Content uploaded by Michael Rosemann
Author content
All content in this area was uploaded by Michael Rosemann on Jul 05, 2018
Content may be subject to copyright.
Short and long-term phototoxicity in cells expressing genetic
reporters under nanosecond laser exposure
Sven Gottschalk
a
,H
ector Estrada
a
, Oleksiy Degtyaruk
a
, Johannes Rebling
a
,
c
,
Olena Klymenko
b
, Michael Rosemann
b
, Daniel Razansky
a
,
c
,
*
a
Institute for Biological and Medical Imaging (IBMI), Helmholtz Zentrum München, Neuherberg 85764, Germany
b
Institute of Radiation Biology, Helmholtz Zentrum München, Neuherberg 85764, Germany
c
Faculty of Medicine, Technische Universit€
at München, München 81675, Germany
article info
Article history:
Received 16 December 2014
Received in revised form
27 July 2015
Accepted 31 July 2015
Available online 4 August 2015
Keywords:
Nanosecond laser pulses
Optoacoustic imaging
Photobleaching
Phototoxicity
Fluorescent protein
abstract
Nanosecond-duration laser pulses are exploited in a plethora of therapeutic and diagnostic applications,
such as optoacoustic imaging. However, phototoxicity effects of pulsed radiation in living cells, in
particular those expressing genetic reporters, are not well understood. We established a three-
dimensional fluorescent protein expressing cellular model in order to reliably investigate the extent
and major exposure parameters responsible for both photobleaching and phototoxicity under pulsed
laser exposure, unveiling a variety of possible effects on living cells, from reversible photobleaching to
cytotoxicity and cell death. Significant losses of fluorescence levels were identified when exposing the
cells to illumination conditions considered safe under common standards for skin exposure in diagnostic
imaging applications. Thus, the use of photolabile fluorescent proteins and their in vivo exposure pa-
rameters have to be designed carefully for all applications using pulsed nanosecond radiation. In
particular, loss of signal due to bleaching may significantly alter signals in longitudinal measurements,
making data quantification challenging.
©2015 Elsevier Ltd. All rights reserved.
1. Introduction
It has been long understood that the biological effects of pulsed
laser radiation may considerably differ from those observed under
continuous exposure [1]. Among a plethora of therapeutic and
diagnostic methods using nanosecond-duration laser pulses [2],
optoacoustics has recently gained significant momentum amid its
highly compelling advantages as a bioimaging modality. These
include the capacity for rapid volumetric imaging [3,4], intrinsic
sensitivity to functional tissue parameters [5e7] and molecular
optical reporters [8,9] as well as excellent spatial resolution in
imaging optical contrast at scalable depths, from millimeters to
centimeters in scattering living tissue [10e12]. The efficiency of
optoacoustic signal conversion may be altered by photobleaching,
which occurs when a chromophore (e.g. fluorophore or fluorescent
protein) permanently loses its reporting ability [13,14]. Although
effects related to photochemical destruction and melting of inor-
ganic nanoparticles under exposure to nanosecond laser radiation
have been reported recently [15], little is known about the extent of
photobleaching of organic chromophores in various optoacoustic
imaging scenarios.
Genetically encoded fluorescent proteins (FPs) are of particular
interest for in vitro and in vivo imaging, as they can act as reporter
molecules for specific cellular structures, whole cells or tissue in
manifold bio-medical applications [16]. FPs are valuable tools in
biological research to study the structures of cellular components
or to conduct functional studies, where they act as biosensors [17].
For in vivo imaging, the recently developed FPs with excitation
and emission wavelengths in the far-red and near-infrared spectral
windows, so-called near-infrared fluorescent proteins (iRFPs) are of
particular interest since tissue optical scattering and absorption are
relatively low in this wavelength range, thus significantly deeper
imaging depth is achieved [8,18,19]. iRFPs also possess low quan-
tum yields, which makes them ideal candidates for absorption-
based imaging with optoacoustics.
The application of FPs in optoacoustic imaging is not very
common yet as compared to their ample use in fluorescence
*Corresponding author. Helmholtz Zentrum München, Institute for Biological
and Medical Imaging (IBMI), Ingolst€
adter Landstraße 1, Building 56, 85764 Neu-
herberg, Germany.
E-mail address: dr@tum.de (D. Razansky).
Contents lists available at ScienceDirect
Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials
http://dx.doi.org/10.1016/j.biomaterials.2015.07.051
0142-9612/©2015 Elsevier Ltd. All rights reserved.
Biomaterials 69 (2015) 38e44
microscopy and spectroscopy. Nonetheless, FPs have shown great
promise for in vivo optoacoustic imaging in zebrafish [9,20], glioma
visualization [8], multi-contrast optoacoustic imaging [21] and
optoacoustic flow cytometry [22].
Many fluorescent proteins lack photostability [16,18], which
may impede quantitative measurements [23]. To this end,
nanosecond-laser-induced photobleaching has been shown to
occur in isolated FPs [24]. Furthermore, fluorescent proteins are
known to induce cytotoxicity under certain excitation conditions
[25]. Here, we investigated the extent of photobleaching in a three-
dimensional (3D) cellular model stably expressing the mCherry
fluorescent protein, which closely represents normal cellular
functions and thus mimics the in vivo architecture of natural tis-
sues. The cytotoxicity resulting from photobleaching was charac-
terized during and after optoacoustic imaging sessions in a wide
range of possible exposure scenarios.
