Toxicology Letters 200 (2011) 201–210

Contents lists available at ScienceDirect

Toxicology Letters
journal homepage: www.elsevier.com/locate/toxlet

In vitro toxicity evaluation of graphene oxide on A549 cells

Yanli Chang a , Sheng-Tao Yang a,b , Jia-Hui Liu a,b , Erya Dong a , Yanwen Wang a ,
Aoneng Cao a,∗ , Yuanfang Liu a,b , Haifang Wang a,∗
a
Institute of Nanochemistry and Nanobiology, Shanghai University, Shanghai 200444, China
b
Beijing National Laboratory for Molecular Sciences and College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China

a r t i c l e i n f o a b s t r a c t

Article history: Graphene and its derivatives have attracted great research interest for their potential applications in
Received 28 July 2010 electronics, energy, materials and biomedical areas. However, little information of their toxicity and bio-
Received in revised form 11 October 2010 compatibility is available. Herein, we performed a comprehensive study on the toxicity of graphene oxide
Accepted 24 November 2010
(GO) by examining the influences of GO on the morphology, viability, mortality and membrane integrity
Available online 2 December 2010
of A549 cells. The results suggest that GO does not enter A549 cell and has no obvious cytotoxicity. But
GO can cause a dose-dependent oxidative stress in cell and induce a slight loss of cell viability at high
Keywords:
concentration. These effects are dose and size related, and should be considered in the development of
Graphene oxide
Biocompatibility
bio-applications of GO. Overall, GO is a pretty safe material at cellular level, which is confirmed by the
Toxicity favorable cell growth on GO film.
Size effect © 2010 Elsevier Ireland Ltd. All rights reserved.
Cell growth substrate

1. Introduction
Because of their unique physicochemical properties, graphene
and its derivatives have attracted tremendous research interest
(Allen et al., 2010; Geim, 2009; Rao et al., 2009). They hold great
promise in electronics, energy, materials and biomedical areas
(Allen et al., 2010; Geim, 2009; Neto et al., 2009; Rao et al., 2009).
Graphene oxide (GO) is one of the most important graphene deriva-
tives and has been extensively studied in recent years (Park and
Ruoff, 2009). We reported that GO could be used to produce directly
the graphene-based composites (Cao et al., 2010). Beyond that,
GO has been also used in many areas, including hydrogen storage
(Wang et al., 2009b), catalysis (Scheuermann et al., 2009), trans-
parent film (Dikin et al., 2007) and electrode (Eda et al., 2008).
In particular, GO is a potential candidate for biological applica-
tions, such as drug delivery and bio-analysis (Liu et al., 2008; Lu
et al., 2010; Sun et al., 2008; Yang et al., 2009b, 2010; Zhang et al.,
2010a). For example, Liu et al. (2008) found that GO could deliver
doxorubicin into cancer cells for the therapeutic purpose.
Many studies have shown that nanomaterials might have side-
effects on health (Aillon et al., 2009; Oberdörster et al., 2005; Xia
et al., 2009). For instance, we have reported the toxicity and reten-
tion of carbon nanotubes (CNTs) in vitro and in vivo (Deng et al.,
2007; Jia et al., 2005; Wang et al., 2004, 2008, 2009a, 2009c; Yang
et al., 2008a, 2008b, 2009a). The thorough understanding of the
biological behavior of nanomaterials guarantees the sustainable
nanotechnology (Aillon et al., 2009; Hussain et al., 2009; Nel et al.,
2006; Oberdörster et al., 2005; Xia et al., 2009). However, for the
newly developed graphene and its derivatives, such information is
generally lacking to date.
Herein, we performed a systematic study on the toxicity of GO
at cell level. The morphology, viability, mortality and membrane
integrity of A549 cells, a human lung carcinoma epithelial cell line,
were evaluated after GO exposure. The results suggest that, GO has
no obvious toxicity to A549 cells, though GO induces the cellu-
lar oxidative stress even at low concentration and induce a slight
decrease of the cell viability at high concentration. The transmission
electron microscopy (TEM) investigation suggests that GO could
hardly enter cells. The size of GO sheets has effect on the toxicity of
GO at high concentration, that is larger sheets have better biocom-
patibility. The good biocompatibility of GO allows it to be used for
various biomedical purposes in future. Preliminarily, we show the
GO film is a good substrate for cell growth.
2. Materials and methods
2.1. Preparation and characterization of GO

Natural graphite powder (≤30 ␮m, with purity higher than 99.85 wt.%) was
purchased from Sinopharm Chemical Reagent Co., Ltd., China. The preparation
∗ Corresponding authors. Fax: +86 21 66135275. of GO followed the modified Hummer method (Hummers and Offerman, 1958;
E-mail addresses: ancao@shu.edu.cn (A. Cao), hwang@shu.edu.cn (H. Wang). Kovtyukhova et al., 1999), which is described in Supplementary Data. The obtained

