Cytometry

A Novel Method for the Detection of Viable Human

Pancreatic Beta Cells by Flow Cytometry Using

Fluorophores That Selectively Detect Labile Zinc,

Mitochondrial Membrane Potential and Protein Thiols

Sundararajan Jayaraman*

� AbstractImprovement over current methods of beta cell viability assessment is highly warrantedin order to efficiently predict the viability and function of beta cells prior to transplan-tation into type 1 diabetes patients. Dispersed human islet cells were stained with thecell-permeable zinc-selective dye, FluoZin-3-AM, along with the mitochondrial mem-brane potential indicator [(tetramethylrhodamine ethylester (TMRE)] and the thiol-binding dye, monochlorobimane (mBcl), and analyzed by flow cytometry. Islets weresubjected to various experimental conditions to validate the usefulness of this methodto accurately determine the viability and function of beta cells. Staining with FluoZin-3revealed the presence of higher amounts of chelatable zinc ions in beta cells than inlymphoid cells and fibroblasts. An intracellular zinc chelator competitively inhibitedthe binding of FluoZin-3 to zinc ions. Mitochondrial depolarization or oxidative stressminimally affected the binding of mBcl and FluoZin-3, respectively, to thiols and zincions. The combination of FluoZin-3, TMRE, and mBcl was sufficient and necessary forthe determination of the viability and function of beta cells. The data demonstrate theusefulness of the zinc-specific dye and the indicators of mitochondrial function andthiol levels, to accurately estimate the beta cell viability and function. This novel flowcytometry method has implications for islet transplantation in type 1 diabetespatients. ' 2008 International Society for Analytical Cytology

� Key termsbeta cells; labile zinc; mitochondrial membrane potential; thiols; glutathione; caspases;vital dyes

TRANSPLANTATION of islets of Langerhans isolated from cadaver donor pancreata

is an optional treatment in selected type 1 diabetes patients (1–3). However, a num-

ber of hurdles remain to be tackled before this promising cellular therapeutic

approach can become a more successful treatment procedure (4). Accurate determi-

nation of the long-term performance of islets in transplanted recipients will signifi-

cantly improve the success of islet transplantation. This requires precise assessment

of the frequency of live and functional insulin-producing beta cells in isolated islets

prior to transplantation. Although the beta cell content can be reliably determined by

intracellular staining with an antibody against insulin, chemical fixation and perme-

abilization required for intracellular staining do not allow simultaneous viability

assessment. Alternative methods of beta cell identification that permit simultaneous

assessment of viability and function are highly warranted.

Zinc is the second most abundant trace element in the body and is crucial for

cell survival, gene transcription, and for the function of more than 300 enzymes

(5,6). Certain central neurons contain substantial amounts of chelatable zinc ion

(Zn21) mainly in vesicles within excitatory nerve terminals (7,8). In pancreatic beta

Department of Surgery, College ofMedicine, University of Illinois atChicago, Chicago, Illinois 60612

Received 7 September 2007; RevisionReceived 20 December 2007; Accepted 20February 2008

This work was in part presented at theXXIII ISAC International Congress atQuebec City, Canada during May 20–24,2006.

Grant sponsor: Department of Surgery,University of Illinois, Chicago.

*Correspondence to: SundararajanJayaraman, Department of Surgery,University of Illinois at Chicago, Collegeof Medicine, 909 South Wolcott Avenue,COMRB Room 8113, Chicago, IL 60612,USA

Email: [email protected]

Published online 15 April 2008 inWiley InterScience (www.interscience.wiley.com)

DOI: 10.1002/cyto.a.20560

© 2008 International Society forAnalytical Cytology

Original Article

Cytometry Part A � 73A: 615�625, 2008

cells, a fraction of the intracellular Zn21 pool is stored with in-

sulin in vesicles as a complex of Zn21-insulin with a stoichiome-

try of 2:1 and is co-released during exocytosis (9–12). In addi-

tion, free Zn21 is found in the extragranular space of beta cells,

where it may act as a reservoir for granular zinc (13). In contrast

to zinc tightly complexed with proteins, including transcription

factors and metalloenzymes, pools of free and loosely bound

Zn21 can be visualized by using zinc-binding dyes (14,15). The

zinc-binding dye dithizone is routinely used for the determina-

tion of the purity of islet preparations (16). However, this colori-

metric dye cannot be combined with fluorescent dyes used for

the assessment of viability and function of beta cells.

Highly selective Zn21 probes such as TSQ, zinquin, and

TFL-Zn have been introduced to image and measure free

intracellular Zn21 (14,15,17-19). Limited aqueous solubility,

uneven cell loading, requirement for UV excitation, and com-

partmentalization into acidic vesicles are thought to limit the

use of some of these dyes (20). A visible wavelength fluores-

cent Zn21 probe, Newport Green (21–22), has been used for

the detection of beta cells in dispersed human islet cells (23).

More recently, FluoZin-3, a newer visible wavelength fluores-

cent probe (24–26) has been shown to be suitable for imaging

Zn21 co-released with insulin after stimulation of mouse islets

with high concentrations of glucose (24,26). FluoZin-3 binds

Zn21 with very high affinity (KD 5 15 nM) and has much

higher quantum yield than other zinc-sensitive dyes including

Newport Green (22,24,26).

Importantly, Newport Green, but not FluoZin-3, has been

shown to bind peroxynitrite nonspecifically (27). These data

suggest that FluoZin-3 may be a more sensitive and specific

probe for the detection of beta cells than other zinc-binding

dyes including Newport Green. However, the usefulness of

FluoZin-3 for quantitative assessment of human beta cells has

not yet been directly determined. Flow cytometry is ideal for

quantitative measurement of a specific cell type in a heterogene-

ous mixture of cells like islets of Langerhans. Therefore, freshly

isolated human islets were stained with FluoZin-3. To estimate

the viability and function of beta cells, dispersed islet cells were

simultaneously stained with FluoZin-3 and indicators of the

mitochondrial membrane potential (DCm), tetramethylrhoda-

mine ethylester (TMRE) (28) and cellular redox status, mono-

chlorobimane (mBcl; 29) and analyzed by flow cytometry. The

data presented in this communication demonstrate the useful-

ness of these dyes for the accurate determination of live and

functional beta cells in clinical islet preparations.