2. Materials and methods
Unless stated otherwise products and chemicals were pur-
chased from SigmaeAldrich (Schnelldorf, Germany) or Life Tech-
nologies (Darmstadt, Germany).
2.1. Cell preparation and culture
A 3D cell culture model was developed to more closely repre-
sent normal cellular functions and mimic the in vivo architecture of
natural tissues [26]. Murine osteosarcoma (MOS) cells were
established from radiation-induced tumors as described elsewhere
[27]. In short, osteosarcoma was induced by incorporating short
living alpha-emitters (Th227) in a female C3H/Hhg mouse from a
breeding colony established at the Helmholtz Zentrum München.
Pieces of osteosarcoma were dissected under sterile conditions and
placed on BD Falcon 6-well cell culture plates in a 2 mm layer of
Dulbecco's Modified Eagle Medium (DMEM) supplemented with
10% fetal calf serum (FCS) for 10 days. When a sufficient number of
adherent cells seeded from the tumor piece, they were further
passaged and finally single-cell cloned into 96-well plates. Osteo-
blastoid differentiation was validated by alkaline phosphatase (AP)
histochemistry and tumorigeneity tested by subcutaneous trans-
plantation into syngeneic mice. MOS cells were then transfected
with pCAG Kosak-Cherry (chicken beta actin promoter, mCherry
red fluorescent protein cDNA, internal ribosome entry site (IRES),
puromycin resistance) using Lipofectamine 2000, and by applying
5
m
g Plasmid on 10
6
cells. To enable the stable integration of the
mCherry transgene, the expression construct was linearized using
the ScaI enzyme before lipofectamine-mediated transfection. Suc-
cessfully transfected cells were selected bygrowing cells for 14 days
under 5
m
g/ml puromycin followed by single cell cloning in 96-well
plates. The expression of the mCherry protein under the CAG pro-
moter is constitutive and very high [28]. The mCherry expression in
the MOS cells was measured over more than 10 passages (approx.
40 cell divisions) after expanding a single clone and the expression
of the fluorescence protein did not change over time within the
range of detection sensitivity (data not shown).
MOS cells were then grown in 5% CO
2
at 37
C in DMEM/10% FCS
supplemented with 1% Penicillin/Streptomycin (PAA, K€
olbe, Ger-
many). In addition to a single cell layer, the MOS cells grow as
multi-cellular spheroids that form on top of the cell layer. Spheroids
up to 100
m
m in diameter can be directly harvested from the flasks
3e4 days after plating the cells. To obtain larger diameter spher-
oids, they were manually detached and transferred into a new flask
for continued growing. To reach a diameter of 1 mm, it was
necessary to repeat the process two or three times to prevent the
newly-formed cell layer from overgrowing the spheroid.
Alternatively, spheroids were transferred into petri dishes with a
hydrophobic surface (Greiner Bio-One, Frickenhausen, Germany),
thus preventing formation of a cell layer at the bottom. From these
petri dishes spheroids with sizes of up to 500
m
m in diameter could
be collected directly after 4e14 days of culture.
For the bleaching experiments, spheroids of different sizes were
embedded into low melting agar (SeaPrep ultralow gelling Agarose,
Lonza, K€
oln, Germany) with a gelling temperature of 8e17
C. A 3%
(by weight) sterile agar-solution in phosphate buffered saline was
prepared by autoclaving. The dissolved agar was then kept in the
cell incubator at 37
C to prevent gelling. Standard 6-well cell
culture plates were filled with 2 ml of the 3% agar and kept at 4
C
until the agar solidified. A freshly collected solution of spheroids in
medium was carefully mixed at a 1:1 ratio with the 3% agar at 37
C,
and 2 ml of the solution were then pipetted into each well of the 6-
well plate. The plate was then stored at 4
C for a few minutes until
the agar solidified. On top, 2 ml of fresh and warm medium were
pipetted and the spheroids were allowed to recover for at least
another 24 h in the incubator at 37
C and 5% CO
2
before the agar-
embedded spheroids were collected from the 6-well plates and
placed into a 35 mm diameter cell culture dish with low walls (
m
-
Dish low, Ibidi, Munich, Germany) as shown in Fig. 1A. The dish was
then filled to the top with low-melting agar to fix the spheroids in
place. Spheroids placed on the outer edges of the cell culture dish
served as control for indirect light exposure. Only the spheroids
placed in the middle of the dish were directly exposed to the
excitation laser light of the optoacoustic scanner.
2.2. Fluorescence microscopy and image analysis
Fluorescence images were recorded using a Leica DMI3000B
fluorescence microscope (Leica Microsystems, Wetzlar, Germany)
equipped with a Leica DFC360 FX camera. To record mCherry-
fluorescence a filter set with excitation 580/20 nm and emission
630/60 nm was used. Fluorescence images of mCherry were taken
with fixed exposure times before, directly after and 24 h after the
laser light exposure sessions, thus making the fluorescence images
comparable. Fluorescence intensities were analyzed from 2 to 4
single spheroids for each of the placements in the cell culture dish
(Fig. 1A) using the freely available Fiji image processing software
[29]. Every spheroid-containing dish exposed to laser light was
accompanied by a control dish prepared and handled in the same
way, except for the fact that there was no laser exposure. If the
fluorescence levels after 24 h either remained at the same level as
right after exposure or diminished, this was considered to be of
phototoxic exposure. Otherwise, the cells were considered to be
viable and healthy.
