Visualization and isolation of langerhans islets by a fluorescent probe PiY

Article abstract of DOI:10.1002/anie.201302149

Lighting up the pancreas: A pancreatic beta cell selective probe PiY (red, see picture) was developed which stains Langerhans islets in live animals without toxicity or side effects. Through tail-vein injection, PiY allowed for a comparison of the islets between healthy and diabetic mice. PiY also facilitated the isolation of healthy Langerhans islets by using a fluorescence-guided surgical procedure. Copyright

Full text of DOI:10.1002/anie.201302149

Angewandte
Chemie
Communications
Fluorescent Probes
Lighting up the pancreas: A pancreatic
beta cell selective probe PiY (green, see
picture) was developed, which stains
Langerhans islets in live animals without
toxicity or side effects. Through tail-vein
injection, PiY allowed for a comparison of
the islets between healthy and diabetic
mice. PiY also facilitated the isolation of
healthy Langerhans islets using a fluores-
cence-guided surgical procedure.
N.-Y. Kang, S.-C. Lee, S.-J. Park, H.-H. Ha,
S.-W. Yun, E. Kostromina, N. Gustavsson,
Y. Ali, Y. Chandran, H. S. Chun, M. A. Bae,
J. H. Ahn, W. Han, G. K. Radda,
Y.-T. Chang*
&&&& — &&&&
Visualization and Isolation of Langerhans
Islets by a Fluorescent Probe PiY
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
DOI: 10.1002/anie.201302149
Fluorescent Probes
Visualization and Isolation of Langerhans Islets by a Fluorescent Probe
PiY**
Nam-Young Kang, Sung-Chan Lee, Sung-Jin Park, Hyung-Ho Ha, Seong-Wook Yun,
Elena Kostromina, Natalia Gustavsson, Yusuf Ali, Yogeswari Chandran, Hang-Suk Chun,
MyungAe Bae, Jin Hee Ahn, Weiping Han, George K. Radda, and Young-Tae Chang*
Pancreatic Langerhans islets are mainly composed of insulin-
secreting beta cells and glucagon-secreting alpha cells, along
with other minor cell types, and play a central role in the
regulation of blood glucose levels.[1] Because of this, imaging
of viable pancreatic islets is an important component in
research on diabetes both in clinical and experimental
medicine.[2–4] The conventional imaging technique for pan-
creatic islets is antibody-based immunostaining directly on
pancreatic sections,[5] or using transgenic mice with lumines-
cent reporter genes linked to islet-specific promoters.[6]
Among small molecule probes, Newport Green[7] and dithia-
zone (DTZ)[8] have been used for ex vivo fluorescent staining
of pancreatic islets, based on their Zn2+ ion binding affinity,
which are abundant in beta cells in complex with insulin. For
in situ application, fluorescently labeled exendin-4 (a GLP-
1R binding peptide: M.W. is about 5 kDa) has been recently
introduced for the measurement of the mass of pancreatic
islet beta cells.[9] However, small molecule probes for selective
staining of beta cells in pancreatic islets of live animals have
not yet been reported.
We predicted that a diversity-oriented fluorescence
library approach (DOFLA), an expedited bioimaging probe
discovery method using high throughput synthesis and high
contents screening, would be a powerful method to achieve
this goal. Using a similar approach, we have previously
elucidated probes for pluripotent stem cells (CDy1),[10]
muscle cells (CDy2),[11] neuronal stem cells (CDr3),[12] and
pancreatic alpha cells (GY= glucagon yellow).[13]
Although the glucagon-targeting probe GY selectively
stains alpha cells in isolated cell culture, it did not clearly
mark mouse pancreatic islets in tissue, partially owing to the
small population of alpha cells (around 15–20% in mouse
islets).[14] We expected that a fluorescent probe for pancreatic
beta cells (with a larger population of 75–80% in mouse
islets)[15] would be more effective for visualizing pancreatic
islets. As a first step, we synthesized fluorescent small-
molecule libraries composed of 1200 compounds,[16] and
screened them against beta TC-6 cells in comparison to
alpha TC-1 cells and acinar cells (exocrine cells in the
pancreas) as controls.
The three cell types were compared in 384-well plates and
incubated with the library compounds (1 mm) at incubation
times ranging from one to 48 hours. The fluorescence live-cell
images were acquired by an automated imaging microscope
system, ImageXpress Micro.
