A tumor-targeted Ganetespib-zinc phthalocyanine conjugate for synergistic chemo-photodynamic therapy

Article abstract of DOI:10.1016/j.ejmech.2018.03.077

Therapeutic effects of photodynamic therapy (PDT) are limited by the selectivity of photosensitizer (PS). Herein, a novel tumor-targeted drug-PS conjugate (Gan-ZnPc) which integrated with zinc phthalocyanine (ZnPc) and Ganetespib has been developed. ZnPc is a promising PS with remarkable photosensitization ability. Ganetespib is a heat shock protein 90 (Hsp90) inhibitor with preferential tumor selectivity and conjugated to ZnPc as a tumor-targeted ligand. The multifunctional small molecule conjugate, Gan-ZnPc, could be bound to extracellular Hsp90 and then selectively internalized into the tumor cells, followed by the generation of abundant intracellular reactive oxygen species (ROS) upon irradiation. Besides, Gan-ZnPc can arrest cell proliferation and induce apoptosis by the inhibition of Hsp90. Herein, with combination of the inhibition of Hsp90 and the generation of cytotoxic ROS, Gan-ZnPc implements tumor selectivity, concentrated PDT and chemotherapy in a synergistic manner, which results in highly effective anti-tumor activity in vitro and in vivo.

Full text of DOI:10.1016/j.ejmech.2018.03.077

Accepted Manuscript
A tumor-targeted Ganetespib-zinc phthalocyanine conjugate for synergistic chemo-
photodynamic therapy
Liangfeng Huang, Gaofei Wei, Xiaoqi Sun, Yali Jiang, Zeqian Huang, Yanjuan Huang,
Yifeng Shen, Xiaoyu Xu, Yunhui Liao, Chunshun Zhao
PII:
S0223-5234(18)30320-9
10.1016/j.ejmech.2018.03.077
EJMECH 10342
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 2 February 2018
Revised Date: 8 March 2018
Accepted Date: 26 March 2018
Please cite this article as: L. Huang, G. Wei, X. Sun, Y. Jiang, Z. Huang, Y. Huang, Y. Shen, X.
Xu, Y. Liao, C. Zhao, A tumor-targeted Ganetespib-zinc phthalocyanine conjugate for synergistic
chemo-photodynamic therapy, European Journal of Medicinal Chemistry (2018), doi: 10.1016/
j.ejmech.2018.03.077.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
A Tumor-targeted Ganetespib-Zinc Phthalocyanine
Conjugate for Synergistic Chemo-Photodynamic Therapy
Liangfeng Huang,‡ Gaofei Wei,‡* Xiaoqi Sun, Yali Jiang, Zeqian Huang, Yanjuan
Huang, Yifeng Shen, Xiaoyu Xu, Yunhui Liao and Chunshun Zhao*
School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, 510006,
People’s Republic of China
*Corresponding Author:E-mail: zhaocs@mail.sysu.edu.cn, Tel: +86 20 39943118
Graphical abstract

ACCEPTED MANUSCRIPT
A Tumor-targeted Ganetespib-Zinc Phthalocyanine
Conjugate for Synergistic Chemo-Photodynamic Therapy
Liangfeng Huang,‡ Gaofei Wei,‡* Xiaoqi Sun, Yali Jiang, Zeqian Huang, Yanjuan
Huang, Yifeng Shen, Xiaoyu Xu, Yunhui Liao and Chunshun Zhao*
School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, 510006,
People’s Republic of China
*Corresponding Author:E-mail: zhaocs@mail.sysu.edu.cn, Tel: +86 20 39943118
‡These authors contributed equally to this work.

ACCEPTED MANUSCRIPT
Abstract
Therapeutic effects of photodynamic therapy (PDT) are limited by the selectivity of
photosensitizer (PS). Herein, a novel tumor-targeted drug-PS conjugate (Gan-ZnPc)
which integrated with zinc phthalocyanine (ZnPc) and Ganetespib has been
developed. ZnPc is a promising PS with remarkable photosensitization ability.
Ganetespib is a heat shock protein 90 (Hsp90) inhibitor with preferential tumor
selectivity and conjugated to ZnPc as a tumor-targeted ligand. The multifunctional
small molecule conjugate, Gan-ZnPc, could be bound to extracellular Hsp90 and then
selectively internalized into the tumor cells, followed by the generation of abundant
intracellular reactive oxygen species (ROS) upon irradiation. Besides, Gan-ZnPc can
arrest cell proliferation and induce apoptosis by the inhibition of Hsp90. Herein, with
combination of the inhibition of Hsp90 and the generation of cytotoxic ROS,
Gan-ZnPc implements tumor selectivity, concentrated PDT and chemotherapy in a
synergistic manner, which results in highly effective anti-tumor activity in vitro and in
vivo.
Key words: Zinc phthalocyanine; Heat shock protein 90; Tumor-targeted;
Photodynamic therapy; Chemotherapy

