Nanomedicine: Nanotechnology, Biology, and Medicine

20 (2019) 102011

nanomedjournal.com

Gold nanoclusters as a contrast agent for image-guided surgery of head

and neck tumors
Cindy Colombé, MD a, b , Xavier Le Guével, PhD a , Angela Martin-Serrano, PhD c ,
Maxime Henry a , Estelle Porret a , Clothilde Comby-Zerbino d , Rodolphe Antoine, PhD d ,
Ihab Atallah, MD, PhD b , Benoit Busser, PharmD, PhD a, b , Jean-Luc Coll, PhD a ,
Christian Adrien Righini, MD, PhD a, b , Lucie Sancey, PhD a,⁎
Cancer Targets & Experimental Therapeutics, Institute for Advanced Biosciences (IAB), University of Grenoble Alpes- INSERM U1209 – CNRS UMR 5309,
a

Grenoble, France
b
Grenoble Alpes University Hospital, Grenoble, France
c
Research Laboratory and Allergy Service, IBIMA, Regional University Malaga Hospital, UMA, 29009 Malaga, Spain and Andalusian Center for
Nanomedicine and Biotechnology - BIONAND, Málaga, Spain
d
Institut lumière matière, UMR5306, Université Claude Bernard Lyon1-CNRS, Université de Lyon, Villeurbanne cedex, France
Revised 23 April 2019

Abstract

With the objective to evaluate the potential of ultra-small gold (Au) nanoclusters (NCs) for optical image-guided surgery, we synthesized
and characterized AuNCs shelled by zwitterionic or pegylated ligands. The toxicity of the different AuNCs was evaluated on the Head and
Neck Squamous Cell Carcinoma (HNSCC) CAL-33 and SQ20B cell lines in vitro. The safer AuNCs were administrated intravenously to
mice for the determination of the pharmacokinetic properties. Biodistributions were performed on orthotopic CAL-33 HNSCC-bearing mice.
Finally, the AuNCs were used for image-guided surgery, allowing the increase of the survival time vs. control animals, and the number of
animals without any local recurrence.
© 2019 Elsevier Inc. All rights reserved.

Keywords: Gold nanoclusters; Image-guided surgery; Head and Neck; Carcinoma; Nanomedicine

Head and neck cancers are frequently observed at the late function of the margin’s size, i.e. the quantity of healthy tissue
stage when the 5-year survival does not exceed 25%. At late removed in addition to the pathologic area. In this context,
stages, the surgery of Head and Neck Squamous Cell Carcinoma nanotechnology might bring a strong support to overcome these
(HNSCC) remains often fairly complex due to the numerous issues, especially in the field image-guided surgery. 2,3 When the
local structures that should be sustained such as small muscles, surgery is assisted by real-time imaging, the tumor resection
nerves and tendons. The preservation of these healthy structures might be improved with the removal of small residues in
is fundamental in the perspective of restorative surgery. 1 In particular and sufficient margin, the healthy tissues might be
addition, the total removal of tumor should be ensured by the spared, and the time of surgery might be reduced. 4–6 NIR image-
presence of healthy surrounding tissues around the lesion. This guided surgery has already been validated in small animal
step is critical for the post-surgery treatment, which varied as studies 5,7 and is currently under investigation in clinical trials. In
particular, antibody directed against Epidermal Growth Factor
Receptor, i.e. cetuximab-IRDye800CW, is under evaluation in a
clinical trial for the intraoperative assessment of tumor margins
We acknowledge the Auvergne Rhône-Alpes region for the support
CLARA-Oncostarter (CVPPRCAN000162), Plan Cancer (C18038CS), ARC during surgical treatment of HNSCC. 8 In addition, the
(RAC17012CCA and R17157CC) for their support. indocyanine green (ICG) is used to identify the sentinel lymph
⁎Corresponding author. nodes, which are associated with locoregional recurrence in the
E-mail address: lucie.sancey@univ-grenoble-alpes.fr. (L. Sancey). most aggressive HNSCC. 9 Recently, Warram's group illustrated
https://doi.org/10.1016/j.nano.2019.04.014
1549-9634/© 2019 Elsevier Inc. All rights reserved.
2 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011