[*] Dr. N.-Y. Kang, Dr. S.-C. Lee, Dr. S.-J. Park, Dr. S.-W. Yun,
Dr. E. Kostromina, Dr. N. Gustavsson, Dr. Y. Ali, Y. Chandran,
Dr. W. Han, Dr. G. K. Radda, Prof. Y.-T. Chang
Singapore Bioimaging Consortium, Agency for Science,
Technology and Research (A*STAR)
One compound from the BDNCA series (Scheme 1),
BDNCA-325 (labs/lem = 558/585 nm, extinction cooefficient
e = 58,000mꢀ1 cmꢀ1, quantum yield F = 0.06) was chosen as
the most selective for the beta TC-6 cells (Figure 1a) in
comparison to the two control cell types in terms of relative
fluorescence intensity. The BDNCA library was prepared
from a BODIPY-aniline (BDN) series by chloroacetylation.
While BDN has very low fluorescence emission (less than 1%
quantum yield) owing to photoinduced electron transfer
(PET), by converting the amine to an amide, the fluorescence
of BDNCA was moderately increased (F = 5–10%). There-
fore, this amide motif is a modulator of the fluorescence
intensity of the BDNCA series through interaction with the
surrounding environment or binding partner.
138667, Singapore (Singapore)
E-mail: chmcyt@nus.edu.sg
Homepage: http://ytchang.science.nus.edu.sg
Prof. Y.-T. Chang
Department of Chemistry and MedChem Program of Life Sciences
Institute, National University of Singapore
117543 Singapore (Singapore)
Prof. H.-H. Ha
Department of Pharmacy, Sunchon National University
Sunchon (Korea)
Dr. H. S. Chun, Dr. M. A. Bae, Dr. J. H. Ahn
Drug Discovery Division, Korea Research Institute of Chemical
Technology
Yuseong-Gu, Daejeon, 305-600 (Korea)
[**] We thank Siti Hajar and Chew Yan Tuang for excellent technical
support in cell culture and screening, and Clement Khaw (SBIC-
Nikon Imaging Centre) for confocal microscopy. This study was
supported by intramural funding from A*STAR (Agency for Science,
Technology and Research, Singapore) Biomedical Research Council,
and a Ministry Of Education MOE2010-T2-1-025 grant to Y.-T.C.
from National University of Singapore.
However, when we injected BDNCA-325 intravenously
into a mouse, the pancreatic islets were not selectively stained
at various incubation times and concentrations (data not
shown). Because BDNCA-325 contains a chemically reactive
chloroacetyl group, we hypothesized that the compound
might have reacted with other tissues in the animal before
reaching the pancreatic islets. Thus, BDNCA-325 was modi-
fied by removing the reactive alpha-chloride and also by
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201302149.
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
tion was the optimal staining time with the lowest background
signal and highest contrast (Supporting Information, Fig-
ure S1a). On the other hand, the current standard dye for
pancreatic islet staining in vitro, Newport Green DCF diac-
etate did not show selective staining in a similar injection
experiment (Supporting Information, Figure S1b). To con-
firm the beta cell selectivity of PiY, we sorted out PiY stained
cells from the islets using FACS and analyzed the mRNA
expression level of insulin, which is a marker for beta cells.
The top 15% of the PiYBri and PiYDim populations were
isolated and analyzed for their relative mRNA levels by real-
time PCR, which clearly showed that PiYBri cells were highly
enriched with beta cells compared to PiYDim cells (Supporting
Information, Figure S2). This is the first demonstration of
a fluorescent small-molecule probe that can selectively stain
pancreatic islets by injection into live animals.
Measurement of the mass of healthy beta cells is an
important index for diagnosing diabetes. PiY staining clearly
showed fewer islets (also with lower intensity) in the
streptozotocin (STZ)-treated type 1 diabetic mouse model
compared to healthy mouse islets (Figure 2c). There was
a good correlation between PiY staining and insulin antibody
immunostaining (Figure 2a,b). This PiY staining method is
much simpler and faster than insulin-antibody-based immu-
nostaining, which requires several steps of sample handling
and thus is not practical for statistical analysis.