ACCEPTED MANUSCRIPT
1. Introduction
Photodynamic therapy, which has been approved for clinical use, is a promising
noninvasive treatment for cancer. It combines three intrinsically nontoxic components
(photosensitizer, light and tissue oxygen) to generate cytotoxic reactive oxygen
species which leads to oxidative damage of the cells construction and triggered an
apoptotic or necrotic response[1-3]. PDT reactions can be controlled externally
because these biological responses only occur in where the PS located and is
restricted by the light irradiation to the diseased tissue[4, 5]. Although PDT has been
extensively studied during recent years and preclinical practices have shown
promising results, it is still less commonly used for cancer therapy in clinic. The main
reason is that PS without tumor selectivity is difficult to accumulate in malignant
tissue[6], which causes undesirable adverse effects, such as skin photosensitivity and
photoallergic reactions during PDT[7, 8]. Thus, it is critical to selectively localize PS
in tumor to enhance PDT efficiency and minimize collateral damage to normal tissue.
Therefore, development of PS with improved targeting specificity to avoid off-target
sensitization has been proven challenging.
Previously, for achieving enhanced PS selectivity, nanomedicine was commonly
applied strategy to increase the specific accumulation at the target site. The major
underlying mechanism of nanomedicines is the enhanced permeability and retention
(EPR) effect, whereby defects or gaps in the tumor vasculature allow the
extravasation and accumulation of larger particles that would be retained within the

ACCEPTED MANUSCRIPT
blood vessels of healthy tissues[9]. Unfortunately, further penetration of these same
large particles is often hampered by the dense extracellular matrix, leading to
problems with the delivery of nanoparticles to cells deep within a tumor mass[10, 11].
However, the recent clinical development of multiple low-molecular weight
ligand–drug conjugates suggests that these small molecule conjugates could be a
promising strategy to delivery drugs to target site. Besides, these ligand-drug small
molecule conjugates are often advantaged in their antigenicity, in vivo and in vitro
stability, ability to penetrate tumors and ease and cost of manufacturing[12, 13].
Moreover, a ligand-drug conjugate only with tumor selectivity may be not efficient
enough for anti-tumor therapy. Recently, there has been considerable interest in
combining different anticancer therapeutic methods. The combination acting on
different disease pathways has shown several advantages and considerably enhanced
therapeutic efficacy[14, 15].
In the present work, a novel tumor-targeted drug-PS conjugate was developed
(scheme 1). This conjugate could selectively bind to Hsp90, a target receptor
overexpressed in tumor cells, leading to enhanced accumulation of PS in tumor cells
for enhanced PDT and chemotherapy. As a proof-of-concept study, zinc
phthalocyanine (ZnPc) was utilized as a characteristic PS, which has a strong
absorption in the tissue-penetrating red-light wavelength range and high efficiency in
generating ROS, but in lack of tumor selectivity[16, 17]. As a tumor-targeted drug,
Ganetespib (termed STA-9090), an attractive Hsp90 inhibitor, was chosen because of

ACCEPTED MANUSCRIPT
its high affinity for activated Hsp90 in tumor cells and low affinity for latent Hsp90 in
normal cells[18-20]. Hsp90 is known to be a molecular chaperone with intracellular
and extracellular overexpression in malignant tumor cells and highly restricted
expression within normal tissue[21, 22]. Therefore, cell-surface expression of Hsp90
and the high affinity with its inhibitors in tumor cells allow Hsp90-targeted drugs to
concentrate in tumor tissue without harming healthy tissue[13]. Also, Hsp90
participates in many key processes in oncogenesis such as self-sufficiency in growth
signals, stabilization of mutant proteins, angiogenesis, and metastasis[23, 24]. As a
Hsp90 inhibitor, Ganetespib can potently arrest cell proliferation and induce apoptosis
in a wide variety of human tumor cell lines[25]. With the conjugation of Ganetespib
and ZnPc, Gan-ZnPc could concentrate in tumor cells due to binding to activated
extracellular Hsp90 secreted by tumor cells and promote the cellular internalization of
Gan-ZnPc by an active endocytotic process[26-28]. The concentration of PS inside
the cells was of great benefit to the generation of intracellular ROS with the
irradiation, thereby an enhanced PDT effect could be expected[29]. Furthermore, A.
Ferrario et al. had previously reported that PDT alone was found to induce expression
of various pro-survival and angiogenic signaling molecules such as HIF-1α, survivin
and VEGF, etc. in tumor tissue and weaken its efficacy. However, inhibition of
activated Hsp90 in tumor cells resulting in the blockade of multiple signal
transduction cascades was a beneficial adjuvant for PDT in a synergistic way[30, 31].

ACCEPTED MANUSCRIPT
The combination of two or more therapeutic modalities was helpful in overcoming
limitations encountered by each therapy when used alone[32].
According to the results in this study, Gan-ZnPc was demonstrated preferential
internalization into MCF7 tumor cells mediated by Hsp90 via receptor blocking
experiments. The dark- and photo-toxicity against MCF7 cells and mice bearing 4T1
xenograft revealed that Gan-ZnPc exhibited potent in vitro and in vivo anti-tumor
effect due to the combination of tumor-targeted chemotherapy and concentrated PDT
in tumor tissue. To the best of our knowledge, this study is the first example of a PS
conjugated with a Hsp90 inhibitor to enhance tumor selectivity and implement PDT
and chemotherapy for synergistic anti-tumor therapy.
Scheme 1. Chemical structure and schematic illustration of Gan-ZnPc for
synergistic tumor-targeted chemotherapy and PDT. Gan-ZnPc was firstly bound to
extracellular Hsp90 and then selectively internalized into the tumor cells. The
accumulation of Gan-ZnPc could inhibit the action of Hsp90 and generate
intracellular ROS upon irradiation. With the combination of these therapeutic

