Tunable phosphatase-sensitive stable prodrugs of 5-aminolevulinic acid for tumor fluorescence photodetection

Article abstract of DOI:10.1016/j.jconrel.2016.05.047

5-Aminolevulinic acid (5-ALA) has been at the forefront of small molecule based fluorescence-guided tumor resection and photodynamic therapy. 5-ALA and two of its esters received marketing authorization but suffer from several major limitations, namely low stability and poor pharmacokinetic profile. Here, we present a new class of 5-ALA derivatives aiming at the stabilization of 5-ALA by incorporating a phosphatase sensitive group, with or without self-cleavable linker. Compared to 5-ALA hexyl ester (5-ALA-Hex), these compounds display an excellent stability under acidic, basic and physiological conditions. The activation and conversion into the 5-ALA is controlled and can be structure-tailored. The prodrugs display reduced acute toxicity compared to 5-ALA-Hex with superior dose response profiles of protoporphyrin IX synthesis and fluorescence intensity in human glioblastoma cells in vitro. Clinically relevant fluorescence kinetics in vivo shown in U87MG glioblastoma spheroid tumor model in chick embryos provide a solid basis for their further development and translation to clinical fluorescence guided tumor resection and photodynamic therapy.

Full text of DOI:10.1016/j.jconrel.2016.05.047

Journal of Controlled Release 235 (2016) 155–164
Contents lists available at ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel
Tunable phosphatase-sensitive stable prodrugs of 5-aminolevulinic acid
for tumor fluorescence photodetection
⁎
Andrej Babič , Viktorija Herceg, Imène Ateb, Eric Allémann, Norbert Lange
School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, 1211 Geneva, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
5-Aminolevulinic acid (5-ALA) has been at the forefront of small molecule based fluorescence-guided tumor re-
section and photodynamic therapy. 5-ALA and two of its esters received marketing authorization but suffer from
several major limitations, namely low stability and poor pharmacokinetic profile. Here, we present a new class of
5-ALA derivatives aiming at the stabilization of 5-ALA by incorporating a phosphatase sensitive group, with or
without self-cleavable linker. Compared to 5-ALA hexyl ester (5-ALA-Hex), these compounds display an excellent
stability under acidic, basic and physiological conditions. The activation and conversion into the 5-ALA is con-
trolled and can be structure-tailored. The prodrugs display reduced acute toxicity compared to 5-ALA-Hex
with superior dose response profiles of protoporphyrin IX synthesis and fluorescence intensity in human glio-
blastoma cells in vitro. Clinically relevant fluorescence kinetics in vivo shown in U87MG glioblastoma spheroid
tumor model in chick embryos provide a solid basis for their further development and translation to clinical fluo-
rescence guided tumor resection and photodynamic therapy.
Received 10 March 2016
Received in revised form 20 May 2016
Accepted 21 May 2016
Available online 25 May 2016
Keywords:
5-Aminolevulinic acid
Prodrugs
Alkaline phosphatase
Tumor photodetection
Photodynamic therapy
© 2016 Published by Elsevier B.V.
1. Introduction
and proved effective in clinical trials [14]. Today, 5-ALA is approved
and commercialized under the tradename of Gliolan® [15] despite the
5-Aminolevulinic acid (5-ALA) is a natural precursor of protopor-
phyrin IX (PpIX) in heme biosynthesis (Fig. 1). Exogenous administra-
tion of 5-ALA causes temporary overproduction and accumulation of
fluorescent PpIX by circumventing the negative feedback mechanism
of heme biosynthesis [1]. Preferential accumulation of PpIX in rapidly
proliferating cells as compared to healthy cells provides excellent inher-
ent selectivity. This gives 5-ALA great potential for use in fluorescence
photodetection (FPD) [2,3], fluorescence-guided tumor resection
(FGR) [4,5], sono-(SDT) [6,7] and photodynamic therapy (PDT) [8,9].
In this respect 5-ALA is an exceptional molecule as there have been
only a handful of small molecules successfully used for these purposes
[10].
The overwhelming clinical need for better detection and treatment
options is perhaps best exemplified by malignant gliomas [11], a
group of highly aggressive brain neoplasms characterized by an ex-
tremely infiltrative growth and median survival rate of 15 months
[12]. The optimal treatment option for glioma patients is the complete
resection of malignant tissue followed by adjuvant radiotherapy and
chemotherapy [13]. However, complete tumor resections often fail be-
cause of undistinguishable differences between tumors and the sur-
rounding healthy tissue during surgery. To address this issue, 5-ALA-
induced PpIX fluorescence was tested for FGR in malignant gliomas
fact that only 0.1% of the administered dose reaches the brain.
The latter can be explained by 5-ALA's poor bioavailability caused by
its zwitterionic character under physiological conditions. The resulting
polarity, thus, limits 5-ALA's ability to pass biological barriers. This,
along with its short half-life, small volume of distribution and accumu-
lation in the liver and kidneys [16], makes 5-ALA non suitable for sys-
temic administration [17]. To overcome these shortcomings, different
strategies aimed at improving the drug-likeness and the pharmacoki-
netic profile have been employed. An optimal clinical outcome can be
expected when a balance is reached between the stability of the mole-
cule, ability to pass biological barriers, passive or active entry [18], low
toxicity, and most importantly efficient induction of PpIX-mediated
fluorescence.
Until today, research on improving 5-ALA-induced PpIX focused on
increasing 5-ALA's lipophilicity mostly through the modifications of car-
boxylic group of 5-ALA resulting in a vast variety of structurally diverse
esters [19]. The major reasoning is that increased lipophilicity potential-
ly improves the passage through biological membranes thus increasing
intracellular substrate concentration for heme biosynthesis. Recently,
substantial efforts have also been directed towards nanocarriers loaded
[20–24] or conjugated [25–28] with 5-ALA. These approaches tackle
some of the pharmacokinetic and specificity drawbacks of 5-ALA but
have so far failed to produce a clinical candidate for the improved deliv-
ery of 5-ALA mostly due to high 5-ALA loading and delivery required for
this type of FPD or PDT.
⁎
Corresponding author.
E-mail address: andrej.babic@unige.ch (A. Babič).
http://dx.doi.org/10.1016/j.jconrel.2016.05.047
0168-3659/© 2016 Published by Elsevier B.V.

156
A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
pseudodipeptides conjugated to 5-ALA's which induced the PpIX pro-
duction in only limited number of cancer cell lines [32–34].
Albeit endowed with poor physico-chemical properties the limited
clinical success of 5-ALA and its ester derivatives provides a proof of
the translational potential for new 5-ALA prodrugs that manage to over-
come these drawbacks. Since 5-ALA-induced PpIX is selective for tu-
mors, we hypothesized that the introduction of an N-terminal group
that can be activated by ubiquitously expressed enzymes would circum-
vent this problem. Therefore, we have focused our efforts on the design
and synthesis of novel classes of 5-ALA derivatives modified at the
amino group that can be activated in vivo by alkaline phosphatases
[35], ubiquitous enzymes present in all tissues and known to be
overexpressed in certain types of tumors [36–39]. To the best of our
knowledge this is the first time that the 5-amino group has been suc-
cessfully protected using a phosphate group. These modifications hold
promise to overcome barriers for systemic administration of conven-
tional 5-ALA esters.
Fig. 1. Exogenous applications of non-fluorescent 5-ALA leads to the accumulation of
fluorescent PpIX in cancer cells.
The only clinically successful solution to improve 5-ALA remains a
simple esterification of the 5-ALA's carboxylic group with aliphatic alco-
hols. 5-ALA-methyl ester (Metvix®) is approved for the topical treat-
ment of actinic keratosis and basal cell carcinoma [9], and 5-ALA-Hex
(Hexvix®) has been marketed for the FPD of bladder cancer [19]. Al-
though improved in the terms of lipophilicity and permeability, 5-
ALA-Hex is acutely toxic with equally poor pharmacokinetic profile as
5-ALA hence not providing a potential alternative for systemic adminis-
tration [17,29].