surgery assisted by real-time imaging. They demonstrated the the lab. 23,24 Both types of AuNC were prepared with low and
valuable adjunct of imaging to the oncologic surgeon, before, high ligand density by increasing the amount of the bidentate
during, and after the surgery, facilitating the detection of occult thiolated organic molecules during the synthesis. For the
lesions in particular. 10 For all these reasons, image-guided zwitterionic ligand, it corresponded to the molar ratio Au:
surgery can be of benefit for patients in terms of post-treatment ZwEt2:NaBH4 = 1:2:2 and 1:5:2. For the PEG ligand, it
procedures, overall survival, side effects, and quality of life. corresponded to the molar ratio Au:PEG:NaBH4 = 1:3:2 and
Gold nanoclusters (AuNCs) are ultra-small particles (b5 nm) 1:5:2. In a typical synthesis for AuPegCOOH 1:3, 26.4 mg of
that emerged as promising renal clearable nanocompounds for PegCOOH was dissolved in 15 mL water at pH 10 (80 μL
cancer diagnostic and therapy. 11–13 Thanks to their photolumi- NaOH (2 M)) under vigorous stirring and 200 μL of HAuCl4
nescence properties in the NIR/SWIR region, 14,15 a spectral (50 mM) was added. After 10 min, 200 μL NaBH4 (100 mM)
window favorable for in vivo monitoring, these probes could be was added dropwise and the solution was kept under mild
tracked in real-time imaging, in various organs. 16 AuNC size stirring for 15 h. Freshly prepared AuPegCOOH was filtered 3
between 10 and 25 gold atoms was recently shown by Zheng's times (Amicon filters 3 kDa) to remove the excess of free ligand
team to strongly affect the pharmacokinetic and the tumor and diluted in PBS at the required concentration. Each stock
accumulation in mice. 17 The physico-chemical properties of the solution was filtered (0.22 μm) before use.
ligands protecting the metal core are other key parameters that The well-described AuNC made of 25 gold atoms stabilized
highly influence the pharmacokinetics and the biodistribution of by the peptide glutathione Au25SG18 was synthesized in two
AuNCs in animals. For instance, short PEG chains stabilizing steps as described by Dugourd and coworkers, 25 and diluted in
AuNCs have enabled to enhance the half-life from 5 to 45 min PBS at the required concentration. Later in this manuscript, the
compared to AuNCs protected by glutathione. 18 Different in AuNCs are cited by the name of their ligands.
vivo studies have demonstrated the potential of AuNCs in
tumor's detection, including brain 15,16,19,20 or breast tumors, 18,21 NC characterization
but to our knowledge only Au NPs (i.e. N10 nm) have been used Absorbance measurements on diluted samples were recorded
for Head and Neck tumors' detection. We previously observed an on an Evolution 201 (ThermoScientific) UV–vis spectropho-
increase of the blood circulation for AuNCs stabilized by a tometer between 300 and 950 nm. Fluorescence spectra were
bidentate zwitterion ligand compared to the well-known NCs recorded on a Perkin Elmer LS45 fluorescence spectrometer.
Au25SG18 from 1.5 to 6.5 min. 15 This led to a stronger Hydrodynamic diameters of the AuNCs were determined by
accumulation in glioblastoma mice models by passive uptake dynamic light scattering (DLS), and their surface charges by zeta
with a high tumor retention over 24 h. 16 potential measurements, both analyses using Malvern Instru-
With the objective to evaluate the potential of AuNCs for ment. For mass spectrometry measurements, a commercial
image-guided surgery of orthotopic HNSCC-bearing mice, we quadrupole time-of-flight mass spectrometer (micro-qTOF,
synthesized AuNCs with different ligands, and performed their Bruker-Daltonics, Bremen, Germany, mass resolution 10,000)
in vitro and in vivo biological investigations. First, the toxicity was used as previously reported 26,27 (see also supplementary
and the cell uptake properties of the different formulations were information).
evaluated on 2 HNSCC cell lines, i.e. CAL-33 and the SQ20B
cells. Then, the most favorable formulations, i.e. safer and well- Cell lines
uptaken by the cells, were administrated intravenously to mice-
bearing tumors for the determination of the biodistribution and This study was conducted using two human head and neck
the pharmacokinetics properties of the selected NCs. Finally, two squamous carcinoma cell lines. The CAL-33 (Leibniz Institute,
AuNCs were used for image-guided surgery, and animals were Braunschweig, Germany) cell line was derived from a tongue
followed-up until tumor recurrence. The image-assisted removed carcinoma and was used for in vitro experiments, and for
residues were also analyzed for tumor invasion by histology. orthotopic models. The second cell line used was the larynx
carcinoma SQ20B cell line (Institute for Advanced Biosciences,
Grenoble, France). The cells were maintained in a monolayer
Methods culture in DMEM (Dulbecco's Modified Eagle's medium)
supplemented with 10% of heat-inactivated FBS (Fetal Bovine
Materials Serum) (v/v) in a humidified incubator (Sanyo, Japan) at 37 °C in
an atmosphere containing 5% CO2. The sub-culturing was
Chemical reagents were purchased from Sigma-Aldrich performed twice a week.
(France) and deionized water was used during the synthesis.
Viability assay
Gold nanoclusters
Cell viability assays were used to evaluate the cytotoxicity of
AuNCs stabilized by bidentate thiolated organic molecules AuNC in vitro. CAL-33 and SQ20B cells were seeded in a 96-
were synthesized following protocols reported earlier. 22 In well plate at a density of 10 5 cells per well and cultured
particular, we chose a zwitterionic ligand LA-ZwEt 2 overnight. Then cells were incubated with 50 or 100 μL of
(C18H36N2O4S3; 440.68 g mol −1) and a short pegylated ligand AuNC/mL of medium for 24 h before medium renewal for the
with a negatively charged terminal group LA-PegnCOOH next 48 h incubation. We used staurosporine as negative control
(C38H74N2O16S2; nPEG = 12; 879.13 g mol −1) synthesized in that induces massive cell death, and no AuNC as positive control
 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011 3

