An HSP90 inhibitor based fluorescent probe for selective tumor targeting

Article abstract of DOI:10.1016/j.dyepig.2021.109783

Fluorescent imaging has been widely used in the detection of diseases, especially cancer. Among them, biomarker-specific ligand-based imaging strategies have attracted much attention, due to their improved image contrast and diagnostic accuracy. In this w

Full text of DOI:10.1016/j.dyepig.2021.109783

Dyes and Pigments 196 (2021) 109783
Contents lists available at ScienceDirect
Dyes and Pigments
journal homepage: www.elsevier.com/locate/dyepig
An HSP90 inhibitor based fluorescent probe for selective tumor targeting
Shulei Zhu, Yingxin Lu, Jiyu Jin*, Jiahui Yu, Wei Lu**
Shanghai Engineering Research Center of Molecular Therapeutics and New Drug Development, School of Chemistry and Molecular Engineering, East China Normal
University, 3663 North Zhongshan Road, Shanghai, 200062, PR China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Fluorescent imaging has been widely used in the detection of diseases, especially cancer. Among them,
biomarker-specific ligand-based imaging strategies have attracted much attention, due to their improved image
contrast and diagnostic accuracy. In this work, an HSP90 inhibitor based fluorescent probe (Probe 1) was
synthesized possessing the potential of selective tumor targeting, which was constructed with an HSP90 inhibitor
and 4-hydroxy-1,8-naphthalimide derivative. Probe 2, a resorcinol-blocked probe was also synthesized as a
control, which prevented binding with the HSP90 ATP binding pocket. Biological studies showed that Probe 1
could be selectively accumulated by tumor cells, while Probe 2 not, suggesting that Probe 1 is a promising
targeted tumor imaging agent for early diagnosis.
HSP90
Fluorescent probe
Selective tumor targeting
No major organ-related toxicities
1. Introduction
which serve as active agents to enhance the fluorescent signal. Among
them, HSP90 has attracted much attention.
Over the past decades, fluorescent imaging has attracted much
attention as a non-invasive modality for studying diseases at the mo-
lecular level in vivo, especially cancer [1–3]. They provide dynamic in-
formation concerning the localization and quantity of the molecules of
interest, without the need of genetic engineering of the sample [4,5].
Small molecule organic dyes have been widely used as fluorescent
probes in diagnose tests and molecular imaging for improving
contrast-enhancement [6,7]. Among these organic dyes, 1,8-naphthali-
mide derivatives are a well-known class of compounds that have a
wide range of applications, due to their high fluorescence quantum
yields and photostability [8–11]. Moreover, these derivatives are also
considered to be good candidates for fluorescent dichroic dyes, fluo-
rescent polymeric paints or plastics and fluorescent markers in medicine
and biology [8,12–14].
HSP90, is an ATP-controlled machine that facilitates the correct
folding and conformational maturation of denatured proteins [28]. In-
hibition of HSP90 ATP hydrolysis activity resulted in systemic degra-
dation of client proteins, among which most are oncogenic [29–31].
Overexpressed HSP90 has been implicated in a variety of tumors, which
is critical for the survival, adaptation, proliferation and metastasis as
well as angiogenesis of cancer cells, and thus led to cancer cell growth
inhibition and apoptosis [32–34]. Therefore, small molecules targeting
HSP90 folding machinery have shown therapeutic success in solid and
hematological malignancies, providing an opportunity for tumor tar-
geting [35].
In this work, an HSP90 inhibitor was synthesized as a tumor-
targeting ligand and incorporated with 4-hydroxy-1,8-naphthalimide
fluorescent molecule to synthesize a tumor-targeting fluorescent probe
Probe 1. This tumor-targeting fluorescent probe was further charac-
terized and its property in vitro was also evaluated. Finally, its fluores-
cent imaging property was investigated in vivo. Probe 2, an inactive
resorcinol-blocked analog of Probe 1, was also synthesized to confirm
the tumor selectivity of HSP90 scaffold.
Conjugation of small fluorophores and biomarker-specific ligands
has become one of the main strategies for fluorescent tumor imaging.
These biomarker-specific ligands include peptide, antibody, sugar,
folate or other small molecules, which could significantly improve the
image contrast and diagnostic accuracy [15,16]. Small molecules anti-
cancer drugs ligands have been the subject of biological and trans-
lational research for many years [17,18]. There have been many small
molecules anticancer drugs based fluorescent probes, including BTK
inhibitor [19,20], HSP90 inhibitor [21–23], FAP inhibitor [24–27], etc,
2. Materials and methods
All reagents were obtained from commercial sources and dried prior
* Corresponding author.
** Corresponding author.
E-mail addresses: jyjin@chem.ecnu.edu.cn (J. Jin), wlu@chem.ecnu.edu.cn (W. Lu).
https://doi.org/10.1016/j.dyepig.2021.109783
Received 12 June 2021; Received in revised form 3 August 2021; Accepted 1 September 2021
Available online 4 September 2021
0143-7208/© 2021 Elsevier Ltd. All rights reserved.

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
Scheme 1. Reagents and conditions: (a)
BnBr (2.2 equiv), K2CO3, DMF, 150 ◦C,
12 h, 95%; (b) Br2 (1.3 equiv), NaOH,
1,4-dioxane, 0 ◦C- rt, 12 h, 80%; (c)
oxalyl chloride (1.5 equiv), anhydrous
DCM, 0 ◦C, 2 h; (d) 5-nitroindole (1.2
equiv), Et3N, anhydrous DCM, 0 ◦C-rt, 3
h, 60%; (e) 10% Pd/C, MeOH, rt, 12 h,
85%; (f) succinic anhydride (1.2 equiv),
toluene, 100 ◦C, 12 h, 85%; (g) com-
pound 10 (1.1 equiv), HATU, DIEA,
anhydrous DMF, rt, 3 h, 55%; (h) CH3I
(5.0 equiv), K2CO3, anhydrous DMF,
0
◦C, 1 h, 60%; (i) LiOH (1.2 equiv),
THF/MeOH/H2O (v: v: v = 1: 1: 1); (j)
compound 10 (1.1 equiv), HATU, DIEA,
anhydrous DMF, rt, 3 h, 50%.
spectrometer. High-resolution mass spectra (HRMS) were obtained on
an electron spray injection (ESI) Thermo Fisher Scientific LTQ FTICR
mass spectrometer. Purity of all tested compounds was ≥95%, as esti-
mated by HPLC analysis performed on Agilent Diamonsil C18 (250 mm
× 4.6 mm). The resulting slides were mounted and observed with a
LEICA TCS SP8 fluorescence microscopy.