2.3. Laser exposure conditions
To investigate the fluence-dependent effects of nanosecond
pulsed laser exposure, the MOS cells were stably transfected to
express the mCherry fluorescent protein and agar-embedded
spheroids were subjected to either unfocused or focused illumi-
nation conditions (Fig. 1B). Laser pulses of 10 ns duration were
generated by a diode-pumped Nd:YAG Q-switched Laser (IS8II-E,
EdgeWave GmbH, Würselen, Germany). The wavelength was tuned
to the peak absorption of mCherry of 594 nm by means of a dye
laser (Credo, Sirah Lasertechnik GmbH, Grevenbroich, Germany).
The unfocused multimode beam, created using a liquid light-guide
(Lumatec Series 2000, Deisenhofen, Germany), had a diameter of
approximately 6 mm at the height of the agar-embedded spheroids.
This represented typical illumination conditions in acoustic reso-
lution optoacoustic microscopy [11] and tomography [3,10] sys-
tems, which use low light fluence levels but require long exposure
S. Gottschalk et al. / Biomaterials 69 (2015) 38e44 39
times of the entire imaged area.
The second type of illumination scheme employed a focused
light beam with a full width at half maximum (FWHM) of 20
m
min
the focal spot, with the Agar-embedded spheroids placed in the
focal region. This resembles the typical illumination conditions in
optical resolution optoacoustic microscopy (OAM) [30]. In OAM, a
focused beam with a high optical fluence is scanned over the
sample, with each sample spot being exposed to only a single pulse
or a few pulses. In our experiments, the focused illumination was
produced by a photonic crystal fiber (LMA20, NKT Photonics A/S,
Birkerod, Denmark), pigtailed with a Gradient-Index (GRIN) lens
(Grintech GmbH, Jena, Germany). The fiber-lens assembly forms a
diffraction limited focal spot of 20
m
m (FWHM) at a distance of
7.1 mm from the transducer surface. The assembly is inserted
coaxially through a 0.8 mm diameter opening in the center of a
spherically-focused polyvinylidene fluoride (PVDF) ultrasound de-
tector having an active diameter of 6 mm and a focal length of
7.8 mm (Precision Acoustics, Dorchester, United Kingdom).
Unless stated otherwise, in the results section both illumination
schemes (unfocused and focused) were performed using 10,000
pulses delivered at a 500 Hz pulse repetition frequency (PRF), thus
exposing the spheroids to nanosecond laser pulses for 20 s.
2.4. Optoacoustic microscopy scanning and image analysis
In the case of focused illumination, in addition to the fluores-
cence images, we also acquired 3D optically-resolved OAM [30]
images of the agar-embedded MOS spheroids. For this, the fast
scanning setup employed a coaxially aligned scanning head,
comprising the focused illumination and ultrasound detector
which scans across a planar field of view of 4 4mm
2
within 25 s.
The laser PRF was set at ~6.2 kHz. Three-dimensional image vol-
umes are then rendered using nearest neighbor interpolation of the
zig-zag trajectory of the scanning head into a regular two-
Fig. 1. Experimental setup and characterization of mCherry-expressing spheroids. (A) Spheroids embedded in agar. Spheroids were grown in 1.5% density agar in standard 6-well
cell culture plates and harvested using a cut pipette tip. As indicated in the schematic image to the right, the harvested spheroid-agar mix was placed in a 35 mm diameter cell
culture dish with low walls. The dish was then filled to the top with low-melting agar in order to fix the spheroids in place and subsequently subjected to fluorescence microscopy
and scanning optoacoustic microscopy (OAM). (B) Schematic of the setups used for bleaching experiments; Left esetup for the OAM experiments using focused illumination; Right
esetup for the unfocused illumination studies. (C) From left to right: Bright field, fluorescence and overlay images of typical spheroidal clusters embedded in agar.
S. Gottschalk et al. / Biomaterials 69 (2015) 38e4440
dimensional grid [31]. The effect of oscillations in the energy of the
laser pulses is corrected using the signals of a photodiode (DET10A/
M, Thorlabs, Newton, New Jersey, USA) placed in the optical path.
Optoacoustic data was analyzed using MatLab (version 2013a,
Mathworks Inc., Natick, MA, USA), as well as open-source Mayavi
4.1.0 software [32] for 3D visualization and matplotlib 1.2.1 [33] for
two-dimensional plotting running on python 2.7.5 (https://www.
phyton.org). Low frequency offsets were removed by means of a
high-pass filter with cut-on frequency at 1 MHz. Due to the high
fluence attained by the focused illumination, small impurities, such
as dust particles in the agar may produce strongly localized opto-
acoustic signals and hence cause cavitation-related signal amplifi-
cation [34] which may hinder the visualization of optoacoustic
signals generated by the spheroids (data not shown). However, no
such interference signals were observed as originating from within
the spheroids. Due to its shot-noise-like characteristics, this inter-
ference could easily be removed using a 5 5 pixels median filter.