The additional drawback of 5-ALA derivatives with unprotected 5-
amino group is their inherent chemical instability. At physiological pH
they undergo irreversible degradation to 3,3′-(3-amino-1H-pyrrole-
2,4-diyl)dipropionic acid derivatives and predominantly 2,5-(β-
carboxyethyl) dihydropyrazine followed by oxidation to fully aromatic
2,5-(β-carboxyethyl)pyrazine (Fig. 2). This interferes with the forma-
tion of porphobilinogen which is the product of enzymatically catalyzed
asymmetrical 5-ALA condensation [30,31]. This adds to the complexity
for the clinical use of 5-ALA and its derivatives since the pharmacologi-
cal response will vary according to pH, buffer composition, concentra-
tion, and the timeframe between the preparation of the
pharmaceutical form and its use [31]. The obvious chemical solution
to this problem is the protection of 5-ALA's terminal amino group. How-
ever, the introduction of most groups at the 5-amino its end either
greatly reduces or completely diminishes pharmacological activity. To
the best of our knowledge, the only exception is a series of
2. Materials and methods
Chemicals were purchased from Sigma-Aldrich and Acros and used
as received. Ultrapure water was prepared using a Millipore MilliQ Ad-
vantage A10 System (Millipore, Bedford, USA). Alkaline phosphatase
from calf intestine was obtained from Applichem (AxonLab AG, Switzer-
land). Buffers were freshly prepared according to standard Geigy buffer
tables. Reactions were performed under inert argon conditions. Dry di-
chloromethane (CH2Cl2) was freshly distilled over CaH2 under nitrogen.
Dry tetrahydrofuran (THF) was distilled from sodium/ benzophenone
under nitrogen. Analytical thin layer chromatography was performed
using TLC Silica gel 60 F254 alumina plates and visualized at 254 nm
and 366 nm.
2.1. Purification
Normal phase Flash chromatography was done using PuriFlash 4100
system from Interchim (Montluçon, France) equipped with a diode
array detector and fraction collector. Preparative reverse-phase HPLC
was performed on PuriFlash 4100 system from Interchim or Waters
delta 600 HPLC system equipped with Waters 2487 dual wavelength
detector and Macherey Nagel Nucleodur C18-Htec 5 μm 250 × 21 mm
semi-preparative column or Waters Symmetry C4 5 μm, 19 × 150 mm
semi-preparative column.
2.2. Characterization
UPLC-MS analyses were performed using Thermo Fisher Accela sys-
tem and Macherey Nagel C18 Nucleodur Gravity 1.8 μm 50 × 2 mm
UPLC columns. 1H, 13C, 31P, COSY, HMBC and HSQC NMR spectra were
acquired at 298 K on Varian Anova 300 MHz and Agilent Varian Inova
500 MHz NMR spectrometer (Palo Alto, CA, USA) 500 MHz spectrome-
ters at the School of Pharmaceutical Sciences and Bruker Avance 300
and Bruker Cryo-Avance III 500 MHz spectrometers at the Chemistry
NMR facility at the University of Geneva, Switzerland. All chemical shifts
are reported in the standard notation of parts per million using the peak
of residual proton and carbon signals of the solvent as internal refer-
ences. NMR peaks are referred to as singlet (s), doublet (d), doublet of
doublets (dd), triplet (t), broad singlet (br s), or multiplet (m). Coupling
constants (J) are reported in Hertz. Low-resolution mass spectroscopy
using electrospray ionization (ESI) were performed on API 150EX
from AB/MDS Sciex. High-resolution mass spectroscopy measurements
were performed on QSTAR Pulsar XL from AB/MDS Sciex. Chemical
structures were drawn and named according IUPAC nomenclature
using ChemBioDraw version 14.0.0.117 software package. NMR spectra
were processed with Mnova version 8.0.2 software package.
Fig. 2. The stability-activity paradox of ALA. Free 5-amino ALA and its ester derivatives
(R1) have improved properties but low stability at physiological pH. Derivatization of
the amino group of 5-ALA (R2) results in stable ALA products that have greatly
diminished or no pharmacological activity.

A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
157
2.3. Synthesis
4.01 (d, J = 5.3 Hz, 2H), 3.90 (t, J = 6.8 Hz, 2H), 2.71 (t, J = 6.4 Hz,
2H), 2.63 (t, J = 6.4 Hz, 2H), 1.65–1.53 (m, 2H), 1.35–1.22 (m, 6H),
2.3.1. Hexyl 5-amino-4-oxopentanoate 1
0.88–0.84 (m, 3H). LRMS, ESI: m/z 400.6 [M + , 417.1
H]+
Hexyl 5-amino-4-oxopentanoate was synthesized according to pub-
lished procedure [40]. Briefly, thionyl chloride (8.20 g, 68.9 mmol) was
added dropwise to 1-hexanol (60.0 mL, 48.8 g, 475 mmol) in an ice-
bath. After the solution was stirred for 10 min the ice-bath was re-
moved, 5-aminolevulinic acid hydrochloride (8.20 g, 49.0 mmol) was
added. The reaction mixture was stirred at ambient temperature for
1 h, followed by reflux at 80 °C for 2 h. After cooling down, ether
(150 mL) was added and the precipitate filtered off and washed with
ether (2 × 50 mL). Extensive drying in vacuo gave colorless solid
(10.6 g, 42.1 mmol, 86%).
[M + NH4]+, 422.1 [M + Na]+.
Crude hexyl 5-(((iodomethoxy)carbonyl)amino)-4-oxo-pentanoate
was dissolved in toluene (50.0 mL) followed by the addition of silver
dibenzyl phosphate (346 mg, 0.90 mmol). The suspension was heated
at 50 °C for 1 h in the dark when the starting material was consumed.
The brown precipitate was filtered off and the solvent evaporated
under reduced pressure. The crude product was purified by Flash chro-
matography using DCM/MeOH gradient yielding colorless oil (139 mg,
0.25 mmol, 42% yield, 2 steps).
1H NMR (300 MHz, DMSO-d6) δ 8.35 (s, 3H), 3.97 (t, J = 6.7 Hz, 2H),
3.92 (s, 2H), 2.78 (t, J = 6.5 Hz, 2H), 2.52 (t, J = 6.5 Hz, 2H), 1.61–1.43
(m, 2H), 1.37–1.14 (m, 2H), 1.05–0.73 (m, 1H). 13C NMR (75 MHz,
DMSO-d6) δ 203.33, 172.74, 64.77, 47.19, 34.94, 31.55, 28.72, 27.76,
25.68, 22.67, 14.58. LRMS, ESI: m/z 216.4 [M + H]+. HRMS: m/z calculat-
ed for C11H22NO3 216.1594 [M]+, observed 216.1597.
2.3.4.2. Procedure B. 2 (813 mg, 2.65 mmol) and silver dibenzyl phos-
phate 3 (1.22 g, 3.18 mmol) were suspended in dry toluene (50.0 mL)
and heated at 60 °C. After 5 h the dark brown suspension was suction
filtered. The solvent was evaporated under reduced pressure and the
crude product purified by Flash chromatography using a DCM/MeOH
gradient yielding a colorless oil (1.04 g, 1.90 mmol, 72% yield). The spec-
tral properties of the product were identical using both procedures. 1H
NMR (300 MHz, CDCl3) δ 7.33 (s, 10H), 5.60 (d, J = 13.8 Hz, 2H), 5.06
(d, J = 7.9 Hz, 4H), 4.16–3.98 (m, 4H), 2.82–2.50 (m, 4H), 1.67–1.51
(m, 2H), 1.40–1.21 (m, 6H), 0.88 (t, J = 5.9 Hz, 3H). 13C NMR
(75 MHz, CDCl3) δ 203.23, 172.60, 154.07, 135.70, 128.77, 128.13,
83.86, 83.79, 77.67, 77.25, 76.82, 69.80, 69.73, 65.33, 50.60, 34.57,
31.61, 29.91, 28.72, 28.02, 25.74, 22.73, 14.21. LRMS, ESI: m/z 550.3
[M + H]+, 567.3 [M + NH4]+, 572.5 [M + Na]+. HRMS: m/z calculated
for C27H36NO9P 550.2201 [M + H]+, observed 550.2205.
