly caused by

recruitment of cells into active synthesis and, to a lesser extent, by higher biosynthetic

activity per recruited cell. Cells from the glucose-sensitive subpopulation were larger, and

presented a threefold higher density of a pale secretory vesicle subtype, which is thought to

contain unprocessed proinsulin. It is concluded that intercellular differences in metabolic

responsiveness result in functional heterogeneity of the pancreatic B cell population.

Research Article

Find the latest version:

Differences

in

Glucose Recognition by Individual Rat Pancreatic

B

Cells

Are

Associated with Intercellular Differences in Glucose-induced Biosynthetic

Activity

Rita Kiekens, Peter In'tVeld,* Tania Mahler,* Frans Schuit,t Marnix Van DeWinkel, and DanielPipeleers Department of Metabolism and Endocrinology, *Department of Experimental Pathology, andtDepartmentofBiochemistry,

VrijeUniversiteit Brussel, B-1090 Brussels, Belgium

Abstract

Invitro incubatedratislet B cells differ in their individualrates

ofprotein synthesis. The number of cells in biosynthetic activ-ity increases with the glucose concentration. Flow cytometric monitoring of the cellular redox states indicated that islet B

cellsdiffer in theirindividual metabolic responsivenessto glu-cose. A shift from basal to increased NAD(P)H fluorescence occurred for 18% of the cells at 1 mM glucose, for 43% at 5 mM, and for 70%at20 mM. The functional significance of this metabolic heterogeneity was assessed by comparing protein

synthesis in metabolically responsive and unresponsive sub-populations, shortly after their separation by autofluorescence-activated cellsorting. Theglucose-sensitive subpopulation

ex-hibited four-tofivefold higherratesof insulin synthesis during 60-min incubations at2.5-10 mM glucose. Its higher biosyn-thetic activity was mainly caused by recruitment of cells into active synthesis and,toalesserextent,by higher biosynthetic activityperrecruited cell.Cells from theglucose-sensitive sub-population werelarger, and presentedathreefold higher den-sity of a pale secretory vesicle subtype, which is thought to

containunprocessed proinsulin. It is concluded that intercellu-lar differences in metabolic responsiveness result in functional heterogeneity of the pancreatic B cell population. (J. Clin.

In-vest. 1992. 89:117-125.) Key words: cellular heterogeneity. insulin synthesis * islet cells*endocrine pancreas *metabolic redoxstate

Introduction

Glucoseexertsatightcontrolonthe release of insulin from the endocrine pancreas. The sugar has been found to stimulate severalstepsinthe formation of insulin and its release by

pan-creaticBcells(1-3). Not all cellsappear,however, equally in-volved in theprocess ofglucose-induced insulin biosynthesis

(4). At low glucose levels, only a minority of islet B cells is actively involved in protein synthesis (4). Glucose dose-depen-dently increases the number ofactive cells and hence therate at

which insulin is synthesized by the total population (4). The sigmoidal shape of the doseresponse curvesinglucose-exposed

Partof this workwaspresentedatthe 25th Congress of theEuropean

DiabetesAssociation, Lisbon, 1989.

AddresscorrespondencetoDaniel Pipeleers, Department

ofMetab-olism and Endocrinology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium.

Receivedforpublication6May 1991andinrevisedform28August 1991.

B cell preparations may thus result, at least in part, from a

dose-dependent recruitment of cells intoabiosyntheticactivity (4).The recognition of intercellular differences in the biosyn-theticactivity ofglucose-exposedBcells alsosuggeststhatthe cellsmarkedly differ inthesensitivity of their glucose-driven

metabolic pathways. The present study examines this possibil-ity and assesses whether variations in cellular thresholds for glucose metabolism can explain the intercellular differences in glucose-induced protein synthesis. The experiments were

con-ducted onpurified rat islet B cells that are isolated by autofluo-rescence-activatedcell sorting (5, 6). This model offers the

pos-sibilityofmeasuring theglucose-inducedchanges in the meta-bolic redox state of individual cells (7). It can be used to

separateB cell(sub)populations which differ in their metabolic

handlingofglucoseand which can then be compared for their

respective biosyntheticactivities.

Methods

Preparation ofrat pancreatic beta cells. Pancreatic islets were isolated

from male adult Wistar rats (200-300 g) according to a modification of thecollagenase digestionmethod ofLacy and Kostianovsky (6, 8).

They weredissociatedin a

Ca2"-free

medium containingtrypsinand DNase(9).Theisletcellsuspensionswerethensubmittedto autofluo-rescence-activated cell sorting, usingcellularlightscatterand flavin adenine dinucleotideautofluorescenceasseparationparameters(5,6). Thistechniqueallowspurification ofsingleinsulin-containingBcells with ayieldof 2-3.I05cells per adult rat pancreas (6).

Flowanalysis of metabolic alterations in glucose-exposedBcells.

We have previously described how the metabolic redox state can be measured inpurifiedBcells (7). Samples of 5.I04cells were incubated for 15 minat370Cin the presence of different glucoseconcentrations.