ACCEPTED MANUSCRIPT
modalities, Gan-ZnPc was a multifunctional molecular containing a promising
synergistic chemo-photodynamic therapy.
2. Results and discussions
2.1 Synthesis of Gan-ZnPc
Gan-ZnPc was synthesized through conjugating sulfonic ZnPc and Ganetespib with
an aliphatic chain. Ganetespib was synthesized according to previous reports[33] and
ZnPc was sulfonated for its later modification. A propane aliphatic linkage was
attached to Ganetespib with exposure of an amino group. After sulfamide reaction,
Gan-ZnPc was successfully synthesized and used for further experiments. W. Ying et
al. have confirmed that Ganetespib bound to Hsp90 through important hydrogen
bonding interactions involving the resorcinol hydroxyl group, carbonyl group of
triazolone and the N2 of triazolone[25]. In our study, ZnPc was conjugated to
aminoindole of Ganetespib to maintain the affinity to Hsp90. But, the phthalocyanine
1
derivatives in our study were not able to be characterized by H NMR and HPLC
presumably due to their aggregation[34]. Therefore, we used MALDI-TOF and FTIR
spectrum to characterize it. As showed in Figure S1, the FTIR spectrum of Gan-ZnPc
demonstrated that the stretching vibrations of -C=C-, -C=N-, -C-N- and -CH in
benzene are around 1629, 1329, 1087 and 2958 cm-1 respectively, which were the
representative peaks of ZnPc segment. Moreover, the stretching vibrations of -OH,
-NH- and -C=O are around 3389, 3219 and 1704 cm-1 respectively, which were the
representative peaks of Ganetespib segment. Comparing the FTIR spectrum of

ACCEPTED MANUSCRIPT
compound S6 and Gan-ZnPc, the appearance of stretching vibrations of -OH around
3389 cm-1 suggested the hydrolysis of the benzyl. Herein, the FTIR spectrum of
Gan-ZnPc suggested that Ganetespib was successfully conjugated to ZnPc.
2.2 Spectroscopy property of Gan-ZnPc
Most PS have a great absorbance at Q-band which enabled them to be excited by
long-wavelength light with high penetrability. Herein, the UV/vis absorbance of
ZnPc, Ganetespib and Gan-ZnPc were detected and compared (Figure 1A). ZnPc
demonstrated the expected characteristic absorption peaks at 672 nm in DMSO. On
the contrary, Ganetespib has none absorbance at Q-band. When ZnPc was covalently
conjugated to Ganetespib, the slight red shift of the absorbance to 678 nm at Q-band
was noticed. This change may be caused by the modification of ZnPc with electron
withdrawing group. The differences between ZnPc and Gan-ZnPc in fluorescence
spectroscopy were similar with the UV/vis spectroscopy. As shown in Figure 1B, the
maximum excitation wavelength of ZnPc was at 672 nm while Gan-ZnPc was at 678
nm, which was corresponded to the maximum absorption wavelength. As for
emission fluorescence (Figure 1C), the maximum emission wavelength of ZnPc and
Gan-ZnPc was at 678 nm and 688 nm respectively. These results indicated that
Gan-ZnPc had a similar spectroscopy property to ZnPc and suggested that Gan-ZnPc
may have a parallel photosensitivity of ZnPc.

ACCEPTED MANUSCRIPT
2.3 ROS generation efficiency
ROS generation was measured by using DPBF as a ROS indicator. DPBF could
capture and react with ROS and reduce its absorbance at 416 nm. Through recording
the degradation rate of DPBF, the generation efficiency of ROS was measured (Figure
1D). Obviously, there was no significant degradation of DPBF without the PS, which
suggested that ROS can’t be generated without PS upon irradiation. However, DPBF
was degraded rapidly when ZnPc or Gan-ZnPc was existed, which demonstrated that
ZnPc and Gan-ZnPc generated plenty of ROS after laser irradiation at 671 nm. More
than 90% of DBPF was degraded after 80 s accumulatively irradiation and the
degradation rate was rapidly decreased probably due to the reducing concentration of
DBPF. Obviously, there is no significant influence on ROS generation after the
modification of ZnPc, revealing the promising ROS generation efficiency of
Gan-ZnPc.
2.4 Photosensitization ability in aqueous solutions
Although Gan-ZnPc can dissolve in aqueous solutions containing 1% DMSO in a
certain concentration, aggregation of PS was still a problem resulting in losing their
photosensitization ability such as fluorescence property and generation of ROS.
However, it has been reported that the aggregation of PS can be disassociated by
proteins under physiological environment[35]. To quantitatively measure the
monomer percentage in aqueous solutions, the fluorescence emission comparison
method was used since the monomers emit fluorescence while the emission of dimers
or multimers can be neglected. As ZnPc can dissolve and disperse absolutely in