Table 1
Data for compound Probe 1 and Probe 2: Binding on
HSP90 (FP Assay).
Comp.
IC50a (nM ± SD)
Probe 1
Probe 2
AT13387b
10c
77.24 ± 14.28
>1000
7.32 ± 1.10
>1000
2.1. Synthesis routes of probe 1 and probe 2
a
Data presented is the mean ± SD value of three
2.1.1. 1-(2,4-bis(benzyloxy)-5-isopropylphenyl)ethan-1-one (2)
Benzyl bromide (29.0 mL, 271.0 mmol) was added to a mixture of 1
(25.0 g, 129.0 mmol) and potassium carbonate (45.0 g, 322.0 mol) in
DMF (300 mL). The mixture was heated to 150 ◦C for 12 h. After
completion of the reaction, the mixture was cooled to room temperature,
quenched with 4 M HCl and extracted with EtOAc. The combined
organic layers were then washed with saturated brine, dried over
anhydrous Na2SO4, and concentrated in vacuo to afford the crude
product. The crude product was then beated with PE: EA (10:1) to give a
pale yellow solid 42.1 g. Yield: 86%. 1H NMR (400 MHz, CDCl3) δ 7.76
(s, 1H),7.44–7.32 (m, 10H), 6.51 (s, 1H), 5.09 (d, J = 4.8 Hz, 4H), 3.28
(dt, J = 13.8, 6.9 Hz, 1H), 2.56 (s, 3H), 1.22 (d, J = 6.9 Hz, 6H). LCꢀ MS
(ESI) m/z (M + H)+: 375.5.
independent determinations.
b
Used as positive control.
c
Unconjugated probe used as a control.
to use unless noted otherwise. Reactions were monitored by thin-layer
chromatography (TLC) on 0.25 mm silicon gel plates and visualized
under UV light. 1H NMR and 13C NMR spectra were recorded at 23 ◦C in
CDCl3 and DMSO‑d6 on a Bruker DRX-400 (400 MHz) using TMS as the
internal standard. Chemical shifts were reported as δ (ppm) and signal
splitting patterns were described as singlet (s), doublet (d), triplet (t),
quartet (q), quintet (quint), or multiplet (m), with coupling constants (J)
in hertz. The mass spectra (MS) were recorded on a Waters mass
Fig. 1. In vitro cytotoxicity evaluation of Probe 1 and Probe 2 on A549 cells and HCT116 cells.
2

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
Fig. 2. In vitro cytotoxicity evaluation of Probe 1 on HL-7702 (A) and HepG2 (B) cells for 24 h.
Fig. 3. Cellular uptake and subcellular distribution of Probe 1 and Probe 2 in A549 and HCT116 cells. (Scale bar = 75 μm).
2.1.2. 2,4-bis(benzyloxy)-5-isopropylbenzoic acid (3)
chromatography with DCM:MeOH (20:1) to give a pale white solid 14.9
g. Yield: 80%. 1H NMR (400 MHz, DMSO‑d6) δ 12.23 (s, 1H), 7.60 (s,
1H), 7.52 (d, J = 7.4 Hz, 2H), 7.49–7.31 (m, 8H), 6.90 (s, 1H), 5.21 (s,
4H), 3.19 (s, 1H), 1.15 (d, J = 6.8 Hz, 6H). LC-MS (ESI) m/z (M + H)+:
377.5.
Bromine (3.86 g, 24.3 mmol) was added to a mixture of 2 (3.86 g,
24.3 mmol) in 1,4-dioxane (100 mL) and a solution of sodium hydroxide
(2.08 g, 52.0 mmol) in water (30 mL) at 0 ◦C. The mixture was stirred at
room temperature for 12 h. After completion of the reaction, the mixture
was quenched with saturated NaHSO3 and extracted with EtOAc. The
combined organic layers were then washed with saturated brine, dried
over anhydrous Na2SO4, and concentrated in vacuo to afford the crude
product. The crude product was then purified with column
2.1.3. 2,4-bis(benzyloxy)-5-isopropylbenzoyl chloride (4)
Oxalyl chloride (10 mL, 39.8 mmol) was added dropwise to a
mixture of 3 (10.0 g, 39.8 mmol) in anhydrous DCM (100 mL) at 0 ◦C.
3

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
Fig. 4. The cellular uptake of tumor cells treated with AT13387 mixed with Probe 1 (25 μM, 100:1, n/n) or Probe 1 mixed with vehicle (25 μM). (Scale bar =
50 μm).
Fig. 5. Time-dependent cellular uptake of Probe 1 by A549 cells. (Scale bar = 50 μm).
Fig. 6. Drug accumulation and biodistribution on HCT116 xenografts.
Catalytic amount of DMF (0.1 mL) was then added and stirred at room
temperature for 2 h. After completion of the reaction, solvent and excess
(COCl)2 were removed in vacuo to afford the crude product 4. The crude
product was used for the following reaction without further purification.
anhydrous DCM (100 mL), followed by Et3N (9.3 mL, 66.0 mmol). The
mixture was stirred at room temperature for 3 h. After completion of the
reaction, the mixture was quenched with water and extracted with DCM.
The combined organic layers were then washed with saturated brine,
dried over anhydrous Na2SO4, and concentrated in vacuo to afford the
crude product. The crude product was then purified with column
chromatography with DCM:MeOH (20:1) to give a yellow solid 8.3 g.
Yield: 60%. 1H NMR (400 MHz, CDCl3) δ 7.99 (s, 1H), 7.33 (d, J = 58.3
2.1.4. (2,4-bis(benzyloxy)-5-isopropylphenyl)(5-nitroindolin-1-yl)
methanone (5)
5-Nitroindoline (4.36 g, 26.5 mmol) was added to a mixture of 4 in
4

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
Fig. 7. CLSM assay of ex vivo mice tumor sections and pharmacokinetic assessment for different compounds. (Scale bar = 50 μm).