MATERIALS AND METHODS

Cell Lines and Treatment

Insulin-producing mouse beta cell line NIT-1 (ATCC,

American Type Tissue Culture, Manassas, VA) was cultured in

F12-K media as described earlier (28). NIT-1 cells (1 3 106/

ml) were treated overnight with 200 lM H2O2 or in media

adjusted to pH 5.9. The rat insulinoma cell line RINm5F

(obtained from Louis Phillipson, University of Chicago) and

mouse fibroblast line L929 (obtained from David Ucker, Uni-

versity of Illinois at Chicago) were grown in RPMI (Roswell

Park Memorial Institute, Invitrogen Corp., Carlsbad, CA)

media supplemented with antibiotics and 10% fetal bovine se-

rum. The mouse insulinoma cell line Min6 (obtained from

Louis Phillipson) and bTc3 (obtained from Jose Oberholzer,

University of Illinois at Chicago) were cultured respectively in

DMEM (Dulbecco’s modified Eagle medium, Invitrogen) and

RPMI media containing high glucose. Cells were treated with

0.05% trypsin-EDTA (Invitrogen) for 5 min at 378C and neu-

tralized with media containing 10% fetal bovine serum. The

human T-cell leukemia cell line E6-1 was cultured in RPMI

media as described (28). Viability was determined by Trypan

blue dye exclusion.

Treatment of Human Pancreatic Islets

Islets isolated from cadaver human pancreata (n 5 12) at

the University of Illinois at Chicago, University of Pennsylvania,

Washington University in St. Louis and City of Hope, Duarte

were made available for this study as a part of the ICR Basic

Science Islet Distribution Program. Islets were cultured at 378Cin CMRL 1066 media (Mediatech, Inc., Herndon, VA) adjusted

to pH 7.3 and supplemented with antibiotics, nicotinimide, in-

sulin and 0.25% human serum albumin. A single cell suspen-

sion was prepared by treatment with 0.05% trypsin-EDTA for

5 min at 378C and neutralized by media containing 10% fetal

bovine serum. The cell suspension was then filtered through a

cell strainer (40 lm, Becton-Dickinson, San Jose, CA) to

remove debris and clumps, and resuspended in culture media

(pH 7.3). Dispersed islet cells were cultured at 378C overnight

at 1 3 106 cells/ml in the presence of 200 lM H2O2 or in

media adjusted to pH 5.9.

Flow Cytometry Methods

Dispersed human islet cells, insulinoma cells, lymphoma

cells, and fibroblasts were resuspended at 1 3 106/ml concen-

tration in appropriate complete media. Mitochondrial mem-

brane potential was assessed by incubating cells with 50 nM

TMRE (Invitrogen) as described earlier (28). Cells were incu-

bated in complete media for 30 min at 378C with indicated

concentrations of FluoZin-3-AM dissolved in equal amounts

of DMSO (Sigma) and 10% Pluronic F127 in water (Invi-

trogen), known to facilitate loading of cells with organic dyes

without causing side effects (22). Intracellular redox status

was assessed by incubating cells for 30 min at 378C with 100

lM mBcl (Invitrogen) dissolved in DMSO as described (29).

In some experiments, stained cells were treated for 30 min at

378C with freshly prepared 3.8% formaldehyde solution in

complete media as we described earlier (28). The stock was a

38% (W/W) formaldehyde solution purchased from Fisher

Scientific, which contains methanol and water. A 5% stock

paraformaldehyde (Sigma) solution was prepared by dissol-

ving paraformaldehyde powder in phosphate buffered saline

(PBS, Invitrogen). For fixation, cells were treated with freshly

prepared 1% paraformaldehyde in PBS for 10 min at room

temperature as described earlier (28). After treatment, cells

were resuspended in the flow analysis buffer (Dulbecco’s PBS

1 0.2% BSA(bovine serum albumin, Sigma)) and kept in the

dark on ice until analysis.

ORIGINAL ARTICLE

616 Beta Cell Detection by Flow Cytometry

To determine the specificity of FluoZin-3 binding to che-

latable Zn21, NIT-1 cells were first incubated for 30 min with

indicated concentrations of N,N,N0,N0-tetrakis-(2-pyridyl-methyl) ethylenediaminedetermine (TPEN, Biomol, Plymouth

Meeting, PA) followed by incubation with FluoZin-3 in the

continued presence of the intracellular zinc chelator. In order

to determine the effect of mitochondrial membrane depolari-

zation, dispersed human islet cells were preincubated with 100

lM carbonylcyanide 3-chlorophenylhydrazone (CCCP, Bio-

mol) for 15 min at 378C and then incubated with FluoZin-3,

TMRE, and mBcl in the continued presence of CCCP for

another 30 min. Cell death was determined after incubation

with 10 lg/ml of 7-AAD or PI on ice for 10 min.

Intracellular activated caspases were determined as

described earlier (30). Briefly, cells were incubated with 10 lMFITC-VAD-FMK (Promega) at 378C for 45 min. PI was added

at 10 lg/ml concentration just before analysis.

Cells were analyzed on a Becton Dickinson LSR flow

cytometer equipped with the following filter sets: 530/28 BP

(FL1), 575/26 BP (FL2), 670 LP (FL3), and 510/20 DF (FL4).

Fluorescence spectrum of mBcl is not available. Fluorescence

spectrum of FluoZin-3 is similar to that of FITC (Refs. 21 and

22). Fluorescence spectra of FITC, TMRE, 7-AAD, and PI can

be found at several biotechnology company websites including

www.invitrogen.com. mBcl was excited with a UV laser (325

nm) and the fluorescence emission collected at 510 nm (FL4).

FluoZin-3, FITC-VAD-FMK, TMRE, PI, and 7-AAD were

excited by the 488-nm blue line. The emission maxima of

FluoZin-3 and FITC-VAD-FMK were collected by the FL1

(530 nm) detector. Emission of TMRE was collected at 575

nm (FL2), and PI and 7-AAD at 670 nm (FL3). Data were

acquired using CellQuestPro and analyzed by FlowJo 6.4.1

software (Tree Star).

Statistical Analysis

All experiments were performed more than three times,

and representative dot plots and histograms are shown. Stand-

ard deviation of the mean of multiple samples was calculated

using GraphPad Prism 4.0c software.

RESULTS

FluoZin-3 Binds to Chelatable Zn21 in Viable Human

Beta Cells with High Affinity and Selectivity

It was shown earlier that DCm, as measured by the reten-

tion of TMRE in the mitochondrial matrix, is an excellent in-

dicator of viability in the insulinoma cell line NIT-1 (28).