2.3.2. Hexyl 5-(((chloromethoxy)carbonyl)amino)-4-oxopentanoate 2
1 (251.0 mg, 1.00 mmol) was dissolved in dry DCM (20.0 mL) and
cooled to −20 °C under argon atmosphere. Chloromethyl
chloroformate (141.8 mg, 1.10 mmol) was added under stirring in one
portion followed by dropwise addition of triethylamine (417 μL,
3.00 mmol) dissolved in dry DCM (5.0 mL). The reaction mixture was
stirred for 1 h at −20 °C and allowed to warm up to ambient tempera-
ture. After quenching with water (5.0 mL) the reaction mixture was ex-
tracted with DCM (2 × 20 mL). The organic phase was washed with
diluted HCl (2 × 10 mL) and saturated NaHCO3 solution and dried
over Na2SO4. The product was purified by Flash chromatography using
DCM/MeOH gradient giving colorless oil (272 mg, 0.886 mmol, 89%
yield). 1H NMR (300 MHz, CDCl3) δ 6.11 (t, J = 5.3 Hz, 1H), 5.61 (s,
2H), 4.01 (d, J = 5.3 Hz, 2H), 3.90 (t, J = 6.8 Hz, 2H), 2.61 (t, J =
6.4 Hz, 2H), 2.49 (t, J = 6.4 Hz, 2H), 1.45 (q, J = 6.9 Hz, 2H), 1.21–
1.07 (m, 6H), 0.83–0.66 (m, 3H). 13C NMR (75 MHz, CD3OD) δ 203.96,
172.67, 153.9, 70.81, 65.11, 50.49, 34.43, 31.49, 28.59, 27.89, 25.62,
2.3.5.
Triethylammonium
salt
of
hexyl
4-oxo-5-
((((phosphonooxy)methoxy)carbonyl)amino)pentanoate 5
(600 mg, 1.09 mmol) and triethylamine (455 μL, 3.30 mmol) were
dissolved in absolute ethanol (50.0 mL). The reaction flask was flushed
with argon before palladium on charcoal (50 mg) was added. Argon was
exchanged with hydrogen and the reaction mixture stirred under hy-
drogen at 1 bar for 3 h. The catalyst was filtered off and washed with ab-
solute ethanol (2 × 10 mL). The product as colorless oil was obtained
after evaporation of the solvent and extensive drying of the product in
vacuo (506 mg, 3.05 mmol, 99.0% yield). 1H NMR (300 MHz, CD3OD) δ
5.49 (d, J = 12.6 Hz, 2H), 4.17–3.96 (m, 4H), 3.17 (q, J = 7.3 Hz, 6H),
2.75 (t, J = 6.3 Hz, 2H), 2.57 (t, J = 6.3 Hz, 2H), 1.61–1.58 (m, 2H),
1.45–1.17 (m, 15H), 0.99–0.81 (m, 3H). 13C NMR (75 MHz, CD3OD) δ
205.49, 173.20, 156.52, 83.39, 83.33, 64.66, 49.73, 48.69, 48.41, 48.12,
47.84, 47.56, 47.27, 46.99, 46.28, 33.82, 31.43, 28.49, 27.48, 25.51,
22.60, 14.07. LRMS, ESI: m/z 308.4 [M + H]+, 330.1 [M + Na]+
.
HRMS: m/z calculated for C13H22ClNO5 330.1079 [M + Na]+, observed
330.1078.
2.3.3. Silver dibenzyl phosphate 3
Dibenzyl phosphate silver salt was synthesized according to the
published procedure [41]. Briefly, dibenzyl phosphate (1.00 g,
3.59 mmol) was dissolved in ethanol/water mixture (50/50 V/V,
10 mL). The solution of sodium hydroxide (3.59 mL, 1.00 M,
3.59 mmol) was added and the solution stirred for 5 min. Then silver ni-
trate (610 mg, 3.59 mmol) dissolved in hot water (5.0 mL) was added. A
colorless precipitate was formed immediately. After 15 min the suspen-
sion was suction filtered and the precipitate washed with cold water
(5.0 mL) and ethanol (5.0 mL) and dried in vacuo to yield colorless
solid (1.15 g, 3.00 mmol, 83% yield). 1H NMR (300 MHz, DMSO-d6) δ
7.32–7.20 (m, 10H), 4.74 (s, 2H), 4.72 (s, 2H). 13C NMR (75 MHz,
DMSO-d6) δ 140.07, 139.96, 128.73, 127.75, 66.88, 66.80. LRMS, ESI:
m/z 277.0 [M − H]−, 555.3 [2M − H]−, 833.7 [3M − H]−.
22.44, 13.20, 7.93. LRMS: m/z 368.0 [M − H]−, 737.3 [2M − H]−
,
1106.7 [3M−H]−. HRMS: m/z calculated for C13H23NO9P 368.1116
[M − H]−, observed 368.1115.
2.3.6.
Triethylammonium
salt
of
hexyl
5-
((hydroxy(phenoxy)phosphoryl)amino)-4-oxopentanoate 6
Phenyl dichlorophosphate (407 mg, 1.93 mmol) was dissolved in
dry THF (20.0 mL) and cooled to 0 °C on ice. Hexyl 5-amino-4-
oxopentanoate (491 mg, 1.95 mmol) was added to the cooled solution
followed by dropwise addition of dry triethylamine (600 μL,
4.31 mmol). The resulting suspension was stirred at 0 °C for 1 h and
then allowed to warm-up to ambient temperature. After 2 h at ambient
temperature the reaction mixture was filtered and the supernatant
quenched with water (5.0 mL) and triethylamine (650 μL, 4.71 mmol).
After 4 h at ambient temperature water (20 mL) and ethylacetate
(40 mL) were added. The organic phase was discarded and the aqueous
phase re-extracted with DCM (2 × 20 mL), washed with brine and dried
with sodium sulfate. The crude product was purified by Flash chroma-
tography using DCM/MeOH (+0.1% TEA) gradient giving colorless oil
(542 mg, 1.15 mmol, 58% yield). 1H NMR (300 MHz, CD3OD) δ 7.34–
7.14 (m, 4H), 7.11–6.96 (m, 1H), 4.02 (t, J = 6.6 Hz, 2H), 3.82 (d, J =
8.3Hz, 2H), 3.18 (q, J = 7.3 H, 6H), 2.73 (dd, J = 7.2, 5.6 Hz, 2H), 2.53
2.3.4.
Hexyl
5-(((((bis(benzyloxy)phosphoryl)oxy)methoxy)
carbonyl)amino)-4-oxopentanoate 4
2.3.4.1. Procedure A. 2 (440 mg, 0.60 mmol) was added to a suspension of
sodium iodide (440 mg, 2.93 mmol) in acetone (20.0 mL) and refluxed
at 60 °C for 2 h. The solvent was evaporated under reduced pressure and
water (10 mL) added. The product was extracted with ether (3 × 20 mL)
and dried over Na2SO4. After the solvent was evaporated colorless oil
was obtained and used immediately in the next step without purifica-
tion. 1H NMR (300 MHz, CDCl3) δ 5.94 (s, 2H), 5.70 (t, J = 4.6 Hz, 1H),

158
A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
(dd, J = 7.2, 5.5 Hz, 2H), 1.67–1.48 (m, 2H), 1.31 (m, 17H), 0.96–0.84
(m, 3H). 13C NMR (75 MHz, CD3OD) δ 208.05 (d), 173.34, 153.34 (d),
129.12, 123.01, 120.66, 120.60, 64.66, 51.41, 48.77, 46.58, 33.94, 31.44,
28.51, 27.64, 25.52, 22.46, 13.30, 8.12. LRMS, ESI: m/z 370.4 [M − H]−,
741.5 [2M − H]−, 1112.7 [3M − H]−, 372.4 [M + H]+, 743.7
[2 M + H]+, 1114.8 [3 M + H]+. HRMS: m/z calculated for
C17H26NO6P 372.1571 [M + H]+, observed 372.1573.
Tecan). The percentage of cell survival was calculated with respect to
controls incubated with the complete medium or a solution of DMSO
(50%) in complete medium, as follows: % cell viability = [A (test-
conc.) − A (100% dead) / A (100% viable) − A (100% dead)] × 100.
2.9. Preparation and treatment of U87MG tumor spheroids
HEMA coated plates were prepared by dissolving Poly-HEMA in eth-
anol. In a laminar flow box, 50 μL was dispensed in each well of a round-
bottom 96-well plate. The 96-well plates were left under the laminar
flow box to evaporate the solvent. After 24 h they were covered with
lids, protected by aluminum foil and kept at room temperature until
the experiments. For spheroid preparation, 200 μL/well of U87MG cell
suspension at the density of 0.5 × 104 cells/mL was dispensed into coat-
ed 96-well plates using a multichannel pipette. In order to obtain opti-
mal three-dimensional structures 3% of Basement Membrane
Matrigel® Matrix (356234, Corning) was added into the cell suspension
prior to seeding. Plates were incubated at 37 °C in a humidified 95% air
and 5% CO2 atmosphere.
2.4. Chemical stability
The stability was assayed at pH 4.00 (acetate buffer), 7.40 (phos-
phate buffer) and 8.90 (borate buffer). The concentration of all buffers
was 25 mM. The compound of interest was dissolved in buffer solution
at 5 mM concentration and pH adjusted if necessary. The vial was placed
in a temperature-controlled LC-MS autosampler preheated at 37 °C and
injected at regular time points. The chromatograms were integrated and
AUC of the compounds was determined and plotted as a function of
time.