ACCEPTED MANUSCRIPT
DMSO, the monomer percentage was defined to 100% and the fluorescence emission
intensity of ZnPc dissolved in aqueous solution was normalized. As shown in Figure
1E, Gan-ZnPc dissolved in PBS containing 1% DMSO emitted inappreciable
fluorescence compared to dissolve in pure DMSO. This phenomenon indicated that
most Gan-ZnPc dissolved in aqueous solutions was in a state of self-aggregation and
lost photosensitization ability. When fetal bovine serum (FBS) was added to the
solution, the emission intensity was increased. The monomer percentage was
increased to 19% and 36% when 5% and 10% FBS contained in PBS respectively.
The recuperative fluorescence with addition of FBS indicated that Gan-ZnPc may
interact with FBS and disperse in monomer state which resulted in recovering their
photosensitization ability. By the way, an inconspicuous blueshift of fluorescence
peak of Gan-ZnPc could be observed due to the interaction between the Gan-ZnPc
and proteins, which is similar to previously reports[35]. The further confirmation of
ROS generation of Gan-ZnPc in aqueous solution was measured. As shown in Figure
1F, Gan-ZnPc dissolved in aqueous solution containing higher concentration of FBS
possessed higher ROS generation efficiency with the laser irradiation, which was
consistent with the fluorescence emission intensity. Futhermore, comparison of ROS
generation efficiency between Gan-ZnPc and ZnPc in aqueous solution showed
insignificant difference (Figure S4), indicating they have an equivalent ROS
generation efficiency in aqueous solution. The considerable percentage of monomers
of Gan-ZnPc and considerable ROS generation efficiency in serum protein contained

ACCEPTED MANUSCRIPT
systems implied Gan-ZnPc could be used for PDT under protein-riched physiological
environment.
2.5 Photostability
Furthermore，a PS used in PDT needs to be stable enough after irradiation, so that it
can generate enough cytotoxic ROS during the treatment. For confirming the
photostability of Gan-ZnPc and ZnPc, their characteristic maximal absorption in
Q-band was used to quantify their concentration and their photobleaching property
was demonstrated. As shown in Figure S5, after 20 min irradiation at a laser
irradiation intensity of 100 mW cm-2 at 671 nm, only about 10% of Gan-ZnPc and
ZnPc was photobleached, which indicating that Gan-ZnPc was stable enough to be
used in PDT.
Figure 1. Spectroscopy and photosensitization ability of Gan-ZnPc. (A) UV/Vis
absorption spectra of Ganetespib, ZnPc and Gan-ZnPc in DMSO (5 µM). Excitation

ACCEPTED MANUSCRIPT
(B) and Emission (C) fluorescence spectroscopy of ZnPc and Gan-ZnPc in DMSO (3
µM). (D) ROS generation of ZnPc and Gan-ZnPc in DMSO (1 µM) under laser
irradiation indicated by the degradation rate of DBPF (normalized absorbance at 416
nm). Emission fluorescence spectroscopy (E) and ROS generation (F) of Gan-ZnPc in
pure DMSO and aqueous solution containing 1% DMSO and different concentration
of FBS.
2.6 Cellular internalization and cellular ROS generation
Cellular internalization profiles were evaluated on MCF7 and NIH/3T3 cell using
confocal laser scanning microscopy (CLSM). MCF7 is human breast cancer cell with
high Hsp90 expression while NIH/3T3 is murine embryonic fibroblast cells used as
normal cell model. As expected, MCF7 cells incubated with Gan-ZnPc for 2 h and 4 h
showed stronger fluorescence compared to ZnPc, indicating that more Gan-ZnPc was
internalized into MCF7 cells (Figure 2A & Figure S6A). Comparing the fluorescence
intensity of 2 h and 4 h incubation, both ZnPc and Gan-ZnPc displayed a
time-dependent cellular internalization. To confirm cellular internalization was
mediated by Hsp90, 100 folds molar of free Ganetespib, which can block the binding
between Hsp90 and Gan-ZnPc, was mixed in culture medium with Gan-ZnPc. It was
obviously that the fluorescence was decreased after incubation for 2 h or 4 h
compared with Gan-ZnPc, suggesting that cellular uptake of Gan-ZnPc was decreased
significantly when Hsp90 was occupied by free Ganetespib. To further confirm the
increased cellular internalization of Gan-ZnPc was achieved by binding to Hsp90

ACCEPTED MANUSCRIPT
expressed on the cell surface, NIH/3T3 was used as a negative control. As shown in
Figure 2B & Figure S6B, it was obviously that weak fluorescence was observed in
both ZnPc and Gan-ZnPc after 2 h or 4 h incubation, and there is no significant
difference between them, indicating a small amount of ZnPc or Gan-ZnPc was
internalized. The difference in cellular internalization of ZnPc and Gan-ZnPc between
the MCF7 and NIH/3T3 cells demonstrated the preferential tumor cellular
internalization of Gan-ZnPc, which may be attributed to the expression of Hsp90 on
the surface of tumor cells. To quantify the cellular internalization, flow cytometry
(FCM) analysis was carried out on MCF7 cells and NIH/3T3 cells and showed in
Figure 2 C&D. Gan-ZnPc showed significantly more cellular internalization
compared to ZnPc or the mixture of Ganetespib and Gan-ZnPc, which was consistent
with the results of CLSM. The high cellular internalization of Gan-ZnPc mediated by
Hsp90 indicated that Gan-ZnPc could concentrate in tumor cells with high Hsp90
expression, which was potential to improve PDT effect.