Since islets contain 68% beta cells, 21% alpha cells, 5% delta

cells, and 6% pancreatic peptide-producing F cells (31), iden-

tification of viable beta cells requires simultaneous staining

with a beta cell-specific probe and a viability indicator such as

TMRE. Since FluoZin-3 binds free Zn21 in vitro at 1:1 ratio

with higher affinity (15 nM) than Newport Green and other

zinc-binding dyes (22,24,26), it appears to be ideal for detect-

ing beta cells. This was tested using 12 different human islet

preparations obtained from various islet isolation centers in

the US. Within a day or two of receipt, islets were dispersed

using trypsin, stained with the cell-permeable format of the

dye, FluoZin-3-AM, and analyzed by flow cytometry.

A significant proportion (36%) of dispersed human islet

cells displayed forward angle light property characteristic of

intact cells and a majority (78%) of these cells were stained

brightly with the zinc-binding dye FluoZin-3 (Fig. 1A). In

contrast, only a small fraction (3%) of islet cells was stained

with FITC-conjugated normal mouse Ig, which can bind

to nonviable cells and emit fluorescence similar to FluoZin-3.

Titration of the dye revealed that as little as 50 nM FluoZin-3

was sufficient to stain a significant proportion of human beta

cells (Fig. 1B) and optimal loading of cells was observed with

500 nM concentrations of FluoZin-3. The frequency of islet

cells stained with optimal concentrations (500 nM) of Fluo-

Zin-3 varied among individual islet preparations (30–90%,

n 5 12). This could be attributed to a number of factors

including the age and health status of the donors of the pan-

creata, cold ischemia of pancreata prior to isolation of islets,

the lots of enzymes used for islet isolation, and various other

conditions including the pH during isolation.

The mouse insulinoma cell line, NIT-1, also displayed a

binding pattern similar to that of human islet cells (Fig. 1B).

However, other insulinoma cell lines such as Min-6, RINm5F,

and b-Tc3 bound FluoZin-3 with intermediate affinity. At 500

nM concentrations, only 30% of these insulinoma cells were

stained positively with FluoZin-3 (data not shown), reflecting

varying levels of free Zn21 in these cells. In contrast to beta

cells, human lymphoma cells and mouse fibroblasts required

higher concentrations (1–2 lM) of FluoZin-3 for comparable

level of staining. Similarly, only 10% of normal human periph-

eral blood leukocytes were stained with extremely high con-

Figure 1. Analysis of live beta cells by flow cytometry. (A) Freshly

isolated human islets were dispersed using trypsin and stained

with normal mouse Ig conjugated with FITC (NMIg-FITC) or 500

nM FluoZin-3-AM. Cells were then analyzed on a flow cytometer

for forward vs. side scatter light properties. Debris was gated out

and intact cells were analyzed for emission of green fluorescence.

(B) Human islet cells, the Jurkat clone E6-1, mouse insulinoma

cell line NIT-1, and mouse fibroblast L929 cells were incubated

with indicated concentrations of FluoZin-3 and analyzed (n 5 3).

ORIGINAL ARTICLE

Cytometry Part A � 73A: 615�625, 2008 617

centrations (10 lM) of FluoZin-3 (data not shown). These

results indicate that, in contrast to beta cells, other types of

cells tested failed to bind FluoZin-3 significantly at optimal

(500 nM) concentrations and required very high concentra-

tions (1–10 lM) of the dye for significant staining. This indi-

cates that nonbeta cells contain less chelatable Zn21 than beta

cells, prostate cells, and neuronal cells (5–6), and therefore

they require larger amounts of FluoZin-3 for comparable

binding. Collectively, these results demonstrate that the use of

lower concentrations (500 nM) of FluoZin-3 can permit the

clear identification of live beta cells in heterogeneous islet cell

populations without ambiguity, because of the presence of

higher levels of chelatable Zn21 in beta cells (9–12).

To determine the selectivity of FluoZin-3 binding to

Zn21 in intact beta cells, the cell-permeant zinc chelator

TPEN that has high affinity for zinc (3.8 3 1015 M21) (32)

was included during staining. Substantial reduction in Fluo-

Zin-3 binding to Zn21 in NIT-1 cells was seen in the presence

of TPEN (Fig. 2A). This was not simply due to decreased via-

bility as evidenced by low levels of PI uptake in TPEN-treated

cells. The data indicate that the dramatic increase in fluores-

cence (high quantum yield) is due to the formation of

the Zn21:FluoZin-3 adduct in intact beta cells, as observed in

solution (24–26).

Generation of ATP depends on DCm and is essential for

insulin secretion and beta cell viability (33). Consistently, the

anionic fluorescent dye TMRE that traverses cell membranes

and accumulates in the mitochondrial matrix (34) is retained

only by viable insulinoma cells (28). In addition to FluoZin-3

and TMRE, dispersed islet cells were costained with mBcl, a

glutathione S-transferase substrate that is used to assay the

pools of intracellular anti-oxidant, glutathione (GSH; 29).

This was done because dissipation of DCm has been shown to

be transient under certain conditions in some cell types (35–

36). In addition, HL-60 cells and neuronal cells have been

shown to undergo cell death without DCm dissipation (37–

38). If beta cells can also depolarize mitochondria reversibly or

without accompanying cell death, measurement of DCm alone

may not provide a true indication of viability and function of

beta cells. Inasmuch as the levels of GSH are critical for cell

survival (39–42), determination of the redox status may pro-

vide a better indication of the viability of beta cells. Data

shown in Figure 2B.i indicate that most beta (FluoZin-31)

cells (77%) displayed polarized mitochondria as indicated by

the retention of TMRE. A majority of beta cells (80%) also

displayed high levels of intracellular GSH, as revealed by mBcl

binding (Figure 2B. iv). Identical staining pattern was seen in

insulinoma cell lines, NIT-1 and RINm5F (data not shown).