Scheme 1. a) MeMgBr, THF, ꢀ788C!258C, overnight; b) NaBH4,
08C!258C, THF/MeOH (10:1), 2 h; c) 2,4-dimethyl pyrrole, InCl3,
CH2Cl2; d) 1) DDQ in CH2Cl2, 2) BF3-Et2O, TEA; e) Pd/C, hydrazine
monohydrate, EtOH, reflux, 2 h; f) 2-chlorotrityl resin, pyridine; g) pyr-
rolidine, RCHO*, 300 W microwave, 2 minutes NMP/nBuOH (4:1);
h) 0.5% TFA in CH2Cl2; i) ClAcCl, NaHCO3, acetonitrile/CH2Cl2 (1:1);
THF=tetrahydrofuran, DDQ=2,3-dichloro-5,6-dicyanobenzoquinone,
TEA=triethylamine, NMP=N-methyl-2-pyrrolidone, TFA=trifluoroace-
tic acid, *RCHO is from commercially available number 237 of
aldehyde building block (Supporting Information, Table S1).
changing the chain length of the acryl group. Among the
various derivatives of BDNCA-325 tested, the acetyl deriv-
ative showed clear pancreatic islet staining upon tail-vein
injection (Figure 1b), and thus was named pancreatic islet
yellow (PiY; labs/lem = 558/585 nm, e = 56,300mꢀ1 cmꢀ1, F =
0.05). PiY also could be used in vivo for staining the islets
of Langerhans, by showing the three-dimensional structures
along with staining of the encircled blood vessel with FITC-
dextran (Figure 1c). By incubating PiY for different times
following a 50 mm injection, we found that one hour incuba-
Finally, we applied PiY staining to harvest pancreatic
islets from the total pancreas. The isolation of pancreatic islets
usually requires a high level of technical experience and
labor-intensive methods, partly because of the difficulty in
visual identification of the islets. In situ staining of the islets
by PiY made visual inspection of the islets extremely easy and
the isolation of islets using the fluorescence surgical proce-
dure was feasible, even for a minimally trained researcher
(Supporting Information, Figure S3a). To confirm the safety
Figure 1. Visualizing pancreatic islets using PiY staining in vivo. a) Chemical formula of BDNCA-325 and the fluorescent cell culture images of
staining by BDNCA-325 and Hoechst 33342 (nuclear stain). Upper panels: fluorescent images using a TRITC filter for BDNCA-325; lower panels:
fluorescent images using a DAPI filter for Hoechst 33342. Scale bar=10 mm. b) Chemical formula of PiY and pancreas tissue section images of
PiY stained islets. 50 mm of PiY was administrated to the mouse tail vein for one hour. Scale bar=200 mm. c) 3D structure processed image of
a pancreatic islet. The fluorescent images of PiY (red) and FITC-dextran (green) were taken using TRITC and FITC channels, respectively. DAPI
(blue) stain shows nuclei. Scale bar=50 mm.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Figure 2. Selective staining of healthy islets and in vivo toxicity study of PiY. a) Fluorescence images of a PiY stained healthy islet using a TRITC
filter. Subsequently, the same islet images were confirmed by immunostaining with an insulin antibody using a Cy5 filter. Scale bar=20 mm.
b) Fluorescence images of PiY stained STZ-induced mouse islet. 150 mgkgꢀ1 of streptozotocin was injected as a single dose intraperitoneally and
monitored for 9 days. Subsequently, PiY was injected for one hour. The same PiY stained islet images were confirmed by immunostaining with an
insulin antibody. c) Total number of PiY stained islets and sum intensity of PiY stained islets. n=3 for both CON and STZ mice with 300 sections
per mouse (p<0.01). d) Change of mouse body weight and blood glucose level during 28 days of PiY injection (n=1). Blood was taken from the
mouse tail and tested using a blood glucose meter (Lifescan Inc., USA).
of PiY as an imaging probe for functional islets, we further
tested the effect of the probe on the islets. We tested for
insulin release from the isolated islets and PiY did not affect
the amount of insulin secreted upon glucose stimulation
compared to the empty vehicle (Supporting Information,
Figure S3b). In addition, there was no detectable change by
PiYon body weight or blood glucose levels of mice for at least
a month following PiY injection at various concentrations,
even at high doses up to 150 mm of the probe (Figure 2d).