2.5. Statistics
All experiments were performed with n ¼5e7 for each exper-
imental group. The data are shown as mean values ±standard error
of the mean (SEM). Statistical significance is further described in
the results and in the figure legends.
3. Results
3.1. Fluence-dependent loss of fluorescence
Representative brightfield and fluorescence images of the
established MOS cell line are shown in Fig. 1C, highlighting its
natural tendency to grow into 3D multi-cellular spheroidal struc-
tures. For the focused illumination in the scanning OAM, only
spheroids placed in the middle of the cell culture dish (Fig. 1A) were
directly exposed to laser light while scanning an area of 4 4mm
2
.
Each exposed dish was accompanied bya control dish prepared and
treated in the same way but not exposed to the laser light. Spher-
oids located at the periphery of the dish, that were only exposed to
scattered laser light, did not show a loss of fluorescence for any of
the fluence levels used during the experiments (data not shown).
Also, none of the spheroids in the control dishes (i.e. no laser
exposure) showed any significant loss of fluorescence. Instead most
of the control and indirectly illuminated spheroids exhibited
significantly increased fluorescence-intensity 24 h after the
experiment (data not shown). Since the MOS cells have a doubling
time of ~12 h while their mCherry expression is constitutive and
very high, this increase in fluorescence after 24 h can therefore
most likely be explained by cell proliferation.
The experiments using unfocused illumination for a total
duration of 20 s at 500 Hz PRF (i.e. total of 10,000 pulses) have
induced significant bleaching with 73 ±4% of the baseline fluo-
rescence remaining after exposure to fluence levels as low as
0.49 ±0.03 mJ/cm
2
(Fig. 2A). Fluorescence readouts further
decreased to 44 ±1% from the baseline when the per-pulse fluence
was increased to 1.23 ±0.01 mJ/cm
2
(Fig. 2A). See also the repre-
sentative fluorescence images taken before and after the experi-
ments (Fig. 2B).
We have also investigated the photobleaching effects under
exposure levels considered safe under the common maximal
permissible exposure (MPE) standards established for human skin
[35]. For this, experiments using unfocused illuminations were
done with different fluence and average intensity levels while
keeping the total number of pulses constant (10,000 pulses deliv-
ered for each experiment). Since all the exposure experiments for
the case lasted for more than 10 s, both the per-pulse energy and
average intensity levels had to be kept below 20 mJ/cm
2
and
200 mW/cm
2
, respectively, in order to meet the MPE standards. The
corresponding results presented in Fig. 3 clearly demonstrate that
pronounced photobleaching of the mCherry protein is imposed in
nearly all experiments eonly 17e32% of the fluorescence was left
in all cases where the exposure was kept considerably below both
the per-pulse and average intensity thresholds. On the other hand,
no phototoxic effects were observed as the fluorescence increased
for all the exposure scenarios 24 h after the experiments (data not
shown).
Results from the scanned focused illumination experiments are
presented in Fig. 4, again demonstrating that the photobleaching
increases with the fluence. Surprisingly, for the highest fluence of
967 ±18 mJ/cm
2
, even though the bleaching effects were less
pronounced (Fig. 4A, about 44 ±2% of the fluorescence signal
remaining) as compared to the unfocused illumination results
presented in Fig. 3, the fluorescence intensity did not recover after
24 h, clearly suggesting a phototoxic effect of the illumination (see
also Fig. 6). Clearly, in the case of focused illumination, none of the
above MPE limits are fulfilled so these scenarios are generally
considered unsafe under the ANSI standards for human skin
exposure.
3.2. Cytotoxic effects of nanosecond laser exposure
As mentioned above, when the fluorescence was recorded
directly after exposure to unfocused illumination, photobleaching
effects on the mCherry-protein were already observed at compa-
rably low fluence values (Fig. 2A). However, these cells recovered to
the initial (and even slightly higher) fluorescence intensities after
24 h, thus indicating a full replacement of the destroyed mCherry-
proteins and also further cell proliferation under unfocused
illumination.
To further assess the cytotoxic effects of nanosecond pulsed
laser light exposure, the mCherry-fluorescence of spheroids was
assessed after the unfocused exposure at different light fluence
levels and PRFs. Here we also tested an additional exposure sce-
nario using higher energy deposition rate whilst keeping the total
delivered energy and per-pulse fluence levels constant. For this, the
agar-embedded spheroids were exposed to unfocused illumination
at 5 kHz laser PRF, thus delivering the same amount of energy and
Fig. 2. Fluence-dependent photobleaching of mCherry-spheroids under unfocused
illumination. Spheroids were embedded in agar and exposed to increasing fluence levels
under unfocused illumination. (A) Loss in fluorescence intensity following unfocused
illumination with 10,000 pulses delivered over 20 s for increasing fluence levels. (B)
Representative fluorescence images of spheroids in agar before and after illumination
with a fluence of 1.23 ±0.01 mJ/cm
2
. Bar size is 500
m
m. Statistical significance was
determined with Student's t-test: *P <0.05, **P <0.01, ***P <0.001. Data are
means ±SEM (n ¼5e7) and are at least P <0.01 vs. their respective controls (spheroids
treated in the same way but notexposed to laser light, data not shown). The exponential
fit (solid black) is shown together with the 95% confidence band (dotted lines).