0378-4274/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.toxlet.2010.11.016
202 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210
Go suspension was further heated to 120 ◦ C for 20 min to get GO mixture (m-GO).
After cooling to room temperature, the suspension was centrifuged at 18,000 rpm
for 50 min to obtain the s-GO (supernatant, GO with smaller size) and l-GO (residue,
GO with larger size) samples.
The three GO samples were characterized by TEM (JEM-200CX, JEOL, Japan),
atomic force microscopy (AFM, SPM-9600, Shimadzu, Japan), Fourier transform
infrared spectroscopy (FTIR, Avatar 370, Thermo Nicolet, USA), Raman spectroscopy
(Renishaw Invia Plus laser Raman spectrometer, Renishaw, UK) and X-ray photo-
electron spectroscopy (XPS, AXIS Ultra instrument, Kratos, UK). The particle size
distribution and zeta potential in water were measured by Nanosizer (Zetasizer
3000 HS, Malvern, UK).
GO were dispersed in ultra-pure water to prepare the stock solution (1.0 mg/mL).
The stock solution was sonicated for 1 h (40 kHz, 50 W) and diluted to different
concentrations with F-12K culture medium just prior to the cell exposure.
2.2. Cell culture
A549 cell line is one popular cell line in nanotoxicology studies with a cell cycle
time of 22 h (Pulskamp et al., 2007; Herzog et al., 2007). A549 cells were kindly
provided by Dr. Y. Zhong at Shanghai University, China. A549 cells were cultured in
F-12K culture medium supplemented with 10% (v/v) fetal bovine serum (Lanzhou
National Hyclone Bio-Engineering Co. Ltd., China) at 37 ◦ C in a humidified atmo-
sphere of 5% CO2 /95% air.
2.3. Cell morphology and ultrastructure
A549 cells were plated in the 96-well plates (5 × 103 cells per well) and incu-
bated for 24 h. m-GO, s-GO and l-GO were introduced separately to cells with a
predetermined concentration in culture medium. Cells cultured in the medium
without adding GO were taken as the control. The cell morphology was recorded
under an optical microscopy at 24 h postexposure.
To investigate the cellular ultrastructure of GO-treated A549 cells, thin-sections
of cells were investigated under TEM. A549 cells were plated in the 6-well plates
(1 × 105 cells per well) and incubated for 24 h. GO samples were introduced to the
cells with a final concentration of 200 ␮g/mL. Cells without GO exposure were taken
as the control. After 24 h exposure, the cells were washed with ice-cold PBS for three
times. After the centrifugation (4000 rpm × 10 min), cells were collected, prefixed
with 2.5% glutaraldehyde, post-fixed in 1% osmium tetroxide, dehydrated in a graded
alcohol series, embedded in epoxy resin, and cut with an ultramicrotome. Thin-
sections poststained with uranyl acetate and lead citrate were inspected with TEM.
2.4. Cell viability
The cell viability was evaluated by CCK-8 assay (Dojindo Molecular Technolo-
gies, Inc.). A549 cells were plated in the 96-well plates (5 × 103 cells per well) and
incubated for 24 h. m-GO, s-GO and l-GO were introduced separately to cells with
different test concentrations (10, 25, 50, 100 and 200 ␮g/mL) in culture medium.
Cells cultured in the medium without adding GO were taken as the control. After
24, 48 and 72 h incubation, the cells were washed with D-Hanks buffer solution.
Two hundred microlitres of CCK-8 solution was added to each well and incubated
for an additional 3 h at 37 ◦ C. The optical density (OD) of each well at 450 nm was
recorded on a Microplate Reader (Thermo, Varioskan Flash). The cell viability (%
of control) is expressed as the percentage of (ODtest − ODblank )/(ODcontrol − ODblank ),
where ODtest is the optical density of the cells exposed to GO sample, ODcontrol is the
optical density of the control sample and ODblank is the optical density of the wells
without A549 cells.
In a separate experiment, to test the effect of the adsorption of culture medium
by GO on the toxicity, the GO samples (10, 25, 50, 100 and 200 ␮g/mL) were incu-
bated in culture medium (cell-free) at 37 ◦ C under 5% CO2 /95% air for 24 h. Then, the
mixtures were centrifuged at 4000 rpm for 5 min to remove precipitate (GO). The
GO free supernatants were collected and introduced to A549 cells (5 × 103 cells per
well). After 24 h incubation, the cell viability was assayed by CCK-8 assay.
2.5. Cell mortality
The cell mortality was evaluated by Trypan blue assay (Beyotime Institute of
Biotechnology, China). A549 cells were plated in the 6-well plates (1 × 105 cells per
well) and incubated for 24 h. Then, GO was introduced to cells with different concen-
trations (10, 25, 50, 100 and 200 ␮g/mL) in culture medium. Cells cultured in the free
medium were taken as the control. Twenty-four hours later, the supernatant was
collected and the cells were detached with 300 ␮L trypsin–EDTA solution. The mix-
ture of the supernatant and detached cells was centrifugated at 1200 rpm for 5 min.
Then, the residue was added with 800 ␮L Trypan blue solution and dispersed. After
5 min staining, cells were counted using cytometer. The dead cells were stained with
blue color. Cell mortality (%) is expressed as percentage of the dead cell number/the
total cell number.
2.6. Membrane integrity
LDH test-kit (CytoTox 96® Non-Radioactive Cytotoxicity Assay, Promega Co.)
was used to assess the cell membrane integrity. A549 cells were plated in the 96-well
plates (5 × 103 cells per well) and incubated for 24 h. GO samples were introduced
separately to the cells with different concentrations (10, 25, 50, 100 and 200 ␮g/mL)
and incubated for another 24 h. The positive control was prepared by adding 10 ␮L
of lysis solution to the control cells at 45 min prior to the centrifugation. Then,
the centrifugation (1200 rpm × 5 min) was performed. One hundred microlitres of
supernatant was taken out from each well for LDH assay following the instruction
of the kit. The absorbance at 490 nm was recorded on a Microplate Reader (Thermo,
Varioskan Flash). The LDH leakage (% of positive control) is expressed as the percent-
age of (ODtest − ODblank )/(ODpositive − ODblank ), where ODtest is the optical density of
the control cells or cells exposed to GO, ODpositive is the optical density of the positive
control cells and ODblank is the optical density of the wells without A549 cells.
2.7. Apoptosis assay
Apoptosis kit (FITC Annexin V Apoptosis Detection Kit I, BD Biosciences, USA)
was employed to detect apoptotic and necrotic cells. The manual of the kit was
strictly followed. Briefly, A549 cells were plated in the 6-well plates (1 × 105 cells
per well) and incubated for 24 h. The GO samples were introduced to the cells at
different concentrations (10, 100 and 200 ␮g/mL) and incubated for another 24 h.
The positive control was prepared by culturing the control cells in medium con-
taining 200 mM H2 O2 for 30 min. A549 cells were collected, washed twice with cold
D-hanks buffer solution, and re-suspended in binding buffer (1 × 106 cells/mL). After