Theywerethenanalyzed for theirlight scatter and

NAD(P)H-auto-fluorescence, usinga FACS4-IVflowcytometer (Becton Dickinson &

Co., Mountain View, CA) equippedwith an argon laser(Spectra Phys-ics Inc., Mountain View, CA). Excitation occurred at 351-363 nm,

emissionwasmeasured between400 and 470nm(7). TheFACSO-data

arerepresentedasdotplots with the relative cellularautofluorescence intensityin theordinate and the relative cellular light scatter in the abcissa(Fig. 1). Thetwo internal controls consistedofislet B cells incubated at 1 mM or at 20mMglucose; their respective dot plots were usedtoselect theglucose-unresponsiveand theglucose-responsivecells

accordingto their distinctautofluorescence intensities (see Results). Windows were set on the basis of meanautofluorescenceintensities±2 SD andonthebasis of the forwardlightscatterofsingleislet B cells (see Results). FACSO analysis determined the percent particles thatwere

collected in eachofthe twopreselected windows.Atotal of5.104cells

wasanalyzedper experiment, each condition being examined in at

least six independent experiments. Results were expressed as

mean+SD,andstatisticallyanalysedusing pairedStudent'sttesting. Comparative studiesonisletBcells withdifferentmetabolic

respon-siveness toglucose.Purified B cells were incubated for 15 min at37°C

in the presence of7.5 mMglucose. Theywerethenanalyzedfor their

light scattering propertiesandcellularNAD(P)H-autofluorescence

in-tensity. Cells appearing in the preset windows of high and low NAD(P)H cells were isolated and consideredassubpopulations of, re-J. Clin.Invest.

©TheAmericanSocietyforClinicalInvestigation, Inc.

0021-9738/92/01/01 17/09 $2.00

Glucose 7.5 mM

U)

z

Forward Scatter

Figure 1. DotplotsafterFACSO-analysis of singlepurifiedBcellswhichhave beenexposed for 15 minto1,20,and 7.5 mM glucose. The cells wereexaminedfortheirNAD(P)Hautofluorescenceintensity (ordinate)andtheir forwardscatter(abcissa).Inthe 1 and 20 mMcondition,the windowsof, respectively, low andhigh NAD(P)H autofluorescencewere set asdescribed under Results(solidline).Thesewindowswerethen used to determine thepercentunresponsive(low)andresponsive (high)cells in each condition(brokenline).

spectively,7.5mMglucose-responsiveandunresponsivecells. Thetwo

subpopulationswerecomparedfortheirstructural and functional

prop-erties.Cellviabilitywasexaminedbyvitalstainingwith neutral red(6),

the percentlivingcellsbeingcounted immediately after isolationas well as afteraI0-d cultureperiodinpolylysine-coated microtitercups. Structuralintegritywasalsoassessedbyelectronmicroscopyon glutar-aldehyde-fixedcells(6).Themeaninsulincontentoftheisolated cells

wasmeasured inaceticacidextractsof samplesof104cells(6).The

biosynthetic activity ofthecellswasdeterminedover60-min incuba-tions,immediatelyafter their isolation.

Analysis ofprotein biosynthesisinisolated isletBcells. Samples of 5.1O islet B cellswereincubatedfor60 minatdifferentglucose

con-centrationsin 200

ql

Earle's-Hepes mediumcontaining 1%(wt/vol) BSA(fractionV, RIAgrade;SigmaChemicalCo.,St.Louis, MO)and 50,uCiL-(3,5-3H)-tyrosine (spact50Ci/mmol;totaltyrosine

concen-tration5

jiM;

AmershamInternational, Amersham, Bucks,UK)(10). Under theseconditionstherateofinsulinbiosynthesiscanbeexpressed

asfmol/h (10).Attheendoftheincubation,thecellswerewashed in Earle's-Hepesbuffercontaining1 mMunlabeledL-tyrosineand then extracted in acetic acidorprocessed forautoradiography.

Acidextracts werepreparedin 2 Maceticacid containing0.25% BSA.Samplesweredried andassayedfor3H-proteinaswellasfortotal and 3H-labeled insulinimmunoreactivity (4).Thecontentin

3H-pro-teinwasobtainedbycountingthe TCAprecipitable radioactivity.The 3H-labeled insulinimmunoreactivitywasassayedby addinganexcess

ofguinea pig antiinsulinserum andprecipitatingthe immune com-plexes withprotein A-Sepharose;nonspecific bindingwasmeasured in samples receiving normal guinea pig serum instead of antiinsulin

serum (4). The 3H-labeled insulin immunoreactivity was extracted from theimmunoprecipitated complexes with 2M acetic acid and counted. Theacid cellextracts werealsofurther analyzed by gel chro-matography. Sampleswereeluted onBiogel P-10columns(Bio-Rad Laboratories, Richmond, CA) with 2 Macetic acid, 0.25% BSA as

running solution (11). Elution fractions were counted and plotted

against '251-insulin standards; the two 3H-labeled peaks were com-pletelyimmunoprecipitable withaninsulin antibody and the second

peakcoincided with the insulin immunoreactive fraction of the cell

extract.

Autoradiographsweredevelopedasdescribed by Schuit et al. (4).