Figure 2. FluoZin-3 binding selectivity and its resistance to treatment with formaldehyde and paraformaldehyde. (A) NIT-1 cells were pre-

treated with indicated concentrations of TPEN for 30 min and then stained with 500 nM FluoZin-3 in the continued presence of TPEN for

another 30 min. Cells were then analyzed for FluoZin-3 fluorescence. Dead cells were monitored following incubation with 10 lg/ml of PI.Representative data from three independent experiments are shown. (B) Dispersed human islet cells were stained with FluoZin-3 1 TMRE

or FluoZin-3 1mBcl and analyzed on a flow cytometer. Background staining was determined using unstained cells and those stained with

NMIg-FITC as described in Fig. 1. Cells were analyzed for staining with FluoZin-3 1 TMRE (i–iii) or FluoZin-3 1 mBcl (iv–vi). Stained isletcells were either treated with paraformaldehyde (PF) (ii, v) or formaldehyde (FA) (iii, vi) as described under Materials and Methods. Note

that paraformaldehyde treatment increased the autofluorescence of unstained islet cells (not shown), explaining the apparent increase in

FluoZin-3 fluorescence (ii and v). Representative data from three independent experiments are shown. (C) NIT-1 cells were stained with

FluoZin-3 1 TMRE and then treated with PF or FA. FluoZin-3 and TMRE signals were analyzed as described (n 5 4).

ORIGINAL ARTICLE


Thus, TMRE and mBcl together may provide a better tool for

the delineation of beta cell viability.

It was previously demonstrated that DCm, as assessed by

TMRE retention, is sensitive to treatment with oxidizing

agents in insulinoma cells (28). However, the susceptibility of

the antioxidant GSH and labile Zn21 to oxidative stress has

not been elucidated. To evaluate this, dispersed human islet

cells were first stained with FluoZin-3 1 TMRE or FluoZin-3

1 mBcl and then treated with paraformaldehyde or formal-

dehyde and analyzed by flow cytometry. The data obtained

indicate that the frequency of FluoZin-31 1 TMRE1 cells

diminished considerably, from 77% in controls to 43% after

paraformaldehyde treatment (Fig. 2Bi vs. ii; upper right

quadrants). As shown earlier in NIT-1 cells (28), formalde-

hyde treatment induced dissipation of DCm in a majority of

islet cells (77% in controls vs. 7% in treated group, Figure

2Bi vs. iii; upper right quadrants). Similarly, the frequency of

beta cells containing detectable levels of GSH (FluoZin-31 1mBcl1) was modestly reduced by parformaldehyde (from

80% in controls to 63% in treated cells, Fig. 2Biv vs. v; upper

right quadrants) and substantially after fomaldehyde treat-

ment (40%, Fig. 2Biv vs. vi; upper right quadrants). These

results are consistent with the previous observation that

formaldehyde is more effective than paraformaldehyde in

dissipating the DCm in beta cells (28). This could be due to

the generation of more oxidants when cells were incubated

with formaldehyde at 378C as opposed to incubation with

paraformaldehyde at room temperature (see Materials and

Methods). Methanol present in the stock solution of formal-

dehyde may also be responsible for increased oxidative stress.

Nevertheless, these results indicate that DCm is more sensi-

tive than GSH to increased oxidative stress induced by treat-

ment with formaldehyde solution.

In contrast to Newport Green, the binding of FluoZin-3

to Zn21 in beta cells was not affected by formaldehyde or

paraformaldehyde treatment. The total frequency of FluoZin-

31 cells (upper right quadrant 1 lower right quadrant) did

not significantly differ from untreated human islet cells (com-

pare Fig. 2B. i and iv, with ii, iii, v and vi). Similarly, treatment

with paraformaldehyde or formaldehyde did not reverse the

binding of FluoZin-3 to Zn21 in NIT-1 insulinoma cells (Fig.

2C). Interestingly, paraformaldehyde treatment appeared to

increase the intensity of FluoZin-3 fluorescence in comparison

with untreated cells (Fig. 2B. ii and v). This apparent increase

in FluoZin-3 binding to Zn21 is partly due to an increase in

autofluorescence after paraformaldehyde treatment (data not

shown). Taken together, the data indicate that DCm is more

sensitive than intracellular GSH to an increase in oxidative

stress while the amount of chelatable Zn21 is barely affected

by enhanced oxidative stress.

Viability Assessment Does Not Require

Supravital Dyes

Human islet preparations contained varying levels of

nonviable cells (range: 12–55%, n 5 12) based on PI uptake.

The emission spectrum of PI overlaps with that of TMRE sig-

nificantly (22), and therefore these dyes cannot be used to-

gether. Since 7-AAD, a commonly used vital dye (43), has a

narrow emission spectrum compared to PI (22), it was next

evaluated whether inclusion of 7-AAD would be useful for the

accurate estimation of viable beta cells in human islet prepara-

tions containing varying proportions of dead cells. To this

end, dispersed islet cells were stained with various dye combi-

nations and analyzed on a flow cytometer. Debris was

excluded from the analysis by gating strategy as described in

Figure 1. This particular islet preparation contained more

than 94% beta (FluoZin-31) cells (Fig. 3, row 1; column A),

and they were viable as indicated by staining with TMRE (Fig.

3, row 2, column B) and mBcl (Fig. 3, row 4, column D). It

should be noted that when islet cells were stained with one of

these three dyes—FluoZin-3, TMRE, and mBcl—fluorescence

emission was detected only at the expected wavelengths, indi-

cating effective elimination of spectral overlap between these

fluorochromes by color compensation. In addition, double

staining with FluoZin-3 1 TMRE (Fig. 3, row 5, columns A

and B), FluoZin-3 1 7-AAD (Fig. 3, row 6, columns A and

C), or FluoZin-3 1 mBcl (Fig. 3, row 7, columns A and D)

resulted in fluorescence emission only at the expected wave-

lengths. Unexpectedly, staining with TMRE and 7-AAD to-

gether led to an increase in the frequency of 7-AAD1 cells

(31%, Fig. 3, row 8, column C) in contrast to those stained

with 7-AAD alone (12%, Fig. 3, row 3, column C). This was

seen consistently in all human islet preparations tested, albeit

at varying levels.

Further analysis revealed that simultaneous staining with

TMRE and 7-AAD resulted in a cell population that stained

with both of these dyes without changing the frequency of 7-

AAD1 cells (Fig. 4C; upper right quadrant vs. upper left quad-

rant). This was accompanied by a concomitant decrease in

TMRE1 cells, from 75% in TMRE stained cells (Fig. 4A, lower

right quadrant) to 48% in cells stained with both TMRE 1 7-

AAD (Fig. 4C; upper right quadrant). Similar changes in

staining patterns were evident when islet cells were stained

with FluoZin-3 and mBcl in addition to TMRE and 7-AAD

(Fig. 4I). Other combinations of fluorochromes neither

increased the double-positive cells nor decreased TMRE1 cells

(Fig. 4D and G). Since it is well documented that TMRE is

retained only by cells displaying both plasma membrane and

inner mitochondrial membrane potentials (28,34), the finding

that a majority of TMRE1 islet cells also allow the entry of 7-

AAD raises questions about the validity of the use of 7-AAD

to define dead cells when used in combination with TMRE.