100 ␮L of A549 cells was transferred to a tube, 5 ␮L of FITC-conjugated Annexin V
(Annexin V-FITC) and 5 ␮L of propidium iodide (PI) were added followed by incuba-
tion for 15 min at room temperature in the dark. The stained A549 cells were diluted
by the binding buffer and directly analyzed by the fluorescence-activated cell sort-
ing method (FACS, FACSCalibur, BD Biosciences, USA). The cells were set as positive
depending on the fluorescence intensity of Annexin V-FITC or PI. The positive of
Annexin V-FITC indicates the out-releasing of phospholipid phosphatidylserine (PS),
which happens in the early stage of apoptosis. The positive of PI indicates the dam-
age of cell membrane, which occurs either in the end stage of apoptosis, in necrosis
or in dead cells. Therefore, the apoptotic cells were identified as Annexin V-FITC+
and PI− . The nonviable cells were identified as Annexin V-FITC+ and PI+ and viable
cells as Annexin V-FITC− and PI− .
2.8. Reactive oxygen species (ROS) assay
The oxidant-sensitive dye DCFH-DA was used for ROS detection (Reactive Oxy-
gen Species Assay Kit, Beyotime Institute of Biotechnology, China). A549 cells were
plated in the 96-well plates (5 × 103 cells per well) and incubated for 24 h. GO sam-
ples were introduced to the cells with different concentrations (10, 25, 50, 100
and 200 ␮g/mL) and incubated for another 24 h. The positive controls were pre-
pared by culturing the normal cells with culture medium containing 200 mM H2 O2
at 1 h prior to the addition of DCFH-DA probe. Then, the culture medium for all
cells was replaced by 100 ␮L of new culture medium containing 20 ␮M DCFH-DA.
The cells were washed with D-Hanks buffer solution for three times 1 h later. After
adding 100 ␮L of D-Hanks buffer solution to each well, the fluorescence intensity
was monitored by a Microplate Reader. The ROS level is expressed as the ratio of
(Ftest − Fblank )/(Fcontrol − Fblank ), where Ftest is the fluorescence intensity of the cells
exposed to GO or the positive control, Fcontrol is the fluorescence intensity of the
control cells and Fblank is the fluorescence intensity of the wells without A549 cells.
In order to test the ROS generation of GO in culture medium (cell free), the GO
samples (0, 10, 25, 50, 100 and 200 ␮g/mL) were incubated in F-12K medium sup-
plemented with 10% (v/v) fetal bovine serum for 24 h. The ROS level was measured
following the protocol developed by Lu et al. (2009). Briefly, 0.1 mM DCFH-DA was
chemically hydrolyzed to 2 , 7 -dichlorofluorescein (DCFH) at pH 7.0 with 0.01 M
NaOH at room temperature (30 min in the dark). The chemical reaction was stopped
by adding 200 ␮L PBS. After adding 50 ␮L DCFH solution to GO in culture medium,
the mixture was incubated at 37 ◦ C for 1 h and centrifuged at 4000 rpm for 5 min.
The fluorescence generated by the DCFH oxidation was measured on a Microplate
Reader.
2.9. Cell growth on GO films
GO films were prepared by evaporating 4 mL of m-GO (1 mg/mL) in a 35 mm
culture dish under 80 ◦ C. A549 cells were plated in the GO coated dish (2 × 105 cells
per dish) and incubated for 24 h for morphology observation. A549 cells cultured in
normal dishes were taken as the control.
2.10. Statistical analysis
Except ultrastructural investigation, six parallel tests were conducted for each
sample. All data are presented as the mean with the standard deviation (mean ± SD).
Significance has been calculated using Student’s t-test. * denotes a statistical signif-
icance (*≤0.05 and **≤0.01) vs the control.
 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210
Fig. 1. Characterization of GO samples. (a–c) AFM images of l-GO (a), s-GO (b) and m-GO (c); (d) Raman spectra of l-GO, s-GO and m-GO.
3. Results
3.1. Characterization of GO
Except for the size, the three GO samples, m-GO, l-GO and s-GO,
are very similar. Fig. 1 shows the representative AFM images of the
three GO samples. Most of GO sheets exist as single or few layers.
The thickness of the GO layer is around 0.9 nm according to AFM
measurement (Fig. S1). Both large and small sheets are presented
in m-GO (430 ± 300 nm). The size of l-GO sheets (780 ± 410 nm) is
larger than that of s-GO sheets (160 ± 90 nm). In aqueous suspen-
sion, the average hydrodynamic diameter (Dh ) is 588 nm for m-GO,
556 nm for l-GO, 148 nm for s-GO, according to Nanosizer measure-
ments (Table S1). Three GO samples have very similar FTIR spectra
(Fig. S2). The broad absorption at 3400 cm−1 suggests the existence
of –COOH and –OH groups. The absorption at 1720 cm−1 corre-
sponds to C O bonds. The oxygen contents based on XPS analysis
are 33.1% for l-GO, 37.0% for s-GO and 35.8% for m-GO (Fig. S3). The
ID /IG values in Raman spectra, which indicate the defect content,
are very close among the three samples (Fig. 1d and Table S1).
3.2. Cell morphology
The morphology is one important indicator of the status of cells.
The cell morphological changes after GO exposure were recorded to
demonstrate the effect of GO on A549 cells directly (Fig. 2). There is
no obvious difference between the GO-treated cells and the control
cells. Most cells adhere to the substrate tightly and are in normal
spindle-shape.
203
We also checked the influence of GO on the cell attachment fol-
lowing Wang et al.’s method with some modifications (Wang et al.,
2010). Compared with control cells, GO-treated cells do not show
any difference in their adhesion to the culture dish (Fig. S7).
3.3. Cell viability
The cell viability is assayed to estimate the toxicity of GO sam-
ples quantitatively by CCK-8 assay (Fig. 3), in which the formation
of formazan dye depends on the mitochondria activity. As a whole,
the viability loss is dose-related. At higher GO concentrations, the
viability loss is observed. Size is another factor on viability. The
influence of l-GO and m-GO on the viability of A549 cells is tiny.
Even at the highest concentration of 200 ␮g/mL, more than 80% of
the cell viability remains. However, s-GO induces more viability
loss than l-GO and m-GO. At 200 ␮g/mL of GO, the cell viability is
67% at 24 h postexposure. Culture period has little influence on the
viability. Similar results were obtained from 24, 48 and 72 h expo-
sure (Figs. 3, S4 and S5). Therefore, the followed experiments were
performed by 24 h exposure.
Nutrient depletion induced by nanomaterial adsorption is a well
recognized reason for the nanotoxicity. Therefore, we tested the
influence of culture medium adsorption on the toxicity of GO. F-12K
medium was pre-treated with GO samples separately for 24 h and
the supernatants were collected for A549 cell culture. If the cells
did not survive in the GO-pretreated culture medium, we could
conclude that the adsorption of nutrients on GO contributes to the
toxicity. But the cells grew just as well as the control cells. The cell
viability does not decrease along with GO concentrations (Fig. 4).
204 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210