for viability. After washing with PBS, the cell viability was Surgical study
determined using the Presto-Blue viability kit (i.e. 72 h after NCs
exposure). According to the manufacturer's instructions, the Animals were randomized 21 days after tumor implantation
fluorescence was recorded on a microplate reader after 4 h of to perform tumor resection under visual macroscopic guidance
incubation, selecting the fluorescence excitation/emission couple (n = 8), using near infrared (NIR) optical imaging after AuNC
560 nm/590 nm, avoiding any cross-absorption from the gold PegCOOH 1:3 (n = 8, one mouse died from anesthesia and was
atoms. removed from the study reducing to 7 the number of mice in this
group) or ZwEt2 1:5 (n = 8) administration. For all animals, the
surgeon first removed the tumor masses under conventional light
FACS analysis (macroscopic resection), and image-guided surgery was used to
check the remaining fluorescence in the AuNC's groups and to
A total of 500.000 cells (CAL-33 and SQ20B) were plated in assist the surgeon when residual fluorescence was observed. The
a six-well plate with DMEM with 10% FBS. After 24 h, different animals were randomized in the different groups to minimize the
AuNC were added at 50 μL/mL for 4 h at 37 °C and 4 °C, 8 h
differences in the means and standard deviations of the tumor
and 24 h at 37 °C. At the end of incubation, the cells were
volumes. Surgery was performed under general anesthesia
prepared for flow cytometry analysis. The NCs' uptake was
(intraperitoneal injection of a mix of medetomidine (0.2 μg/g) /
measured with LSRII (Becton Dickinson, Le Pont de Claix,
ketamine (0.1 mg/g)) associated with a subcutaneous injection of
France) and photon emission was selected using long-pass
buprenorphine (0.1 mg/kg)), as described in the supplementary
filters, and recorded at 730/45 and 780/60-nm. Data were
information section.
analyzed with the BD FACSDiva 6.3.1 software. One should
note that the fluorescence emission of the NCs, in the NIR region Histology and confocal imaging
prevents the internalization study using conventional
microscopy. The removed residues were embedded into OCT and sliced at
8 μm. The sections were then H&E stained for histology.
Adjacent slices were mounted with a coverglass for confocal
Mice and pharmacokinetic study microscopy analysis. The microscopy of the frozen remaining
tissues was carried out using a confocal microscope (LSM510,
Animal experiments were conducted in accordance with
Zeiss Jena, Germany) with a 10× objective. The AuNCs were
protocols approved by the Ethical Committee of Grenoble, and
excited by a 488-nm, or 405-nm emission laser and their
the French Minister in agreement with the European directive
emission was collected using a long-pass filter at 600 nm.
2010/63/UE. The pharmacokinetic (PK) study was conducted on
12 SWISS mice divided into 4 groups (n = 3/group). The
solutions of 200 μL AuNC were administrated IV at 2.5 mg/mL. Results
The PK of the Au25SG18 was performed previously by our group
in similar condition, and was not performed again for ethical Library of nanoclusters
reasons. 15 Approximately 20 μL of blood was collected in
AuNCs are ultra-small gold particles with core size between 1
heparinized vials from the tail vein at the indicated times after
and 3 nm. To protect the body from the reactive gold species, the
injection. The vials were centrifuged for 20 min to recover the
core of the NCs can be shelled by various ligands. In this study, 5
plasma serum. Then, a drop of 10 μL of serum was imaged
NCs were synthesized: 2 with a lipoic zwitterionic ligand (i.e.
(Optimal, Grenoble, France) and black and white pictures were
ZwEt2), 2 with the pegylated ligand PegCOOH and the last one
acquired with a black-thinned CCD camera at −80 °C (ORCAII-
with glutathione (SG). Thickness of the zwitterionic and
BT-512G, Hamamatsu, Massy, France). Image display and
pegylated layers covalently bonded to the gold core could be
analysis were obtained with the Wasabi® software (Hamamatsu,
tuned during the synthesis while keeping a similar core size as
Massy, France).
represented in the Figure S1. The increase of the organic shell
leads to an obvious increase of the molecular weight of AuNCs,
Head and neck orthotopic tumor model which is confirmed by electron-spray ionization mass spectrom-
etry (ESI-MS). For instance AuZwEt2 1:5 has an average
In total, 65 female athymic NMRI nude mice (5-6 weeks old) molecular weight around 17.5 kDa, which is 22% bigger than
were purchased from Janvier laboratories (Le Genet sur Isle, AuZwEt2 1:2 with a molecular weight around 14.5 kDa (see
France), and used for this study. The complete tumor model can Figure S2). Surprisingly, AuPegCOOH 1:3 is smaller than
be found in the supplementary section, and in Atallah et al. 7 AuZwEt2 1:2 with a molecular weight estimated at ~9 kDa but
an increase of the pegylated ligand during the synthesis for
Biodistribution study AuPegCOOH 1:5 leads to a molecular weight (MW) around
~22 kDa heavier than AuZwEt2 1:5. Despite a similar lipoic acid
After tumor growth, AuNCs were injected IV (200 μl at moiety to anchor the gold core, the difference in size between
[Au] = 5 mg/mL, n = 24 for biodistribution and n = 12 for zwitterionic and pegylated ligands might be related to the
imaging). The nanoparticles' biodistribution was investigated electrostatic interaction between the ligands during the AuNC's
using in vivo fluorescence imaging (Optimal, Grenoble, France), formation in solution. This could influence the number of ligand
as previously described. 16 per AuNCs considering no significant change of the metal core
4 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011