ACCEPTED MANUSCRIPT
Figure 2. Cellular internalization evaluation by CLSM and FCM. CLSM images of
MCF7 cells (A) and NIH/3T3 cells (B) after incubation with ZnPc, Gan-ZnPc or
Gan-ZnPc/Ganetespib mixture for 4 h and stained with DAPI. Flow cytometry
analysis of MCF7 cells (C) and NIH/3T3 cells (D) after different incubation for 4 h.
(scale bar = 50 µm)
Additionally, the intracellular ROS generation ability of Gan-ZnPc in living cells
was detected using DCFH-DA as an indicator. The higher DCF fluorescence was
corresponded to the higher intracellular ROS generation. As shown in Figure 3A,
compared to ZnPc or the mixture of Ganetespib and Gan-ZnPc, Gan-ZnPc exhibited a
greater intracellular ROS generation in MCF7 cells, while showed negligible
difference in NIH/3T3 cells (Figure 3B). The results of intracellular ROS generation
determined by DCF fluorescence was consistent with the cellular internalization

ACCEPTED MANUSCRIPT
studies, indicating that the more cellular internalization, the more intracellular ROS
generation, which may result in the higher cytotoxicity.
Figure 3. Intracellular DCF fluorescence imaged of MCF7 cells (A) and NIH/3T3
cells (B) after incubation with ZnPc, Gan-ZnPc or Gan-ZnPc/Ganetespib mixture for
4 h and irradiation. (scale bar = 50 µm)
2.7 In vitro Cytotoxicity Study
The in vitro cytotoxicity of Gan-ZnPc against MCF7, 4T1, NIH/3T3 was evaluated by
MTT assay. Firstly, cytotoxicity of Gan-ZnPc against MCF7 was evaluated for totally
24 h, 48 h and 72 h incubation. For totally 24 h incubation, cells were incubated with
drug solution for 6 h first and exposed to irradiation for 10 min at 6 h, 9 h and 12 h,
and then incubated till reached 24 h. Similarly, for 48 h group, cells were incubated
with drug solution for 12 h first and exposed to irradiation at 12 h, 18 h and 24 h. For
72 h group, cells were incubated with drug solution for 24 h first and exposed to
irradiation at 24 h, 30 h and 36 h. As illustrated in Figure 4, cell viability decreased
along with the increasing concentration, which demonstrated the cytotoxicity of
Ganetespib, ZnPc and Gan-ZnPc were dose-dependent. Upon irradiation,
Ganetespib/ZnPc equimolar mixture was relatively more potent than Ganetespib or
ZnPc alone, which may benefit from the combination of PDT and inhibition of

ACCEPTED MANUSCRIPT
Hsp90. However, cell viability was inhibited to less than 20% by 100 nM of
Gan-ZnPc with irradiation in 24 h (Figure 4A) indicating that Gan-ZnPc possessed the
higher cytotoxicity against MCF7 cells than Ganetespib/ZnPc equimolar mixture.
This may due to the more internalization of Gan-ZnPc into MCF7 cells than ZnPc
which resulted in enhancing PDT efficacy. In addition, the in vitro cytotoxicity
evaluation was carried out without irradiation (Figure 4D), the free ZnPc didn’t show
significant cytotoxicity during the assay. As for Ganetespib, there is no significant
difference in the absence and presence of irradiation, which proving that irradiation
played a negligible role in cell viability without the PS. However, in the absence of
irradiation, Gan-ZnPc also exhibited certain dark toxicity which may be attributed to
the inhibition of Hsp90 in MCF7 cells. But it was noticeable that the dark toxicity of
Gan-ZnPc is slightly lower than Ganetespib, which was likely due to the bulky PS of
Gan-ZnPc and the reduced affinity to Hsp90. Furthermore, the cytotoxicity of
Gan-ZnPc was evaluated in 48 h (Figure 4 B&E) and 72 h (Figure 4 C&F). To further
characterize that Gan-ZnPc was an “all in one” molecule, in vitro cellular cytotoxicity
was also evaluated on 4T1 cells. As showed in Figure S7, Gan-ZnPc was more
potential than Ganetespib/ZnPc equimolar mixture and these results was similar to the
evaluation on MCF7 cells. Therefore, the in vitro cytotoxicity on MCF7 and 4T1 cells
suggested that Gan-ZnPc was a potential molecule used for synergistic
chemo-photodynamic therapy.

ACCEPTED MANUSCRIPT
Figure 4. Comparison of the in vitro cytotoxicity of ZnPc, Ganetespib,
Ganetespib/ZnPc (equimolar mixture) and Gan-ZnPc on MCF7 cells for 24 h (A), 48
h (B) and 72 h (C) in the presence of irradiation and for 24 h (D), 48 h (E) and 72 h
(F) in the absence of irradiation. (red-light LED lamp, 100 mW cm-2, 660-670 nm).
Data are expressed as mean ± standard deviation (SD). (n=6)
To confirm the cell cytotoxicity of Gan-ZnPc differed from tumor cells and
normal cells, in vitro cytotoxicity assay was also carried out on the 4T1 and NIH/3T3
cells for 24 h (Figure 5). In the present of irradiation, the aggressive 4T1 cells were
more impressionable compared to the non-oncogenic NIH/3T3 cells due to the high
affinity to activated Hsp90 and high internalization through binding to the
extracellular Hsp90. However, Gan-ZnPc possessed certain phototoxicity to NIH/3T3
cells upon irradiation which may be attributed to the free diffusion into the cells. The
negligible toxicity to NIH/3T3 cells in the absent of irradiation suggested that Hsp90