S. Gottschalk et al. / Biomaterials 69 (2015) 38e44 41
number of pulses over 2 s instead of 20 s. The results presented in
Fig. 5 clearly show a greater extent of photobleaching induced by
the higher energy deposition rate, i.e. higher average intensity with
the same per-pulse fluence.
The extent of long-term phototoxicity was similarly assessed for
the scanning focused illumination scenario. Likewise, reversible
photobleaching was observed when the fast scanning was per-
formed with fluence levels of 242 ±5 mJ/cm
2
(Fig. 6). In contrast,
fluence values of 967 ±18 mJ/cm
2
not only had a stronger photo-
bleaching effect, but also caused lower recovery rates of the cells.
No recovery of fluorescence was observed 24 h after illuminating
the cells with 967 ±18 mJ/cm
2
, most probably due to irreversible
cytotoxicity. It must be noted that it was not technically possible to
pre-calibrate and reproduce exactly the same fluence levels for
each experiment (see the inset in Fig. 8A).
3.3. Loss of optoacoustic signal amplitude
Next, the effects of scanning focused illumination on mCherry-
fluorescence were compared to the corresponding loss in the
recorded optoacoustic signal intensity. For this, agar-embedded
spheroids were repeatedly scanned in the OAM setup at 594 nm
with fluorescence recorded in between (Fig. 7). Analysis of the
three representative spheroids (Fig. 7A) shows a loss in fluores-
cence intensities of 69%, 74% and 71%, respectively after the first
scan (Fig. 7Cvs.Fig. 7B, leftmost columns). The corresponding de-
cays in optoacoustic signal intensities between the first and second
scan (Fig. 7Cvs.Fig. 7B, rightmost columns) were in good agree-
ment with the decreasing fluorescence intensities (69%, 62% and
72% for the spheroids numbered 1e3, respectively).
To better understand the dynamics of the mCherry-protein
bleaching under continuous laser exposure, single agar-
embedded spheroids were placed at the focus of the illumination
beam and subjected to a total of 10,000 consecutive nanosecond
laser pulses with a PRF of 500 Hz without beam scanning, hence the
cells in the beam focus were continuously exposed to focused laser
pulses for a total of 20 s. A laser fluence level of 800 mJ/cm
2
was
selected for this experiment as a compromise between detectable
optoacoustic signal levels and a reasonably slow bleaching rate.
Note also the considerable per-pulse fluence variations at these
high fluence levels (see inset in Fig. 8A). As can be seen, after ~20 s
of continuous stationary exposure the recorded optoacoustic signal
levels diminished by more than 99% and reached the noise floor of
the measurement system (Fig. 8A). The fluorescence images taken
before (Fig. 8B) and after (Fig. 8C) the exposure similarly show
diminished signal from the exposed areas surrounding the focused
beam.
4. Discussion and conclusions
In this work, we have established a 3D mCherry-expressing cell
culture model, in order to reproducibly study the short and long-
term phototoxicity of nanosecond laser exposure in cells express-
ing fluorescent genetic reporters. By using various pulsed nano-
second laser exposure scenarios, utilized in common optoacoustic
microscopy and tomography systems, it has been shown that
photobleaching of the reporter proteins may occur already at
Fig. 3. Photobleaching effects under illumination conditions considered safe under
ANSI exposure standards for skin exposure. Agar-embedded spheroids were exposed
to unfocused illumination conditions with different fluence and average intensity
levels fulfilling the ANSI standards, i.e. the per-pulse energy and average intensity
levels were kept below 20 mJ/cm
2
and 200 mW/cm
2
, respectively [35]. All exposures
lasted for more than 10 s while the total number of pulses was kept constant, i.e.
10,000 pulses delivered for each experiment. Each black dot represents the average
from multiple spheroids taken from two dishes. The plotted numbers correspond to
the normalized fluorescence (%from baseline) obtained after exposing the cells to the
10,000 pulses.
Fig. 4. Fluence- and absorption-dependent photobleaching of mCherry-spheroids
under focused illumination. Spheroids were embedded in agar and exposed to
increasing fluence levels under focused illumination. (A) Loss in fluorescence intensity
(% from the baseline fluorescence recorded before the experiment) For statistical sig-
nificance and details of the fitting curve see Fig. 2(B) Representative fluorescence
images of the agar-embedded spheroids before and after the scan. Red line indicates
the border of the scanned area. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)
******
**
100
50
0
150
High intensity
(5 kHz PRF)
Low intensity
(500 Hz PRF)
Normalized fluorescence
(%from baseline)
Unfocused Illumination Low fluence
High fluence Control
**
Fig. 5. Effects of the energy deposition rate for the unfocused illumination scenario.