size determined by microscopy measurements (data not shown). the ligands in the NCs' shell did not affect the safety profile of the
However, this observation needs to be considered carefully NCs: both ligand densities were equally safe, except for the
because of the quite broad MW distribution obtained by ESI-MS SQ20B cells exposed to AuPegCOOH (1:3), as described in the
for these AuNCs. In addition and as previous reported, 28 ESI- Figure 1, B.
MS analysis also demonstrated that the effective charges (zeff) The cell uptake of the NCs was evaluated using flow
extracted from the measured Zeta-potential (ζ) of NPs in solution cytometry on CAL-33 cells. As summarized in the Figure 2, after
were partially correlated with the average values of charge 8 h of incubation, AuNCs were moderately internalized. Only
Zaverage of NPs in gas phase. Interestingly, this correlation seems the fluorescence signal of the AuZwEt2 (1:5) and AuPegCOOH
also valid for AuNCs investigated in this study, where a nice (1:3) was recorded inside the CAL-33 cells. Percentages of cells
correlation between the Zeta-potential and the average charge positively stained by the AuNCs reach 17% and 38% for the
state of Zaverage is observed (see Table 1). The atomically precise AuPegCOOH (1:3) and the AuZwEt2 (1:5), respectively (see
AuNCs stabilized by glutathione Au25SG18 have a molecular Figure 1, D). We hypothesized that the weak AuPegCOOH (1:5)
weight of 10.4 kDa. uptake may be in part attributed to a higher contribution of the
Hydrodynamic diameters measured by dynamic light scatter- Peg coating as compared to AuPegCOOH (1:3), which exhibits
ing (DLS) show a size below 10 nm for all types of AuNCs with anti-fouling effect. 32
an increase of diameter for thicker organic shell. From zeta
potential measurements, all AuNCs were negatively charged at In vivo distribution and behavior of the NCs
pH 7, and less charged for the zwitterionic species AuZwEt2.
The main physico-chemical characteristics and properties of the The distribution of the NCs was evaluated in mice bearing
different NCs are summarized in the Table 1. Absorbance subcutaneous CAL-33 tumors, using optical imaging. Organs
profiles for AuZwEt2, AuPegCOOH, and Au25SG18 present were sampled 5 h post-intravenous (IV) administration and
similar pattern with a decrease of signal at longer wavelength analyzed. The results are summarized in the Figure 3, and Table
(Figure S3). All types of AuNCs emit in the red/near-infrared S1. As previously reported, most of the AuNCs were rapidly
region with a blue-shift and an increase of their brightness for eliminated through the urine, i.e. after renal filtration, due to their
thick organic shell (Figure S3, Table 1). This behavior has been small size. 15,16 The remaining NCs were mainly distributed in
extensively reported and is correlated to the nature, the density, the hepato-biliary system (liver and spleen), and the kidneys. We
and the rigidity of the ligands stabilizing the metal core of noticed some uptake in the stomach for AuZwEt2 (1:2) while the
AuNCs. 29–31 other AuNCs were not specifically accumulated in this organ.
This accumulation might be due to its hepato-biliary elimination.
Toxicity profiles and cell internalization of the different NCs The AuNCs were then mostly found in the skin and the tumor,
5 h post injection. Interestingly, the AuZwEt2 1:5, and the
The toxicity of the different NCs was determined at 50 and pegylated AuNCs were rapidly eliminated from the body, with
100 μg gold/mL on two different human HNSCC cell lines, very moderate to low liver and kidney accumulations. Therefore,
namely CAL-33 and SQ30B cells (Figure 1). The viability assay the most promising AuNCs were selected for next PK and
was performed after a 24-h exposure followed by a further 48-h imaging investigations. Due to its long liver retention, the
incubation, i.e. 72 h in total. Due to the natural blue-color of the AuZwEt2 (1:2) was not selected for the in vivo studies. AuZwEt2
AuNCs, the presto-blue assay was performed in fluorescence (1:5), and the pegylated AuNCs were injected to healthy mice to
mode, to avoid any overlapping due to the solutions themselves. determine their plasma half-lives, and compared to Au25SG18.
If most of the tested NCs were not or weakly toxic, the main The blood samples were collected directly after the intravenous
toxicities were measured for the AuPegCOOH, with a particular administration, and over the next 2 h. As indicated in the
susceptibility of the SQ20B cells for AuPegCOOH (1:3) at high Figure 4, the plasma half-lives are longer than that of Au25SG18
concentration (i.e. 100 μg/mL). For these AuNCs, the density of (see also Suppl. Table S2), with prolonged exposure. The nature

Table 1
Main physicochemical characteristics of the AuNCs.
AuNC Size by DLS (nm) Size by ESI-MS Zaverage a Density (kg/m 3) Zeta potential λem. (nm) QY (%) b
(kDa) (mV) at pH 7
AuZwEt2 1:2 3.1 ± 0.7 14.5 3.5 1543.4 −18.0 ± 0.5 780 5.7
AuZwEt2 1:5 8.6 ± 1.1 17.5 3.5 87.2 −18.0 ± 0.2 720 7.1
AuPegCOOH 1:3 7.2 ± 1.6 8-10 5 76.4 −31.4 ± 1.0 780 8.0
AuPegCOOH 1:5 10.5 ± 2.4 20-24 7.5 60.2 −35.5 ± 2.5 720 9.0
Au25GSH18 1.4 ± 0.2 10.4 - - −32.0 ± 2.5 800 0.5
The Zeta-potential was measured at pH 7, and the maximum emission was obtained for any excitation from 350 to 600 nm.
a
The average charge
P state Zaverage, is calculated from the observed charge state distribution in ESI-MS spectra (plotted in m/z) using the
Ni zi
formula:Zaverage ¼ P where Ni corresponds to the number of ions of charge state zi. 28
Ni
b
The quantum yield as determined at the maximum emission wavelength using QD705 as reference.
 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011 5

Figure 1. Cell viability assay performed 72 h after incubation with the NCs at 50 and 100 μg gold/mL on HNSCC CAL-33 (A), and SQ20B (B). Staurosporine
was used as a cell-death inducer (red control diagram). Results are expressed as the mean of cell viability (% vs. control) ± standard deviation (SD).

of the ligand and the thickness of this organic layer stabilizing pegylated AuNCs were also retained in the kidneys around
the gold core have a significant effect on the pharmacokinetics. 10% of ID/g until 48 h post injection. In our case, the density of
All the AuNCs presented plasma half-life fitting with two-phase the peg should be higher, leading to more extended structure of
decay equation, which is a combination of 2 plasma half-lives, the peg around the gold core and to the increase of the final size
i.e. a fast and a slow half-life. Indeed, zwitterionic ligand Et2 of the AuNC. However, in our case, the AuPegCOOH is well
(1:5) prolonged the biopersistence of the AuNCs in the plasma, eliminated from the kidneys, and the blood. The dose-normalized
reaching almost 4 min and 11.5 min for fast and slow half-lives AUC indicated a strong enhancement of the bio-availability of
respectively. For pegylated AuNCs, the fast half-lives were very the AuNCs stabilized by organic ligands, with an increase by a
short with 0.45 and 0.32 min for AuPegCOOH (1:3), and factor 4 to 6.7 for the pegylated, and 10.8 for the zwitterionic
AuPegCOOH (1:5), respectively. The slow half-lives were ligand, respectively, as compared to Au25SG18. Such increased
longer, reaching almost 10 min for AuPegCOOH (1:3), and availability should favor the amount of AuNCs reaching the
more than 13 min for AuPegCOOH (1:5). Such compounds with tumor region.
both fast and slow half-lives are partially absorbed in organs, In the perspective of image-guided surgery, the imaging
such as the liver, before being released from this one into the properties of these AuNCs were determined, in terms of Tumor/
plasma, and available for clearance. Such two-phase decay Skin ratios. This investigation was necessary for the selection of
equation was expected, and previously reported by Zheng and the best AuNCs, able to reach the highest Tumor/Skin imaging
collaborators with pegylated AuNCs. 18 One should note that the ratios in the context of orthotopic tumors, and for the adjustment
Pegylated AuNC from this group was ≈5.5 nm size, but of the surgery protocol, i.e. the time between the injection of the
possesses very long half-lives of 56 min and 9 h. Their AuNC and the surgery itself, for optimized distribution. As
6 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011