ACCEPTED MANUSCRIPT
in normal cells was in a latent and uncomplex state with low affinity to the Gan-ZnPc.
The in vitro cytotoxicity results suggested that covalent conjugation of Ganetespib to
ZnPc interestingly promoted the cellular internalization of PS to enhance PDT effect
and made Gan-ZnPc a multifunctional molecule with high cytotoxicity against tumor
cells while sparing normal cells.
Figure 5. Cytotoxicity evaluation of Gan-ZnPc on 4T1 (A) and NIH/3T3 (B) cells for
24 h with and without irradiation. (red-light LED lamp, 100 mW cm-2, 660-670 nm).
Data are expressed as mean ± standard deviation (SD). (n=6)
2.8 In vivo and ex vivo biodistribution
Next, in vivo biodistribution of Gan-ZnPc and ZnPc was evaluated by fluorescence
imaging on unhairing area using 4T1 bearing mice. As shown in Figure 6A, after
single intravenous administration, ZnPc was distributed without significant tumor
selectivity and the fluorescence of ZnPc was decreased during 24 h. On the contrast,
Gan-ZnPc obviously accumulated in the tumor tissue after intravenous injection.
Besides, although the fluorescence of Gan-ZnPc of normal tissue near tumor was
decreased rapidly during 24 h, the fluorescence in the tumor region was decreased
slowly and even maintained a strong fluorescence in 24 h compared to ZnPc. The
accumulation of Gan-ZnPc in tumor tissue was similar to Ganetespib, which

ACCEPTED MANUSCRIPT
possessed a half-life of 58.3 h in tumor[18], indicating that Gan-ZnPc was
preferentially distributed to tumor and maintained a high concentration in tumor.
To investigate the distribution of Gan-ZnPc and ZnPc in major organs, mice were
sacrificed at 24 h post-injection and the major organs were excised for imaging
(Figure 6B). Ex vivo fluorescence images showed a higher fluorescence intensity at
the tumor site than other organs in mice which were treated with Gan-ZnPc. Although
Gan-ZnPc was distributed in liver and lung, the controllable irradiation to tumor
tissue may restrict the adverse effect to the major organs. Additionally, mean
fluorescence intensity (MFI) of excised major organs and tumor was shown in Figure
6C. The MFI of Gan-ZnPc in excised tumor was dramatically 2.3 times higher than
ZnPc at 24 h post injection. These results confirmed that Gan-ZnPc could successfully
accumulate and remain in tumor, which may effectively enhance the PDT efficacy
with irradiation in vivo.

ACCEPTED MANUSCRIPT
Figure 6. In vivo and ex vivo biodistribution studies on balb/c mice bearing 4T1
xenografts. (A) In vivo fluorescence images on unhairing area of 4T1 bearing mice at
different time after single intravenous injection of ZnPc or Gan-ZnPc. (B) Ex vivo
fluorescence images of excised major organs and tumor at 24 h post-injection (C)
MFI of excised major organs and tumor at 24 h post-injection. Data are expressed as
mean ± standard deviation (SD). (n=3, ***p < 0.001)
2.9 In vivo antitumor efficacy
The in vivo antitumor efficacy of Gan-ZnPc was investigated on balb/c mice bearing
4T1 xenografts. When the tumor grew to about 100 mm3, they were separated
randomly into eight groups and intravenously injected with Saline, Ganetespib, ZnPc
and Gan-ZnPc respectively. The volume of tumor was measured every two days. As
illustrated in Figure 7A, the tumors were completely suppressed for the treatment of
Gan-ZnPc with laser irradiation. Such a superiority should be benefited from the
selective accumulation of Gan-ZnPc in tumor due to the Hsp90 mediated cellular

ACCEPTED MANUSCRIPT
internalization, following by the combination of efficient PDT and the inhibition of
Hsp90, the more PS accumulated in tumor, the more efficient PDT was. Furthermore,
even in the absence of laser irradiation, Gan-ZnPc exhibited a certain anticancer
activity, which was ascribed to the inhibition of Hsp90 action in the tumor cells. On
the contrary, ZnPc possessed negligible anticancer effect without laser irradiation
compared to with laser irradiation. Upon completion of the PDT treatment, all the
mice were sacrificed and all the tumors were excised and gathered for imaging
(Figure 7B) and weighing (Figure 7C). During the PDT treatment, body weight was
also monitored and no significant weight changes were observed compared to the
control groups, suggesting that Gan-ZnPc did not induce significant adverse effects on
the mice (Figure 7D).
Figure 7. In vivo anti-tumor evaluation on balb/c mice bearing 4T1 xenografts. (A)
The relative tumor volume (V/V0) during the treatment of Saline, ZnPc, Ganetespib
and Gan-ZnPc with or without laser irradiation (500 mW cm-2, 671 nm). Photograph
(B) and weight (C) of the tumors excised from the mice bearing 4T1 tumor after PDT

ACCEPTED MANUSCRIPT
treatment. (D) The relative body weight of mice during the PDT treatment. Data are
expressed as mean ± standard deviation (SD). (solid line for irradiation, dash line for
no irradiation, n=6, ***p < 0.001)
3. Conclusions
In summary, Gan-ZnPc, as an “all in one” molecule, could combine Hsp90 inhibition
with PDT. Gan-ZnPc could identify and bind to Hsp90 secreted on the tumor cells
surface, then entered the cells by Hsp90 mediated cellular internalization. Since
Gan-ZnPc concentrating in tumor cells, abundant cytotoxic ROS could generate
intracellularly and damage susceptible organelles while triggered by the light.
Besides, Gan-ZnPc could inhibit Hsp90 activity in the tumor cells, leading to cell
apoptosis. Consequently, under irradiation, Gan-ZnPc was a promising PS with
preferable cellular internalization, enhanced tumor accumulation and combination of
PDT and chemotherapy. All these superior characteristics of Gan-ZnPc resulted in
facilitating anti-tumor efficiency and showed that Gan-ZnPc was a highly promising
tumor-targeted drug-PS conjugate for synergistic tumor-targeted chemotherapy and
PDT to breast cancers. Our study inspired that conjugating Hsp90 inhibitor with a PS
for synergistically implementing tumor-targeted PDT and chemotherapy is an
effective means for cancer therapy. Furthermore, the detailed mechanisms of
synergistic action of PDT and Hsp90 inhibition of this conjugate are under
investigation.

