Ytterbium-based bioprobes for near-infrared two-photon scanning laser microscopy imaging

Article abstract of DOI:10.1002/anie.201202212

Cleanly separated by two photons: The design and photophysical characterization of a new highly stable macrocyclic ytterbium complex featuring a two-photon antenna ligand is described (see picture). The biphotonic sensitization of the near-infrared ytterbium(III) luminescence and the conception of an unconventional near-infrared biphotonic microscope allow performing in-depth imaging of thick tissues. Copyright

Full text of DOI:10.1002/anie.201202212

.
Angewandte
Communications
DOI: 10.1002/anie.201202212
Bioimaging
Ytterbium-Based Bioprobes for Near-Infrared Two-Photon Scanning
Laser Microscopy Imaging**
Anthony DꢀAlꢁo, Adrien Bourdolle, Sophie Brustlein, Teddy Fauquier, Alexei Grichine,
Alain Duperray, Patrice L. Baldeck, Chantal Andraud,* Sophie Brasselet,* and Olivier Maury*
Dedicated to Dr. Hubert Le Bozec on the occasion of his 60th birthday
For decades, optical microscopy has been an essential tool for
biological imaging, and more recently luminescence-based
techniques have gained widespread use for medical analyses
and diagnostics.[1] Conventional one-photon microscopy using
commercial bioprobes or fluorescent proteins generally
proceeds using excitation wavelength in the UV or visible
and detection in the visible spectral range. These microscopy
configurations will be referred to as UV-to-visible or visible-
to-visible according to the excitation-to-detection spectral
ranges. Since biological tissues strongly absorb and scatter
UV/visible light, such configurations are restricted to surface
bioimaging experiments, for example, 2D cell imaging. On the
other hand, the transparency of biological tissues in the near-
infrared (NIR) between 700 and 1200 nm, a region called
biological window, allows in-depth imaging in this spectral
range.[2] Therefore, numerous academic and industrial
research endeavors are currently focused on the improvement
of microscopy techniques and on the design of new lumines-
cent bioprobes featuring both excitation and emission in this
NIR spectral range. Microscopy in this NIR-to-NIR config-
uration will enable in depth imaging in thick tissues and
several bioprobes (cyanine, (aza)-bodipy) combining NIR
excitation and emission have been developed and commer-
cialized this last decade.[3] However in these cases, the small
Stokes shift between the excitation and the optimal collection
range of emitted light is a real technical drawback for
microscopy because of the need to cleanly separate the
emission from the excitation. Nonlinear biphotonic excita-
tion, that is the simultaneous absorption of two photons of
half energy typically in the NIR region, inherently introduces
a larger Stokes shift and is therefore an elegant way to
circumvent this drawback.[4] However, up to now, all the
designed chromophores exhibit an emission in the visible
spectral range, and the currently available biphotonic micro-
scopes work in this two-photon NIR-to-visible configuration
with a detection wavelength shorter than the incident laser
wavelength.[4,5]
In this context, lanthanide complexes and particularly
NIR emitters like ytterbium and neodymium are known to
exhibit very large pseudo-Stokes shift,[6a,b] and are therefore
potentially well-suited for such two photon NIR-to-NIR
imaging purpose. In spite of their generally low luminescence
quantum yield, such complexes have already been used for
NIR bioimaging applications but with a classical one-photon
excitation generally localized in the UV/visible up to 550–
600 nm.[6,7] The sensitization of lanthanide luminescence by
two-photon absorption (TPA) is currently a challenging field
of research but so far, most endeavors were focused on
terbium and europium emitting in the green and red spectral
region, respectively.[6a,8] With regard to NIR emitters, the
proof of concept of ytterbium two-photon sensitization has
been first described in the early 2000s by Lakowicz and co-
workers.[9] Recently, Wong and co-workers have reported an
ytterbium complex that exhibit exceptional luminescence
properties in water with a remarkable two-photon cross-
section.[10] Interestingly, this complex was successfully used as
bioprobe to image HeLa cells by a two-photon microscopy
technique, working in the classical NIR-to-visible configura-
tion with the detection centered in the residual emission of
the ligand.