Figure 2. NCs' uptake in CAL-33 cells after 8 h of NCs exposure at 50 μg/mL. (A-C) Smooth histograms of the cells normalized to the control condition. The
positive cells were gated with the marker M1, assuming that less than 1% of the control cells induce autofluorescence signal. The percentage of isolated cells
positively stained for the NCs is presented in the table (D).

represented in the Figure 5, the pegylated AuNCs presented After orthotopic CAL-33 tumor engraftment, the tumors were
similar behaviors, i.e. a rapid tumor uptake reaching a Tumor/ measured and the mice were split into 3 groups: (i) control mice whose
Skin ratio value of 1.5 to 2 since the first 30 min post-injection tumors were removed without any contrast agent (n = 8), mice treated
(see Figure 5, B). The zwitterionic AuZwEt2 exhibited a stable with the AuNC, either (ii) AuPegCOOH (1:3) (n = 7), or (iii)
Tumor/Skin ratio of 1.5 from 1 to 5 h post-injection, and AuZwEt2 (1:5) (n = 8), before surgery. The surgeon first removed the
favorable Tumor/Muscle ratio up to 3, similarly to the one main tumor mass using the white light (macroscopic tumor resection)
measured for the pegylated AuNCs (see Figure S4). In view of in all groups, and for the groups (ii) and (iii), used the optical imaging
these data, we decided to use the NCs AuPegCOOH 1:3 and the as a quality control. If any fluorescent mass persisted in the cheek of
AuZwEt2 1:5 for the following image-guided surgery investiga- the animal, the surgeon used the optical live imaging system to localize
tions; however, as both compounds presented similar results, the the tumor residue and to remove it (Figure 6). The residues removed
data were merged for these two AuNCs. by the optical imaging were counted, and further analyzed by

Figure 3. Distribution of the different AuNCs, 5 h after intravenous administration in mice bearing sub-cutaneous CAL-33 tumor. The results are expressed as
the mean of the fluorescence intensity/pixel/100 ms of 3 mice/group, and the error bar represents the SD.
 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011 7

Figure 4. Blood PK of selected NCs after intravenous injection. The NCs concentrations were determined from the fluorescence intensity values of the plasmas
collected just after the injection, and at 5, 10, 15, 30, 60, and 120 min, for AuZwEt2 (1:5) (A), AuPegCOOH (1:3) (B), AuPegCOOH (1:5) (C), and Au25SG18
*
(D). Two PKs parameters are summarized in the table (E). DN AUC0-end corresponds to the area under the curve taken from the injection time to the end of blood
*
plasma sampling, normalized to the dose of 1 mg of gold. DN AUC0-end is expressed as min*mg/mL. T1/2 is the blood plasma half-life expressed in minutes. The
AuNCs were best fitted with two-phase decay equation, generating fast and slow half-lives (see also Table S2 with goodness of fit).

histology to ensure the presence of tumor cells, and by confocal versus 170 mg and 167 mg for the AuNC groups, respectively.
imaging to localize the AuNCs within the tissues (Figure 6, F-G). Such information confirmed that before surgery, the tumor
As shown in Figure 7, A, the main tumor masses masses were statistically similar. Following the surgery, the
macroscopically removed were almost similar in the different animals were monitored for tumor relapse. All the animals of the
groups, reaching a median value of 180 mg for the control group, control group relapsed rapidly and were euthanized for ethical

Figure 5. Kinetic of the tumor-to-skin ratios obtained from the optical analysis. (A) The Tumor/Skin ratios were measured at 0.5, 1, 2.5, and 5 h post injection in
mice-bearing orthotopic CAL-33 tumors. The Tumor/Skin ratios are expressed as the mean of the fluorescent ratios normalized to T = 0 ± SD (n = 3/group).
(B) Example of pictures obtained after administration of AuPegCOOH (1:3).
8 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011

Figure 6. Illustration of the image-guided surgery principle. The mouse was injected with AuPegCOOH (1:3) 30 min before the fluorescence imaging (A). Then,
the main tumor mass was removed (B), and the animal was imaged to check the presence of any tumor residue. In (C), the naturally fluorescent tendon is clearly
seen (white arrow), and a tumor residue is also observed (yellow arrow, and D). This residue presents tumor cells (yellow arrows). (E) Histology image of a
representative removed residue. The yellow arrows indicate the tumor areas. Confocal microscopy confirmed the presence of AuNCs in the outer (F) and in the
center (G) of the tumor residue. Scale bar: 10 μm.

reasons within 40 days, reaching a mean survival time of 34.8 ± sponding to 29.9% and 35.3% increase in life span (ILS) vs.
5.95. The image-guided surgery allowed the increase of the mean control animals, respectively for AuPegCOOH (1:3) and
survival time to 45.1 ± 21.8 days and 47 ± 27.2 days, corre- AuZwEt2 (1:5). The quality of the image-guided surgery