ACCEPTED MANUSCRIPT
4. Materials and methods
4.1 Materials
Zinc phthalocyanine (ZnPc), chlorosulfonic acid, 2’,7’-Dichlorofluorescin diacetate
(DCFH-DA), 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT),
Fetal Bovine Serum (FBS), Dulbecco's modified Eagle's medium (DMEM),
phosphate buffered saline (PBS), 1-Boc-5-Aminoindole, 4-isopropylresorcinol,
Ganetespib (STA-9090) were purchased from Sigma-Aldrich (USA). Phenyl
chloroformate, 3-Bromo-1-propanol, 4',6-Diamidino-2-phenylindole Dihydrochloride
(DAPI), 1,3-Diphenylisobenzofuran (DBPF), hydrazine monohydrate, thionyl
chloride, palladium on activated charcoal (Pd/C), potassium ferricyanide was obtained
from Aladdin (China). Other chemicals were of analytical grade and used without
further purification.
4.2 Synthesis of Gan-ZnPc
Detailed synthesis procedure was seen in supplementary data.
4.3 Characterization of Gan-ZnPc
UV−vis Spectroscopy. ZnPc, Ganetespib, Gan-ZnPc were dissolved in DMSO at 5
µM. The UV-visible absorption spectra from 200 nm to 900 nm were acquired with
an UV-vis spectrophotometer (TECHCOMP, UV2600 spectrophotometer).
Fluorescence Spectroscopy. ZnPc and Gan-ZnPc were dissolved in DMSO at 3
µM and their fluorescence spectra were acquired with a spectrofluorophotometer
(Fluoromax-4, Horiba, Japan). The fluorescence spectroscopy of Gan-ZnPc in
aqueous solution was obtained similarly. Briefly, Gan-ZnPc was dissolved in PBS

ACCEPTED MANUSCRIPT
containing 1% (V/V) DMSO, then the exact volume of FBS (fetal bovine serum) was
added. The fluorescence spectroscopy in aqueous solution was expressed by
normalized emission intensity. The fluorescence excitation spectra of ZnPc and
Gan-ZnPc were acquired through monitoring emission wavelength at 780 nm. The
fluorescence emission spectra were acquired by exciting at 560 nm.
Detection of ROS. The generation of ROS was detected using DBPF as a ROS
indicator. DBPF can capture and react with ROS rapidly, leading the decreased
absorbance at 416 nm. Therefore, the generation of ROS was demonstrated by the
characteristic absorbance decrease of DPBF at 416 nm. The stock solution of DBPF
(60 µM) in DMSO was added blank DMSO as control, ZnPc or Gan-ZnPc at a final
concentration at 1 µM. Then, the mixture was exposed to laser irradiation at 671 nm
(20 mW cm-2) in an interval of 120 s and the absorbance at 416 nm was detected per
10 s. The generation of ROS of Gan-ZnPc in aqueous solution was acquired using the
same method. In brief, DBPF and Gan-ZnPc was dissolved in PBS containing 1%
(V/V) DMSO, then the exact volume of FBS was added. The mixture was exposed to
laser irradiation and the absorbance at 416 nm was detected per 20 s. The DBPF
degradation was calculated as (A0-At)/A0 × 100% (At: the absorbance of the mixture
solution at specified time after irradiation; A0: the absorbance of the mixture solution
before irradiation)
Photostability. The photostability of ZnPc and Gan-ZnPc after irradiation was
detected by the decrease of their characteristic absorption. ZnPc and Gan-ZnPc was

ACCEPTED MANUSCRIPT
dissolved in DMSO at 5 µM and exposed to laser irradiation at 671 nm (100 mW
cm-2) in an interval of 20 min. The characteristic absorption, 672 nm for ZnPc and
678 nm for Gan-ZnPc, was detected per 2 min. And the photostability was calculated
as At /A0. (At: the characteristic absorption of the solution at specified time after
irradiation; A0: the characteristic absorption of the solution before irradiation)
4.4 Cellular assay
Cell culture. MCF7 human breast cancer cells, 4T1 murine breast cancer cells,
NIH/3T3 murine embryonic fibroblast cells were obtained from Laboratory Animal
Center of Sun Yat-sen University (Guangzhou, China) and cultured in RPMI 1640 or
DMEM containing 10% FBS and antibiotics (penicillin 100 U mL-1 and streptomycin
100 µg mL-1) under a humidified atmosphere at 37 °C with 5% CO2. During the cell
assay experiments, drugs were dissolved in culture medium containing 0.5% DMSO
and 10% FBS.
Cellular Uptake (determined by confocal CLSM). MCF7 or NIH/3T3 cells were
seeded at glass base dish and incubated overnight. The culture medium was removed
and rinsed twice with PBS, then the cells were incubated with a solution of ZnPc or
Gan-ZnPc, Gan-ZnPc/Ganetespib (1/100, n/n) at a concentration of 1 µM for 2 h or 4
h under the same conditions. After that, the solution was removed and the cells were
washed twice with PBS and dyed with DAPI. Then, cells were washed 5 times with
PBS and imaged with CLSM (LSM 710, Zeiss, Germany).