Briefly,the3H-labeledcellpreparationswerefixed for 30 min in 2%

p-formaldehyde, washedwith PBScontaining0.1% BSA,anddried

overnight on poly(L-lysine)-coated glass slides. The cells were then

postfixed for 15 min in 2%glutaraldehydeand washed with distilled water. The slides were exposed for 90 minto an autoradiographic

emulsion (L-4;IlfordLtd.,Basildon,Essex,UK),whichwasdeveloped for7min in ID- 1 (Ilford)at20'C and then fixed for another 7 min in Hypam(Ilford).Thecellpreparationswerecounterstainedwith

hema-toxylin.Atotalofatleast125 cellswereanalyzedpercondition. Analy-sisconsistedincountingthe numberof silvergrainsper cellat a screen

magnificationof5,800usingcontrast-enhancedvideomicroscopy.As

itwas notpossibleto countthegrainsin cellswithahigh graindensity, cellswithmorethan40grainswereconsideredas onegroup. Withinan experimentacellwasconsideredtobebiosyntheticallyactivated when

100

5-c

* oA

-0 =_

_i _

;4

C'

s6) s

50k

0

0 10 20

Glucose,

mM

Figure 2.Dose-responsecurvefor theeffect ofglucoseonthe percent

Bcells withhigh NAD(P)Hautofluorescence. Data representmean values±SEM of 6to 18experiments.

118 R.Kiekens,P. In'tVeld, T.Mahler,F.Schuit,M.VanDeWinkel,andD.

_Pipeleers

>.

*,

-g*;

;tmt'.,'s@. a\

*

\~~~~~~~*

s. 0- -* &

*9-*m wq ) * \ * *sA * @ *

C. tzu a*-~~~~~~~~~~~~~~~~~~~~~~~1* *S* C

etC . K.,

i4$

~~_'-

'p3

t

*~ ~a~ ~~~*

*.t *r;

FS2

Mt

.::;

* L..t . em.v

pAsgeA

*.W-C . ' '*,.u'

4.

.9;X9'%

I'

'mm 5t e** ~'w~

.*

*

K

v*

**

s

a

4

.W~t.%

.1';.4.

a

*-

cm,;;

.-.: .,;'

.;.;

; .*

+Q

JU,

a

.,.<j->,6:>-R.-.i;;

4

g

D~~~~~~~~~~~~~~~~~~~~~C

74p

7.

'

mm.

gluo.e

*h

cell meo

bohsbouainsehbta

nac lrsrcue

t *

..jj,~ ~ ~ im *

.

'*. 3

Cv'~~~*'A.~ ~ ~~~~~~~O..AC

*

_-41;;

m*.k

4 2.10

VI

._4

10

sC s

_.

_

r u 4

r. f 10

._ O

a E

so

0

0~~~

V

so-._

u

.C a ;Ib 0 0 v

*- C

.= _- 25

0 0

c

I--0~~~

V

0 5

Glucose, mM

Figure 4. Total protein (top) and (pro)insulin biosynthesis (bottom) in

the twoisletBcellsubpopulations separatedonthebasisoftheir low

(o) andhigh(.)NAD(P)H autofluorescenceat7.5 mMglucose.

Im-mediately after separation,the twopreparationswereincubated for

60 min at2.5, 5,or10 mMglucose.Data represent mean val-ues±SEMofeighttonineexperiments.Lowandhigh NAD(P)Hcells werecompared at the same glucoseconcentrationbypairedStudent's

ttest.*P<0.001;$P<0.005;P<0.01.

its numberofsilvergrainsexceeded the mean+2 SDofthe number that was found in thelow NAD(P)H cells labeled at 1.4 mM glucose.

Electronmicroscopic analysis. The two subpopulations that were

separatedonthe basis of their(un)responsivenessto7.5 mMglucose

were fixed in 2.5%glutaraldehyde(preparedin0.1 Msodium

cacodyl-atepH 7.4), postfixed in 1% osmiumtetroxide,and stained en bloc with 1%uranylacetate beforebeing dehydratedand embedded in Spurr's resin. Sectionsof 1 ,umthicknesswerestainedwithmethyleneblue and thenusedformeasuring totaland nuclear cell surface areasby

semiau-tomaticimageanalysis. Cytoplasmic surfaceareas werecalculatedas thedifferencebetween total and nuclear surface areas. The measure-ments werecarriedout on aVideoplan(Zeiss, Oberkochen, Germany)

at a screenmagnificationof5,800.Foreachsubpopulation,atotal of 400 cell profileswereanalyzed from fourindependent experiments

(100 cells persubpopulationand perexperiment). Ultrathin sections

wereexamined inaZeiss 9S2electronmicroscope, and micrographs

were taken at afinal magnification of17,300. Themicrographswere alsosubmittedtosemiautomaticimage analysis forthecounting ofthe numberofsecretory vesiclespercytoplasmic surfacearea.Analysiswas

restrictedtovesicles withasectionedgranule core. Adistinctionwas made betweenvesiclescontainingdark andpale granulecores. Apale secretory granulewasdefined byalargecore,occupyingatleast

two-thirds ofthevesiclearea,and byanelectrondensity thatwas compara-ble to, or lower than,that ofthesurrounding cytoplasm. Foreach

subpopulation,atotalof100cellswith sectionednucleuswere

ana-lyzedfrom fourindependent experiments (i.e.,25 cells per subpopula-tionand perexperiment).