2.5. Alkaline phosphatase assay
2.10. Confocal laser-scanning microscopy of spheroids
The compounds (5.0 mM) were dissolved in borate buffer (25 mM)
at pH 8.90 containing MgCl2 (0.5 mM). Alkaline phosphatase from calf
intestine (Applichem-Axon Lab) dissolved in the same buffer was
added and the vial was placed in a temperature-controlled LC-MS
auto-sampler preheated at 37 °C and injected at regular time points.
The chromatograms were integrated and AUC of the starting material
was determined and plotted as a function of time.
On day 2 of spheroid tumor development the spheroids were treated
with 0.33 mM concentration of compounds of interest and 5-ALA-Hex
as positive control. Non-treated spheroids incubated with cell medium
were used as a negative control. Spheroids were kept at 37 °C in humid-
ified 95% air and 5% CO2 atmosphere for 4 h after which they were
washed with cell medium. The non-fixed U87MG spheroids were incu-
bated for 1 h with Hoechest 33342 stain (62249, Thermo Fisher Scientif-
ic) to achieve the staining of nuclei. Prior to imaging, spheroids were
transferred to a glass bottom microwell dishes (35 mm petri dish,
14 mm Microwell, MatTek Corporation). Detection and localization of
PpIX fluorescence in intact spheroids was done with the Zeiss LSM
780 inverted confocal microscope. The z-stack images were taken
with a 405 nm laser using 415–493 nm and 615–659 nm filters, pin-
holes of 27 μm, and a 10× objective. During the image acquisition
time, spheroids were kept at 37 °C in a humidified chamber (INUBTF-
WSKM-F1, Tokai Hit). Obtained z-stack fluorescence images were ana-
lyzed using Zeiss ZEN lite software to produce the maximal intensity
projection image for each sample.
2.6. Cell culture
Human glioblastoma cells U87MG (ATTC® HTB-14™) were grown
in monolayers and maintained in Minimum Essential Media (31095-
029, Thermo Fisher Scientific) supplemented with 10% fetal calf serum
(CVFSVF00-01, Eurobio), 100 μL/mL streptomycin and 100 IU/mL peni-
cillin (15140-122, Thermo Fisher Scientific). Cells were cultivated at 37 °
C in humidified 95% air and 5% CO2 atmosphere and routinely main-
tained by serial passage in a new medium every 5 days.
2.7. PpIX fluorescence time profiles
U87MG cells (10,000 cells/well) were seeded in a 96-well plate
(clear bottom black plate, 3603, Corning). The following day the medi-
um in the wells was replaced with medium containing increasing con-
centrations (0.033, 0.1, 0.33, 1.00, 2.00 and 3.33 mM) of compounds of
interest. 5-ALA-Hex was used as positive control. PpIX fluorescence
was recorded with a plate reader (Safire, Tecan) at different time points
(1, 2, 4, 6, 8 and 24 h). Excitation wavelength was set at 405 nm and
emission wavelength at 635 nm. Cells were maintained at 37 °C for
the duration of the assay.
2.11. Chicken chorioallantoic membrane (CAM) spheroid model
2.11.1. Egg incubation
Fertilized chicken eggs (University animal facility, University of Ge-
neva, Geneva, Switzerland) were placed in the incubator (MG 200,
Savimat, Chauffry, France) and incubated at 37 °C and 65% relative hu-
midity. The eggs were placed with the narrow apex oriented down-
wards. During the first three days, eggs were rotated gently using the
automatic rotation mode of the incubator. On embryonic development
day 3 (EDD3) a small hole (≈3 mm) was made at the narrow apex of
the eggshell and closed with an adhesive tape. Eggs were returned to
the incubator and placed so that the narrow apex was directed upward.
Rotation was no longer necessary after this point until the end of the
experiments.
2.8. Dark toxicity experiments
PSC-ALA-Hex, P-ALA-Hex, and 5-ALA-Hex at increasing concentra-
tions (0.033, 0.10, 0.33, 1.00, 2.00 and 3.33 mM) were incubated in
the dark at 37 °C in humidified 95% air and 5% CO2 atmosphere. After
24 h U87MG cells were washed with DPBS (14190-094, Thermo Fisher
Scientific) and fresh medium was put into each well. They were then
kept in the dark for another 24 h at 37 °C in humidified 95% air and 5%
CO2 atmosphere. Cell viability was measured using a 3-(4,5-dimethyl-
thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (M2128,
Sigma). Cells were washed with DPBS (100 μL) and MTT (50 μL,
0.65 mg/mL) in complete medium was added into each well. Three
hours later, formazan crystals were dissolved by adding DMSO
(100 μL) (D/4121/PB15, Thermo Fisher Scientific) into each well. The
absorption was measured at 570 nm with a plate reader (Safire,
2.11.2. Nodule inoculation into CAM
U87MG spheroids were prepared as described above using
2.5 × 104 cells/mL. Spheroids grew 4 days before inoculation into the
CAM. On EDD7, the hole in the eggshell was enlarged to allow access
to the CAM vasculature, using a needle (25 Gauge) a hole was drilled
in the CAM. A binocular lens Leica M80 (Wetzlar, Germany) was used
to perform the inoculation. Once the hole was pierced in the CAM, the
spheroid was placed in the cavity beneath the hole. Cell medium
(20 μL) was placed on the spheroid to allow the tumor to attach to the

A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
159
CAM membrane and develop. Then, the eggs were sealed with parafilm
3. Results and discussion
and returned to the incubator to let spheroids grow until EDD13.
5-ALA is currently marketed for several applications. In cancer detec-
tion and management its use is sparse with the notable exception of
FGR for glioblastoma. 5-ALA-Hex is a hexyl ester derivative of 5-ALA in-
duces saturation of fluorescence induction in tumor cell lines several-
fold higher compared to 5-ALA in vitro. Its superiority in fluorescence in-
duction at lower doses and shorter time resulted in its approval for top-
ical application in FGR of bladder cancer. However, the toxicity profile of
5-ALA-Hex narrows its therapeutic index meaning it cannot be admin-
istered safely into systemic circulation. Furthermore, 5-ALA and its
hexyl ester suffer from inherent instability coming from the free
amino group which results in the irreversible dimerization rendering
them inactive (Fig. 2). To our knowledge, no 5-ALA derivative modified
at the amino group showed robust fluorescence profile in vitro or in vivo
[43]. The reason for the biological inactivity of the amino-modified de-
rivatives is the high substrate specificity of porphobilinogen synthase
coming from its reaction mechanism through the formation of Schiff
base linkage between free amino group of 5-ALA and the lysine residues
in the enzyme's active site [44].
We set out to investigate whether it would be possible to solve this
5-ALA paradox, meaning the modification of the 5-amino end of the
molecule that will render the molecule stable and impede the dimeriza-
tion degradation but not at the expense of biological activity. To this end
the protecting group attached to the 5-amino group needs to be very ef-
ficiently cleaved off either chemically or enzymatically in vitro and in
vivo. We present 2 new classes of amino modified 5-ALA prodrugs
that both tackle this problem (Fig. 3).
The first class contains 5-ALA and a phosphate group connected via a
short chemical linker that is designed to yield 5-ALA or its ester upon
the action of alkaline phosphatase. After the enzymatic cleavage the
subsequent hydrolysis of the self-immolative linker [45] is spontaneous.