127.30, 119.12, 115.05, 113.96, 112.23, 101.86, 48.07, 27.13, 26.65,
22.38. LC-MS (ESI) m/z (M + H)+: 313.4.
2.1.6. 4-((1-(2,4-dihydroxy-5-isopropylbenzoyl)indolin-5-yl)amino)-4-
oxobutanoic acid (7)
Succinic anhydride (774 mg, 7.69 mmol) was added to a mixture of 6
(2.0 g, 6.41 mmol) in toluene (20 mL). The mixture was heated to reflux
for 12 h. After completion of the reaction, solvent was removed in vacuo
to afford the crude product. The crude product was then purified with
column chromatography with DCM:MeOH (10:1) to give a yellow solid
2.24 g. Yield: 85%. 1H NMR (400 MHz, DMSO‑d6) δ 9.89 (s, 1H), 9.72 (s,
1H), 9.19 (s, 1H), 7.65 (d, J = 0.6 Hz, 1H), 7.56 (dd, J = 7.5, 1.5 Hz, 1H),
7.34 (q, J = 1.1 Hz, 1H), 7.30 (d, J = 7.5 Hz, 1H), 6.53 (s, 1H), 4.30–4.05
(m, 2H), 3.16 (dtd, J = 18.9, 7.0, 0.9 Hz, 3H), 2.68–2.59 (m, 2H),
2.59–2.48 (m, 2H), 1.22 (d, J = 6.8 Hz, 6H). 13C NMR (100 MHz,
DMSO‑d6) δ 175.04, 170.94, 167.33, 161.76, 158.59, 138.62, 137.57,
130.15, 127.67, 126.80, 119.68, 117.32, 116.56, 113.66, 101.77, 47.57,
31.77, 28.84, 27.00, 22.43. LC-MS (ESI) m/z (M + H)+: 413.5.
Fig. 8. Pharmacokinetic assessment of Probe 1, Probe 2 and AT13387.
Hz, 12H), 6.53 (s, 1H), 5.08 (s, 4H), 4.26–3.98 (m, 2H), 3.33 (s, 1H),
3.05 (s, 2H), 1.22 (d, J = 6.8 Hz, 6H).
2.1.7. Methyl 4-((1-(5-isopropyl-2,4-dimethoxybenzoyl)indolin-5-yl)
amino)-4-oxobutanoate (8)
2.1.5. (5-Aminoindolin-1-yl)(2,4-dihydroxy-5-isopropylphenyl)methanone
(6)
Iodomethane (1.5 mL, 24.25 mmol) was added to a mixture of 7 (2.0
g, 4.85 mmol) and potassium carbonate (2.2 g, 19.4 mmol) in anhydrous
DCM (30 mL) at 0 ◦C. The mixture was then stirred at room temperature
for 2 h. After completion of the reaction, the mixture was quenched with
1 M HCl and the combined organic layers were then washed with
saturated brine, dried over anhydrous Na2SO4, and concentrated in
vacuo to afford the crude product. The crude product was then purified
with column chromatography with DCM:MeOH (30:1) to give a pale
yellow solid 1.7 g. Yield: 78%. 1H NMR (400 MHz, CDCl3) δ 9.72 (s, 1H),
7.73 (s, 1H), 7.61 (dd, J = 7.5, 1.5 Hz, 1H), 7.42–7.18 (m, 2H), 6.71 (s,
1H), 4.22 (t, J = 7.1 Hz, 2H), 3.95 (s, 3H), 3.82 (s, 3H), 3.62 (s, 3H), 3.17
10% Pd/C (1.5 g) was added to a mixture of 5 (5.0 g, 9.58 mmol) in
methanol (100 mL) under N2 at room temperature. The mixture was
purged with H2 and stirred at room temperature for 4 h. After comple-
tion of the reaction, the mixture was filtered through celite and washed
with DCM. The combined organic layers were concentrated in vacuo to
afford the crude product. The crude product was then purified with
column chromatography with DCM:MeOH (10:1) to give a yellow solid
2.5 g. Yield: 85%. 1H NMR (400 MHz, DMSO‑d6) δ 9.89 (s, 1H), 9.77 (s,
1H), 7.65 (d, J = 0.6 Hz, 1H), 7.27 (d, J = 7.5 Hz, 1H), 6.82–6.66 (m,
2H), 6.52 (s, 1H), 4.66 (dd, J = 35.0, 5.3 Hz, 2H), 4.21 (t, J = 7.1 Hz,
2H), 3.31–2.97 (m, 3H), 1.22 (d, J = 6.8 Hz, 6H). 13C NMR (100 MHz,
DMSO‑d6) δ 168.12, 161.88, 157.58, 145.00, 135.15, 131.62, 128.07,
(td, J = 7.1, 1.0 Hz, 3H), 2.79–2.55 (m, 4H), 1.25 (d, J = 6.8 Hz, 6H). 13
C
NMR (100 MHz, CDCl3) δ 170.86, 165.13, 159.37, 157.86, 138.53,
Fig. 9. Representative HE images of ex vivo tumor sections. (Scale bar = 100 μm).
5

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
Fig. 10. Representative TUNEL images of ex vivo tumor sections. (Scale bar = 100 μm).
136.42, 129.80, 127.55, 119.93, 117.83, 116.58, 56.00, 51.59, 47.58,
MHz, DMSO‑d6) δ 9.89 (s, 1H), 9.77 (s, 1H), 9.64 (s, 1H), 8.53 (d, J =
8.3 Hz, 1H), 8.47 (d, J = 7.2 Hz, 1H), 8.35 (d, J = 8.2 Hz, 1H), 7.92 (dt, J
= 23.2, 5.6 Hz, 2H), 7.75 (t, J = 7.8 Hz, 1H), 7.56 (s, 1H), 7.26 (d, J =
8.5 Hz, 1H), 7.15 (d, J = 8.2 Hz, 1H), 6.98 (s, 1H), 6.41 (s, 1H), 4.04 (t, J
= 7.1 Hz, 2H), 3.96 (t, J = 8.3 Hz, 2H), 3.39 (t, J = 6.1 Hz, 2H), 3.35 (t, J
= 5.9 Hz, 2H), 3.17 (dq, J = 17.4, 6.0 Hz, 5H), 3.09 (q, J = 6.8 Hz, 1H),
3.03 (t, J = 8.3 Hz, 2H), 2.51 (s, 4H), 2.41 (t, J = 7.1 Hz, 2H), 2.16 (t, J =
7.7 Hz, 2H), 1.86 (p, J = 7.4 Hz, 2H), 1.26 (d, J = 6.8 Hz, 2H), 1.13 (d, J
= 6.9 Hz, 6H). 13C NMR (100 MHz, DMSO‑d6) δ 172.88, 170.61, 168.16,
165.80, 162.76, 157.98, 154.42, 140.47, 138.77, 133.91, 128.40,
128.15, 127.36, 125.15, 124.49, 122.97, 119.77, 118.86, 116.52,
112.97, 99.61, 70.00, 69.22, 46.47, 44.93, 41.37, 35.89, 34.71, 31.92,
28.13, 26.97, 23.17. HR-MS (m/z) (ESI): calcd for C44H50N5O11 [M +
Na]+: 846.3326; found 846.3295.