Results

Metabolic

responsiveness

ofindividual islet

B

cells

to

glucose.

A

15-minexposure to20mMglucose increased themeancellular NAD(P)H-autofluorescenceintensity threefoldoverthe levels

measured at 1 mMglucose (78±14 relative fluorescence units at 20mM glucose vs. 26±6 at 1 mMglucose; mean±SD, P

<

0.001).

Thesemeancellular

NAD(P)H

autofluorescence

in-tensities±2 SD were taken to delineate the areas characteristic for islet B cells with a high (Fig. 1 b, high) and with a low

NAD(P)Hcontent

(Fig.

1 a,

low).

The low

NAD(P)H

window contained 82±7

(mean±SD,

n =

18)

percentof the cellsat 1 mM glucose, the

high NAD(P)H window

collected

70±8

Figure 5. Elution pattern of3H-labeled pro-teinssynthesized in thelow(o)andhigh(-)

NAD(P)H subpopulationsat10mMglucose. Cellularproteinswereextracted in acetic acid

__ andelutedon aBiogel PlOcolumn. The

100 counts perfractionwereexpressed per103

cells. The elutionpositionsofproinsulinand insulin were determined as inref. 11.

120 R.Kiekens,P.In 'tVeld, T.Mahler,F.Schuit,M. Van DeWinkel, and D.Pipeleers

proinsul

[in

insulin

3

2.10

ceg

1:W 3 no 10

0

ci

Po

0

50

fraction

number

(mean±SD,n = 7)percent at20mM.Analysis of the individ-ual cellresponses to20 mM glucose

indicated

thatnotall glu-cose-exposed B cells showedanincrease in NAD(P)H content: 30±8percent(mean±SD) remained inthewindow ofcellswith

low cellularNAD(P)H levels (Fig. 1 b, broken line). On the

other hand, 18±7 (mean±SD)percentofthe cells analyzed at 1 mM glucose appeared within thewindow ofcells with high

cellularNAD(P)Hcontent(Fig. 1 a,brokenline).These results

indicate

theexistenceofaheterogeneity inthemetabolicredox stateof

individual

Bcells, whetherthesecellsareexaminedat1 or at20 mM glucose. They also demonstrate that glucose

in-creasesthe number of islet B cells

with high

NAD(P)H content. Todetermine the dose responsiveness of thisglucoseeffect,we measuredthepercentofpancreaticB cellswithhigh NAD(P)H

autofluorescence after

15 minexposure tofour intermediate glucose

concentrations.

Itwas found that glucose increased,

dose-dependently,

the number of cells falling within the high NAD(P)H window, the sharpest rise being noticed between 2.5

0-*

*@ 0

and 5 mM

glucose (Fig. 2).

At 7.5 mM

glucose,

- 50%ofthe cellswererecovered within the

high NAD(P)H

window. This

percentage was

slightly higher

at 10 mMglucose (56±7) and

rose to 70±3 percent at 20 mM. It was not tested whether

glucose

concentrationsabove 20 mMwereabletofurther

in-creasethispercentage.In additiontoincreasingthe number of cells with elevated NAD(P)H

level,

glucose also raised the

NAD(P)H

fluorescence

intensity

ofmetabolically

responsive

cells. This effectwas mostmarked above 7.5 mM

glucose

as

evidencedbythehighermeanautofluorescence intensityofthe

"high"

cells at 20 mM than those at 1 or 7.5 mM

glucose

(Fig. 1).

Preparation

ofislet

B

cells

subpopulations with

different

met-abolic responsiveness

to

glucose.

To

compare

the functional

propertiesof islet B cells withdifferentmetabolic

responsive-ness to glucose, we separated the cells according to their

NAD(P)H-autofluorescence intensityat7.5 mMglucose.This concentration was selected in view of theequalsize of the

sub-0

l

'I

* Af

4

9

,

t

O

41

*-"ll

F

"&

_C

populations collected in the low andhigh NAD(P)H windows (Figs. 1 and2).Underthis condition,

sufficient

cell numbers

werecollectedfrom the lowandhighwindowstoundertake the

proposedexperiments. This wouldnothave beenpossible in the 1 mM or20mM condition. On the other

hand,

thetwo

subpopulations,

separatedat7.5 mMand

further

denotedas

lowandhigh NAD(P)H cells,

differ

less intheirmean

fluores-cence intensity, than the low and

high

subpopulation

at 20

mM.

This is

less

optimal for

detecting functional differences

thatare

associated

tocellular differences in redoxstate.

Immunocytochemistry

of the low and

high

NAD(P)H cells showed that both

fractions contained

>95%

insulin-positive

cells(datanot

shown).

This

degree

of

purity

was

confirmed

by electron

micrographs

illustrating ultrastructurally intact

and wellgranulatedcells in bothpreparations

(Fig.

3).Staining with neutral red

indicated

acomparable

viability

(>90%) of both

subpopulations

immediately

after their

isolation;

thetwocell

groupssurvived equally well duringaI0-d cultureperiod (data

notshown).

Comparison

ofglucose-inducedprotein biosynthesis

in

islet

B

cell

subpopulations with

different metabolic responsiveness

to

glucose.