To further scrutinize the use of 7-AAD to delineate live

beta cell population, 7-AAD1 dead/dying cells were excluded

from the analysis by electronic gating. In islet preparations

that had few dead/dying (7-AAD1) cells, exclusion of 7-

AAD1 cells did not substantially alter the frequency of total

beta cells (Fig. 5B vs. E) or their viability, as indicated by stain-

ing with TMRE and mBcl (Fig. 5C vs. F). When the same islet

preparation was analyzed after 2 days of culture, the numbers

of viable cells were reduced (Fig. 5D vs. J). Exclusion of 7-

AAD1 cells in this preparation appeared to increase the fre-

quency of beta (FluoZin-31) cells (Fig. 5H vs. K) and their vi-

ORIGINAL ARTICLE


Figure 3. Examination of the compatibility of 7-AAD with indicators of labile Zn21, DCm and thiols. Dispersed human islet cells were

stained singly or doubly with indicated dyes and analyzed on a flow cytometer. Fluorescence emissions detected by different detectors are

shown as histograms. Unstained cells and those stained with NMIg-FITC were used to determine the background fluorescence as indi-

cated in Fig. 1. Note that when cells were stained with FluoZin-3, TMRE or 7-AAD alone, the appearance of fluorescence signals in inap-

propriate channels was avoided by manual color compensation. However, the appearance of mBcl signal excited by the UV laser, in other

detectors could not be compensated due to the instrument/software constraints. Data shown are from a single experiment (n 5 6).

ORIGINAL ARTICLE


ability (Fig. 5I vs. L). These results suggest that the elimination

of 7-AAD1 cells, especially in islet preparations with higher

proportion of cells with compromised plasma membrane in-

tegrity may lead to overestimation of the frequency and viabil-

ity of beta cells.

Since 7-AAD does not appear to be suitable for the discri-

mination of dead cells when used in conjunction with TMRE,

it was next examined whether the antioxidant GSH could be

used as a marker of viability in addition to TMRE. Since the

dissipation of DCm can occur transiently under certain condi-

tions and even without the loss of cell viability (35–38), the

relationship between DCm, GSH levels and chelatable Zn21

was then examined in islet cells. To this end, dispersed islet

cells were preincubated with CCCP, a potent uncoupler of mi-

tochondrial oxidative phosphorylation (44) for 15 min fol-

lowed by incubation with the dyes in the continued presence

of CCCP. Data obtained indicate that CCCP treatment

induced the loss of TMRE signal, indicating the dissipation of

DCm (Fig. 6C vs. G, upper right quadrant). However, only a

modest decrease in GSH levels was seen in CCCP-treated cells

(from 72% in controls to 54% in CCCP-treated group, upper

left 1 right quadrants). Importantly, CCCP treatment did not

drastically alter FluoZin-3 binding to labile Zn21 (Fig. 6B vs.

F). Treatment with CCCP also increased the plasma mem-

brane permeability of cells to 7-AAD (Fig. 4D vs. H). Taken

together, the data indicate that the combination of FluoZin-3,

TMRE, and mBcl is sufficient and necessary to determine the

frequency of viable and functional beta cells and obviate the

use of vital dyes like 7-AAD. Importantly, the inclusion of

mBcl assures the precise estimation of the viability of beta cells

even in the presence of mitochondrial depolarization without

accompanying cell death.

Figure 4. Incompatibility between TMRE and 7-AAD for live cell determination. Dispersed human islet cells were stained with indicated

fluorochromes and analyzed on a flow cytometer. The quadrants were set based on fluorescence emitted by unstained cells and by cells

stained with NMIg-FITC as described in Fig. 1. The fluorescence emissions of TMRE and 7-AAD were detected respectively by FL2 (575 nm)

and FL3 (670 nm) detectors. Color compensation was performed using cells stained alone with FluoZin-3, TMRE, 7-AAD, or mBcl as

described in Fig. 3. Representative data from four independent experiments are shown.

ORIGINAL ARTICLE


Determination of Beta Cell Viability Under

Pathological Conditions

The validity of the flow cytometry method using Fluo-

Zin-3, TMRE, and mBcl for the determination of viable beta

cells was then tested under pathological conditions. To this

end, human islets were either treated with 100 lM H2O2, a

physiological oxidant (42) or cultured in acidic media over-

night. Treated cells were then stained with FluoZin-3, TMRE,

and mBcl simultaneously. Both of these treatments substan-

tially reduced the frequency of viable beta cells-FluoZin-3-

gated, TMRE1 1 mBcl1 cells (Fig. 7A vs B and C; upper right

quadrants).

Treatment of NIT-1 cells with H2O2 induced more cell

death than exposure to extracellular acidosis as indicated by

the reduction in live cells (TMRE1 1 mBcl1, Fig. 7D vs. E

and F, upper right quadrants). However, the frequency of apo-

ptotic cells as indicated by the incorporation of FITC-conju-

gated pancaspse inhibitor VAD-FMK (30) was higher in cells

that were exposed to extracellular acidosis than those treated

with H2O2 (Fig. 7G vs. H and I). This could be due to the fact

that H2O2 may predominantly trigger caspase-independent

cell death in insulinoma cells. Treatment with BAPTA that pre-

vents intracellular Ca21 mobilization (45) or with the inhibi-

tor of phosphatidyl inositol 3 kinase, wortmannin (46)

induced the activation of intracellular caspases and reduced

the viability of insulinoma cells concurrently (manuscript

under preparation).

DISCUSSION

The data presented herein demonstrate that viable beta

cells can be assessed by using a combination of three fluores-

cent dyes with unique characteristics: the zinc-binding dye

FluoZin-3, the mitochondrial potential indicator, TMRE, and

the indicator of intracellular thiol, mBcl. Although TMRE

(28,34) and mBcl (29) have been used respectively to assay

DCm and GSH in various cell types, simultaneous monitoring

of these parameters in conjunction with the detection of labile

Zn21 is novel and has implications for the enumeration of live

beta cells in clinical islet preparations.