Figure 7. Image-guided surgery and animal survival. (A) The animals were split in 3 groups: without (control) and with image-guided surgery using
AuPegCOOH and AuZwEt2. The tumors were macroscopically removed and weighted. The results are presented as a box-plot with the median value in red. (B)
Animal survival for control and AuNCs groups. (C) The survival time of the animals is expressed as the mean ± the standard deviation in days, but also as the
median survival time (MeST). The increase in life-span (ILS) is expressed as the percentage versus the life-span of control group. The number of mice with
image-guided tumor residue removal is indicated, as the percentage of tumor positive residues.
 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011 9

facilitated the removal of the tumor residues for 8 mice, In our case, both AuNCs presented similar Tumor/Muscle ratio
including 3 mice with prolonged survival (1 for the group justifying their use for image-guided surgery.
AuPegCOOH (1:3) and 2 for the group AuZwEt2 (1:5)) without The optical properties of the AuNCs are due to their ultra-
any local tumor recurrence until the end of the experiment on day small size, and the nature of the coating ligand. 24,41,42 Such
90. All the removed residues were analyzed by histology and properties might be tuned to be helpful for medical applications,
revealed the presence of tumor cells (Figure 6, E, see also Table in particular for image-guided surgery. In HNSCC, margin
S3). In addition, the presence of AuNCs was observed by delineation impacts tumor recurrence and patients' quality of life
confocal imaging on selected residues. The AuNCs were (preservation of vital structures as nerves, swallowing and
heterogeneously distributed in the tumor mass, with a particular speech functions). 43 In this study, the AuNCs allowed to identify
accumulation in some cells into small and cytoplasmic vesicles small remaining tumor lesions, after surgical tumor removal.
(Figure 6, F-G). However, in the NIR optical window, some parts of the cheek as
the tendons are autofluorescent. Therefore, the optimization of the
fluorescence quantum yield would facilitate the detection of the
residues versus the nerves and tendons. In this study, both the two
Discussion types of AuNCs selected showed a tumor-to-healthy tissue
ratio N 1.5 within the first hour after administration allowing a
AuNCs with different surface chemistries were synthesized, rapid, and prolonged tumor detection. The image-guided surgery
characterized and biologically evaluated in this study. The nature was helpful for half of the mice (53% of the animals), allowing the
of the biocompatible ligands protecting the gold metal core and detection and removal of undetected tumor residues. Such strategy
their density have been tuned to construct 4 different structures increased the mean survival time of the animals of ≈30/35% as
that were compared to the well-defined Au25SG18 NCs. An compared to the control group. However, improvement of the
increase of zwitterionic or pegylated coating on the gold core AuNCs formulation might increase such favorable results, 7 in
leads a photoluminescence enhancement of the AuNCs due to a particular with augmentation of the fluorescent quantum yield
more efficient energy transfer of the gold shell, and a better leading to higher sensitivity of detection, and tumor accumulation.
protection from the environment. 29,30 As expected, a thicker Most of the AuNCs studied in vivo are stabilized by the peptide
ligand coating on AuNCs that still remains quite small glutathione. 11,13,17 These AuNCs exhibit an efficient renal clearance,
(b22 kDa) seems to enhance the interaction with cell surface, but possess a short blood half-life (b6 min), preventing a strong and
through stronger cell internalization with zwitterionic ligand and prolonged tumor uptake. We recently demonstrated how the
antifouling effect for charged pegylated ligand. selection of thioctic sulfobetaine ligand as stabilizers could slightly
In this study, we confirmed the efficient renal clearance of enhance AuNCs blood circulation time to 6.5 min, and increase
these AuNCs as already reported for such small sized nano- tumor retention in glioblastoma-bearing mice. 15,16 Here, we moved
objects. 16,33 In addition, their rapid renal clearance should to other constructs based on bidentate thiol ligand with charged
contribute to limit their toxicity, if any. However, a small part of pegylated moiety and hydrophobic zwitterionic moiety to prolong
the AuNCs was also trapped by the liver and the spleen, as the AuNC's half-life, while keeping a high renal clearance
already reported by us and other teams. 15,16,34–36 As expected, accompanied by a favorable tumor accumulation and retention.
the pharmacokinetics, i.e. AUC, indicated that smaller AuNCs as From the 2 tested AuNCs, both presented similar properties in terms
Au25SG18 were rapidly eliminated from the blood (half-life of absence of toxicity, cell uptake, and quantum yield. However the
b2 min) while the 17.5 kDa zwitterionic AuNCs could reach a AuZwEt2 presents a better circulation time, with a 2.6 extended
half-life around 10-15 min, which is 2 times longer than other AUC versus AuPegCOOH (1:3), leading to a favorable Tumor/
zwitterionic AuNCs recently reported. 12,15 One should note the Healthy tissue ratio.
pegylated AuNCs were cleared from the blood stream by a two- To conclude, the protection of the gold core by small
phase decay kinetics, with a first fast clearance (b1 min). In biocompatible ligands as zwitterionic or pegylated moieties helped
absence of specific targeting ligand at the surface of the AuNC, to improve the optical properties of the AuNCs, and to prolong their
the NCs are weakly internalized in monolayer cell cultures. In plasma half-life while conserving their renal elimination. One of the
vivo, they are able to reach the tumor site after intravenous major interests of AuNCs is their multiple uses for both optical
administration, and furthermore are internalized inside the tumor imaging, enhanced radiation therapy, 13,36 similarly to larger Au
cells, into small vesicles as demonstrated by fluorescent confocal nanoparticles, or photothermal therapy. 44,45 The combination of
imaging (Figure 6). From the preliminary experiments, 2 AuNCs image-guided surgery with a curative modality could reinforce the
appear promising for image-guided surgery investigations: the treatment efficacy. Such combining modalities would be of great
zwitterionic ZwEt2 (1:5) and the PegCOOH (1:3). Coating interest in the next development of the AuNCs for the therapeutic
nanoparticles with pegylated molecules has shown to improve care of Head and Neck pathologies. A particular interest would be
the anti-fouling effect, leading to the significant increase in blood the development of SWIR emitted AuNCs with good quantum yield
circulation, 37,38 and sometimes the tumor accumulation by EPR for better tumor to healthy tissue ratios.
effect. 39,40 However, there are still some debates about the tumor
cell uptake and the strong liver accumulation of pegylated NPs.
Acknowledgments
In addition, zwitterionic AuNCs may have better Tumor-to-
Kidney ratio than pegylated AuNCs, 18 which questions the most
We thank Maria-Isabel Montañez for the fruitful discussion
suitable surface chemistry on ultra-small particle such as AuNCs.
on the organic molecule design.
10 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011