32.64, 28.79, 26.99, 25.27, 23.13. LC-MS (ESI) m/z (M + H)+: 454.5.
2.1.8. 4-((1-(5-isopropyl-2,4-dimethoxybenzoyl)indolin-5-yl)amino)-4-
oxobutanoic acid (9)
Lithium hydroxide (52 mg, 2.2 mmol) was added to a mixture of 8
(500 mg, 1.1 mmol) in mixed solvents THF (2 mL), MeOH (2 mL) and
H2O (2 mL). The mixture was heated to reflux for 2 h. After completion
of the reaction, the pH of the mixture was adjusted to ~2 with 1 M HCl
and then extracted with EtOAc. The combined organic layers were
washed with saturated brine, dried over anhydrous Na2SO4, and
concentrated in vacuo to afford a pale yellow solid 411 mg. Yield: 85%.
1H NMR (400 MHz, DMSO‑d6) δ 9.72 (s, 1H), 7.78 (d, J = 0.6 Hz, 1H),
7.58 (dd, J = 7.5, 1.5 Hz, 1H), 7.34 (q, J = 1.2 Hz, 1H), 7.21 (d, J = 7.5
Hz, 1H), 6.70 (s, 1H), 4.28–4.10 (m, 2H), 3.88 (d, J = 57.9 Hz, 6H), 3.17
(td, J = 7.1, 1.0 Hz, 3H), 2.69–2.50 (m, 4H), 1.25 (d, J = 6.8 Hz, 6H). 13
C
2.1.10. N1-(2-(2-(2-(4-(6-hydroxy-1,3-dioxo-1H-benzo[de]isoquinolin-2
(3H)-yl)butanamido)ethoxy)ethoxy)ethyl)-N4-(1-(5-isopropyl-2,4-
dimethoxybenzoyl)indolin-5-yl)succinamide (probe 2)
NMR (100 MHz, DMSO‑d6) δ 170.94, 164.82, 159.37, 157.88, 138.53,
136.42, 129.80, 127.53, 119.97, 117.88, 116.69, 96.25, 55.82, 47.64,
31.75, 28.84, 26.98, 25.43, 23.15. LC-MS (ESI) m/z (M + H)+: 441.5.
HATU (31 mg, 0.083 mmol) and DIEA (35 μL, 0.192 mmol) was
added to a mixture of 7 (28 mg, 0.064 mmol) in anhydrous DMF 2.5 mL
at 0 ◦C. After stirring in an ice bath for 30 min, 10 (42 mg, 0.099 mmol)
was added. The mixture was stirred at room temperature for 8 h. After
completion of the reaction, the mixture was diluted with water and
extracted with EtOAc. The combined organic layers were then washed
with HCl (0.1 M), saturated brine, dried over anhydrous Na2SO4, and
concentrated in vacuo to give the crude product. The crude product was
then purified with column chromatography to give a yellow solid 35 mg.
Yield: 65%. 1H NMR (400 MHz, DMSO‑d6) δ 9.75 (s, 1H), 9.54 (s, 1H),
8.65 (dd, J = 7.5, 1.6 Hz, 1H), 8.37 (dd, J = 7.5, 1.5 Hz, 1H), 8.07 (d, J =
7.5 Hz, 1H), 7.85 (dd, J = 7.5, 1.5 Hz, 1H), 7.77 (d, J = 0.7 Hz, 1H),
7.52–7.43 (m, 4H), 7.34 (q, J = 1.1 Hz, 1H), 7.17 (d, J = 7.5 Hz, 1H),
6.68 (s, 1H), 4.26 (t, J = 7.1 Hz, 2H), 3.88 (d, J = 46.7 Hz, 6H), 3.74 (t, J
= 7.1 Hz, 2H), 3.63 (s, 4H), 3.58 (td, J = 7.1, 2.8 Hz, 4H), 3.40 (q, J =
7.0 Hz, 4H), 3.24–3.06 (m, 3H), 2.56 (dd, J = 7.4, 6.8 Hz, 2H), 2.37 (t, J
2.1.9. N1-(1-(2,4-dihydroxy-5-isopropylbenzoyl)indolin-5-yl)-N4-(2-(2-
(2-(4-(6-hydroxy-1,3-dioxo-1H-benzo[de]isoquinolin-2(3H)-yl)
butanamido)ethoxy)ethoxy)ethyl)succinamide (probe 1)
HATU (31 mg, 0.083 mmol) and DIEA (35 μL, 0.192 mmol) was
added to a mixture of 7 (26 mg, 0.064 mmol) in anhydrous DMF 2.5 mL
at 0 ◦C. After stirring in an ice bath for 30 min, N-(2-(2-(2-aminoethoxy)
ethoxy)ethyl)-4-(6-hydroxy-1,3-dioxo-1H-benzo[de]isoquinolin-2(3H)-
yl)butanamide (10) (42 mg, 0.099 mmol) was added. The mixture was
stirred at room temperature for 8 h. After completion of the reaction, the
mixture was diluted with water and extracted with EtOAc. The com-
bined organic layers were then washed with HCl (0.1 M), saturated
brine, dried over anhydrous Na2SO4, and concentrated in vacuo to give
the crude product. The crude product was then purified with column
chromatography to give a yellow solid 37 mg. Yield: 70%. 1H NMR (400
6

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
= 7.0 Hz, 2H), 2.28 (t, J = 7.1 Hz, 2H), 1.97 (p, J = 7.0 Hz, 2H), 1.25 (d,
J = 6.8 Hz, 6H). 13C NMR (400 MHz, DMSO‑d6) δ 173.62, 172.88,
170.65, 168.16, 163.68, 158.45, 157.53, 154.42, 140.90, 138.77,
137.11, 129.77, 128.40, 127.36, 126.88, 125.15, 120.43, 119.77,
118.86, 113.06, 93.32, 70.09, 56.48, 46.47, 44.93, 42.27, 35.59, 34.71,
31.92, 28.13, 26.83, 23.17. HR-MS (m/z) (ESI): calcd for C46H54N5O11
[M + Na]+: 874.3639; found 874.3618.