[*] Dr. A. D’Alꢀo, Dr. A. Bourdolle, Dr. P. L. Baldeck, Dr. C. Andraud,
Dr. O. Maury
University Lyon 1, ENS Lyon, CNRS UMR 5182
46 allꢀe d’Italie, 69364 Lyon (France)
E-mail: chantal.andraud@ens-lyon.fr
olivier.maury@ens-lyon.fr
Dr. S. Brustlein, Dr. S. Brasselet
Institut Fresnel, CNRS UMR 6133, Universitꢀ Aix Marseille III
Ecole Centrale de Marseille. Domaine Universitaire St Jꢀrꢁme
13397 Marseille cedex 20 (France)
E-mail: sophie.brasselet@fresnel.fr
Dr. A. Grichine, Dr. A. Duperray
Institut Albert Bonniot, INSERM-U823-CHU Grenoble-EFS
Universitꢀ Joseph Fourier—Grenoble I
BP170, 38042 Grenoble (France)
Dr. T. Fauquier
Institut de Gꢀnomique Fonctionnelle de Lyon, Universitꢀ de Lyon
CNRS, INRA, Ecole Normale Supꢀrieure de Lyon
46 allꢀe d’Italie, 69364 Lyon (France)
[**] We are grateful to Drs Y. Guyot and A. Brenier (LPCML, University of
Lyon) for their help with the near-infrared luminescence decay
measurements. We also thank F. Albrieux, C. Duchamp, and N.
Henriques (University of Lyon) for the assistance and access to the
high-resolution mass spectrometry facility.
Here, we report on the proof-of-concept of two-photon
NIR-to-NIR microscopy. To that end, we designed water-
soluble ytterbium complexes, containing extended p-conju-
gated skeleton suitable for two-photon excitation and mean-
while, we built up an unconventional two-photon NIR-to-
NIR microscopy set-up.
Supporting information for this article, including synthetic proce-
dures and characterization, spectroscopic and microscopic exper-
imental details, and the preparation of the biological samples, is
available on the WWW under http://dx.doi.org/10.1002/anie.
201202212.
6622
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 6622 –6625

Angewandte
Chemie
(980 nm) transition in the NIR spectral range and 2) the
broad residual ILCT emission around 600 nm, indicating that
the energy transfer to the central metal ion is not complete.
The luminescence lifetime associated to the NIR transition
was found to be perfectly mono-exponential, with a value of
0.34 and 3 ms for Yb1 and Yb2, respectively (Figure S3 in the
Supporting Information), this later value being in the range of
best complexes already reported in water.[10,15] The strong
improvement between the two complexes Yb1 and Yb2
featuring similar antenna can be ascribed to the increased
stability of the macrocyclic derivative.[13] Under femtosecond
Ti:Sa laser irradiation in the 700–900 nm spectral range
(Figure S4), an identical luminescence profile is obtained
(Figure 2d) with NIR emission observed at a longer wave-
length relative to that of the incident laser wavelength (i.e.
unusual spectral configuration for TPA-induced luminescence
with ldet > lex). The quadratic dependence of both residual
ILCT and YbIII emission intensities versus the laser power
(Figure 2e) unambiguously established that the NIR YbIII
emission is sensitized by a two-photon antenna effect. The
two-photon cross-sections could not be accurately determined
because of the weak luminescence quantum yields of both
residual ILCT and YbIII emissions (< 1%), and to the lack of
calibration of the TPA experimental setup in the NIR spectral
range. However in a first approximation, these cross-sections
should lie in the same range as that of the europium or
lutetium analogous estimated to 775 and 500 GM at 740 nm,
respectively.[11] Therefore the TPA-induced NIR emission of
ytterbium is particularly interesting for in-depth bioimaging
experiments because both excitation (750–850 nm) and
emission (980 nm) are localized in the biological transparency
window (NIR-to-NIR configuration).
Figure 1. Structure of the target complexes.
The target complexes (Figure 1) need to fulfill some
requirements to be used as bioprobe, ideally efficient two-
photon antenna, high thermodynamic and kinetic stability,
and good photophysical properties. For initial spectroscopy
and microscopy experiments, we chose to use the previously
described[11] Yb1 complex containing bis-PEGamino-(phenyl-
ethynyl)-dipicolinic ligands in spite of its limited water
stability.[12] In a second time, the analogous macrocylic
complex, Yb2 based on the triaza-cyclononane plateform[13]
was designed to increase the stability in biological medium
(see the Supporting Information for synthetic details[14] and
characterization).