Appendix A. Supplementary data 17. Du B, Jiang X, Das A, Zhou Q, Yu M, Jin R, et al. Glomerular barrier
behaves as an atomically precise bandpass filter in a sub-nanometre
regime. Nat Nanotechnol 2017;12(11):1096-102.
Supplementary data to this article can be found online at
18. Liu J, Yu M, Ning X, Zhou C, Yang S, Zheng J. PEGylation and
https://doi.org/10.1016/j.nano.2019.04.014. zwitterionization: pros and cons in the renal clearance and tumor
targeting of near-IR-emitting gold nanoparticles. Angew Chem Int Ed
Engl 2013;52(48):12572-6.
References 19. Chen D, Li B, Cai S, Wang P, Peng S, Sheng Y, et al. Dual targeting
luminescent gold nanoclusters for tumor imaging and deep tissue
1. Iseli TA, Lin MJ, Tsui A, Guiney A, Wiesenfeld D, Iseli CE. Are wider therapy. Biomaterials 2016;100:1-16.
surgical margins needed for early oral tongue cancer? J Laryngol Otol 20. Hu H, Huang P, Weiss OJ, Yan X, Yue X, Zhang MG, et al. PET and
2012;126(3):289-94. NIR optical imaging using self-illuminating (64)Cu-doped chelator-free
2. van Driel PB, van der Vorst JR, Verbeek FP, Oliveira S, Snoeks TJ, gold nanoclusters. Biomaterials 2014;35(37):9868-76.
Keereweer S, et al. Intraoperative fluorescence delineation of head and 21. Liu J, Yu M, Zhou C, Yang S, Ning X, Zheng J. Passive tumor targeting of
neck cancer with a fluorescent anti-epidermal growth factor receptor renal-clearable luminescent gold nanoparticles: long tumor retention and fast
nanobody. Int J Cancer 2014;134(11):2663-73. normal tissue clearance. J Am Chem Soc 2013;135(13):4978-81.
3. Rosenthal EL, Warram JM, de Boer E, Chung TK, Korb ML, 22. Fernandez TD, Pearson JR, Leal MP, Torres MJ, Blanca M, Mayorga C, et al.
Brandwein-Gensler M, et al. Safety and tumor specificity of Intracellular accumulation and immunological properties of fluorescent gold
cetuximab-IRDye800 for surgical navigation in head and neck cancer. nanoclusters in human dendritic cells. Biomaterials 2015;43:1-12.
Clin Cancer Res 2015;21(16):3658-66. 23. Mei BC, Susumu K, Medintz IL, Mattoussi H. Polyethylene glycol-
4. van den Berg NS, Miwa M, KleinJan GH, Sato T, Maeda Y, van Akkooi based bidentate ligands to enhance quantum dot and gold nanoparticle
AC, et al. (Near-infrared) fluorescence-guided surgery under ambient stability in biological media. Nat Protoc 2009;4(3):412-23.
light conditions: a next step to embedment of the technology in clinical 24. Porret E, Sancey L, Martin-Serrano A, Montanez M, Seeman R, Yahia-
routine. Ann Surg Oncol 2016;23(8):2586-95. Ammar A, et al. Hydrophobicity of gold nanoclusters influences their
5. Atallah I, Milet C, Coll JL, Reyt E, Righini CA, Hurbin A. Role of near- interactions with biological barriers. Chem Mater 2017;29
infrared fluorescence imaging in head and neck cancer surgery: from animal (17):7497-506.
models to humans. Eur Arch Otorhinolaryngol 2015;272(10):2593-600. 25. Hamouda R, Bertorelle F, Rayane D, Antoine R, Broyer M, Dugourd P.
6. Atallah I, Milet C, Quatre R, Henry M, Reyt E, Coll JL, et al. Role of Glutathione capped gold AuN(SG)M clusters studied by isotope-
near-infrared fluorescence imaging in the resection of metastatic lymph resolved mass spectrometry. Int J Mass Spectrom 2013;335:1-6.
nodes in an optimized orthotopic animal model of HNSCC. Eur Ann 26. Waszkielewicz M, Olesiak-Banska J, Comby-Zerbino C, Bertorelle F,
Otorhinolaryngol Head Neck Dis 2015;132(6):337-42. Dagany X, Bansal AK, et al. pH-induced transformation of ligated Au25
7. Atallah I, Milet C, Henry M, Josserand V, Reyt E, Coll JL, et al. Near- to brighter Au23 nanoclusters. Nanoscale 2018;10(24):11335-41.
infrared fluorescence imaging-guided surgery improves recurrence-free 27. Truillet C, Lux F, Tillement O, Dugourd P, Antoine R. Coupling of
survival rate in novel orthotopic animal model of head and neck HPLC with electrospray ionization mass spectrometry for studying the
squamous cell carcinoma. Head Neck 2016;38(Suppl 1):E246-55. aging of ultrasmall multifunctional gadolinium-based silica nanoparti-
8. de Boer E, Warram JM, Tucker MD, Hartman YE, Moore LS, de Jong JS, et cles. Anal Chem 2013;85(21):10440-7.
al. In vivo fluorescence immunohistochemistry: localization of fluorescently 28. Girod M, Antoine R, Dugourd P, Love C, Mordehai A, Stafford G. Basic
labeled cetuximab in squamous cell carcinomas. Sci Rep 2015;5:10169. vapor exposure for tuning the charge state distribution of proteins in
9. Prabhu RS, Hanasoge S, Magliocca KR, Hall WA, Chen SA, Higgins negative electrospray ionization: elucidation of mechanisms by fluores-
KA, et al. Lymph node ratio influence on risk of head and neck cancer cence spectroscopy. J Am Soc Mass Spectrom 2012;23(7):1221-31.
locoregional recurrence after initial surgical resection: implications for 29. Le Guevel X, Tagit O, Rodriguez CE, Trouillet V, Pernia Leal M,
adjuvant therapy. Head Neck 2015;37(6):777-82. Hildebrandt N. Ligand effect on the size, valence state and red/near
10. Moore LS, Rosenthal EL, Chung TK, de Boer E, Patel N, Prince AC, et infrared photoluminescence of bidentate thiol gold nanoclusters. Na-
al. Characterizing the utility and limitations of repurposing an open-field noscale 2014;6(14):8091-9.
optical imaging device for fluorescence-guided surgery in head and neck 30. Pyo K, Thanthirige VD, Kwak K, Pandurangan P, Ramakrishna G, Lee
cancer patients. J Nucl Med 2017;58(2):246-51. D. Ultrabright luminescence from gold nanoclusters: rigidifying the Au
11. Yu M, Zheng J. Clearance pathways and tumor targeting of imaging (I)-thiolate shell. J Am Chem Soc 2015;137(25):8244-50.
nanoparticles. ACS Nano 2015;9(7):6655-74. 31. Wu Z, Jin R. On the ligand's role in the fluorescence of gold
12. Yu M, Zhou C, Liu L, Zhang S, Sun S, Hankins JD, et al. Interactions of nanoclusters. Nano Lett 2010;10(7):2568-73.
renal-clearable gold nanoparticles with tumor microenvironments: vascu- 32. Lee S, Tong X, Yang F. Effects of the poly(ethylene glycol) hydrogel
lature and acidity effects. Angew Chem Int Ed Engl 2017;56(15):4314-9. crosslinking mechanism on protein release. Biomater Sci 2016;4(3):405-11.
13. Zhang XD, Luo Z, Chen J, Song S, Yuan X, Shen X, et al. Ultrasmall 33. Zhang XD, Wu D, Shen X, Liu PX, Fan FY, Fan SJ. In vivo renal
glutathione-protected gold nanoclusters as next generation radiotherapy clearance, biodistribution, toxicity of gold nanoclusters. Biomaterials
sensitizers with high tumor uptake and high renal clearance. Sci Rep 2012;33(18):4628-38.
2015;5:8669. 34. Chen F, Goel S, Hernandez R, Graves SA, Shi S, Nickles RJ, et al.
14. Chen Y, Montana DM, Wei H, Cordero JM, Schneider M, Le Guevel X, Dynamic positron emission tomography imaging of renal clearable gold
et al. Shortwave infrared in vivo imaging with gold Nanoclusters. Nano nanoparticles. Small 2016;12(20):2775-82.
Lett 2017;17(10):6330-4. 35. Ning X, Peng C, Li E, Xu J, Vinluan R, Yu M, et al. Physiological
15. Shen D, Henry M, Trouillet V, Comby-Zerbino C, Bertorelle F, Sancey stability and renal clearance of ultrasmall zwitterionic gold nanoparti-
L, et al. Zwitterion functionalized gold nanoclusters for multimodal near cles: ligand length matters. APL Materials 2017;5(5)053406.
infrared fluorescence and photoacoustic imaging. APL Mater 36. Zhang XD, Chen J, Luo Z, Wu D, Shen X, Song SS, et al. Enhanced
2017;5053404. tumor accumulation of sub-2 nm gold nanoclusters for cancer radiation
16. Le Guevel X, Henry M, Motto-Ros V, Longo E, Montanez M, Pelascini therapy. Adv Healthc Mater 2014;3(1):133-41.
F, et al. Elemental and optical imaging evaluation of zwitterionic gold 37. Cheng J, Teply BA, Sherifi I, Sung J, Luther G, Gu FX, et al.
nanoclusters in glioblastoma mouse models. Nanoscale 2018;10 Formulation of functionalized PLGA-PEG nanoparticles for in vivo
(39):18657-64. targeted drug delivery. Biomaterials 2007;28(5):869-76.
 C. Colombé et al / Nanomedicine: Nanotechnology, Biology, and Medicine 20 (2019) 102011 11