performed to detect the fluorescence intensity of all tested compounds
inside the cells. Cells were grown in 6-well plates at appropriate den-
sities and incubated for 24 h. Afterwards, different solutions at a con-
centration of 25 mmol Lꢀ 1 were added to the above cells for desired
time. Then, the cells were trypsinized, washed three times with PBS
along with centrifugation of 5000 rpm for 5 min, resuspended in 0.5 mL
PBS and measured using a FACSCalibur flow cytometer (BD Biosciences,
USA).
2.2. HSP90 binding affinity assay
2.7. In vivo biodistribution and drug accumulation
Various concentrations of Probe 1 and Probe 2 and FITC-labeled
geldanamycin were added to each reaction well. Recombinant HSP90
protein was diluted with the reaction buffer (20 mM 4-(2-hydroxyethyl)-
1-piperazineethanesulfonic acid pH 7.3, 50 mM KCl, 5 mM MgCl2, 20
Studies were performed in female nude mice bearing HCT116 tumors
(5 weeks old). After intravenous injection of test compounds via lateral
tail veins at a dose of 15 mg kgꢀ 1 for the desired time periods, the tumor-
bearing mice were sacrificed, and the heart, liver, spleen, lung, kidney,
and tumor tissues were excised, washed with phosphate buffered saline
(pH 7.4), weighed after drying and homogenized in 0.5 mL of DMSO. For
the determination of concentration of the test compounds, an aliquot of
plasma or the tissue homogenate was acidified with aqueous phosphoric
acid (0.15 M) and then extracted with chloroform: methanol (10:1 v/v).
After centrifugation, the chloroform: methanol layer was directly
analyzed by high performance liquid chromatography (SHIMADZU,
Japan).
mM Na2MoO4, 0.01% Triton X-100, 100 μg/mL bovine serum albumin,
and 2 mM dithiothreitol) and added to each well to get the final con-
centration of 22.2 nM. The enzyme reaction was initiated as soon as the
purified HSP90 was added. After incubation for 60 min at 4 ◦C, the
fluorescent polarization of the samples was determined by a microtiter-
plate reader (Envision, PerkinElmer).
2.3. In vitro cytotoxicity assay
Cells were seeded in 96-well plates at appropriate densities in growth
media. After incubation for 24 h, a range of concentrations of the test
compounds were added and the plates were incubated at 37 ◦C for 24 h,
48 h and 72 h. Cell proliferation was determined by MTT assay. The
cytotoxicity of all tested compounds is expressed as a percentage of the
control using the following equation:
2.8. Pharmacokinetics studies
BALB/c mice were randomly divided into three groups (n = 6). The
solutions of all tested compounds were intravenously injected via a tail
vein at a dose of 15 mg kgꢀ 1. Blood samples were collected from the tails
at time intervals (5 min, 30 min, 1 h, 4 h, 8 h and 12 h), and then placed
in heparinized tubes and centrifuged to obtain plasma. The amounts of
all tested compounds were calculated from standard curves previously
obtained, using high performance liquid chromatography (SHIMADZU,
Japan).
Cell viability (%) = (AS - AB)/(AControl - AB) * 100
(1)
where AS, AB and AControl are the absorbance of the sample, blank and
control wells.
2.4. Cellular uptake and subcellular distribution
2.9. Histological analysis
Cells were plated in 20 mm glass bottomed culture dishes at appro-
After the first injection, the mice (n = 3) with proper average tumor
size (9.78 mm × 6.73 mm) and weight (20.0 g) were sacrificed by CO2
inhalation. Then, tumors and major organs were harvested 2 h after the
final treatment and immediately flash-frozen in ꢀ 80 ◦C. Tumors and
major organs were embedded in paraffin for sectioning, and the sections
were subjected to histological staining analysis.
priate densities and incubated for 24 h. Afterwards, 25
μ
mol Lꢀ 1 test
compounds were added to the cells and the mixture was incubated for
desired time periods. The cells were washed with PBS and stained with
4% paraformaldehyde and subsequently imaged using a Leica TCS SP8
confocal fluorescence microscope by excitation at λex 488 nm
(Germany).
For the purpose of investigating the subcellular distribution, cells
were grown in 6-well plates at appropriate densities and incubated for
24 h. Afterwards, cells were treated with 10 mM compounds at 37 ◦C for
6 h. Then, the cells were washed with PBS three times and stained with
4% paraformaldehyde for 10 min and stained with Hoechst 33258 (1
mg mLꢀ 1) for 5 min. Subsequently, cells were imaged using a Leica TCS
SP8 confocal fluorescence microscope by excitation at λex 405 nm and
488 nm (Germany).
3. Results and discussion
3.1. Synthesis of probe 1 and probe 2
As shown in Scheme 1, the synthesis route of Probe 1 and probe 2
was started from compound 1 (1-(2,4-dihydroxy-5-isopropylphenyl)
ethan-1-one), a commercially available chemical intermediate. Firstly,
both phenolic hydroxyls of 1 were protected by benzyl bromide to give 2
with a high yield. Next, compound 2 was oxidized at the presence of
bromine by converting aldehyde to carboxylic acid to afford 3, which
was subsequently reacted with oxalyl chloride to generate 4 as a reactive
intermediate, without further purification. Compound 4 was further
treated with 5-nitroindole to generate 5 in a moderate condition, which
was fully debenzylated and reduced using 10% Pd/C in methanol in a
hydrogenation vessel to afford amine 6. Introduction of the carboxylic
acid group was conducted in the presence of the succinic anhydride to
afford the desired key acid 7 in a relatively high yield. To construct the
inactive resorcinol-blocked analog, acid 7 was reacted with methyl io-
dide at 0 ◦C to block the resorcinol group to afford the inactive scaffold,
which was then hydrolyzed in the presence of lithium hydroxide to give
the desired key acid 9. Finally, Probe 1 and Probe 2 were achieved by
condensation of 7 and 9 with 10, respectively, in the presence of HATU
2.5. Competitive experiments
Cells were plated in 20 mm glass bottomed culture dishes at proper
densities and incubated for 24 h. Afterwards, Probe 1/AT13387 (1/100,
n/n) and Probe 1 in PBS solution at a concentration of 25
μ
mol Lꢀ 1 were
added to the above cells and the mixtures were incubated for 4 h. The
cells were washed with PBS and stained with 4% paraformaldehyde and
subsequently imaged using a Leica TCS SP8 confocal fluorescence mi-
croscope by excitation at λex 488 nm (Germany).