Comparison of bleaching caused by unfocused illumination when the same amount of
energy was delivered over a shorter period of time. On the left, the effects of exposure
with lower average intensities (i.e. 10,000 pulses delivered at 500 Hz for a total
duration of 20 s) at two different fluence levels (0.49 ±0.03 mJ/cm
2
and 1.06 ±0.02 mJ/
cm
2
) are shown. On the right, data is shown from spheroids exposed to high intensity
exposures (i.e. 10,000 pulses delivered at 5 kHz for a total duration of 2 s). The cor-
responding fluence were 0.51 ±0.02 mJ/cm
2
and 1.01 ±0.02 mJ/cm
2
in this case. For
statistical significance see Fig. 2.
S. Gottschalk et al. / Biomaterials 69 (2015) 38e4442
relatively low laser fluence levels of ~0.5 mJ/cm
2
when delivered for
a total duration of 20 s. In general, according to the standards of
human exposure to prolonged pulsed laser radiation in the visible
spectrum, the fluence on the surface of the imaged object should
not exceed 20 mJ/cm
2
and the average delivered light intensity has
to remain below 200 mW/cm
2
. In our experiments, fluence and
average intensity levels well below the MPE thresholds caused
pronounced photobleaching with only 17e32% of the fluorescence
intensity remaining ea constant number of 10,000 pulses was
delivered for each experiment. While this photobleaching was
reversible, typical optoacoustic imaging systems would usually
expose the object to significantly higher illumination levels for
extended durations [3,36], which may indeed additionally lead to
cellular toxicity and a complete loss of fluorescence and therefore
optoacoustic intensity when imaging genetically expressed
markers. Our cellular model clearly consists of a thin layer of semi-
transparent cells, which were directly exposed to laser radiation.
When applied to large tissue specimen, the light intensity is
strongly attenuated as it penetrates into opaque tissues [37], thus
the actual light fluence at the given imaging depth should be
considered instead for the exposure estimations.
The laser pulse repetition frequency (PRF), and hence the
delivered averagelight intensity, proved tobe equally important for
the extent of bleaching. For the unfocused illumination conditions
at relatively low fluence levels between 0.5 and 1 mJ/cm
2
, photo-
bleaching was much more pronounced when increasing the PRF
(while keeping the fluence constant). This indicates certain non-
linearities in fluorescence recovery since the same total amount of
energy is essentially delivered in a shorter period of time. This has
important implications for fast optoacoustic imaging systems, as
higher imaging frame rates can only be achieved by an equal in-
crease of the laser pulse repetition rate and therefore by an equal
increase in light intensity. The higher peak fluence values have also
invoked nonlinear effects since the extent of photobleaching
exhibited an exponential relation to the increasing fluence values
(Fig. 4A). Nonlinear behavior was also shown before to occur for
continuous wave exposure intensities above 10 W/cm
2
[23].
Furthermore, the loss in fluorescence intensity under the un-
focused illumination exposure was in agreement with the previ-
ously reported values in isolated proteins for similar exposure
conditions [24], anticipating that the effects observed herein can be
242 ± 5 mJ/cm
***
**
100
50
0
After bleaching 24 h after bleaching
Normalized fluorescence
(% from baseline)
967 ± 18 mJ/cm
Focused scanning illumination
***
ns
Fig. 6. Fluorescence recovery and phototoxicity under the focused illumination sce-
nario. Spheroids were embedded in agar and exposed to the scanning OAM experi-
ments at various fluence levels. Nearly full fluorescence recovery was observed after
exposure to a fluence of 242 ±5 mJ/cm
2
while no recovery occurred with 967 ±18 mJ/
cm
2
fluence, indicating laser-induced phototoxicity. For statistical significance see
Fig. 2.
Fig. 7. Correlation between fluorescence and optoacoustic signal intensity following
bleaching. Spheroids embedded in agar were subjected to repeated OAM scans while
fluorescence images were taken before and in between the scans. (A) Left: Digital
photograph of the spheroids embedded in agar (view from below, the ultrasound
transducer can be seen in the background as a dark circle). Right: 3D projection view of
the acquired optoacoustic data. Note that some spheroids are located out of focus in
the fluorescence images whereas OAM delivers true high resolution 3D data. (B) Left:
Fluorescence image taken directly before the first OAM scan. The jagged arrow in the
middle indicates the chronology of the experiment. Right: Maximum amplitude pro-
jection (MAP) along the depth-direction of the volumetric OAM data acquired at the
first scan. (C) Left: Fluorescence image taken after the first optoacoustic scan dem-
onstrates an average decrease in fluorescence intensity to ~71% compared to the first
scan. Right: MAP image from the second OAM scan, also showing loss of optoacoustic
signal intensity to ~68% as compared to the initial scan. Fluence levels for the first and
second optoacoustic scan were ~900 mJ/cm
2
and ~840 mJ/cm
2
, respectively.
Fig. 8. Rapid bleaching under continuous exposure to a focused beam. A single
spheroid embedded in agar was subjected to a total of 10,000 laser pulses with a
fluence of 800 mJ/cm
2
and a repetition rate of 500 Hz, i.e. total exposure duration of
20 s. (A) Loss of optoacoustic signal intensity over time. Note that the signals are
originating from the focal volume of the laser beam. The red line represents a moving
average over 200 pulses. The insets show a digital photograph of the spheroid and the
measured fluence distribution over all laser shots. The corresponding fluorescence
images of the selectively bleached spheroids directly before and after the scan are
shown in (B) and (C), respectively. Loss of fluorescence in a circular area around the
focal area of the light beam can be clearly identified. Bar indicates 200
m
m. (For
interpretation of the references to color in this figure legend, the reader is referred to
the web version of this article.)