This class was therefore named phospho-self-immolative 5-ALA (PSI-
ALA). The second class, called phospho-5-ALA (P-ALA), is also activated
by alkaline phosphatase but gives 5-ALA directly after enzymatic action
without the use of a self-immolative spacer. Both classes were designed
to display improved stability since the amino group is protected and at
the same time reduce the acute toxicity of 5-ALA-Hex which can at least
partially be explained by the free positively charged amino group. The
differences in amine-protecting groups were also expected to give dif-
ferent sensitivity towards the target enzyme. We decided to synthesize
the hexyl esters of both new classes because the 5-ALA-Hex demon-
strates vastly superior profile in fluorescence induction as compared
to 5-ALA. The two classes are hence represented by hexyl ester
2.11.3. Fluorescence imaging of CAM tumors
On EDD13, the compounds were injected intravenously into the
CAM vasculature. The hole enlarged on EDD7 was further widened if
necessary to allow more access to vessels and to facilitate handling. In-
jection (10–40 μL) was performed preferentially into a large blood ves-
sels using a needle (gauge 33, 51 mm, tap N type) attached to a 100 μL
syringe (Hamilton, Reno, NV). Fluorescence imaging of tumor nodules
was done with 12-bit monochrome CDD camera (Retiga EX Q-Imaging
Canada) connected to fluorescence Eclipse E600 FN microscope with a
CFI achromat objective characterized by magnification 4×, numerical
aperture of 0.10 and a working distance of 30 mm (Nikon, Tokyo,
Japan). A conical holder was used to place the eggs during experimenta-
tion under the objective of fluorescence microscope. A mercury arc
lampHBO 103/W/2 (MS Scientific Berlin, Germany) was used to provide
illumination. For the detection of PpIX a BV-2A cube (Nikon, Tokyo,
Japan) with an excitation filter: 400–440 nm, a dichroic mirror
(455 nm) and an emission filter (470 nm) was used. A hollow slider fil-
ter 650/50 nm was added. Band pass filter 560/40 was used to image the
autofluorescence of tumors. All pictures were taken with a 90 ms expo-
sure and gain set at 50. Tumor imaging was performed at time 0, 30, 60,
120 and 300 min after injection. Autofluorescence images were taken
before injection. Images were processed using Openlab software version
3.1.5. Image resolution was 678 × 518 pixels with a binning set at 2 and
the brightness and contrast for all images. The surface-averaged fluores-
cence signal of an area of interest (AOI) was determined within the
tumor nodules and outside of the tumors. Data was analyzed using mul-
tiple t-tests and statistical significance determined at each time point.
2.11.4. Chick embryo acute toxicity
The experiments were performed according to published proce-
dures and standard animal treatment practices [42]. Escalating doses
of compounds dissolved in milliQ water and pH adjusted to 7.4 were ad-
ministrated to chick embryos on EDD13 via intravenous injection. The
viability of chick embryos has been evaluated after 24 h after injection.
2.11.5. Whole blood stability assay
Whole blood was taken from chick embryos on EDD13 using an in-
sulin 500 μL syringe and 25 gauge needle (0.5 × 16 mm). Whole blood
(300 μL) was collected from each embryo and the anticoagulant and
preservative solution added (45 μL) according to Ph. Eur. 8.8. Briefly,
the anticoagulant solution was prepared by dissolving sodium citrate
(2.20 g), citric acid monohydrate (0.80 g) and glucose monohydrate
(2.45 g) in water for injections (100 mL). The citrated whole blood
was stored at 4–6 °C and used within 72 h of collection.
Compounds PSI-ALA-Hex (5) and P-ALA-Hex (6) were dissolved in
whole blood giving the final concentration of 10 mM. The vial was
placed in a temperature-controlled LC-MS auto-sampler preheated at
37 °C and aliquots of plasma were injected (2 μL) at regular time points.
The full gradient (ammonium formate/MeOH) chromatograms of were
integrated and AUC of the compounds were determined and plotted as a
function of time.
2.12. Statistical analysis
GraphPad Prism 6.0.1 for Windows (GraphPad Software, version
6.05, San Diego, CA, USA) was used for biological data analysis and treat-
ment. All results are presented as means standard deviations. Where
applicable one-way ANOVA or multiple t-tests was used to analyse the
results. P values b0.05 were considered as statistically significant.
Fig. 3. Bioactivation mechanism of phosphatase-sensitive prodrugs of ALA. Alkaline
phosphatase enzymatic cleavage of phosphate group is followed by spontaneous
conversion of ALA-hemiacetal into ALA (PSI-ALA, left) or affords ALA ester directly (P-
ALA, right).

160
A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
expressed hydrolases can also be extended to other enzymes such as
glycosidases, nucleases, oxyreductases, and ureases.
3.1. Synthesis
The synthesis of PSI-ALA class of compounds was done in four or five
synthetic steps depicted in Fig. 4. 5-ALA was used as starting material
and was firstly esterified by 1-hexanol to give 2 in excellent yield. The
self-cleavable linker between the phosphate group and 5-ALA moiety
was introduced with chloromethyl chloroformate and triethylamine
as base at −20 °C to keep the dimerization side reaction to a minimum.
3 was obtained in very good yield. The introduction of the phosphate
group was attempted in two steps. The Finkelstein reaction using an ex-
cess of sodium iodide in refluxing acetone gave firstly more electrophilic
iodide derivative. It was used in the following step without purification
as it proved to be poorly stable and degrading even stored at −30 °C
under argon. The following step gave the protected phosphate deriva-
tive 5 by reacting 4 with silver salt of dibenzyl phosphate at room tem-
perature in moderate yield for the 2 steps (42%). The low yield was
presumably due to the degradation of the poorly stable
iodomethyloxycarbonyl derivative. We decided to attempt a direct con-
version of 2 to protected phosphate 4 by reacting the same chloro deriv-
ative 2 with the dibenzyl phosphate silver salt 3 by refluxing in toluene
at 60 °C which gave a much better 72% yield. Hydrogenolysis using Pd/C
as catalyst in absolute ethanol and excess of TEA gave the final product 5
(PSI-ALA-Hex) in quantitative yield. It is noteworthy that the presence
of base is essential for the outcome of the deprotection step. The self-
immolative linker is acid-sensitive and is spontaneously cleaved under
these reaction conditions in the absence of triethylamine as base.
Contrary to the PSI-ALA-Hex, the P-ALA-Hex was synthesized in a
straightforward 2-step conversion from 5-ALA. The esterification step
yielding 2 was the same as described above for PSI-ALA-Hex. Reacting
2 with 1 eq of phenyl dichlorophosphate at 0 °C to increase the yield
by reducing the dimerization rate of 5-ALA-Hex under basic conditions
Fig. 4. Synthesis of phosphatase-sensitive ALA prodrugs. PSI-ALA-Hex 5 contains a self-
immolative linker whilst P-ALA-Hex 6 does not contain a chemical linker.
derivatives compound 5 (PSI-ALA-Hex) and compound 6 (P-ALA-Hex)
and compared to 5-ALA-Hex. It did not escape our notice that the con-
cept of 5-ALA derivatives that can be activated by ubiquitously
Fig. 5. Different stability profiles of 5-ALA-Hex (top), PSI-ALA-Hex, (bottom left), and P-ALA-Hex (bottom right) at 37 °C at different pH values: pH 4.0 (circles), 7.4 (squares), and 8.9
(triangles). Each time point represents the mean SD (n = 3).

A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
161
Fig. 6. Structure-controlled activation of phosphatase-sensitive prodrugs in vitro. PSI-ALA-Hex (left) in the presence of 1 U alkaline phosphatase at pH 8.90 of is cleaved very rapidly. P-ALA-
Hex (right) displays a much slower release profile in the presence of 100 U of alkaline phosphatase. Each time point represents the mean SD (n = 3).
gave the final product 6 as TEA salt after extraction and chromatograph-
ic purification.
class of enzymes [47]. This finding led to a 5-ALA prodrug class that is
activated much more slowly. Therefore, fine-tuning the structure of
phosphatase-sensitive 5-ALA prodrugs, allows for the modulation of
5-ALA release.
3.2. Chemical stability
Both classes were designed to improve the chemical stability profile
of 5-ALA and 5-ALA-Hex. Fig. 5 (top) shows that the stability profiles at
36 °C the 5 mM solution of 5-ALA-Hex in different buffers displays pH
dependent stability curves. At pH 4 the molecule is stable for prolonged
periods of time because the amino group is fully protonated (pKa = 8.9
at 25 °C). However, at pH 7.40 a significant increase in deprotonation of
the amino group facilitates fast and irreversible dimerization. The mol-
ecule has a half-life of 180 min under accelerated conditions at 37 °C.
This phenomenon is even more pronounced at pH 8.9. Contrary, PSI-
ALA-Hex 5 shows much improved stability. At neutral and basic pH
the degradation rate is very slow and the stability is even better at
pH = 4.00. The P-ALA-Hex 6 showed excellent stability similar to PSI-
ALA-Hex. At pH 4.00, 7.40 and 8.90 the molecule stays intact for hours
compared to very fast degradation of 5-ALA-Hex which occurs in mi-
nutes. The stability profiles are vastly superior to that of 5-ALA-Hex as
is expected from the protection of the amino group. These observation
are also in agreement with reports on the stability of 5-ALA itself [31,
46].