Fig. 2. Optical microscopy images of GO-treated A549 cells. (a) l-GO; (b) s-GO; (c) m-GO; (d) the control.

This reveals that the adsorption of nutrients on GO sheets from

medium does not affect the status of cells under our experiment
condition.

3.4. Cell mortality

While viability shows the activity of cell mitochondria, the

mortality indicates the death of cell. Here, the cell mortality is
monitored by Trypan blue exclusion assay, in which the dead cells
are stained into blue while the live ones remain unchanged. The
mortality is expressed by the ratio of dead cells in all cells. While
there is the viability loss induced by GO exposure, compared to
the control, no mortality increase of A549 cells is observed after
GO treatment (Fig. 5). The mortality remains around 1.5% upon the
exposure, nearly the same as that of the control (1.4%).

3.5. Membrane integrity

When the membrane is damaged, the intracellular LDH

Fig. 3. The viability of A549 cells after exposed to GO for 24 h.
molecules would be released into the culture medium. Therefore,
LDH level out of cells reflects the cell membrane integrity. Inter-
estingly, GO exposure does not induce, but restrains LDH leakage
(Fig. 6). The LDH levels of GO-treated cells are even slightly lower
than that of the control cells (7.5%). For example, at GO concen-
tration of 200 ␮g/mL, the LDH leakage level is around 6% for the
GO-treated cells. In contrast, the size of GO samples has ignorable
influence on the LDH leakage.
 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210 205

Fig. 4. The viability of A549 cells after exposed to the supernatant of GO pre-
incubated medium for 24 h. Fig. 6. The influence of GO on the membrane integrity of A549 cells.

3.6. Cell apoptosis

GO does not induce any apoptosis or necrosis of A549 cells

(Fig. 7). The apoptosis level is not relevant to the dose or the size of
the GO samples. At the concentration of 200 ␮g/mL, the apoptosis
rates (1.1–2.4%) are still comparative with that of the control cells
(1.5%). In contrast, the positive control cells, which treated with
200 mM H2 O2 for 30 min, show much serious apoptosis (32.4%) and
necrosis (54.3%).

3.7. Ultrastructure investigation

The ultra-section of A549 cells was observed under TEM for the
uptake of GO and the changes of ultrastructure. All GO treated cells
show similar structures to the control cells (Fig. 8). GO exposure
does not have any obvious impact on the ultrastructure of A549
cells. We did not find any GO sheets inside cells, either.

3.8. ROS level

The ROS generation is one commonly proposed toxicological

mechanism of nanoparticles. The GO exposure induces oxidative
stress in A549 cells even at low concentrations (Fig. 9). GO with
higher concentrations induces more ROS. Among the three GO sam-
ples, s-GO causes the most serious oxidative stress. For example, at
200 ␮g/mL, the ROS level for s-GO treated cells is 3.9 times of con-
trol, while it is 2.6 for l-GO treated cells and 2.1 for m-GO treated
cells. However, for the positive control, the ROS level is 12.0 times
of control, much higher than that of GO-treated cells under the
same condition. There is no meaningful difference between the cells
exposed to l-GO and m-GO.
Fig. 5. The influence of GO on the mortality of A549 cells. In the F-12K medium (cell free), GO induces the GO dose-
dependent ROS generation. Higher GO dose brings on the higher
level of ROS (Fig. 10). However, it is l-GO, not s-GO or m-GO,
206 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210

Fig. 7. FACS results of the Annexin V-FITC and PI assay. (a–d) Scatter diagrams of cells exposed to 200 ␮g/mL of l-GO (a), s-GO (b), m-GO (c) and the negative control (d). (e
and f) The summary of the apoptosis rate (e) and necrosis rate (f) of A549 cells after exposed to GO for 24 h.

shows high ability in generating ROS. At 200 ␮g/mL, the fluores-
cence intensity from l-GO sample is 50% higher than that from s-GO
or m-GO.
3.9. Cell growth on GO substrate
The cells grow very well on the GO film. The density and mor-
phology of the cells cultured on GO film are comparative to those
of the cells cultured in normal culture dish (Fig. 11). The thickness
of the GO film is around several tens of micrometers. Due to the
dark brown color of GO film, the contrast of Fig. 11a is not as good
as that of the control.
4. Discussion
Nanomaterials have unique physicochemical properties and are
applied in various areas. However, their biological properties in
organisms will finally determine their destiny in future. Com-
pared to available results of carbon based nanomaterials, such as
fullerene, CNT, carbon nanofibre and carbon nanoparticle (Jia et al.,
 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210 207

Fig. 9. The influence of GO on the ROS level of A549 cells.