Thetwo

islet

Bcell

subpopulations

thatwereseparated

according

to

their distinct

NAD(P)H-autofluorescence

at 7.5

mMglucose,werecompared

for their

protein

biosynthetic

ac-tivity

during

a60-min

incubation

at

different glucose

concen-trations.

At2.5mM

glucose,

the

high NAD(P)H

cells

incorpo-rated

fourfold

more

3H-tyrosine into

total

protein

than the low NAD(P)H cells

(Fig.

4).

Their

rate

of insulin biosynthesis

was

almostfivefold

higher (7.2±1.5

fmol/103

cellsperhvs.1.5±0.4

fmol/103

cells per h in the low NAD(P)H

preparation,

mean±SEM, n = 9, P<0.005)

(Fig.

4).

Raising

the glucose concentrationto5and 10mM

resulted,

in both

preparations,

ina

dose-dependent increase

in the

respective

rates

of

protein

and

proinsulin

biosynthesis

(Fig.

4).

The

biosynthetic

ratesin

high

NAD(P)H cells

remained

threeto

fourfold higher

than those in low

NAD(P)H

cells. Inboth

subpopulations,

60%

of

the 3H-total

protein

fraction

at 10 and20mM

glucose,

corre-sponded to

newly formed

(pro)insulin.

When the 3H-labeled material waselutedover a

Biogel

P-10

column,

most of the newly formed

proteins

wererecovered inthe

proinsulin

peak

(Fig.

5). No

difference

wasobserved in the

respective

propor-tions

of the

insulin peaks:

11±2% of 3H-IRI in the low NAD(P)H cells and 12±3% in the

high NAD(P)H

cells

(Fig. 5).

The3H-labeled Bcell

preparations

werealso

examined

in

autoradiographs

to

determine

thepercentcells

which

have

con-tributed

tothe measured

biosynthetic activities (Fig. 6).

At 1.4

mMglucose, 20±4%

of

high NAD(P)H-cells

were

involved

in protein synthesis, versus only

7±2%

of low NAD(P)H cells (mean±SEM,n =_4,NS)

(Fig.

7). Addition

of

glucose dose-de-pendently increasedthe numberofactivecells in both

prepara-tions. This

glucose-induced

recruitment occurred almost lin-early between 1.4and 7.5 to 10mMglucose. The sharpest rise

wasobservedinhighNAD(P)H cells,wheretheaddition of3.6 mMglucosetothe basalmedium (1.4 mM) recruited 54±6% of

the cellsinto

protein synthesis;

in low NAD(P)H cells, this rise

in glucoserecruitedonly28±10% ofthecells

(mean+SEM,

n

= 4,P<0.05). The maximal level of cell recruitmentwas ob-tained at 7.5 mMglucose for the high NAD(P)H cells. This

resulted in a totalof87±3% activecells. In low NAD(P)H cells,

aglucose concentration of10mMwasrequired for maximal cell

recruitment,

which resulted in 65±6% active cells

100

L

W

C ;P

y_ *° s0

.9

t 0 § 10 Glucose, mM 20

Figure 7.Effect ofglucose on the percentbiosynthetically activated

cells inautoradiographs oflow(o)andhigh(.)NAD(P)H

subpopu-lationswhich had been separated at 7.5 mM glucose. Data represent meanvalues±SEMof fourexperiments.Lowandhigh

NAD(P)H-cells werecompared at the sameglucose concentrationbypaired

Student's ttesting.

*P

<0.005;

§P

<0.01;

11P

<0.05.

(mean±SEM, n =4),a

significantly

lowernumber than in the high NAD(P)H

subpopulation

(P<0.01)

(Fig.

7).

Comparison of

hormone stores in

islet

B

cell subpopulations

with

different

metabolic responsiveness to glucose. The

obser-vation ofalowerbiosyntheticactivity in cellswithalow

meta-bolicresponsivenessto

glucose

ledus tocomparethe hormone

storesin thetwo

separated

Bcell

subpopulations.

By

immuno-assay,the

high NAD(P)H

cellswerefoundto

contain

a

higher

amount of insulin immunoreactive material

(68±6 ng/103

cells) than the low

NAD(P)H

cells

(49±4

ng/103

cells,

mean±SEM, n= 1 1, P <0.01). The

high

NAD(P)H cells

pre-sented also a

larger

cytoplasmic

surface area than the low NAD(P)Hcells (Table I),

while

the number

of secretory

vesi-clesperunit

cytoplasmic

area, was

similar

in both

subpopula-tions (Table

I).

However, when the

relative

amountofa

pale

secretory

vesicle

subtypewas

compared

in both

preparations,

markedly higher numbers per unit

cytoplasmic

area were

counted in

high

NAD(P)H

cells than in low

NAD(P)H

cells

(P

<0.05, Table I and

Fig.

8).

Pale

granules

were

noticed

in 12%

of

the secretory vesicles in the

high

NAD(P)H subpopulation,

versusonly4% inthe lowNAD(P)H subpopulation (Table I).

Discussion

We havepreviously reported that in vitro incubated isletBcells

do notexhibitan identical stateof

protein

synthesis (4).