In both cases, the UV/visible absorption spectra show
a broad intense transition at 400 nm, assigned to an intrali-
gand charge-transfer (ILCT) transition from the dialkylamino
donor part to the chelated dipicolinic electron-withdrawing
fragment (Figure 2 and Figure S2 in the Supporting Informa-
tion). As already observed,[11] the steady-state luminescence
spectra obtained by excitation in this ILCT transition is
composed of two emission bands: 1) the characteristic
Conventional one- and two-photon microscopy imaging
experiments were carried out using commercially available
epifluorescence and confocal microscopes with either a one-
photon excitation in the visible or a two-photon excitation at
760 nm (femtosecond Ti:Sapphire laser). The complex Yb1
was incubated with fixed T24 human cancer cells. The one-
photon microscopy image (Figure 3a) reveals that the
complex successfully stained the cell and localized
preferentially in the perinuclear areas or nucleoli as
already observed for related europium complexes.[12] The
conventional two-photon imaging experiments (Fig-
ure 3c) were carried out in the NIR-to-visible config-
uration using a visible detection at a shorter wavelength
than that of the TPA excitation (ldet < lex). As expected,
this image is very similar to the one-photon one, but the
signal-to-noise ratio is higher because the background
fluorescence is dramatically reduced using a two-photon
excitation. Unfortunately, it was not possible to record
any image at the 950–1050 nm emission band because
conventional biphotonic microscopes contain optical
filtering schemes which do not allow the NIR-to-NIR
configuration (that is detection of ldet > lex with both lex
and ldet in the NIR range). Additionally, the sensitivity of
the standard photomultiplier tube (PMT) detectors of
the confocal microscope is very low at 980 nm. To tackle
these limitations, we developed our own biphotonic
microscopy setup based on adequate optical filtering.
2
2
ytterbium (III) emission, arising from the F5/2! F7/2
Figure 2. Photophysical data of Yb1 in water solution: a) absorption spectrum
(scale on the left), b) residual CT fluorescence, and c) NIR emission
(lex =380 nm; scale on the right), and d) uncorrected two-photon-induced
luminescence spectrum (lex =760 nm; scale on the right). Inset e) shows the
variation of the two-photon luminescence intensity with the incident laser
&
*
power at 573 ( ) and 981 nm ( ).
Angew. Chem. Int. Ed. 2012, 51, 6622 –6625
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6623

.
Angewandte
Communications
tion. The variation of the normalized emission intensity
in the visible (600 nm) and in the NIR (1000 nm) was
measured simultaneously as a function of the depth of
the incident laser-focusing position (Figure 5a). As
anticipated, the visible emission is more affected by
scattering than the NIR emission: at a depth of 100 mm
in a strongly scattering medium, almost no visible light
is detected whereas more than 20% of the NIR light
remains available for imaging purpose. These data are
in agreement with recent results comparing the influ-
ence of the detection wavelength for a given TPA
excitation[16] obtained in a classical TPA configuration
(lex > ldet). The data shown in Figure 5 are normalized,
and usually the NIR detection range exhibits fewer
signal than the visible range, because of the lower
efficiency of both the imaging setup and the lumines-
cence quantum yield of the complex in this spectral
range. Nevertheless, all these experiments unambigu-
ously emphasize the interest of the NIR-to-NIR con-
figuration for in-depth imaging purposes.
Figure 3. Imaging of the fixed cells stained with Yb1 using a conventional
microscope: a) one-photon image using an inverted wide-field microscope
(lex =450–490 nm, visible detection), b,d) one- and two-photon emission
spectra (normalized one-photon (b) and two-photon (d) excited lumines-
cence), and c) two-photon NIR-to-visible laser scanning image, (lex =760 nm,
detection in the 491–673 nm range).
To check the potentialities of this two-photon NIR-
to-NIR configuration, thick tissue imaging experiments
were undertaken. Mouse brain capillary vessels were
imaged in depth using the Yb2 as luminescent probes.[17]
A phosphate buffer solution of the complex ([C] of
about 10ꢀ4 molLꢀ1) was directly perfused in the heart of
a mouse. Brains were quickly dissected, post-fixed, and slices
(100 mm thickness) were cut and conserved between two glass
The two-photon NIR-to-NIR imaging microscopy setup
(Figure 4a) consists in focusing incident pulsed Ti-sapphire
laser light (100 fs, 80 MHz, 760 nm wave-
length) through a high numerical aperture
objective (NA 1.15, water immersion). Images
are formed using a galvalnometric scanning
over typical regions of 100 mm ꢀ 100 mm in the
sample plane. To reject as much as possible
the incident laser light in the detection
channel around 1000 nm, the laser is reflected
on a dichroic mirror which transmits the
fluorescence emission in the epi descanned
detection path. The emission is further fil-
tered by two interference filters, and focused
on an avalanche photodiode working in the
photon-counting mode. Validation experi-
ments were first carried out using a thin film
of Yb1 spread on a glass plate as substrate
(Figure 4b). Regions where the film is present
are clearly visible with a high signal-to-back-
ground ratio relative to the glass substrate.