38. Cheng Y. A CS, Meyers JD, Panagopoulos I, Fei B, Burda C. Highly efficient 42. Zhang L, Wang E. Metal nanoclusters: new fluorescent probes for
drug delivery with gold nanoparticle vectors for in vivo photodynamic sensors and bioimaging. Nano Today 2014;9:132-57.
therapy of cancer. J Am Chem Soc 2008;130(32):10643-7. 43. Liao CT, Chang JT, Wang HM, Ng SH, Hsueh C, Lee LY, et al. Analysis
39. Verhoef JJ, Anchordoquy TJ. Questioning the use of PEGylation for of risk factors of predictive local tumor control in oral cavity cancer. Ann
drug delivery. Drug Deliv Transl Res 2013;3(6):499-503. Surg Oncol 2008;15(3):915-22.
40. Dufort S, Sancey L, Coll JL. Physico-chemical parameters that govern 44. Gamaleia NF, Shton IO. Gold mining for PDT: great
nanoparticles fate also dictate rules for their molecular evolution. Adv expectations from tiny nanoparticles. Photodiagn Photodyn 2015;12
Drug Deliv Rev 2012;64(2):179-89. (2):221-31.
41. Jin R, Zeng C, Zhou M, Chen Y. Atomically precise colloidal metal 45. Gao X, Yue Q, Liu Z, Ke M, Zhou X, Li S, et al. Guiding brain-tumor
Nanoclusters and nanoparticles: fundamentals and opportunities. Chem surgery via blood-brain-barrier-permeable gold nanoprobes with acid-
Rev 2016;116(18):10346-413. triggered MRI/SERRS signals. Adv Mater 2017;29(21).