2.6. Flow cytometry analysis
To quantitatively analyze cellular uptake, flow cytometry was
7

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
and Hunig’s base under moderate conditions.
cellular uptake, competitive experiments were also conducted. A549
and HCT116 cells were incubated with 100-fold molar free AT13387
mixed with Probe 1 (25 μM, 100:1 n/n) for 6 h, which can block the
3.2. HSP90 binding affinity assay
binding between Probe 1 and HSP90. Vehicle mixed with Probe 1 in
culture medium (25 M) was also incubated at the same condition as a
μ
First, both synthesized compounds were tested for their potential
HSP90 inhibitory properties, conducted by a fluorescence polarization
(FP) assay using AT13387 and unconjugated probe 4 as a positive
control at the concentration of 1000 nM, 333.3 nM, 111.1 nM, 37.0 nM,
12.3 nM, and 4.1 nM in duplicate. As presented in Table 1, Probe 1
showed significant differences compared with Probe 2. It was apparent
that Probe 1 showed appreciable binding to recombinant human
contrast. As shown in Fig. 4, brighter green fluorescence was observed in
both cell lines incubated with vehicle and Probe 1 than the AT13387
group, indicating a higher Probe 1 cellular uptake. In the normal
HL7702 cell groups, both test groups showed no obvious fluorescence
(Fig. 4C), demonstrating the HSP90-dependent cellular uptake mecha-
nism. Quantitative analysis was further performed. Notably, the mean
fluorescence intensity (MFI) values of Probe 1 uptake in the vehicle
groups were approximately 4.2- and 4.2-fold stronger than those of the
AT13387 groups, further confirm the results.
HSP90
that of AT13387, while Probe 2 and unconjugated probe 10 were shown
to have a negligible impact on HSP90 that was much weaker than
α protein with an inhibition potency that was slightly less than
α
AT13387. These results showed that the conjugation of AT13387 and
fluorescent probe did not affect the binding affinity of the hybrids to
HSP90, which indicated that Probe 1 with active HSP90 scaffold was
suitable for constructing HSP90 inhibitor conjugates by this method,
suggesting the possibility of higher selectivity in tumor cells. Thus, these
results inspired us to conduct a further evaluation of compound Probe 1.
3.6. Time-dependent cellular uptake and cellular accumulation of probe 1
in A549 cells
After validating the mechanism, a fluorescence experiment was also
performed to demonstrate whether Probe 1 could increase intracellular
drug accumulation in a time-dependent manner. The cellular uptake
behavior was investigated in A549 cells by CSLM. A549 cells were
3.3. In vitro cytotoxicity assay of probe 1 and probe 2
incubated with 25 μM Probe 1 for different incubation time periods (1 h,
Next, the in vitro anti-proliferative activities of Probe 1 and Probe 2
were evaluated by MTT assay in A549 cells and HCT116 cells. First, the
cytotoxicity of both compounds in A549 and HCT116 were evaluated for
totally 24 h, 48 h and 72 h. As illustrated in Fig. 1, cell viability was
decreased with increasing drug concentrations and exposure, which
demonstrated the cytotoxicity of Probe 1 was dose-dependent and time-
dependent and may be attributed to the inhibition of HSP90 in both cell
lines. As for Probe 2 with reduced HSP90 binding affinity, there was
almost negligible toxicity at all concentrations and treatment durations
during the assay, demonstrating that HSP90 affinity had a non-
negligible role in cell viability and may lead to cumulative probe
accumulation.
2 h, 4 h, and 6 h). As shown in Fig. 5A, when A549 cells were incubated
with Probe 1 for 1 h, green fluorescence in the cytoplasm could be
observed, indicating the successful internalization of Probe 1. In addi-
tion, with the increase in incubation time from 1 h to 6 h, stronger
fluorescence appeared in A549 cells. A further quantitative measure-
ment of fluorescence intensity was undertaken in A549 cells by cell lysis
analysis and demonstrated that the fluorescence intensity of Probe 1
within A549 cells was greatly enhanced (Fig. 5B). The MFI of the group
treated with Probe 1 for 6 h exhibited about a 3-fold increase compared
with the 1 h group. Taken together, the results indicated that HSP90
binding may result in a time-dependent increase in the accumulation of
higher concentrations of probes or drugs in tumor cells for imaging.
To confirm that the cytotoxicity of Probe 1 differed between tumor
cells and normal cells, in vitro cytotoxicity assays were performed in HL-
7702 cells and HepG2 cells for 24 h. Because of the high affinity of
activated HSP90, HepG2 cells were more affected compared with the
non-oncogenic HL-7702 cells. As shown in Fig. 2, the covalent conju-
gation of AT13387 to 4-hydroxy-1,8-naphthalimide displayed high
selectivity and high cytotoxicity against tumor cells while sparing
normal cells, indicating Probe 1 as a cancer-specific bioimaging probe.
3.7. In vivo biodistribution and drug accumulation of probe 1 and probe 2
To further investigate the biological properties of Probe 1 and Probe
2 in vivo, biodistribution and drug accumulation profiles in major organs
were evaluated in HCT116 xenografts. After a single intravenous
administration of 15 mg kgꢀ 1 of both compounds, the mice were sacri-
ficed and the major organs were excised and homogenized in 0.5 mL
DMSO. The concentration of all samples was examined by HPLC. As
shown in Fig. 6, biodistribution analysis indicated that Probe 1 showed
more enhanced accumulation at the tumor site than other organs, and
the concentration of Probe 1 in excised tumor tissue was 4.5-fold higher
than Probe 2 at 12 h post-injection, suggesting favorable tumor selec-
tivity. Additionally, with the increase of administration time from 1 h to
12 h, the concentration of Probe 1 in the 12 h group exhibited 4.7-fold
higher than the 1 h group. Overall, these results indicate the therapeutic
superiority of Probe 1, demonstrating preferential accumulation in
tumor tissue.