Both the spectrum of emission (Figure 4c)
and the incident intensity dependence (Fig-
ure 4d) ascertain the detection of NIR light
around 1000 nm in the two-photon mode.
As preliminary to the imaging experi-
Figure 4. a) Experimental setup for two-photon NIR-to-NIR imaging. O: objective, T:
telescope, M: protected silver mirror, GM: galvanometric mirrors, D760: dichroic mirror
(FF720-SDi01, Semrock), F800 and F1000: interference filters (800DF50 and 1000DF50,
Omega Optical), L: lens, and APD: avalanche photodiode. b) Two-photon scanning image
of a thin film of Yb1 on a glass substrate, using a detection wavelength of 1000 nm .
c) Spectrum measured in a bright region of the image of (b). A high pass filter at 850 nm
is used before the spectrometer. d) Incident power dependence of the fluorescence signal
ments, the influence of this NIR-to-NIR
configuration on the depth of penetration in
a scattering sample was studied. To that end,
the complex Yb1 was dissolved in intralipid
solutions of different concentrations, mimick-
ing the scattering ability of the biological
media and irradiated by a biphotonic excita- recorded at 1000 nm.
6624 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 6622 –6625

Angewandte
Chemie
Received: March 20, 2012
Published online: May 23, 2012
Keywords: bioimaging · biphotonic microscopy ·
.
lanthanide complexes · ytterbium
[1] V. Ntziachristos, Nat. Methods 2010, 7, 603 – 614.
[2] J. V. Frangioni, Curr. Opin. Chem. Biol. 2003, 7, 626 – 634.
[3] a) K. Kiyose, H. Kojima, T. Nagano, Chem. Asian J. 2008, 3, 506 –
515; b) S. Achilefu, Angew. Chem. 2010, 122, 10010 – 10012;
Angew. Chem. Int. Ed. 2010, 49, 9816 – 9818; c) H. S. Choi, K.
Nasr, S. Alyabyev, D. Feith, J. H. Lee, S. H. Kim, Y. Ashitate, H.
Hyun, G. Patonay, L. Strekowski, M. Henary, J. V. Frangioni,
Angew. Chem. 2011, 123, 6382 – 6387; Angew. Chem. Int. Ed.
2011, 50, 6258 – 6263.
[4] a) W. R. Zipfel, R. M. Williams, W. W. Webb, Nat. Biotechnol.
2003, 21, 1369 – 1377; b) K. Konig, J. Microsc. 2000, 200, 83 – 104.
[5] a) H. M. Kim, B. R. Cho, Acc. Chem. Res. 2009, 42, 863 – 872;
b) G. S. He, L.-S. Tan, Q. Zheng, P. N. Prasad, Chem. Rev. 2008,
108, 1245 – 1330.
[6] a) S. V. Eliseeva, J.-C. G. Bꢁnzli, Chem. Soc. Rev. 2010, 39, 189 –
227; b) J.-C. G. Bꢁnzli, S. V. Eliseeva, J. Rare Earth 2010, 28,
824 – 842; c) S. Pandya, J. Yu, D. Parker, Dalton Trans. 2006,
2757; d) S. Faulkner, S. J. A. Pope, B. P. Burton-Pye, Appl.
Spectrosc. Rev. 2005, 40, 1.
[7] For recent examples see: a) H. He, L. Si, Y. Zhong, M. Dubey,
Chem. Commun. 2012, 48, 1886 – 1888, and references therein.
[8] For reviews see: a) C. Andraud, O. Maury, Eur. J. Inorg. Chem.
2009, 4357 – 4371; b) Y. Ma, Y. Wang, Coord. Chem. Rev. 2010,
254, 972 – 990.
[9] G. Piszczek, I. Gryczynski, B. P. Maliwal, J. R. Lakowicz, J.
Fluoresc. 2002, 12, 15 – 17.
[10] T. Zhang, X. Zhu, C. C. W. Cheng, W. M. Kwok, H. L. Tam, J.
Hao, D. W. J. Kwong, W. K. Wong, K. L. Wong, J. Am. Chem.
Soc. 2011, 133, 20120 – 20122.