FluoZin-3 has several advantages over the zinc-binding

dye Newport Green, used previously for the determination of

beta cells in human islet preparations (23,47). Although a

direct comparison between Newport Green and Fluozin-3 is

beyond the scope the current study, several recent publications

Figure 5. Comparison of the frequency of viable beta cells between preparations of variable viability. Dot plot of dispersed human islet

cells with high viability is shown in A. These cells were stained with FluoZin-3, TMRE, and mBcl and then analyzed for the frequency of

beta (FluoZin-31) cells (B). Further analysis was done to estimate the frequency of viable beta cells, FluoZin-31 cells coexpressing TMRE1

1mBcl1 (upper right quadrant, C). Cells that were permeable to 7-AAD were eliminated electronically (D) and further analyzed for FluoZin-

3 fluorescence (E) and for costaining with TMRE and mBcl (F). The same islet preparation was cultured for another 2 days and then ana-

lyzed for the presence of intact cells (G) as well as for FluoZin-31 cells (H) coexpressing TMRE and mBcl (upper right quadrant, I). In addi-

tion, cells were analyzed after the addition of 7-AAD to eliminate dead/dying cells (J). Both the frequency (K) and the viability (TMRE1 1mBcl1) of beta cells were determined among cells that excluded 7-AAD (L). Representative data from six independent experiments are

shown.

ORIGINAL ARTICLE


indicate that Newport Green has severe drawbacks and there-

fore cannot be reliably used for the detection of beta cells.

These include low quantum yield, poor selectivity to Zn21

(22,25) and nonspecific binding to peroxynitrite (27). In con-

trast, FluoZin-3 binds zinc with higher affinity, has high quan-

tum yield (22,24–26,48) and does not bind peroxynitrite (27).

The data presented herein demonstrate for the first time that

FluoZin-3 can be used to distinguish live beta cells from

other types of islet cells and from lymphomas, leukocytes

and fibroblasts (Fig. 1B). This is because beta cells store

larger amounts of labile Zn21 than other cell types in the

body, except hippocampal cells and prostate cells (6–9).

Moreover, FluoZin-3 binding is competitively inhibited by

the zinc chelator TPEN in intact beta cells (Fig. 2A), indicat-

ing the selectivity of FluoZin-3 binding to labile Zn21, as

shown in neuronal cells (25).

We have recently shown that the combination of Fluo-

Zin-3 and TMRE allowed the determination of live beta cells

in a large number of ([35) human islet preparations (S.

Jayaraman et al., 2008, manuscript under revision). Although

TMRE is a good indicator of the inner mitochondrial mem-

brane potential in cells including insulinomas (28,34), it may

not serve as the sole indicator of beta cell viability under all

conditions. This is because beta cells, like other cell types (37–

38), may not dissipate DCm under certain conditions when

undergoing cell death. In addition, the loss of DCm may not

always correlate with cell death. This was indicated by the lack

of beta cell death even when the mitochondria were depolar-

ized (Fig. 6) following treatment with CCCP, a potent uncou-

pler of mitochondrial oxidative phosphorylation (44). There-

fore, it is important to consider other markers of cell viability

that are independent of mitochondrial membrane potential.

Assessment of GSH using mBcl provides an additional

and important parameter of beta cell viability. GSH is crucial

for cell survival, since it is a key antioxidant that maintains

protein thiols in a reduced state and scavenges H2O2 in a reac-

tion catalyzed by glutathione peroxidase (39–42). Since thiol

oxidation induces the loss of ATP/ADP exchange activity and

opening of the permeability transition pore leading to cyto-

chrome c release and cell death (41–42), GSH depletion

primes cells for death by toxic stimuli. Although decreased in-

tracellular and extracellular GSH levels have been implicated

in pathological conditions, the ramifications of depleting in-

tracellular GSH have not been elucidated in beta cells. Expo-

sure to extracellular acidosis and H2O2 leads to the collapse of

DCm and reduction in GSH levels concurrently in human

islets and in insulinoma cells (Fig. 7). These data support the

contention that simultaneous assessment of DCm and GSH

levels can provide an accurate estimation of live beta cells

rather than based on DCm alone. Insofar as the levels of GSH

were not reduced in parallel with the dissipation of DCm fol-

lowing CCCP treatment (Fig. 6), GSH levels may provide an

indication of beta cell viability independent of the functional

status of mitochondria. The data presented herein strongly

support this contention.

Another distinct feature unraveled in the current study is

that the measurement of beta cell viability using FluoZin-3,

TMRE, and mBcl obviates the need for 7-AAD to eliminate

dead cells. Although it is a common practice to use 7-AAD to

eliminate dead/dying cells (43), including beta cells that had

been stained with TMRE (47), the data presented herein

demonstrate that 7-AAD is not compatible with TMRE for the

Figure 6. Mitochondrial membrane depolarization did not impair

the binding efficiency of FluoZin-3 and mBcl. Dispersed human is-

let cells were stained with FluoZin-3, TMRE, and mBcl (A–C). Analiquot of cells were also preincubated with the protonophore

CCCP for 15 min and then stained with FluoZin-3, TMRE, and

mBcl in the continued presence of CCCP for another 30 min (E–G).FluoZin-3 binding cells (B and F) were gated and analyzed for

costaining with TMRE and mBcl (upper right quadrants, C and G).

Control and CCCP-treated cells were also incubated separately

with 7-AAD to determine cells with compromised plasma mem-

brane integrity (D and H). Representative data from four inde-

pendent experiments are shown.

ORIGINAL ARTICLE


determination of viability in beta cells. Simultaneous staining

with TMRE and 7-AAD unexpectedly increased the frequency

of TMRE1 live beta cells that also displayed permeability to 7-

AAD with a concomitant decrease in cells with polarized mito-

chondria (TMRE1 cells, Fig. 6). Additionally, exclusion of 7-

AAD1 cells aberrantly increased the frequency and viability of

beta cells in islet preparations that contained considerable

numbers of nonviable cells (Fig. 5). This was not simply due

to improper color compensation but due to unexpected inter-

action between TMRE and 7-AAD not reported earlier. As

demonstrated herein, the alternative solution is to use the mi-

tochondrial membrane potential sensitive dye TMRE along

with an indicator of the antioxidant GSH, mBcl. The data pre-

sented herein clearly demonstrate that viability determination

based on triple staining provides a useful strategy for the accu-

rate estimation of viability and function of beta cells under

various pathological conditions, which include progressive

loss of viability during in vitro culture (Fig. 5), mitochondrial

membrane depolarization (Fig. 6), H2O2 treatment, and expo-

sure to extracellular acidosis (Fig. 7). These results strongly

support the conclusion that staining with FluoZin-3, TMRE,

and mBcl is a useful strategy to estimate the viability of

human beta cells without the disadvantage of using 7-AAD.