3.4. Cellular uptake and subcellular distribution of probe 1 and probe 2
Next, we employed both probes for the imaging of a tumor micro-
environment to demonstrate whether Probe 1 could be efficiently
internalized into tumor cells. The cellular uptake and subcellular dis-
tribution behaviors of Probe 1 and Probe 2 in HCT116 and A549 cells
was measured by confocal laser scanning microscopy (CLSM). As shown
in Fig. 3A and B, when A549 cells and HCT116 cells were treated with
Probe 1 and Probe 2 for 0.5 h, respectively, the Probe 1 groups were
observed to have significantly stronger green fluorescence in the cyto-
plasm than the Probe 2 groups, in which it was almost undetectable,
indicating that Probe 1 has more effective cellular uptake and imaging
than Probe 2 in both tumor cell lines. Blue Hoechst 33258, which is
known to selectively and strongly stain intact nuclei, was also used to
investigate the subcellular distribution of the probes. Strong blue fluo-
rescence was observed in the nuclei and it is obvious that Probe 1 was
mostly located in the cytoplasm and with very little in the nuclei,
demonstrating effective cellular uptake of the probe.
3.8. Confocal laser scanning microscopy (CLSM) assay of ex vivo mice
tumor sections
Furthermore, sections from mice tumor tissues (12 h-treatment at 15
mg kgꢀ 1) embedded in paraffin were also cut into a section of 8
μm-thick
sections for CLSM assay. As shown in Fig. 7, the tumor sections treated
with Probe 1 exhibited strong fluorescence, suggesting the high tumor
accumulation of Probe 1. These results are consistent with those of the
in vitro and ex vivo biodistribution assays, further elucidating the tumor
cumulative effect of Probe 1.
3.5. Competitive experiments of probe 1
To further confirm the mechanism of HSP90-mediated Probe 1
8

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
3.9. Pharmacokinetics studies
Appendix A. Supplementary data
To further determine the results of previous studies that Probe 1 has
this property in vivo, a pharmacokinetic study was undertaken by
intravenous injection of 15 mg kgꢀ 1 of Probe 1, Probe 2, and free
AT13387 to BALB/c mice of approximately 20 g in weight. As shown in
Fig. 8, Probe 1 was retained at a relatively higher concentration in the
blood for up to 12 h, whereas Probe 2 and free AT13387 were rapidly
cleared from the blood after i.v. administration. Probe 1 with an active
HSP90 inhibitor scaffold could bind with the amino acid residues of
HSP90 in vivo, which could affect the blood concentration and lead to a
relatively longer retention time. However, Probe 2 with inactive HSP90
inhibitor scaffold didn’t have these properties. These results further
confirm the possibility of enhanced probe accumulation via HSP90
binding in the tumor tissues for imaging.
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.dyepig.2021.109783.
References
[1] Ye Y, Chen X. Integrin targeting for tumor optical imaging[J]. Theranostics 2011;1:
102.
[2] Imamura T, Saitou T, Kawakami R. In vivo optical imaging of cancer cell function
and tumor microenvironment[J]. Canc Sci 2018;109(4):912–8.
[3] Zhang RR, Schroeder AB, Grudzinski JJ, et al. Beyond the margins: real-time
detection of cancer using targeted fluorophores[J]. Nat Rev Clin Oncol 2017;14(6):
347.
[4] Boyce M, Bertozzi CR. Bringing chemistry to life[J]. Nat Methods 2011;8(8):
638–42.
[5] Liu Z, Lavis LD, Betzig E. Imaging live-cell dynamics and structure at the single-
molecule level[J]. Mol Cell 2015;58(4):644–59.
[6] Zhou Z, Lu ZR. Molecular imaging of the tumor microenvironment[J]. Adv Drug
Deliv Rev 2017;113:24–48.
3.10. Histological analysis
[7] Elias DR, Thorek DLJ, Chen AK, et al. In vivo imaging of cancer biomarkers using
activatable molecular probes[J]. Canc Biomarkers 2008;4(6):287–305.
[8] Lin HH, Chan YC, Chen JW, et al. Aggregation-induced emission enhancement
characteristics of naphthalimide derivatives and their applications in cell imaging
[J]. J Mater Chem 2011;21(9):3170–7.
Finally, to validate the therapeutic efficacy, heart, liver, spleen, lung,
kidney and tumor issues were excised from HCT116-bearing mice at day
14 after the first injection for further histological analysis. As illustrated
in Fig. 9, typical features of apoptosis, such as nuclear damage, were
observed from haematoxylin and eosin (H&E)- stained sections, and the
results suggest almost no noticeable major organ-related toxicities. In
situ TdT-mediated dUTP nick end labeling (TUNEL) assay was further
employed to observe tumor cell apoptosis, with apoptotic cells repre-
sented in green (Fig. 10). The Probe 1 group showed a relatively higher
positive rate of tumor cell apoptosis than the Probe 2 group and the
control group (Fig. 10), exhibiting the same tendency as in the HSP90
binding assay. All of these results are in good agreement with the bio-
distribution assay, further validating the superior efficacy of Probe 1.
[9] Georgiev NI, Bojinov VB. Design, synthesis and sensor activity of a highly
photostable blue emitting 1, 8-naphthalimide[J]. J Lumin 2012;132(9):2235–41.
[10] Georgiev NI, Bojinov VB, Nikolov PS. The design, synthesis and photophysical
properties of two novel 1, 8-naphthalimide fluorescent pH sensors based on PET
and ICT[J]. Dyes Pigments 2011;88(3):350–7.
[11] Zhu S, Yu X, He Y, et al. Synthesis and fluorescent studies of a low molecular
weight rotor for living cancer cell imaging[J]. Dyes Pigments 2020;178:108353.
[12] Bardajee GR, Li AY, Haley JC, et al. The synthesis and spectroscopic properties of
novel, functional fluorescent naphthalimide dyes[J]. Dyes Pigments 2008;79(1):
24–32.
[13] Shaki H, Gharanjig K, Rouhani S, et al. Synthesis and photophysical properties of
some novel fluorescent dyes based on naphthalimide derivatives[J]. J Photochem
Photobiol Chem 2010;216(1):44–50.