Glu-coseincreased, dose-dependently, thepercent Bcells with

acti-vatedbiosyntheticactivity (4). These results indicatethatislet

Bcellsdifferintheir individual responsivenesstoglucose. The molecularbasis of thiscellularheterogeneityhasnotyet been

identified.Itcould be locatedatthe levelof glucose

handling

or

atthe site where glucose-induced

signals

control translation.

The presentstudypointstotheexistence of intercellular

differ-ences inglucose

handling.

The cellular redoxstate wastakenas

122 R.Kiekens, P. In 't Veld, T.Mahler,F.Schuit,M. Van De Winkel, and D. Pipeleers

11

kss

~~ ~ ~ ~ ~ ~

u-> ; ~ ~ * ;

*i:>*

f tS

-'a

* o*'W t ;,Fe

-

0

>-ff

..

.

.~~~

t~~~~~~~~~~~~~~~~~~~~~~~~s~~~ffi~~~~~~~ta

.

A6

40.

>

4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l

t

. .:..

..

ll. _,

SC

-t

* .7

* **.

12_1

-.~ ~~~~.ftt*

*

.

L~*..

*. pi

A .. _w

4- .t.

,_;.

+F

]-

.

::,

.,,.

Is.

..

.0,

M

.,.

.:

,X

,.

4

,- l*' - _

*i,+-S,X,. ;/?-; X -2 +

.

* _ X

S *0

*.

Table I. Morphometric Comparison ofLow and High NAD(P)H Cells Separatedat 7.5mM Glucose

Low High P

Total cellsurfacearea Am2/cell 106±6 127±5 P<0.001

Cytoplasmic surfacearea(CSA) Om2/cell 81±7 99±6 P<0.01

Numberofsecretoryvesicles

Total

n/,gm2

C.S.A. 1.85±0.26 2.00±0.35 NS "Pale"

n/hAm2

C.S.A. 0.07±0.05 0.25±0.14 P <0.05

%Total 4±2 12±4 P <0.01

Results are expressed as mean±SDof four independentexperiments. The numberof sectionedcells perconditionareindicatedin Results.

Dif-ferencesbetweenthe valuesoflow andhigh NAD(P)HcellsarecalculatedbypairedStudent's t test.

anindex for the level of

glucose

metabolism. Itwasmonitored through the cellular

NAD(P)H-autofluorescence intensity,

which is knownto

increase

during the

fluxes of glycolysis,

the Krebs cycle, and thepentosephosphate shunt.

When

pancreatic

B cell

populations

were

examined

at 1 mMglucose, cellular NAD(P)H levelswere notnormally

dis-tributed, with

almost 20% of the cells

exhibiting

higher NAD(P)H levels than the majority

of

cells

which

wereclustered aroundvalues

considered

asbasal. Shortexposure toglucose dose-dependently

increased

the number

of

Bcells

with

a

higher

NAD(P)H

autofluorescence.

This

glucose-induced

shift in cel-lular

metabolic

redoxstateoccurred for 5%

of

the cells

after

a

rise

from 1 to2.5 mMglucose;anadditional 20%

of

cells

re-spondedto an

increase

from

2.5to5mM, while 25%wasonly

responsive

between 5and 20 mM. PancreaticBcells thusdiffer in

their individual

thresholds for

glucose-induced

changes in the cellular redox state.

Since glucose metabolism

in

islet

B

cells

regulates

the

opening

or

closing of ionic

membrane chan-nels

(12-14),

the intercellular

differences

in

metabolic

thresh-oldare

likely

tobe

associated with

parallel differences

in

glu-cose-induced

depolarization.

The present

observations

can

thusaccount

for

the

earlier described

heterogeneity

in

electri-cally

recorded

signals

from

glucose-exposed islet

cells

(15).

In

view of the well known

parallellism

between therateof

glucose

metabolism

and that

of insulin

synthesis (1

1, 16), the cellular

heterogeneity

in

metabolic

activity

canbe held

responsible

for the

earlier

observed

intercellular differences

in

protein

synthe-sis

(4).

The

glucose-induced recruitment of islet

Bcellsintoa

protein synthetic activity

seemsthusto

depend

onthesugar's

ability

to overcome the cellular thresholds in the

metabolic

pathways which regulate

Bcell

function. This

recruiting effect

was

mainly

observed between 1and 7.5 mM

glucose.

Increas-ing

theglucose

concentration

above 7.5mMratherresulted in

a further rise

of NAD(P)H

levels in

metabolically activated

cells,and may thus be

responsible

foran

amplification

of func-tionsinrecruited cells.

Itis presently unknownwherethe sites of metabolic vari-abilityare located within the insulin-producing B cells. The

ability

to prepare islet Bcell

subpopulations

which differ in metabolicresponsivenesstoglucose, makes it feasibleto com-pareglucose responsive and

unresponsive

subpopulations for

theexpressionandactivitiesofpotentialkeyregulators such as

the

liver-type

glucosetransporter(17),

glucokinase

( 18, 19),or otherglucose-induced signals (20, 21). Suchtypeof investiga-tionmay alsoindicate whyaminority of adultratisletBcells

remain unresponsiveto alltested glucoseconcentrations.