3.4. In vitro PpIX production
We assayed the compounds of interest for in vitro production of PpIX
in U87MG glioblastoma cell line. This phenotypic assay measured the
final fluorescence intensity of PpIX and not each individual step of the
prodrug activation and conversion. The compounds were incubated
with cells at different concentrations and compared to 5-ALA-Hex. 5-
ALA itself induces PpIX up to 100–200 times less efficiently than 5-
ALA-Hex in certain cancer cell lines and was therefore not used as con-
trol our assays [48,49]. Fig. 7 presents the PpIX fluorescence intensity
time profile in U87MG cells at optimal concentration. It is evident that
the PSI-ALA-Hex performs better than 5-ALA-Hex over 24 h. Fig. S1
shows that PSI-ALA-Hex curves at different concentrations are
superimposed which means that the PpIX production is saturated
under our experimental conditions. Despite this and to our surprise,
the 5-ALA-Hex induced fluorescence intensity is lower than that of
PSI-ALA-Hex and P-ALA-Hex for long incubation periods. It was rea-
soned that this is due to the instability of 5-ALA-Hex under physiological
conditions and/or limited dark toxicity at 0.33 mM as depicted in Fig. S2.
In spite of the lower rate of in vitro enzymatic activation P-ALA-Hex
also performed well in this assay. At 0.33 mM fluorescence intensity in-
creases linearly over the 24 h of the assay. Similarly to PSI-ALA-Hex, this
molecule also gave higher fluorescence intensities than 5-ALA-Hex at
every time point, the difference becoming statistically significant at
24 h for both molecules compared to 5-ALA-Hex.
3.3. Enzymatic activation
The increased chemical stability displayed by both new ALA chemi-
cal classes is a consequence of the 5-amino group modification. Howev-
er, the increase of stability might also cause a dramatic decrease in PpIX
production as shown by amino-acid protected 5-ALA derivatives [34].
Therefore, we first tested 5-ALA-Hex release from PSI-ALA-Hex and P-
ALA-Hex in the presence of alkaline phosphatase. PSI-ALA-Hex which
contains the free phosphate group was found to be an excellent sub-
strate for the target enzyme. 1 U of alkaline phosphatase resulted in al-
most full activation and conversion to 5-ALA-Hex in 60 min (Fig. 6, left).
The conversion was not complete presumably because of the experi-
mental conditions. PSI-ALA-Hex was enzymatically converted to 5-
ALA-Hex which in turn formed dimerized dihexyl-2,5-(β-carboxyethyl)
dihydropyrazine very rapidly at pH 8.90. This product is water-insoluble
and precipitates out of solution resulting in enzyme inhibition, causing
the curve to flatten out after 80% conversion. It is noteworthy that the
activation of PSI-ALA-Hex yields equimolar amounts of formaldehyde
which is potentially cytotoxic. This could explain the higher in vitro
dark toxicity of PSI-ALA-Hex compared to P-ALA-Hex at higher concen-
trations (2.00 and 3.33 mM) as depicted in Fig. S2.
P-ALA-Hex, performed quite differently when subjected to enzymat-
ic activation. The conversion was achieved much more slowly in the
presence of 100 U of alkaline phosphatase (Fig. 6, right). It was actually
surprising that P-ALA-Hex is a substrate for alkaline phosphatase at all
but this can be explained by the broad substrate specificity of this
Fig. 7. In vitro PpIX fluorescence kinetics in U87MG cells. PSI-ALA-Hex (black triangles)
and P-ALA-Hex (dark gray squares) outperform 5-ALA-Hex (light gray circles) at
optimal concentration 0.33 mM. * asterisk indicates statistically significant difference in
fluorescence intensity at 24 h between PSI-ALA-Hex or P-ALA-Hex and 5-ALA-Hex, 2-
way ANOVA, p b 0.05.

162
A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
not shown) and that the corresponding methyl esters did not show
any clinically relevant fluorescence levels in vitro. However, our main
focus was on 5-ALA-Hex analogues since 5-ALA-Hex is clinically
approved.
3.5. PpIX production in full U87MG tumor spheroids
The phenotypic fluorescence assay in cell monolayers described
above was used as medium-throughput screen for new 5-ALA prodrugs.
Despite its robustness, it has very limited ability to predict penetration
across several layers of cells typically found in neoplasms. We therefore
performed live U87MG spheroid imaging to determine if the new
prodrugs penetrate into the core of the 400–500 μm glioblastoma
microspheroids (Fig. 8). The fluorescence was visible throughout the
entire spheroids (Fig. S4) and was more pronounced than the fluores-
cence produced by 5-ALA-Hex administration. This was especially the
case for the P-ALA-Hex, which is lipophilic and negatively charged. It
is noteworthy that Hoechst 33342 stained only the cell nuclei of the sur-
face cells proving evidence for the difficulty of passage through several
layers of cells.
3.6. In vivo chick acute toxicity
Encouraged by the robust in vitro fluorescence profiles we assessed
the acute in vivo toxicity of our compounds in living chick embryos
[50]. 5-ALA-Hex is known for its pronounced acute toxicity after intra-
venous injection and we determined the LD50 of 75 μmol/kg. PSI-ALA-
Hex was better tolerated and LD50 values of 200 μmol/kg were
achieved. We postulated that these higher LD50 is a consequence of
rapid conversion into 5-ALA-Hex. Therefore, the very high doses of P-
ALA-Hex that were tolerated by chick embryos came as no surprise.
Since this compound is much more slowly converted into 5-ALA-Hex
(and 5-ALA) in all tissues expressing alkaline phosphatase, the LD50
was 500 μmol/kg which is 5-fold increase in tolerated doses compared
to 5-ALA-Hex. The lower toxicity of P-ALA-Hex could also be explained
by the lower rate of activation in whole blood as compared to PSI-ALA-
Hex.
Fig. 8. Confocal microscopy fluorescence images of U87MG tumor spheroids after
incubation with PSI-ALA-Hex 5 (top left), P-ALA-Hex 6 (top right), 5-ALA-Hex as
positive control (bottom left) and cell medium negative control (bottom right). PpIX
fluorescence is represented in red and nuclei staining with Hoechst 33342 in blue. Scale
bar 100 μm.
Generally speaking, the optimal range of concentrations of both
prodrugs for fluorescence induction were comparable to those required
for 5-ALA-Hex with best results obtained in the range of 0.03–0.33 mM
(Fig. S1). P-ALA-Hex, however, required higher concentrations used
which corresponded to the phosphatase activation profile which was
significantly slower than that of PSI-ALA-Hex. It should be noted that a
larger series of PSI-ALA and P-ALA esters has been synthesized (data
Fig. 9. In vivo fluorescence of CAM U87MG spheroid tumor. Injection of non-fluorescent PSI-ALA-Hex (100 μmol/kg, top), P-ALA-Hex (300 μmol/kg, middle) and fluorescent PpIX
(100 μmol/kg, bottom). Tumor auto-fluorescence images at time 0 (left) and fluorescence images at times 0, 30, 60 and 120 min.

A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
163
3.7. In vivo CAM U87MG spheroid tumor model
Statistically significant fluorescence was observed throughout the
tumor mass within the clinically acceptable time-frame. This work
opens the doors towards new ALA prodrug based FGR with enormous
translational potential for fluorescence-based detection of tumors.
From a clinical perspective, we believe that the faster onset of PSI-
ALA-Hex induced fluorescence in vivo would be an advantage as it
would reduce the time required between the injection and the clinical
intervention. On the other hand, slower-releasing 5-ALA prodrugs
might prove to be a better solution for a long-circulating formulations.
All in all, these proof-of-principle studies of stable 5-ALA derivatives
open the way towards self-assembling 5-ALA nanomicelles.
Using the chorioallantoic membrane (CAM) model U87MG tumor
spheroids were obtained using a modified procedure [51,52]. Spheroids
approximately of 200 μm were implanted and microtumors could be
observed on the CAM several days later. Compounds were injected
and the fluorescence in micro tumors was imaged (Fig. 9). In accordance
with in vitro data, PSI-ALA-Hex started to outline the tumors as early as
30 min post injection followed by an excellent and statistically signifi-
cant signal to background at 1 h and also at 2 h (Figs. 9 and 10). As ex-
pected, P-ALA-Hex displayed a delayed onset of the fluorescence signal
but robust contrast nevertheless.
The fluorescence time profiles of P-ALA-Hex were not fully in accor-
dance with the in vitro enzymatic assays data, so the whole blood stabil-
ity assay was also performed. It showed that both compounds get
slowly metabolized in blood. PSI-ALA-Hex was still a faster ALA-releas-
ing compound, however, the difference in activation rate was much
lower than the one in vitro (Fig. S3). Other enzymes, such as
phosphoamidases and other phosphatases could play a role in the acti-
vation of P-ALA-Hex, thereby reducing the difference in effective activa-
tion rates between the compounds.