[14] Chang SC, Utecht RE, Lewis DE. Synthesis and bromination of 4-alkylamino-N-
alkyl-1, 8-naphthalimides[J]. Dyes Pigments 1999;43(2):83–94.
[15] Ha Y, Choi HK. Recent conjugation strategies of small organic fluorophores and
ligands for cancer-specific bioimaging[J]. Chem Biol Interact 2016;248:36–51.
[16] Sirk SJ, Olafsen T, Barat B, et al. Site-specific, thiol-mediated conjugation of
fluorescent probes to cysteine-modified diabodies targeting CD20 or HER2[J].
Bioconjugate Chem 2008;19(12):2527–34.
4. Conclusions
In conclusion, we have successfully developed an HSP90-mediated
fluorescent probe, Probe 1, for cancer-specific bioimaging. HSP90
binding and in vitro anti-proliferation activities confirmed that conju-
gation of HSP90 inhibitor with 4-hydroxy-1,8-naphthalimide retains
HSP90 affinity. Furthermore, cellular uptake and competitive experi-
ments evaluation revealed distinctly different results for Probe 1 and the
active analog Probe 2. Probe 1 with an active HSP90 inhibitor scaffold
was demonstrated to have HSP90-dependent cellular uptake mechanism
and accumulate in tumor cells in a time-dependent manner. Moreover,
in vivo biological evaluation further revealed that Probe 1 showed
enhanced tumor-targeting behavior, compared with Probe 2. Therefore,
Probe 1 can be considered as a potential tumor targeting probe in
fluorescent imaging with good biocompatibility.
[17] Trapani G, Denora N, Trapani A, et al. Recent advances in ligand targeted therapy
[J]. J Drug Target 2012;20(1):1–22.
[18] Srinivasarao M, Low PS. Ligand-targeted drug delivery[J]. Chem Rev 2017;117
(19):12133–64.
[19] Zuo Y, Shi Y, Li X, et al. A novel 2, 5-diaminopyrimidine-based affinity probe for
Bruton’s tyrosine kinase[J]. Sci Rep 2015;5(1):1–9.
[20] Chen J, Wang X, He F, et al. Development of a selective labeling probe for Bruton’s
tyrosine kinase quantification in live cells[J]. Bioconjugate Chem 2018;29(5):
1640–5.
[21] Crowe LB, Hughes PF, Alcorta DA, et al. A fluorescent Hsp90 probe demonstrates
the unique association between extracellular Hsp90 and malignancy in vivo[J].
ACS Chem Biol 2017;12(4):1047–55.
[22] Taldone T, Rodina A, Gomes E, et al. Synthesis and evaluation of cell-permeable
biotinylated PU-H71 derivatives as tumor Hsp90 probes[J]. Beilstein J Org Chem
2013;9(1):544–56.
[23] Zhu S, Li Y, Huang Y, et al. Optimized HSP90 mediated fluorescent probes for
cancer-specific bioimaging[J]. J Mater Chem B 2020;8(9):1878–96.
[24] Windisch P, Zwahlen DR, Giesel FL, et al. Clinical results of fibroblast activation
protein (FAP) specific PET for non-malignant indications: systematic review[J].
EJNMMI Res 2021;11(1):1–12.
CRediT authorship contribution statement
Shulei Zhu: Conceptualization, Methodology, Writing – original
draft. Yingxin Lu: Investigation, Formal analysis. Jiyu Jin: Investiga-
tion. Jiahui Yu: Validation. Wei Lu: Writing – review & editing,
Supervision.
[25] Zhou S, Zhen Z, Paschall AV, et al. FAP-Targeted photodynamic therapy mediated
by ferritin nanoparticles elicits an immune response against cancer cells and cancer
associated fibroblasts[J]. Adv Funct Mater 2021;31(7):2007017.
[26] Syed M, Flechsig P, Liermann J, et al. Fibroblast activation protein inhibitor (FAPI)
PET for diagnostics and advanced targeted radiotherapy in head and neck cancers
[J]. Eur J Nucl Med Mol Imag 2020;47:2836–45.
Declaration of competing interest
[27] Xing J, Gong Q, Zou R, et al. A novel fibroblast activation protein-targeted near-
infrared fluorescent off–on probe for cancer cell detection, in vitro and in vivo
imaging[J]. J Mater Chem B 2018;6(10):1449–51.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
[28] Cherfaoui B, Guo T, Sun HP, et al. Synthesis and evaluation of 4-(2-hydroxypropyl)
piperazin-1-yl) derivatives as Hsp90 inhibitors[J]. Bioorg Med Chem 2016;24(11):
2423–32.
[29] Kamal A, Thao L, Sensintaffar J, et al. A high-affinity conformation of Hsp90
confers tumour selectivity on Hsp90 inhibitors[J]. Nature 2003;425(6956):
407–10.
Acknowledgements
[30] Garcia-Carbonero R, Carnero A, Paz-Ares L. Inhibition of HSP90 molecular
chaperones: moving into the clinic[J]. Lancet Oncol 2013;14(9):e358–69.
This work was supported by the Fundamental Research Funds for the
Central Universities and the National Natural Science Foundation of
China (No.22077034).
9

S. Zhu et al.
Dyes and Pigments 196 (2021) 109783
[31] Jensen MR, Schoepfer J, Radimerski T, et al. NVP-AUY922: a small molecule
HSP90 inhibitor with potent antitumor activity in preclinical breast cancer models
[J]. Breast Canc Res 2008;10(2):1–12.
[34] Pandey MK, Prasad S, Tyagi AK, et al. Targeting cell survival proteins for cancer
cell death. Pharmaceuticals 2016;9(1):11.
[35] Jafari A, Rezaei-Tavirani M, Farhadihosseinabadi B, et al. HSP90 and co-
chaperones: impact on tumor progression and prospects for molecular-targeted
cancer therapy[J]. Canc Invest 2020;38(5):310–28.
[32] Chatterjee S, Burns TF. Targeting heat shock proteins in cancer: a promising
therapeutic approach[J]. Int J Mol Sci 2017;18(9):1978.
[33] Maloney A, Workman P. HSP90 as a new therapeutic target for cancer therapy: the
story unfolds[J]. Expet Opin Biol Ther 2002;2(1):3–24.
10