Fig. 8. TEM images of the m-GO treated A549 cells (a) and the control cells (b).

2005; Lewinski et al., 2008; Lindberg et al., 2009; Liu et al., 2010;
Tian et al., 2006), our results indicate that GO has pretty good bio-
compatibility to A549 cells.
A systematic study was performed to evaluate the toxic-
ity/biocompatibility of GO to A549 cell, a widely used model cell
line for the toxicity study. The results collectively indicate that GO is
highly biocompatible, which is consistent with the GO drug delivery
studies (Liu et al., 2008; Lu et al., 2010; Sun et al., 2008; Yang et al.,
2009b; Zhang et al., 2010a). In addition to the literature reporting
the good biocompatibility of GO, there is literature reporting that
GO has higher toxicity to cells and animals at high concentrations Fig. 10. GO induced ROS generation in F-12K culture medium.
(Agarwal et al., 2010; Hu et al., 2010; Wang et al., 2010). For exam-
ple, Wang et al. found that GO is toxic to human fibroblast cells at
the concentration of 50 ␮g/mL and higher. The inconsistency might without obvious toxicity. Our study suggests that GO could be used
come from the GO synthesis/film preparation, and the testing mod- as the cell growth substrate.
els. The good biocompatibility of GO sheets is also reflected by the CNT is the closest material of graphene (Geim and Novoselov,
cell growth on GO film. Unlike Agarwal’s report (Agarwal et al., 2007). The toxicity of CNTs is heavily influenced by their function-
2010), we found that A549 cells grew very well on the GO film alization degree (Sayes et al., 2006). For example, carboxylation
208 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210
Fig. 11. The microscopic images of cells grown on the GO film (a) and in normal cell
culture dish (b).
of CNTs makes CNTs abundant in oxygen atoms, and decreases
their toxicity. GO contains many oxygen atoms in the forms of car-
boxyl groups, epoxy groups and hydroxyl groups (Dreyer et al.,
2010). The functionalization degree of GO is generally higher
than that of carboxylated CNTs according to the oxygen content.
Therefore, the good biocompatibility of GO is generally expected
to this regard. Comparing to the very recent toxicity results of
graphene (Zhang et al., 2010b), we find that GO has much lower
toxicity, which is indicated by results of the viability assay and
LDH leakage assay. This supports the phenomenon obtained from
CNTs studies, i.e. functionalization decreases the toxicity of CNTs.
Another aspect might contribute to the high biocompatibility is the
two-dimensional structure of GO. The distinct difference between
GO and CNTs is GO’s two-dimensional structure and CNTs’ one-
dimension (Geim and Novoselov, 2007). Although the effect of
shape on the toxicity is still unknown in detail to date, many pre-
vious results show that the shape affects the biological fate of
nanomaterials (Oh et al., 2010; Simon-Deckers et al., 2009).
In addition, GO is not found in cells by TEM investigation. This
might contribute to the high biocompatibility of GO, too. It is dif-
ficult to investigate the monolayer GO sheet in biological samples
under TEM. However, GO folds and aggregates when being added
into the culture medium. The aggregates in culture medium made
GO distinguishable under TEM (Fig. S6), compared with it in pure
water. Therefore, if there is any in cells, we may find it easily. For
example, Wang et al. observed GO aggregates in human fibroblast
cells (Wang et al., 2010). Based on all experimental observations
in this study, GO is hardly swallowed by A549 cells. The difference
between our results and Wang et al.’s results might come from the
different sample properties and cell lines. In the drug/DNA delivery
studies, GO was found entering the cells with cargo (Liu et al., 2008;
Lu et al., 2010; Sun et al., 2008). But, the size of GO they used is less
than 100 nm, even reached 5 nm. The uptake of carbon nanomateri-
als by cells is a widely observed phenomenon (Lewinski et al., 2008;
Raffa et al., 2010). In particular, the negative charged fullerene and
CNTs are easily swallowed by different cells (Li et al., 2008b; Wang
et al., 2009a, 2009c). However, the uptake of nanomaterials is reg-
ulated by their size. For example, the CNTs accumulation in cells is
length-dependent (Becker et al., 2007; Jin et al., 2009; Raffa et al.,
2008). CNTs with length longer than 2 ␮m can hardly enter cells
(Raffa et al., 2008). Therefore, the size of Go might be the control
factor of inhibiting the endocytosis of GO. The aggregation of GO
in culture medium may take the consequences too (Wick et al.,
2007), though the hydrodynamic diameters of three GO samples
are less than 2 ␮m in pure water. We propose that the shape, size
and aggregation of GO sheets affect the uptake.
Considering that GO is not observed inside the A549 cells,
GO more possibly interact with the cells on the cellular surface
or via other pathway indirectly. The interaction on the cellular
surface may be reflected by the membrane integrity evaluation.