S. Gottschalk et al. / Biomaterials 69 (2015) 38e44 43
generally extrapolated to other genetic reporters having different
degrees of photostability [13,23].
When using rapid scanning of a focused beam, carrying fluence
levels of up to 1000 mJ/cm
2
, single laser pulses were sufficient to
cause irreversible photobleaching, whereas the toxic effects on cells
were more prominent. It is indeed very common that the rapid
scanning optical-resolution optoacoustic microscopy techniques
employ peak fluence values beyond 1000 mJ/cm
2
in the optical
focus [11,30]. Also during fast scanning OAM experiments in the
current study, a fluence of ~900 mJ/cm
2
was necessary in order to
reliably detect the presence of fluorescent proteins in cells and their
photobleaching effects. Interestingly, whilst the photobleaching
levels of the focused illumination with the highest tested fluence
still remained well below the bleaching levels in most unfocused
illumination experiments, no fluorescence recovery was recorded
for the former case.
In conclusion, we were able to identify several parameters that
can influence the extent of photobleaching and phototoxicity of
genetic markers in an in vivo-like cellular model. Since, photo-
bleaching occurs under relatively low and safe exposure conditions,
the in vivo use of photolabile FPs has to be designed carefully for all
applications using nanosecond pulsed laser exposures. This is of
special importance when it comes to extracting quantified data from
the images since loss of signal due to photobleaching may signifi-
cantly alter the measured signal in longitudinal measurements.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgments
The research leading to these results has received funding from
the European Research Council under grant agreement ERC-2010-
StG-260991. The technical assistance of J. Turner is greatly
acknowledged.
References
[1] M.H. Niemz, Laser-tissue Interactions: Fundamentals and Applications, third
ed., Springer Berlin Heidelberg, 2007.
[2] H.P. Berlien, in: H.P. Berlien, G.J. Mueller (Eds.), Applied Laser Medicine, first
ed., Springer Berlin Heidelberg, 2003.
[3] D. Razansky, A. Buehler, V. Ntziachristos, Volumetric real-time multispectral
optoacoustic tomography of biomarkers, Nat. Protoc. 6 (2011) 1121e1129.
[4] X.L. De
an-Ben, D. Razansky, Adding fifth dimension to optoacoustic imaging:
volumetric time-resolved spectrally enriched tomography, Light Sci. Appl. 3
(2014).
[5] P. Beard, Biomedical photoacoustic imaging, Interface Focus 1 (2011)
602e631.
[6] M. Kneipp, J. Turner, S. Hambauer, S.M. Krieg, J. Lehmberg, U. Lindauer, et al.,
Functional real-time optoacoustic imaging of middle cerebral artery occlusion
in mice, PLoS One 9 (2014) e96118.
[7] W. Liu, K.M. Schultz, K. Zhang, A. Sasman, F. Gao, T. Kume, et al., Corneal
neovascularization imaging by optical-resolution photoacoustic microscopy,
Photoacoustics 2 (2014) 81e86.
[8] N.C. Deliolanis, A. Ale, S. Morscher, N.C. Burton, K. Schaefer, K. Radrich, et al.,
Deep-tissue reporter-gene imaging with fluorescence and optoacoustic to-
mography: a performance overview, Mol. Imaging Biol. 16 (2014) 652e660.
[9] D. Razansky, C. Vinegoni, V. Ntziachristos, Imaging of mesoscopic-scale or-
ganisms using selective-plane optoacoustic tomography, Phys. Med. Biol. 54
(2009) 2769e2777.
[10] L.V. Wang, S. Hu, Photoacoustic tomography: in vivo imaging from organelles
to organs, Science 335 (2012) 1458e1462.
[11] J. Yao, L.V. Wang, Sensitivity of photoacoustic microscopy, Photoacoustics 2
(2014) 87e101.
[12] M. Heijblom, D. Piras, W. Xia, J.C. van Hespen, J.M. Klaase, F.M. van den Engh,
et al., Visualizing breast cancer using the Twente photoacoustic mammo-
scope: what do we learn from twelve new patient measurements? Opt. Ex-
press 20 (2012) 11582e11597.
[13] N.C. Shaner, M.Z. Lin, M.R. McKeown, P.A. Steinbach, K.L. Hazelwood,
M.W. Davidson, et al., Improving the photostability of bright monomeric or-
ange and red fluorescent proteins, Nat. Methods 5 (2008) 545e551.
[14] R.I. Ghauharali, G.J. Brakenhoff, Fluorescence photobleaching-based image
standardization for fluorescence microscopy, J. Microsc. 198 (Pt 2) (2000)
88e100.
[15] L. Gao, L. Wang, C. Li, A. Garcia-Uribe, L.V. Wang, Photothermal bleaching in
time-lapse photoacoustic microscopy, J. Biophot. 6 (2013) 543e548.