The fluorescence decreased to background levels in chick embryos in
5 h which is a clinically relevant advantage (Fig. S5). The kinetics of the
fluorescence signal were very important here as this gave an indication
to prodrug activation, PpIX biosynthesis and final conversion to non-
fluorescent heme or clearance out of the tumors. Interestingly, PpIX
that was injected as one of the controls, to demonstrate that the PpIX
was produced within the tumors and not in other developing tissues
in the embryos, showed no preferential accumulation and fluorescence
in the tumors (Fig. 9).
Acknowledgements
We thank the Mass spectrometry center of the University of Geneva
for the high-resolution mass spectroscopy experiments. We would also
like to acknowledge the support of the NMR facility of School of Phar-
maceutical Sciences, and the Chemistry NMR facility at the University
of Geneva, Switzerland. This work was supported, by grants from the
Swiss National Science Foundation (310000-109402, CR32I3_129987,
205320_138309; 205321_126834, CR32I3_147018) and the Swiss Gov-
ernment Excellence Scholarships for Foreign Scholars (2013.0082).
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jconrel.2016.05.047.
References
4. Conclusions
[1] Q. Peng, T. Warloe, K. Berg, J. Moan, M. Kongshaug, K.E. Giercksky, J.M. Nesland, 5-
Aminolevulinic acid-based photodynamic therapy — clinical research and future
challenges, Cancer 79 (1997) 2282–2308.
[2] N. Lange, P. Jichlinski, M. Zellweger, M. Forrer, A. Marti, L. Guillou, P. Kucera, G.
Wagnieres, H. van den Bergh, Photodetection of early human bladder cancer
based on the fluorescence of 5-aminolaevulinic acid hexylester-induced protopor-
phyrin IX: a pilot study, Brit. J. Cancer 80 (1999) 185–193.
[3] P. Hillemanns, H. Weingandt, R. Baumgartner, J. Diebold, W. Xiang, H. Stepp,
Photodetection of cervical intraepithelial neoplasia using 5-aminolevulinic acid-in-
duced porphyrin fluorescence, Cancer 88 (2000) 2275–2282.
FPD and FGR remain perhaps the most promising upcoming modal-
ities in tumor management and therapy. 5-ALA and 5-ALA-Hex have
been at the forefront of small molecules used for FGR. However their
use is limited due to unfavorable stability, pharmacokinetics and toxic-
ity profiles. We have presented 2 new 5-ALA prodrugs: PSI-ALA-Hex
and P-ALA-Hex that exhibit clearly improved chemical stability at acidic,
neutral and basic pH values. They also show designed and structure-de-
pendent sensitivity to alkaline phosphatase. The robust production of
PpIX and red fluorescence in U87MG glioblastoma cancer cell lines is
better than the state of the art fluorescence levels of 5-ALA-Hex. Acute
toxicity was reduced several fold for the PSI-ALA-Hex whilst P-ALA-
Hex shows unprecedented 7-fold increase in LD50 in chick embryos.
Clinically translational fluorescence profiles in CAM implanted U87MG
tumor spheroids were observed after the injection of both compounds.
[4] Q.T. Nguyen, R.Y. Tsien, Fluorescence-guided surgery with live molecular navigation
— a new cutting edge, Nat. Rev. Cancer 13 (2013) 653–662.
[5] M.J. Colditz, R.L. Jeffree, Aminolevulinic acid (ALA)-protoporphyrin IX fluorescence
guided tumour resection. Part 1: clinical, radiological and pathological studies, J.
Clin. Neurosci. 19 (2012) 1471–1474.
[6] T. Ohmura, T. Fukushima, H. Shibaguchi, S. Yoshizawa, T. Inoue, M. Kuroki, K. Sasaki,
S.I. Umemura, Sonodynamic therapy with 5-aminolevulinic acid and focused ultra-
sound for deep-seated intracranial glioma in rat, Anticancer Res. 31 (2011)
2527–2533.
[7] H.J. Chen, X.B. Zhou, Y. Gao, B.Y. Zheng, F.X. Tang, J.D. Huang, Recent progress in de-
velopment of new sonosensitizers for sonodynamic cancer therapy, Drug Discov.
Today 19 (2014) 502–509.
[8] B. Nokes, M. Apel, C. Jones, G. Brown, J.E. Lang, Aminolevulinic acid (ALA): photody-
namic detection and potential therapeutic applications, J. Surg. Res. 181 (2013)
262–271.
[9] D.E.J.G.J. Dolmans, D. Fukumura, R.K. Jain, Photodynamic therapy for cancer, Nat.
Rev. Cancer 3 (2003) 380–387.
[10] C. Chi, Y. Du, J. Ye, D. Kou, J. Qiu, J. Wang, J. Tian, X. Chen, Intraoperative imaging-
guided cancer surgery: from current fluorescence molecular imaging methods to fu-
ture multi-modality imaging technology, Theranostics 4 (2014) 1072–1084.
[11] A. Omuro, L.M. DeAngelis, Glioblastoma and other malignant gliomas a clinical re-
view, JAMA 310 (2013) 1842–1850.
[12] J.T. Huse, E.C. Holland, Targeting brain cancer: advances in the molecular pathology
of malignant glioma and medulloblastoma, Nat. Rev. Cancer 10 (2010) 319–331.
[13] W. Stummer, U. Pichlmeier, T. Meinel, O.D. Wiestler, F. Zanella, H.J. Reulen, A.L.-G.S.
Group, Fluorescence-guided surgery with 5-aminolevulinic acid for resection of ma-
lignant glioma: a randomised controlled multicentre phase III trial, Lancet Oncol. 7
(2006) 392–401.
[14] W. Stummer, H.J. Reulen, T. Meinel, U. Pichlmeier, W. Schumacher, J.C. Tonn, V.
Rohde, F. Oppel, B. Turowski, C. Woiciechowsky, K. Franz, T. Pietsch, A.L.-G.S.
Group, Extent of resection and survival in glioblastoma multiforme: identification
of and adjustment for bias, Neurosurgery 62 (2008) 564–576 (discussion 564-576).
[15] M. Hefti, H.M. Mehdorn, I. Albert, L. Dorner, Fluorescence-guided surgery for malig-
nant glioma: a review on aminolevulinic acid induced protoporphyrin IX photody-
namic diagnostic in brain tumors, Curr. Med. Imaging Rev. 6 (2010) 254–258.
Fig. 10. In vivo CAM U87MG spheroid tumor fluorescence time profiles kinetics (black
circles) as compared to the surrounding chorioallantoic membrane (gray squares).
Injection of non-fluorescent PSI-ALA-Hex (100 μmol/kg) produces PpIX fluorescence;
n = 3, **p b 0.01, ***p b 0.001.

164
A. Babič et al. / Journal of Controlled Release 235 (2016) 155–164
[16] J.T. Dalton, C.R. Yates, D.H. Yin, A. Straughn, S.L. Marcus, A.L. Golub, M.C. Meyer, Clin-
ical pharmacokinetics of 5-aminolevulinic acid in healthy volunteers and patients at
high risk for recurrent bladder cancer, J. Pharmacol. Exp. Ther. 301 (2002) 507–512.
[17] C. Perotti, A. Casas, H. Fukuda, P. Sacca, A. Batlle, ALA and ALA hexyl ester induction
of porphyrins after their systemic administration to tumour bearing mice, Brit. J.
Cancer 87 (2002) 790–795.
[18] L. Rodriguez, A. Batlle, G. Di Venosa, A.J. MacRobert, S. Battah, H. Daniel, A. Casas,
Study of the mechanisms of uptake of 5-aminolevulinic acid derivatives by PEPT1
and PEPT2 transporters as a tool to improve photodynamic therapy of tumours,
Int. J. Biochem. Cell Biol. 38 (2006) 1530–1539.
[19] N. Fotinos, M.A. Campo, F. Popowycz, R. Gurny, N. Lange, 5-Aminolevulinic acid de-
rivatives in photomedicine: characteristics, application and perspectives,
Photochem. Photobiol. 82 (2006) 994–1015.
[20] J.Y. Fang, P.C. Wu, C.L. Fang, C.H. Chen, Intravesical delivery of 5-aminolevulinic acid
from water-in-oil nano/submicron-emulsion systems, J. Pharm. Sci. US 99 (2010)
2375–2385.
protoporphyrin IX production by direct determination of cellular prodrug uptake
and prodrug metabolization, J. Med. Chem. 52 (2009) 4026–4037.