Figure 5. a) Variation of the normalized luminescence intensity of Yb1
at 610 (green) or 1000 nm (red) with the depth of the 760 nm TPA
excitation in two intralipid solutions of different concentrations (10%,
15%, 20%) mimicking different scattering strengths. b) 3D biphotonic
scanning microscopy image using the home-made setup
(lex =760 nm) of mouse brain slice stained with Yb2 with detection at
1000 nm (incident power 25 mW). Sections of the 3D images of the
aforementioned image.
plates. These mouse brain slices were successfully imaged
using a biphotonic excitation (lex = 760 nm) and the 3D blood
capillary network is observed using the NIR-to-NIR config-
uration with a detection wavelength at 1000 nm (Figure 5b).
This image was recorded at higher laser power to compensate
for the lower imaging efficiency in this spectral range. Even
though the recorded signal is weak, it shows stained blood
vessels with a reasonable signal-to-noise ratio up to 80 mm
depths in strongly scattering samples. The integration time of
50 ms per pixel, which is essentially chosen to ensure a high
signal-to-noise ratio, could be even more increased to reach
a larger imaging depth. This would lead to longer time scales
for imaging, however it is possible to reach faster dynamics by
using a higher incident power. This result therefore unambig-
uously establishes the proof-of-concept of two-photon NIR-
to-NIR imaging.
In conclusion, we demonstrate the feasibility of in-depth
imaging of strongly scattering thick tissue, for example, the
vascular network of mouse brain, by two-photon scanning
microscopy in an unprecedented NIR-to-NIR configuration.
To that end, we developed a new biphotonic setup and we
designed a macrocyclic ytterbium complex functionalized by
two-photon antenna combining appropriate stability and
efficiency in water. Further studies are currently conducted
to improve the two-photon brightness of such probes and to
expand the scope of this NIR-to-NIR biphotonic microscope.
[11] a) A. DꢂAlꢃo, A. Picot, A. Beeby, J. A. G. Williams, B. Le Guen-
nic, C. Andraud, O. Maury, Inorg. Chem. 2008, 47, 10258 – 10268;
b) A. DꢂAlꢃo, A. Picot, P. L. Baldeck, C. Andraud, O. Maury,
Inorg. Chem. 2008, 47, 10269 – 10279.
[12] A. Picot, A. DꢂAlꢃo, P. L. Baldeck, A. Grichine, A. Duperray, C.
Andraud, O. Maury, J. Am. Chem. Soc. 2008, 130, 1532 – 1533.
[13] G. Nocton, A. Nonat, C. Gateau, M. Mazzanti, Helv. Chim. Acta
2009, 92, 2257 – 2273.
[14] A. Bourdolle, M. Allali, J.-C. Mulatier, B. Le Guennic, J. Zwier,
P. L. Baldeck, J.-C. G. Bꢁnzli, C. Andraud, L. Lamarque, O.
Maury, Inorg. Chem. 2011, 50, 4987 – 4999.
[15] For mono-metallic complexes see: a) E. G. Moore, J. Xu, S. C.
Dodoni, C. J. Jocher, A. DꢂAlꢃo, M. Seitz, K. Raymond, Inorg.
Chem. 2010, 49, 4156 – 4166; b) A. Nonat, D. Imbert, J. Pecaut,
M. Giraud, M. Mazzanti, Inorg. Chem. 2009, 48, 4207 – 4218;
c) S. Comby, D. Imbert, C. Vandevyver, J.-C. G. Bꢁnzli, Chem.
Eur. J. 2007, 13, 936 – 944; d) M. H. V. Werts, R. H. Woudenberg,
P. G. Emmerink, R. van Gassel, J. W. Hofstraat, J. W. Verhoeven,
Angew. Chem. 2000, 112, 4716 – 4718; Angew. Chem. Int. Ed.
2000, 39, 4542 – 4544.
[16] a) P. M. Allen, W. Liu, V. P. Chauhan, J. Lee, A. Y. Ting, D.
Fukumura, R. K. Jain, M. G. Bawendi, J. Am. Chem. Soc. 2010,
132, 470 – 471; b) D. Kobat, M. E. Durst, N. Nishimura, A. W.
Wong, C. B. Schaffer, C. Xu, Opt. Express 2009, 17, 13354 –
13364.
[17] No images could be obtained using Yb1 as probe certainly
because of its too low stability in the small animal.
Angew. Chem. Int. Ed. 2012, 51, 6622 –6625
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6625