[16] N.C. Shaner, R.E. Campbell, P.A. Steinbach, B.N. Giepmans, A.E. Palmer,
R.Y. Tsien, Improved monomeric red, orange and yellow fluorescent proteins
derived from Discosoma sp. red fluorescent protein, Nat. Biotechnol. 22
(2004) 1567e1572.
[17] D.S. Bindels, J. Goedhart, M.A. Hink, L. van Weeren, L. Joosen, T.W. Gadella Jr.,
Optimization of fluorescent proteins, Methods Mol. Biol. 1076 (2014)
371e417.
[18] R. Weissleder, A clearer vision for in vivo imaging, Nat. Biotechnol. 19 (2001)
316e317.
[19] G.S. Filonov, A. Krumholz, J. Xia, J. Yao, L.V. Wang, V.V. Verkhusha, Deep-tissue
photoacoustic tomography of a genetically encoded near-infrared fluorescent
probe, Angew. Chem. 51 (2012) 1448e1451.
[20] M. Liu, N. Schmitner, M.G. Sandrian, B. Zabihian, B. Hermann, W. Salvenmoser,
et al., In vivo three dimensional dual wavelength photoacoustic tomography
imaging of the far red fluorescent protein E2-Crimson expressed in adult
zebrafish, Biomed. Opt. Express 4 (2013) 1846e1855.
[21] A. Krumholz, D.M. Shcherbakova, J. Xia, L.V. Wang, V.V. Verkhusha, Multi-
contrast photoacoustic in vivo imaging using near-infrared fluorescent pro-
teins, Sci. Rep. 4 (2014) 3939.
[22] D.A. Nedosekin, M. Sarimollaoglu, E.I. Galanzha, R. Sawant, V.P. Torchilin,
V.V. Verkhusha, et al., Synergy of photoacoustic and fluorescence flow
cytometry of circulating cells with negative and positive contrasts, J. Biophot.
6 (2013) 425e434.
[23] K.D. Piatkevich, E.N. Efremenko, V.V. Verkhusha, S.D. Varfolomeev, Red fluo-
rescent proteins and their properties, Russ. Chem. Rev. 79 (2010) 243.
[24] J. Laufer, A. Jathoul, M. Pule, P. Beard, In vitro characterization of genetically
expressed absorbing proteins using photoacoustic spectroscopy, Biomed. Opt.
Express 4 (2013) 2477e2490.
[25] Z.X. Liao, Y.C. Li, H.M. Lu, H.W. Sung, A genetically-encoded KillerRed protein
as an intrinsically generated photosensitizer for photodynamic therapy, Bio-
materials 35 (2014) 500e508.
[26] T.M. Achilli, J. Meyer, J.R. Morgan, Advances in the formation, use and un-
derstanding of multi-cellular spheroids, Expert Opin. Biol. Ther. 12 (2012)
1347e1360.
[27] M. Rosemann, M. Lintrop, J. Favor, M.J. Atkinson, Bone tumorigenesis induced
by alpha-particle radiation: mapping of genetic loci influencing predisposition
in mice, Radiat. Res. 157 (2002) 426e434.
[28] J.Y. Qin, L. Zhang, K.L. Clift, I. Hulur, A.P. Xiang, B.Z. Ren, et al., Systematic
comparison of constitutive promoters and the doxycycline-inducible pro-
moter, PLoS One 5 (2010) e10611.
[29] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, et
al., Fiji: an open-source platform for biological-image analysis, Nat. Methods 9
(2012) 676e682.
[30] H. Estrada, J. Turner, M. Kneipp, D. Razansky, Real-time optoacoustic brain
microscopy with hybrid optical and acoustic resolution, Laser Phys. Lett. 11
(2014), 045601.
[31] H. Estrada, E. Sobol, O. Baum, D. Razansky, Hybrid optoacoustic and ultra-
sound biomicroscopy monitors laser-induced tissue modifications and
magnetite nanoparticle impregnation, Laser Phys. Lett. 11 (2014) 125601.
[32] P. Ramachandran, G. Varoquaux, Mayavi: 3D visualization of scientific data,
Comput. Sci. Eng. 13 (2011) 40e51.
[33] J.D. Hunter, Matplotlib: a 2D graphics environment, Comput. Sci. Eng. 9 (2007)
90e95.
[34] A.B. Karpiouk, S.R. Aglyamov, F. Bourgeois, A. Ben-Yakar, S.Y. Emelianov,
Quantitative ultrasound method to detect and monitor laser-induced cavita-
tion bubbles, J. Biomed. Opt. 13 (2008), 034011.
[35] American National Standards for the Safe Use Lasers ANSI Z136.1, American
Laser Institute, Orlando, FL, 2000.
[36] H.P. Brecht, R. Su, M. Fronheiser, S.A. Ermilov, A. Conjusteau, A.A. Oraevsky,
Whole-body three-dimensional optoacoustic tomography system for small
animals, J. Biomed. Opt. 14 (2009), 064007.
[37] S.L. Jacques, Optical properties of biological tissues: a review, Phys. Med. Biol.
58 (2013). R37.
S. Gottschalk et al. / Biomaterials 69 (2015) 38e4444