Surprisingly, the LDH leakage levels of cells treated with high con-
centration GO are lower than that of the control cells. It could hardly
be regarded that GO exposure improves the membrane integrity,
but most likely the leakage tunnels are partially blocked by GO
covering. This hints that GO might partially block the substance
exchange of A549 cells. The reduced LDH leakage might be a distinct
character of sheet-like GO, since it is not reported in the toxicity
study of fullerene, CNTs and other carbon nanoparticles.
As for the indirect interaction, one possibility is that GO absorbs
the nutrients in culture medium and then the depletion of nutri-
ents induces the oxidative stress and toxicity to A549 cells. Such
toxicity mechanism has been reported in the study of CNTs (Guo
et al., 2008; Liu et al., 2009). Guo et al. reported that the depletion
of nutrients by the absorption onto CNTs led to severe toxicity to
HepG2 cells (Guo et al., 2008). The theoretical calculations have pre-
dicted the absorption of amino acid and other biological molecules
onto graphene (Qin et al., 2010; Rajesh et al., 2009). We mixed GO
and culture medium for 24 h, then centrifuged the mixture to pre-
cipitate GO. The supernatant was used to culture cells. No toxicity
to A549 cells was found (Fig. 4), compared with the cells incu-
bated with the normal culture medium. Therefore, the absorption
of nutrients from the culture medium does not influence A549 cells
under the experimental condition in this study.
Another possibility is that GO influences the cell adhesion ability
of A549 cells. However, GO shows ignorable influence on the cell
adhesion ability of A549 cells (Fig. S7). The unaffected adhesion
ability is also indicated in the GO film evaluation. Our results clearly
suggest that cells adhere to GO membrane steadily.
Although GO hardly enters A549 cells and the mortal-
ity/apoptosis of GO-treated A549 cells is the same as that of the
control cells, GO induces statistically significant ROS generation,
even at low concentration. Oxidative stress is a well recognized
toxicological mechanism of various nanoparticles (Lewinski et al.,
2008; Li et al., 2008a; Pulskamp et al., 2007; Yang et al., 2008b).
 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210
At low dose, GO induces ROS generation, but no obvious toxicity is
observed. Similarly, oxidative stress was observed while no toxic-
ity of CNTs to cells presented in Pulskamp et al.’s study (Pulskamp
et al., 2007). The oxidative stress may contribute to the slight via-
bility decrease of GO at high concentration. The oxidative stress
induced by GO is moderately low when comparing to fullerene and
CNTs (Lewinski et al., 2008). Further, ROS generate when incubat-
ing GO with the culture medium alone (cell free) and ROS level is
GO concentration depended. Hence, the intracellular ROS is most
likely induced by the external ROS. To make such a mechanism
clear, more efforts are required.
5. Conclusions
In summary, the toxicity of GO to A549 cells was evaluated by
various cytotoxicity methods. GO hardly enters cells and shows
good biocompatibility. GO has potential being the substrate for the
cell growth. However, GO arouses oxidative stress, and induces the
slight decrease of the cell viability at high GO dose. The effect of
GO on A549 cells is dose and size related. Our results are essential
for the biomedical applications and safety assessment of GO and
would stimulate more toxicology evaluations of graphene and its
derivatives.
Conflict of interest
There are no conflicts of interest.
Acknowledgements
We acknowledge financial support from the China Nat-
ural Science Foundation (No. 21071094), the National Basic
Research Program of China (973 Program Nos. 2011CB933402 and
2009CB930200), Shanghai MEC (11ZZ82) and Shanghai Leading
Academic Disciplines (S30109).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.toxlet.2010.11.016.
References
Agarwal, S., Zhou, X., Ye, F., He, Q., Chen, G.C.K., Soo, J., Beoy, F., Zhang, H.,
Chen, P., 2010. Interfacing live cells with nanocarbon substrates. Langmuir 26,
2244–2247.
Aillon, K.L., Xie, Y.M., El-Gendy, N., Berkland, C.J., Forrest, M.L., 2009. Effects of nano-
material physicochemical properties on in vivo toxicity. Adv. Drug Deliv. Rev.
61, 457–466.
Allen, M.J., Tung, V.C., Kaner, R.B., 2010. Honeycomb carbon: a review of graphene.
Chem. Rev. 110, 132–145.
Becker, M.L., Fagan, J.A., Gallant, N.D., Bauer, B.J., Bajpai, V., Hobbie, E.K., Lacerda, S.H.,
Migler, K.B., Jakupciak, J.P., 2007. Length-dependent uptake of DNA-wrapped
single-walled carbon nanotubes. Adv. Mater. 19, 939–945.
Cao, A., Liu, Z., Chu, S., Wu, M., Ye, Z., Cai, Z., Chang, Y., Wang, S., Gong, Q., Liu, Y., 2010.
A facile one-step method to produce graphene–CdS quantum dot nanocompos-
ites as promising optoelectronic materials. Adv. Mater. 22, 103–106.
Deng, X., Jia, G., Wang, H., Sun, H., Wang, X., Yang, S., Wang, T., Liu, Y., 2007. Transloca-
tion and fate of multi-walled carbon nanotubes in vivo. Carbon 45, 1419–1424.
Dikin, D.A., Stankovich, S., Zimney, E.J., Piner, R.D., Dommett, G.H.B., Evmenenko,