To examine the effect of the probe in different species, we
further tested PiY in rats by tail-vein injection, and a selective
staining of pancreatic islets was clearly demonstrated after
one and four hours incubation (Supporting Information,
Figure S4). With the promising data from the rats, we carried
out a pharmacokinetic study on rats and all the parameters
are summarized in the Supporting Information, Figure S5a.
When injected, almost all PiY existed as a bound form in
human and rat blood plasma (99.4 ꢁ 0.3 and 99.7 ꢁ 0.1%).
The plasma concentration of PiY reached a maximum at five
hours and slowly decreased with a half-life t1/2 of 15 hours
(Supporting Information, Figure S5b). At 24 hours after
injection of PiY in the rat tail vein, we collected pancreas,
lung, liver, and brain tissues and measured the concentration
of residual PiY. While the concentration in the brain was very
low, lung and liver samples showed higher PiY signal than
pancreas (Supporting Information, Figure S5c). The trend
was similar for a shorter time point (0.5 hours after tail vein
injection), showing that the PiY concentration was higher in
the liver, lungs, and heart than in the pancreas and even lower
in the stomach, intestines, brain, and testes (Supporting
Information, Figure S6). This data showed that PiY is
selective for islet staining among pancreatic cells as designed
for the original screening (alpha, beta, and acinar cells), but
not for other organ cells. For intact animal imaging, PiY could
be used only when combined with another anatomical
modality of imaging such as MRI or X-ray CT.
Finally, the elucidation of the cellular target protein of
PiY would be extremely useful for understanding the
molecular mechanism of the selective staining of beta cells
by PiY. We tried most of the conventional target identifica-
tion techniques including an affinity matrix and activity-based
protein profiling, but could not convincingly identify the
molecular target. Further studies with more sophisticated
techniques are underway to understand the unique binding
properties of PiY.
In conclusion, we developed the first fluorescent small
molecule probe PiY for live-animal pancreatic islet staining.
Through a simple tail-vein injection, PiY stains pancreatic
islets without toxicity or functional disturbance of the beta
cells, demonstrating its potential to quantitatively measure
healthy islets and as a marker for facile islet surgical
separation. In situ islet imaging using PiY will be important
not only for the study of diabetes and the measurement of
healthy islets, but also to further investigate beta cell
development and islet transplantation.
Experimental Section
Injection of PiY for both in vivo staining and isolation of islets was
carried out at 50 mm in the mouse tail vain for one hour. A stock
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
solution in DMSO of PiY (1 mm) was dissolved in PBS with 1%
poly(ethylene glycol) (M.W. = 4600 Da) and 0.1% Tween 20 to make
a 50 mm PiY injection solution. 250 mL of PiY injection solution was
administrated.
[3] A. Hçrnblad, U. Ahlgren, Islets 2009, 1, 163 – 164.
[4] U. Ahlgren, M. Gotthardt, Adv. Exp. Med. Biol. 2010, 654, 39 –
57.
[5] Y. B. Ahn, G. Xu, L. Marselli, E. Toschi, A. Sharma, S. Bonner-
Weir, D. C. Sgroi, G. C. Weir, Diabetologia 2007, 50, 334 – 342.
[6] B. Lukowiak, B. Vandewalle, R. Riachy, J. Kerr-Conte, V. Gmyr,
S. Belaich, J. Lefebvre, F. Pattou, J. Histochem. Cytochem. 2001,
49, 519 – 528.
[7] J. I. Conget, Y. Sarri, J. M. Gonzalez-Clemente, R. Casamitjana,
M. Vives, R. Gomis, Pancreas 1994, 9, 157 – 160.
[8] S. Speier, D. Nyqvist, O. Cabrera, J. Yu, R. D. Molano, A. Pileggi,
T. Moede, M. Kohler, J. Wilbertz, B. Leibiger, C. Ricordi, I. B.
Leibiger, A. Caicedo, P. O. Berggren, Nat. Med. 2008, 14, 574 –
578.
[9] T. Reiner, G. Thurber, J. Gaglia, C. Vinegoni, C. W. Liew, R.
Upadhyay, R. H. Kohler, L. Li, R. N. Kulkarni, C. Benoist, D.
Mathis, R. Weissleder, Proc. Natl. Acad. Sci. USA 2011, 108,
12815 – 12820.
[10] C. N. Im, N. Y. Kang, H. H. Ha, X. Bi, J. J. Lee, S. J. Park, S. Y.