In addition to dispersed islet cells, whole live islets are

amenable for analysis by confocal microscopy after staining

with FluoZin-3, TMRE, and mBcl (manuscript in prepara-

tion). Previous studies did not directly provide evidence for

the binding of Newport Green to beta cells (23,47) by perme-

abilization and staining of Newport Green1 cells with an anti-

body specific to insulin. The sensitivity of Newport Green to

aldehydes would not allow such an analysis. Since FluoZin-3 is

fairly resistant to aldehyde treatment (Fig. 2), it was possible

to visualize labile Zn21 and intracellular insulin in human

beta cells (manuscript in preparation).

The flow cytometry based method described herein is

simple to perform and provides quantitative information on

the proportion of viable beta cells in clinical islet preparations.

This novel method can provide a reliable estimate of func-

tional beta cells in clinical islet preparations prior to trans-

plantation into type 1 diabetes patients. Current methods of

beta cell viability assessment include injection of islets into

immunodeficient mice made diabetic by streptozotocin treat-

Figure 7. Determination of beta cell viability under various experimental conditions. HP: Control human islet cells (A) and those treated

with H2O2 (B) or cultured in pH 5.9 media (C) were stained with FluoZin-3, TMRE, and mBcl. FluoZin-31 cells were gated and analyzed for

costaining with TMRE and mBcl. NIT-1: Control mouse insulinoma cells (D) and those treated with H2O2 (E) or acidic media (F) were exam-

ined for viability after staining with TMRE and mBcl. Intracellular activation of caspases was assessed in control NIT-1 cells (G) and those

induced to undergo apoptosis by H2O2 treatment (H) or exposure to extracellular acidosis (I). PI was added to cells incubated with FITC-

VAD-FMK in order to distinguish late apoptotic cells (upper right quadrant, VAD1 1 PI1) from dead cells (upper left quadrant, PI1). Data

shown are representative of three to five experiments.

ORIGINAL ARTICLE


ment and assessment of diabetes reversal by transplanted

islets. However, this method is technically challenging and

requires several days for completion. Although the flow cyto-

metry method described herein cannot completely replace the

in vivo assessment of islet function, it can nevertheless provide

more accurate information on the viability of beta cells within

a short period of time.

ACKNOWLEDGMENTS

The gift of human islets and the insulinoma cell lines by

several investigators is gratefully acknowledged.

LITERATURE CITED

1. Ricordi C, Lacy PE, Scharp DW. Automated islet isolation from human pancreas.Diabetes 1989;38(Suppl 1):140–142.

2. Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, Warnock GL, Kneteman NM,Rajotte RV. Islet transplantation in seven patients with type 1 diabetes mellitus usinga glucocorticoid-free immunosuppressive regimen. N Engl J Med 2000;343:230–238.

3. Hering BJ, Kandaswamy R, Harmon JV, Ansite JD, Clemmings SM, Sakai T,Paraskevas S, Eckman PM, Sageshima J, Nakano M, Sawada T, Matsumoto I, ZhangHJ, Sutherland DE, Bluestone JA. Transplantation of cultured islets from two-layerpreserved pancreases in type 1 diabetes with anti-CD3 antibody. Am J Transplant2004;4:390–401.

4. Ault A. Edmonton’s islet success tough to duplicate elsewhere. Lancet 2003;361:2054.

5. Vallee BL, Falchuk KH. The chemical basis of Zn21 physiology. Physiol Rev1993;73:79–118.

6. Berg JM, Shi Y. The galvanization of biology: A growing appreciation for the roles ofZn21. Science 1996;271:1081–1085.

7. Chung S-H, Johnson MS. Divalent transition-metal (Cu21 and Zn21) in the brainsof epiletogenic and normal mice. Brain Res 1983;280:323–334.

8. Canzoniero LMT, Sensi, S, Choi DW. Measurement of intracellular free Zn21 inliving neurons. Neurobiol Dis 1997;4:275–279.

9. Formby B, Schmid-Formby F, Grodsky GM. Relationship between insulin release andzinc efflux from rat pancreatic islets maintained in tissue culture. Endocrinology1984;33:229–234.

10. Grodsky GM, Schmid-Formby F. Kinetic and quantitative relationships between insu-lin release and 65Zn efflux from perifused islets. Endocrinology 1985;117:704–710.

11. Aspinwall CA, Brooks SA, Kennedy RT, Lakey JR. Effects of intravesicular H1 andextracellular H1 and Zn21 on insulin secretion in pancreatic beta cells. J Biol Chem1997;272:31308–31314.

12. Qian WJ, Aspinwall CA, Battiste MA, Kennedy RT. Detection of secretion from singlepancreatic beta-cells using extracellular fluorogenic reactions and confocal fluores-cence microscopy. Anal Chem 2000;72:711–717.

13. Figlewicz DP, Hodgson AT, Schmid FG, Grodsky GM. Kinetics of 65zinc uptake anddistribution in fractions from cultured rat islets of Langerhans. Endocrinology1980;29:767–773.

14. Zalewski PD, Forbes I, Betts W, Mahadevan I. Correlation of apoptosis with changein intracellular labile Zn[II] using zinquin [(2-methyl-8-p-toluenesulphoamido-6-quinolyloxy)lacetic acid], a new specific fluorescent probe for Zn[II]. Biochem J1993;296:403–408.

15. Frederickson C. Imaging zinc: Old and new tools. Sci STKE 2003;182:1–4.

16. Latif Z, Noel J, Alejandro R. A simple method of staining fresh and cultured islets.Transplantation 1988;45:827–830.

17. Budde T, Minta A, White JA, Kay AR. Imaging free Zn21 in synaptic terminals inlive hippocapal slices. Neuroscience 1997;79:347–358.

18. Frederickson CJ, Kasarkis EJ, Ringo D, Frederickson RE. A quinoline fluorescencemethod for visualizing and assaying the histochemically reactive Zn21 (boutonZn21) in the brain. J Neurosci Methods 1987;20:91–103.

19. Pearce DA, Jotterand H, Carrico IS, Imperiali B. Derivates of 8-hydroxy-2-methylqui-noline are powerful prototypes for Zn21 sensors in biological systems. J Am ChemSoc 2001;123:5160–5161.

20. Sensi SI, Canzoniero LMT, Yu SP, Ying HS, Koh J-Y, Kerchner GA, Choi DW.Measurement of intracellular free Zn21 in living cortical neurons: Routes of entry.J Neurosci 1997;17:9554–9564.