Al-though

it is difficulttoexclude that this

subpopulation

corre-spondstocells

which

haveundergonesomedamage

during

the

isolation

procedure,wehavenot

found

any

experimental

evi-dence to support

this

possibility.

Most

cells of

this fraction

appear

structurally intact

by

vital

staining

and electron

micros-copy,and

survive

a

10-d

culture

period.

It

is

tobe

examined

whether the glucose

unresponsiveness of

these cells results from the in

vitro conditions

during

their

preparation or

whether itexpressesthe in situ

metabolic

stateofaparticular isletBcell

subpopulation.

The

correlation

between the cellular

metabolic

responsive-nesstoglucose and its

biosynthetic activity

became

evident

by

comparing glucose responsive

and

unresponsive

subpopula-tions.

Metabolically different subpopulations

were

prepared

on

the

basis of

thecellular redoxstateat7.5 mM

glucose.

At

this

concentration,

the

distribution of

cells is

mainly

the result

of

the

sugar's

recruiting action, i.e.,

a

shift from

basalto

increased

NAD(P)H levels. The

majority

of cells witha

rapid metabolic

response to 7.5 mM glucosewere also

recruited into

biosyn-thetic

activity during

a

subsequent

60-min

incubation

at

this

glucose

concentration. Despite this relatively

long

incubation

time, markedly

lesscellswererecruited in the

subpopulation

which

had failedto

respond

toshort 7.5 mM glucoseexposure.

Similar

differences

in

recruitment

werenoticed when both

sub-populations

wereincubated atlower glucose

concentrations.

This

observation

can

explain

themarkedly lowerratesof

pro-tein

synthesis

in the 7.5-mM

unresponsive

Bcell

preparation.

When

this subpopulation

was

incubated

athigher glucose

con-centrations

(10 mM),

itspercent of recruited cells

increased,

but its rate of

protein synthesis

did not rise

proportionally.

Higher glucose

levels canthus succeed in the

recruitment of

cells withpoor

glucose sensitivity

but induce

only

a

relatively

small increase in

their

protein

synthetic activity.

One

explana-tion

for

this

finding

could be a shortage in

preproinsulin

mRNA,

possibly

as aconsequence

of

a

chronic

state

of

reduced glucose metabolism in these cells. Itisthusconceivable that, under basal

physiologic circumstances, only

a

subpopulation

of islet B-cells is

metabolically responsive

toglucose and hence involved inhormone

synthesis.

OtherisletBcells, which are not

responsive

to

physiologic

glucose

concentrations,

seemto

participate

only

marginally

totheprocessof hormone

produc-tion under these

circumstances,

butcould be activated

by

con-ditions of

high

metabolic demand. This view is

supported

by

morphometric comparison

of the 7.5 mM

responsive

and

unresponsive

cells. This

analysis

strongly

suggested

that the two

subpopulations

had been indifferentstatesof

biosynthetic

activity

before their isolation.

High

responsive

cellswere

larger,

and

contained,

both inabsolute and relative terms, a

higher

amount of a palesecretoryvesiclesubtype which is shownto

contain unprocessed hormone (22, 23). We have, so far,not

detectedanyinhibition intherateofproinsulin conversionin the high responsive cells, sothattheirhigher content inpale granules may be related to the higher state ofbiosynthetic

activ-ity in thesecells. Thehigh responsivecells were, on the other

hand,foundtobemoresensitivetoglucosestimulifor release (24).Assuming thatahigher secretory activity impliesahigher release of dark, "mature" granules, this could contribute to a

higher proportion of paleover dark granules in the high re-sponsive cells. Theseobservations favor the hypothesis thatthe

frequencyofpalegranulesmay serve as amarkerfor the

glu-cose-responsivestateofislet cells in situ. Such studymayhelp answering the question whether the in situ

pancreatic

B cell population is also composed ofcells with different

sensitivity

to glucose, and hence of cells in a different stateoffunctional activity.

Acknowledgments

The authors thank thetechnical staff ofthe Departmentsof

Metabo-lismandEndocrinology,andofExperimental Pathology,Vrije

Univer-siteitBrussels, forexpert technicalassistance,and N. VanSlycke for secretarial help.

Thiswork wassupportedbygrantsfromtheBelgianFundfor

Medi-cal Research(3.0075.88and3.0093.90)andfromtheBelgianMinistry ofScience Policy (GekoncerteerdeAktie86/91-102).RitaKiekensand Tania Mahler are researchfellows oftheBelgian NationalFundfor

Scientific Research.

References

1. Permutt, M. A., and D. M.Kipnis. 1972. Insulin biosynthesis. Onthe

mechanism of glucose stimulation.J.Biol.Chem. 247:1194-1199.

2.Steiner,D. F., S. J.Chan,J. M.Welsh, D.Nielsen,J.Michael,H.S. Tager, and A. H.Rubenstein. 1986.Models ofpeptidebiosynthesis:the molecular and cellularbasis of insulin production. Clin.Invest.Med. 9:328-336.

3.Malaisse,W.J.,A.Sener,A.Herchuelz,and J.C. Hutton. 1979. Insulin

release:thefuelhypothesis.Metab.(Clin. & Exp.) 28:373-386.