G., Nguyen, S.T., Ruoff, R.S., 2007. Preparation and characterization of graphene
oxide paper. Nature 448, 457–460.
Dreyer, D.R., Park, S., Bielawski, C.W., Ruoff, R.S., 2010. The chemistry of graphene
oxide. Chem. Soc. Rev. 39, 228–240.
Eda, G., Fanchini, G., Chhowalla, M., 2008. Large-area ultrathin films of reduced
graphene oxide as a transparent and flexible electronic material. Nat. Nanotech-
nol. 3, 270–274.
Geim, A.K., 2009. Graphene: status and prospects. Science 324, 1530–1534.
Geim, A.K., Novoselov, K.S., 2007. The rise of graphene. Nat. Mater. 6, 183–191.
Guo, L., Bussche, A.V.D., Buechner, M., Yan, A., Kane, A.B., Hurt, R.H., 2008. Adsorp-
tion of essential micronutrients by carbon nanotubes and the implications for
nanotoxicity testing. Small 4, 721–727.
209
Herzog, E., Casey, A., Lyng, F.M., Chambers, G., Byrne, H.J., Davoren, M., 2007. A new
approach to the toxicity testing of carbon-based nanomaterials—the clonogenic
assay. Toxicol. Lett. 174, 49–60.
Hu, W., Peng, C., Luo, W., Lv, M., Li, X., Li, D., Huang, Q., Fan, C., 2010. Graphene-based
antibacterial paper. ACS Nano 4, 4317–4323.
Hummers Jr., W.S., Offerman, R.E., 1958. Preparation of graphitic oxide. J. Am. Chem.
Soc. 80, 1339.
Hussain, S.M., Braydich-Stolle, L.K., Schrand, A.M., Murdock, R.C., Yu, K.O., Mat-
tie, D.M., Schlager, J.J., Terrones, M., 2009. Toxicity evaluation for safe use of
nanomaterials: recent achievements and technical challenges. Adv. Mater. 21,
1549–1559.
Jia, G., Wang, H., Yan, L., Wang, X., Pei, R., Yan, T., Zhao, Y., Guo, X., 2005. Cytotox-
icity of carbon nanomaterials: single-wall nanotube, multi-wall nanotube and
fullerene. Environ. Sci. Technol. 39, 1378–1383.
Jin, H., Heller, D.A., Sharma, R., Strano, M.S., 2009. Size-dependent cellular uptake
and expulsion of single-walled carbon nanotubes: single particle tracking and a
generic uptake model for nanoparticles. ACS Nano 3, 149–158.
Kovtyukhova, N.I., Ollivier, P.J., Martin, B.R., Mallouk, T.E., Chizhik, S.A., Buzaneva,
E.V., Gorchinskiy, A.D., 1999. Layer-by-layer assembly of ultrathin composite
films from micron-sized graphite oxide sheets and polycations. Chem. Mater.
11, 771–778.
Lewinski, N., Colvin, V., Drezek, R., 2008. Cytotoxicity of nanoparticles. Small 4,
26–49.
Li, N., Xia, T., Nel, A.E., 2008a. The role of oxidative stress in ambient particulate
matter-induced lung diseases and its implications in the toxicity of engineered
nanoparticles. Free Radic. Biol. Med. 44, 1689–1699.
Li, W., Chen, C., Ye, C., Wei, T., Zhao, Y., Lao, F., Chen, Z., Meng, H., Gao, Y., Yuan, H.,
Xing, G., Zhao, F., Chai, Z., Zhang, X., Yang, F., Han, D., Tang, X., Zhang, Y., 2008b.
The translocation of fullerenic nanoparticles into lysosome via the pathway of
clathrin-mediated endocytosis. Nanotechnology 19, 145102.
Lindberg, H.K., Falck, G.C.-M., Suhonen, S., Vippola, M., Vanhala, E., Catalán, J.,
Savolainen, K., Norppa, H., 2009. Genotoxicity of nanomaterials: DNA damage
and micronuclei induced by carbon nanotubes and graphite nanofibres in human
bronchial epithelial cells in vitro. Toxicol. Lett. 186, 166–173.
Liu, J.-H., Anilkumar, P., Cao, L., Wang, X., Yang, S.-T., Luo, P.G., Wang, H., Lu, F.,
Meziani, M.J., Liu, Y., Korch, K., Sun, Y.-P., 2010. Cytotoxicity evaluations of flu-
orescent carbon nanoparticles. Nano Life 1, 153–161.
Liu, J., Yang, L., Hopfinger, A.J., 2009. Affinity of drugs and small biologically active
molecules to carbon nanotubes: a pharmacodynamics and nanotoxicity factor?
Mol. Pharm. 6, 873–882.
Liu, Z., Robinson, J.T., Sun, X., Dai, H., 2008. PEGylated nanographene oxide for deliv-
ery of water-insoluble cancer drugs. J. Am. Chem. Soc. 130, 10876–10877.
Lu, C.H., Zhu, C.L., Li, J., Liu, J.J., Chen, X., Yang, H.H., 2010. Using graphene to protect
DNA from cleavage during cellular delivery. Chem. Commun. 46, 3116–3118.
Lu, S., Duffin, R., Poland, C., Daly, P., Murphy, F., Drost, E., MacNee, W., Stone, V.,
Donaldson, K., 2009. Efficacy of simple short-term in vitro assays for predicting
the potential of metal oxide nanoparticles to cause pulmonary inflammation.
Environ. Health Perspect. 117, 241–247.
Nel, A., Xia, T., Mädler, L., Li, N., 2006. Toxic potential of materials at the nanolevel.
Science 311, 622–627.
Neto, A.H.C., Guinea, F., Peres, N.M.R., Novoselov, K.S., Geim, A.K., 2009. The electronic
properties of graphene. Rev. Mod. Phys. 81, 109–162.
Oberdörster, G., Oberdörster, E., Oberdörster, J., 2005. Nanotoxicology: an emerging
discipline evolving from studies of ultrafine particles. Environ. Health Perspect.
113, 823–839.
Oh, W.K., Kim, S., Yoon, H., Jang, J., 2010. Shape-dependent cytotoxicity and
proinflammatory response of poly(3,4-ethylenedioxythiophene) nanomateri-
als. Small 6, 872–879.
Park, S., Ruoff, R.S., 2009. Chemical methods for the production of graphenes. Nat.
Nanotechnol. 4, 217–224.
Pulskamp, K., Diabaté, S., Krug, H.F., 2007. Carbon nanotubes show no sign of acute
toxicity but induce intracellular reactive oxygen species in dependence on con-
taminants. Toxicol. Lett. 168, 58–74.