ACCEPTED MANUSCRIPT
Cellular Uptake (determined by FCM analysis). MCF7 or NIH/3T3 cells were
incubated in 24-well plates overnight. The medium was removed and rinsed twice
with PBS, the cells were washed with PBS, then the cells were incubated with a
solution of ZnPc or Gan-ZnPc, Gan-ZnPc/Ganetespib (1/100, n/n) for 4 h under the
same conditions. After removing the solution and rinsing with PBS, cells were
harvested and resuspended in PBS for flow cytometry measurements. (Guava
EasyCyte 6-2L, Merck Millpore).
Intracellular ROS generation. MCF7 cells or NIH/3T3 cells were seeded at glass
base dish overnight. The medium was removed and rinsed twice with PBS, then the
solution of ZnPc or Gan-ZnPc, Gan-ZnPc/Ganetespib (1/100, n/n) was added. After 4
h incubation, the solution was removed and further incubated with DCFH-DA for 30
min and then exposed to red light irradiation (50 mW cm-2, 660-670 nm) for 5 min.
Then, the cells were washed with PBS and dyed with DAPI following imaged with
CLSM.
In vitro Cytotoxicity Study. The in vitro cytotoxicity of Ganetespib, ZnPc and
Gan-ZnPc against MCF7, 4T1 or NIH/3T3 cells for 24 h, 48 h and 72 h were
determined by MTT assay. Ganetespib, ZnPc, Ganetespib/Gan-ZnPc (equimolar
mixture) and Gan-ZnPc were first dissolved in DMSO and diluted with culture
medium to get different concentration solution with contained 0.5% DMSO. Cells
were incubated in 96-well plates (5×104 cells per well) with 100 µL DMEM contain
10% FBS overnight at 37 °C with 5% CO2. The medium was removed and rinsed

ACCEPTED MANUSCRIPT
twice with PBS, then 100 µL of drug solution or culture medium containing 0.5%
DMSO as control was added and incubated for determinate time. For the first group
that total incubation time was 24 h, cells were first incubated for 6 h with drug
solutions. After drug solution was removed, cells were rinsed with PBS and fresh
culture medium was added. Cells were exposed for red light irradiation (100 mW
cm-2,660-670 nm) for 10 min and then continuously incubated in incubator. The same
red-light irradiation (same intensity, same wavelength and same irradiation time) were
then exposed to cells at 9 h and 12 h, and then cells were incubated in incubator for
another 12 h, before MTT solution was added. The total incubation time was 48 h for
the second group, the cells were first incubated for 12 h with drug solutions rinsed
with PBS followed by adding fresh culture medium. The same red-light irradiation
was also exposed to cells at 12 h, 18 h and 24 h. Then the cells ware incubated in
incubator for another 24 h. For the third group with the total incubation time of 72 h,
in which the cells were first incubator for 24 h with drug solutions and then rinsed
with PBS followed by added fresh culture medium. The same red-light irradiation was
exposed to cells at 24 h, 30 h and 36 h. Then the cells ware incubated in incubator for
another 36 h. Cytotoxicity for 24 h, 48 h and 72 h without irradiation was carried out
similarly. In brief, cells were incubated with drug solution and rinsed with PBS at 6 h,
12 h, 24 h, respectively. Then fresh culture medium was added and cells were
continually incubated till the time achieved. When the incubation was completed,
MTT solution was added and incubated for another 4h, then the culture medium was

ACCEPTED MANUSCRIPT
replaced with DMSO. The cell viability was measured with a microplate reader
(ELX800, Bio-Tek, USA) by the percentage of optical density value at 490 nm.
4.5 In vivo assay
Animals and tumor model. Female BALB/c mice (20 ± 2 g, 5–6 weeks) were
obtained from the Laboratory Animal Center of Sun Yat-sen University (Guangzhou,
China). All experimental procedures were approved and supervised by the
Institutional Animal Care and Use Committee of Sun Yat-sen University.
To develop the breast tumor model, the 4T1 cell suspension (1×106 cells) was
subcutaneously injected into the right flank of each mouse. Tumor volumes were
calculated according to the formula: (L×W2)/2, (L is the longest and W is the shortest
tumor diameter (mm)).
In vivo biodistribution. 4T1 bearing BALB/c mice were randomly divided into
two groups (n=3) and received a single intravenous injection of ZnPc and Gan-ZnPc
(both at a dosage of 0.5 µmol kg-1). Mice were anesthetized and imaged on the
unhairing area directly at 1, 3, 6, 9, 12, 24 h via small animal imaging system
(NightOWL LB983, Berthold, Germany, excitation: 630 nm, emission: 680 nm).
After last imaged, they were sacrificed and hearts, livers, spleens, lungs, kidneys and
tumors were excised for imaging. Mean fluorescence intensity (MFI) of excised major
organs and tumor was calculated as fluorescence counts per square millimeter.
In vivo antitumor efficacy. When the tumor volume reached about 100 mm3, 4T1
bearing mice were divided randomly into 8 groups (n = 6) and treated with saline,