Lee, M. Vendrell, Y. K. Kim, J. S. Lee, J. Li, Y. H. Ahn, B. Feng,
H. H. Ng, S. W. Yun, Y. T. Chang, Angew. Chem. 2010, 122,
7659 – 7662; Angew. Chem. Int. Ed. 2010, 49, 7497 – 7500.
[11] Y. K. Kim, J. S. Lee, X. Bi, H. H. Ha, S. H. Ng, Y. H. Ahn, J. J.
Lee, B. K. Wagner, P. A. Clemons, Y. T. Chang, Angew. Chem.
2011, 123, 2813 – 2815; Angew. Chem. Int. Ed. 2011, 50, 2761 –
2763.
In vivo staining of PiY: After PiY injection, the dissected
pancreas was fixed using 4% paraformaldehyde (PFA) for three
hours at 48C and changed with fresh PBS for one hour at 48C.
Subsequently, the fixed pancreas was incubated overnight in 30%
sucrose at 48C. The following day, the pancreas was embedded in an
optimal cutting temperature compound for one hour at ꢀ808C and
followed by incubation in the cryo machine (set below ꢀ208C) for
one hour. The cryo-pancreas block was sectioned into 10 mm thick
slices. The sections on slides were air-dried at room temperature for
30 min and stored at 48C until use. To observe PiY-stained islets, the
pancreatic sections were washed with PBS for 5 minutes and the
sections were observed using fluorescent microscopy on the TRITC
channel for PiY.
Isolation of PiY-stained islets: Extracted pancreas was immedi-
ately transferred into a vial containing cold Krebs–Ringer–Hepes
(KRH) medium (130 mm NaCl, 4.7 mm KCl, 1.2 mm KH2PO4, 1.2 mm
MgSO4, 2.56 mm CaCl2, 1 mgmLꢀ1 BSA, 20 mm Hepes, pH 7.4) and
cut into smaller pieces. Collagenase-P (0.5 mgmLꢀ1; Roche) was
added and the vial was placed into a 378C water bath shaker (180–
200 rpm) for 15–20 minutes. After digestion, cold KRH buffer with
0.1% BSA was added and the vial was inverted several times. The
suspension was washed away after allowing the islets to settle to the
bottom of the vial for 10 minutes and this wash process was repeated
twice more. After picking islets from PiY and control solutions, they
were transferred to a new glass-bottom dish to observe using
fluorescent microscopy.
[12] S. W. Yun, C. Leong, D. Zhai, Y. L. Tan, L. Lim, X. Bi, J. J. Lee,
H. J. Kim, N. Y. Kang, S. H. Ng, L. W. Stanton, Y. T. Chang,
Proc. Natl. Acad. Sci. USA 2012, 109, 10214 – 10217.
[13] N. Y. Kang, H. H. Ha, S. W. Yun, Y. H. Yu, Y. T. Chang, Chem.
Soc. Rev. 2011, 40, 3613 – 3626.
[14] J. S. Lee, N. Y. Kang, Y. K. Kim, A. Samanta, S. Feng, H. K. Kim,
M. Vendrell, J. H. Park, Y. T. Chang, J. Am. Chem. Soc. 2009,
131, 10077 – 10082.
Received: March 14, 2013
Published online: && &&, &&&&
[15] M. Brissova, M. Shiota, W. E. Nicholson, M. Gannon, S. M.
Knobel, D. W. Piston, C. V. Wright, A. C. Powers, J. Biol. Chem.
2002, 277, 11225 – 11232.
[16] M. Vendrell, G. G. Krishna, K. K. Ghosh, D. Zhai, J. S. Lee, Q.
Zhu, Y. H. Yau, S. G. Shochat, H. Kim, J. Chung, Y. T. Chang,
Chem. Commun. 2011, 47, 8424 – 8426.
Keywords: bioimaging · fluorescent probes · imaging agents ·
Langerhans islets · pancreatic beta cells
.
[1] C. M. Benitez, W. R. Goodyer, S. K. Kim, Cold Spring Harbor
Perspect. Biol. 2012, 4, a012401.
[2] N. V. Evgenov, Z. Medarova, J. Pratt, P. Pantazopoulos, S.
Leyting, S. Bonner-Weir, A. Moore, Diabetes 2006, 55, 2419 –
2428.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!