Qin, W., Li, X., Bian, W.W., Fan, X.J., Qi, J.Y., 2010. Density functional theory calcula-
tions and molecular dynamics simulations of the adsorption of biomolecules on
graphene surfaces. Biomaterials 31, 1007–1016.
Raffa, V., Ciofani, G., Nitodas, S., Karachalios, T., D’Alessandro, D., Masini, M.,
Cuschieri, A., 2008. Can the properties of carbon nanotubes influence their inter-
nalization by living cells? Carbon 46, 1600–1610.
Raffa, V., Ciofani, G., Vittorio, O., Riggio, C., Cuschieri, A., 2010. Physicochemical prop-
erties affecting cellular uptake of carbon nanotubes. Nanomedicine 5, 89–97.
Rajesh, C., Majumder, C., Mizuseki, H., Kawazoe, Y., 2009. A theoretical study on the
interaction of aromatic amino acids with graphene and single walled carbon
nanotube. J. Chem. Phys. 130, 124911.
Rao, C.N.R., Sood, A.K., Subrahmanyam, K.S., Govindaraj, A., 2009. Graphene: the new
two-dimensional nanomaterial. Angew. Chem. Int. Ed. 48, 7752–7777.
Sayes, C.M., Liang, F., Hudson, J.L., Mendez, J., Guo, W., Beach, J.M., Moore, V.C., Doyle,
C.D., West, J.L., Billups, W.E., Ausman, K.D., Colvin, V.L., 2006. Functionalization
density dependence of single-walled carbon nanotubes cytotoxicity in vitro.
Toxicol. Lett. 161, 135–142.
Scheuermann, G.M., Rumi, L., Steurer, P., Bannwarth, W., Mülhaupt, R., 2009.
Palladium nanoparticles on graphite oxide and its functionalized graphene
derivatives as highly active catalysts for the Suzuki–Miyaura coupling reaction.
J. Am. Chem. Soc. 131, 8262–8270.
Simon-Deckers, A., Loo, S., Mayne-L’hermite, M., Herlin-Boime, N., Menguy, N., Rey-
naud, C., Gouget, B., Carrière, M., 2009. Size-, composition- and shape-dependent
210 Y. Chang et al. / Toxicology Letters 200 (2011) 201–210
toxicological impact of metal oxide nanoparticles and carbon nanotubes toward
bacteria. Environ. Sci. Technol. 43, 8423–8429.
Sun, X.M., Liu, Z., Welsher, K., Robinson, J.T., Goodwin, A., Zaric, S., Dai, H., 2008. Nano-
graphene oxide for cellular imaging and drug delivery. Nano Res. 1, 203–212.
Tian, F., Cui, D., Schwarz, H., Estrada, G.G., Kobayashi, H., 2006. Cytotoxicity of single-
wall carbon nanotubes on human fibroblasts. Toxicol. In Vitro 20, 1202–1212.
Wang, H., Wang, J., Deng, X., Sun, H., Shi, Z., Gu, Z., Liu, Y., Zhao, Y., 2004. Biodis-
tribution of carbon single-wall nanotubes in mice. J. Nanosci. Nanotechnol. 4,
1019–1024.
Wang, J., Deng, X., Yang, S.T., Wang, H., Zhao, Y., Liu, Y., 2008. Rapid translocation
and pharmacokinetics of hydroxylated single-walled carbon nanotubes in mice.
Nanotoxicology 2, 28–32.
Wang, J., Sun, R.H., Zhang, N., Nie, H., Liu, J.H., Wang, J.N., Wang, H., Liu, Y., 2009a.
Multi-walled carbon nanotubes do not impair immune functions of dendritic
cells. Carbon 47, 1752–1760.
Wang, K., Ruan, J., Song, H., Zhang, J., Wo, Y., Guo, S., Cui, D., 2010. Biocompatibility
of graphene oxide. Nanoscale Res. Lett., doi:10.1007/s11671-010-9751-6.
Wang, L., Lee, K., Sun, Y.Y., Lucking, M., Chen, Z., Zhao, J.J., Zhang, S.B., 2009b.
Graphene oxide as an ideal substrate for hydrogen storage. ACS Nano 3,
2995–3000.
Wang, X., Jia, G., Wang, H., Nie, H., Yan, L., Deng, X., Wang, S., 2009c. Diameter effects
on cytotoxicity of multi-walled carbon nanotubes. J. Nanosci. Nanotechnol. 9,
3025–3033.
Wick, P., Manser, P., Limbach, L.K., Dettlaff-Weglikowska, U., Krumeich, F., Roth,
S., Stark, W.J., Bruinink, A., 2007. The degree and kind of agglomeration affect
carbon nanotube cytotoxicity. Toxicol. Lett. 168, 121–131.
Xia, T., Li, N., Nel, A.E., 2009. Potential health impact of nanoparticles. Ann. Rev. Public
Health 30, 137–150.
Yang, S.T., Fernand, K.A.S., Liu, J.H., Wang, J., Sun, H.F., Liu, Y., Chen, M., Huang, Y.,
Wang, X., Wang, H., Sun, Y.P., 2008a. Covalently PEGylated carbon nanotubes
with stealth character in vivo. Small 4, 940–944.
Yang, S.T., Wang, H., Meziani, M.J., Liu, Y., Wang, X., Sun, Y.P., 2009a. Biode-
functionalization of functionalized single-walled carbon nanotubes in mice.
Biomacromolecules 10, 2009–2012.
Yang, S.T., Wang, X., Jia, G., Gu, Y., Wang, T., Nie, H., Ge, C., Wang, H., Liu, Y., 2008b.
Long-term accumulation and low toxicity of single-walled carbon nanotubes in
intravenously exposed mice. Toxicol. Lett. 181, 182–189.
Yang, W., Ratinac, K.R., Ringer, S.P., Thordarson, P., Gooding, J.J., Braet, F., 2010.
Carbon nanomaterials in biosensors: should you use nanotubes or graphene?
Angew. Chem. Int. Ed. 49, 2114–2138.
Yang, X.Y., Zhang, X.Y., Ma, Y.F., Huang, Y., Wang, Y.S., Chen, Y.S., 2009b. Superpara-
magnetic graphene oxide–Fe3 O4 nanoparticles hybrid for controlled targeted
drug carriers. J. Mater. Chem. 19, 2710–2714.
Zhang, L.M., Xia, J.G., Zhao, Q.H., Liu, L.W., Zhang, Z.J., 2010a. Functional graphene
oxide as a nanocarrier for controlled loading and targeted delivery of mixed
anticancer drugs. Small 6, 537–544.
Zhang, Y., Ali, S.F., Dervishi, E., Xu, Y., Li, Z., Casciano, D., Biris, A.S.,
2010b. Cytotoxicity effects of graphene and single-wall carbon nanotubes
in neural phaeochromocytoma-derived PC12 cells. ACS Nano 4, 3181–
3186.