4.Schuit,F.C.,P. A. In'tVeld,and D. G.Pipeleers.1988. Glucosestimulates proinsulin biosynthesisby a dose-dependentrecruitment ofpancreatic beta cells. Proc.NatL. Acad. Sci. USA.85:3865-3869.

5.Van De Winkel, M., E. Maes, and D. G. Pipeleers. 1982. Islet cell analysis andpurification by lightscatterandautofluorescence. Biochem.Biophys.Res. Commun. 107:525-532.

6.Pipeleers,D.G.,P. A. In'tVeld,M. Van DeWinkel,E.Maes,F.C.Schuit,

andW.Gepts.1985.Anewin vitromodelfor thestudyof pancreaticAandB

cells.Endocrinology. 117:806-816.

7. Van DeWinkel, M.,andD.Pipeleers. 1983.Autofluorescence-activated

cellsortingofpancreaticislet cells:purificationofinsulin-containingB-cells

ac-cordingtoglucose-induced changesin cellular redoxstate.Biochem.Biophys. Res.Commun. 114:835-842.

8.Lacy,P.E.,and H.Kostianovsky. 1967. Method for the isolation of intact isletsof Langerhans fromtheratpancreas.Diabetes. 16:35-39.

9.Pipeleers, D.G.,and M.A.Pipeleers-Marichal. 1981.Amethod for the purificationofsingleA, B and D cells andfortheisolationofcoupledcells from

isolatedratislets.Diabetologia.20:654-663.

10.Schuit,F.,and D.Pipeleers.1990.Measuringthebalancebetweeninsulin biosynthesisandrelease.Diabetologia. 33(Suppl P):331.(Abstr.)

11.Pipeleers,D.G.,M.Marichal,and W. J.Malaisse. 1973.The

stimulus-se-cretioncouplingofglucose-inducedinsulinrelease. XIV. Glucoseregulationof

insular biosynthetic activity. Endocrinology. 93:1001-1011.

12.Henquin,J.C. 1979.Opposite effects ofintracellularCa2l andglucoseon

K+permeabilityofpancreaticisletcells. Nature(Lond.).280:66-68.

13.Ashcroft,F.M.,D. E.Harrison,andS. J. Ashcroft. 1984. Glucose induces closureofsingle potassiumchannels inisolatedratpancreaticbeta cells. Nature (Lond.).312:446-448.

14.Cook,D. L., and C. N. Hales. 1984. Intracellular ATPdirectlyblocks K+ channelsinpancreaticB-cells. Nature(Lond.). 311:271-273.

15.Dean,P. M., and E. K. Matthews.1970. Glucose-inducedelectrical

activ-ityinpancreatic isletcells. J.Physiol. (Lond.). 210:255-264.

16.Lin,B.J.,B. R.Nagy,and R. E. Haist. 1972.Effectof various

concentra-tions of glucoseoninsulin biosynthesis. Endocrinology.91:309-31 1.

17.Thorens,B., H. K.Sarkar,H. R.Kaback, and H. F. Lodish. 1988.Cloning

andfunctionalexpressionin bacteria ofanovelglucosetransporter present in

liver, intestine, kidney,and#-pancreaticisletcells. Cell. 55:281-290. 18.Meglasson,M.D.,P. T.Burch,D.K.Berner,J.Najafi,A.P.Vogin,and F. M.Matschinsky. 1983. Chromatographicresolution and kinetic

characteriza-tion of glucokinase fromislets ofLangerhans.Proc.Natl.Acad. Sci. USA. 80:85-89.

19.lynedjian,P.B.,G.Mobius,H. J.Seitz,C. B.Wollheim,and A. E. Renold. 1986.Tissue-specificexpressionof glucokinase:identification of the geneproduct

inliverandpancreatic islets.Proc.Natl.Acad. Sci. USA.83:1998-2001. 20. McDonald, M.J. 1990. Elusive proximal signals of p-cells for insulin

secretion.Diabetes. 89:1561-1566.

21.Wolf,B. A., J. Florholmen, J.Twek, and M. L. McDaniel. Studies ofCa2"

requirementsfor glucose- and carbachol-induced augmentation of inositol tri-phosphate andinositoltetrakisphosphate accumulation in digitonin-permeabi-lized islets. Evidence foraglucoserecognition site in insulin secretion. 1988. J.

Biol. Chem. 263:3565-3575.

22.Orci, L.,M.Ravazzola,M.Amherdt, 0.Madsen,J.-D.Vassalli, and A. Perrelet. 1985.Direct identification of prohormone conversion site in

insulin-se-cretingcells.Cell. 42:671-681.

23.Orci,L., A. A.Like,M.Amherdt, B.Blondel, Y. Kanazawa, E. B.Marliss,

A. E. Lambert,C. B. Wollheim, and A. E. Renold. 1973. Monolayer cell culture

ofneonatal rat pancreas: an ultrastructural and biochemical study offunctioning endocrinecells.J.Ultrastruct.Res.43:270-297.

24. Van Schravendijk, C. F. H., L. Heylen, R. Kiekens, and D. G.Pipeleers.

1991. Heterogeneityin the secretory responseofpancreatic B cells. Diabetes.