21. Kuhn MA, Haugland RP, Hoyland BM. Methods of sensing with fluorescent conju-gates of metal-chelating nitrogen heterocytes. US Patent Number 5,648,270 (1997).

22. Haugland RP. Handbook of Fluorescent Probes and Research Products. Eugene, Ore-gon: Molecular Probes; 2001.

23. Lukowiak B, Vandewalle B, Riachy R, Kerr-Conte J, Gmyr V, Belaich S, Lefebvre J,Pattou F. Identification and purification of functional human beta-cells by a new spe-cific zinc-fluorescent probe. J Histochem Cytochem 2001;49:519–528.

24. Gee KR, Zhou ZL, Qian WJ, Kennedy R. Detection and imaging of zinc secretionfrom pancreatic beta-cells using a new fluorescent zinc indicator. J Am Chem Soc2002;124:776–778.

25. Gee KR, Zhou ZL, Ton-That D, Sensi SL, Weiss JH. Measuring zinc in living cells. Anew generation of sensitive and selective fluorescent probes. Cell Calcium 2002;31:245–251.

26. Qian W-J, Gee KR, Kennedy RT. Imaging of Zn21 released from pancreatic b-cells atthe level of single exocytotic events. Anal Chem 2003;75:3468–3475.

27. Zhang Y, Wang H, Li J, Jimenez DA, Levitan ES, Aizenman E, Rosenberg PA. Peroxy-nitrite-induced neuronal apoptosis is mediated by intracellular zinc release and 12-lipoxygenase activation. J Neurosci 2004;24:10616–10627.

28. Jayaraman S. Flow cytometric determination of mitochondrial membrane potentialchanges during apoptosis of T lymphocytes and pancreatic beta cells. Comparison oftetramethylrhodamineethylester (TMRE), chloromethyl-X-rosamine (H2-CMX-Ros)and MitoTracker Red580 (MTR580). J Immunol Methods 2005;306:68–79.

29. Hedley DW, Chow S. Evaluation of methods for measuring cellular glutathione con-tent using flow cytometry. Cytometry 1994;15:349–358.

30. Jayaraman S. Intracellular determination of activated caspases (IDAC) by flow cyto-metry using a pancaspase inhibitor labeled with FITC. Cytometry Part A 2003;56A:104–112.

31. Baetens D, Malaisse-Lagae F, Perrelet A, Orci L. Endocrine pancreas: Three-dimen-sional reconstruction shows two types of islets of langerhans. Science 1979;206:1323–1325.

32. Arslan P, Di Virgilio F, Beltrame M, Tsien RY, Pozzan T. Cytosolic Ca21 homeostasisin Ehrlich and Yoshida carcinomas: A new, membrane-permeant chelator of heavymetals reveals that these ascites tumor cell lines have normal cytosolic free Ca21.J Biol Chem 1985;260:2719–2727.

33. Wiederkehr A, Wollheim CB. Minireview: Implication of mitochondria in insulinsecretion and action. Endocrinology 2006;147:2643–2649.

34. Ehrenberg B, Montana V, Wei MD, Wuskell JP, Loew LM. Membrane potential can bedetermined in individual cells from the nernstian distribution of cationic dyes. Bio-phys J 1988;53:785–794.

35. Ward MW, Huber HJ, Weisova P, Dussmann H, Nicholls DG, Prehn JH. Mitochon-drial and plasma membrane potential of cultured cerebellar neurons during gluta-mate-induced necrosis, apoptosis, and tolerance. J Neurosci 2007;27:8238–8249.

36. Halestrap AP, Doran E, Gillespie JP, O’Toole A. The mitochondrial permeability tran-sition: Its molecular mechanism and role in reperfusion injury. Biochem Soc Symp1999;66:181–203.

37. Finucane DM, Waterhouse NJ, Amarante-Mendes GP, Cotter TG, Green DR. Col-lapse of the inner mitochondrial transmembrane potential is not required for apo-ptosis of HL60 cells. Exp Cell Res 1999;251:166–174.

38. Krohn AJ, Wahlbrink T, Prehn JHM. Mitochondrial depolarization is not requiredfor neuronal apoptosis. J Neurosci 1999;19:7394–7404.

39. Hall AG. The role of glutathione in the regulation of apoptosis. Eur J Clin Invest1999;29:238–245.

40. Mytilineou C, Kramer BC, Yabut JA. Glutathione depletion and oxidative stress.Parkinson Relat Disord 2002;8:385–387.

41. Ueda S, Masutani H, Nakamura H, Tanaka T, Ueno M, Yodoi J. Redox control of celldeath. Antioxid Redox Signal 2002;4:405–414.

42. Rahman I, Biswas SK, Jimenez LA, Torres JM, Forman HJ. Glutathione, stressresponses, and redox signaling in lung inflammation. Antioxid Redox Signal 2005;7:42–59.

43. Schmid I, Krall WJ, Uittenbogaart CH, Braun J, Giorgi JV. Dead cell discriminationwith 7-amino-actinomycin D in combination with dual color immunofluorescencein single laser flow cytometry. Cytometry 1992;13:204–208.

44. Ales E, Fuentealba J, Garcıa AG, Lopez MG. Depolarization evokes different patternsof calcium signals and exocytosis in bovine and mouse chromaffin cells: The role ofmitochondria. Eur J Neurosci 2005;21:142–145.

45. Tsien RY. New calcium indicators and buffers with high selectivity against magnesiumand protons: Design, synthesis, and properties of prototype structures. Biochemistry1980;19:2396–2404.

46. Arcaro A, Wymann MP. Wortmannin is a potent phosphatidylinositol 3-kinase inhib-itor: the role of phosphatidylinositol 3,4,5-trisphosphate in neutrophil responses.Biochem J 1993;296:297–301.

47. Ichii H, Inverardi L, Pileggi A, Damaris Molano R, Cabrera O, Caicedo A, MessingerS, Kuroda Y, Berggren P-O, Ricordi C. A novel method for the assessment of cellularcomposition and beta-cell viability in human islet preparations. Am J Transplant2005;5:1635–1645.

48. Devinney II MJ, Reynolds IJ, Dineley KE. Simultaneous detection of intracellular freecalcium and zinc using fura-2FF and FluoZin-3. Cell Calcium 2005;37:225–232.

ORIGINAL ARTICLE


Documents

Cytometry