[35] J.L. Millan, W.H. Fishman, Biology of human alkaline-phosphatases with special ref-
erence Ts cancer, Crit. Rev. Clin. Lab. Sci. 32 (1995) 1–39.
[36] J. Bukowczan, S. Pattman, F. Jenkinson, R. Quinton, Regan isoenzyme of alkaline
phosphatase as a tumour marker for renal cell carcinoma, Ann. Clin. Biochem. 51
(2014) 611–614.
[37] A. Chaux, E. Comperat, J. Varinot, J. Hicks, K. Lecksell, J. Solus, G.J. Netto, High levels of
phosphatase and Tensin homolog expression are associated with tumor progres-
sion, tumor recurrence, and systemic metastases in pT1 urothelial carcinoma of
the bladder: a tissue microarray study of 156 patients treated by transurethral re-
section, Urology 81 (2013) 116–122.
[38] A.A.N.P.M. Dabare, A.M.E. Nouri, H. Cannell, T. Moss, A.K. Nigam, R.T.D. Oliver, Profile
of placental alkaline phosphatase expression in human malignancies: effect of tu-
mour cell activation on alkaline phosphatase expression, Urol. Int. 63 (1999)
168–174.
[21] Z. Li, L.L. Pan, F.L. Zhang, X.L. Zhu, Y. Liu, Z.Z. Zhang, 5-aminolevulinic acid-loaded ful-
lerene nanoparticles for in vitro and in vivo photodynamic therapy, Photochem.
Photobiol. 90 (2014) 1144–1149.
[22] X. Ma, Q.Y. Qu, Y.L. Zhao, Targeted delivery of 5-aminolevulinic acid by multifunc-
tional hollow mesoporous silica nanoparticles for photodynamic skin cancer thera-
py, ACS Appl. Mater. Interfaces 7 (2015) 10671–10676.
[39] E.Y. Yu, F. Saad, A. Londhe, N.D. Shore, H. Van Poppel, D.E. Rathkopf, M.R. Smith, C.
Logothetis, P.L. De Souza, K. Fizazi, P.F.A. Mulders, P.N. Mainwaring, J.D.
Hainsworth, T.M. Beer, S.A. North, E.J. Small, H.I. Scher, T.W. Griffin, M.K. Yu, C.J.
Ryan, Association of alkaline phosphatase (ALP) with clinical outcomes in chemo-
therapy-naive patients (pts) with metastatic castration-resistant prostate cancer
(mCRPC): results from COU-AA-302, J. Clin. Oncol. 32 (2014).
[23] L. Shi, X.L. Wang, F. Zhao, H.S. Luan, Q.F. Tu, Z. Huang, H. Wang, H.W. Wang, In vitro
evaluation of 5-aminolevulinic acid (ALA) loaded PLGA nanoparticles, Int. J.
Nanomedicine 8 (2013) 2669–2676.
[40] A. Casas, A.M. Batlle, A.R. Butler, D. Robertson, E.H. Brown, A. MacRobert, P.A. Riley,
Comparative effect of ALA derivatives on protoporphyrin IX production in human
and rat skin organ cultures, Br. J. Cancer 80 (1999) 1525–1532.
[24] C.W. Chung, K.D. Chung, Y.I. Jeong, D.H. Kang, 5-Aminolevulinic acid-incorporated
nanoparticles of methoxy poly(ethylene glycol)-chitosan copolymer for photody-
namic therapy, Int. J. Nanomedicine 8 (2013) 809–819.
[25] M.K.K. Oo, X. Yang, H. Du, H. Wang, 5-Aminolevulinic acid-conjugated gold nano-
particles for photodynamic therapy of cancer, Nanomedicine UK 3 (2008) 777–786.
[26] H. Xu, C. Liu, J.S. Mei, C.P. Yao, S.J. Wang, J. Wang, Z. Li, Z.X. Zhang, Effects of light ir-
radiation upon photodynamic therapy based on 5-aminolevulinic acid-gold nano-
particle conjugates in K562 cells via singlet oxygen generation, Int. J.
Nanomedicine 7 (2012) 5029–5038.
[41] J.C. Sheehan, V.S. Frank, Peptide syntheses using energy-rich phosphorylated amino
acid derivatives, J. Am. Chem. Soc. 72 (1950) 1312–1316.
[42] C.S. Kue, K.Y. Tan, M.L. Lam, H.B. Lee, Chick embryo chorioallantoic membrane
(CAM): an alternative predictive model in acute toxicological studies for anti-cancer
drugs, Exp. Anim. Tokyo 64 (2015) 129–138.
[43] J. Kloek, G.M.J.B. van Henegouwen, Prodrugs of 5-aminolevulinic acid for photody-
namic therapy, Photochem. Photobiol. 64 (1996) 994–1000.
[44] E.K. Jaffe, The porphobilinogen synthase catalyzed reaction mechanism, Bioorg.
Chem. 32 (2004) 316–325.
[27] S.H. Battah, C.E. Chee, H. Nakanishi, S. Gerscher, A.J. MacRobert, C. Edwards, Synthe-
sis and biological studies of 5-aminolevulinic acid-containing dendrimers for photo-
dynamic therapy, Bioconjug. Chem. 12 (2001) 980–988.
[45] A. Alouane, R. Labruere, T. Le Saux, F. Schmidt, L. Jullien, Self-immolative spacers: ki-
netic aspects, structure-property relationships, and applications, Angew. Chem. Int.
Ed. 54 (2015) 7492–7509.
[28] A. Casas, S. Battah, G. Di Venosa, P. Dobbin, L. Rodriguez, H. Fukuda, A. Batlle, A.J.
MacRobert, Sustained and efficient porphyrin generation in vivo using dendrimer
conjugates of 5-ALA for photodynamic therapy, J. Control. Release 135 (2009)
136–143.
[29] H.F. L'Eplattenier, B. Klem, E. Teske, F.J. van Sluijs, S.A. van Nimwegen, J. Kirpensteijn,
Preliminary results of intraoperative photodynamic therapy with 5-aminolevulinic
acid in dogs with prostate carcinoma, Vet. J. 178 (2008) 202–207.
[30] B. Elfsson, I. Wallin, S. Eksborg, K. Rudaeus, A.M. Ros, H. Ehrsson, Stability of 5-
aminolevulinic acid in aqueous solution, Eur. J. Pharm. Sci. 7 (1999) 87–91.
[31] M. Novo, G. Huttmann, H. Diddens, Chemical instability of 5-aminolevulinic acid
used in the fluorescence diagnosis of bladder tumours, J. Photochem. Photobiol. B
34 (1996) 143–148.
[32] Y. Berger, A. Greppi, O. Siri, R. Neier, L. Juillerat-Jeanneret, Ethylene glycol and amino
acid derivatives of 5-aminolevulinic acid as new photosensitizing precursors of pro-
toporphyrin IX in cells, J. Med. Chem. 43 (2000) 4738–4746.
[46] O.B. Gadmar, J. Moan, E. Scheie, L.W. Ma, Q. Peng, The stability of 5-aminolevulinic
acid in solution, J. Photochem. Photobiol. B 67 (2002) 187–193.
[47] J.L. Millan, Alkaline phosphatases: structure, substrate specificity and functional re-
latedness to other members of a large superfamily of enzymes, Purinergic Signal 2
(2006) 335–341.
[48] Z. Luksiene, I. Eggen, J. Moan, J.M. Nesland, Q. Peng, Evaluation of protoporphyrin IX
production, phototoxicity and cell death pathway induced by hexylester of 5-
aminolevulinic acid in Reh and HPB-ALL cells, Cancer Lett. 169 (2001) 33–39.
[49] E.S. Chu, T.K. Wong, C.M. Yow, Photodynamic effect in medulloblastoma: downreg-
ulation of matrix metalloproteinases and human telomerase reverse transcriptase
expressions, Photochem. Photobiol. Sci. 7 (2008) 76–83.
[50] H. Nishigori, M. Mizumura, M. Iwatsuru, The hen's fertile egg screening test (HEST):
a comparison between the acute toxicity for chick embryos and rodents of 20 drugs,
Cell Biol. Toxicol. 8 (1992) 255–265.
[51] N. De Magalhaes, L.H. Liaw, M. Berns, V. Cristini, Z. Chen, D. Stupack, J. Lowengrub,
Applications of a new in vivo tumor spheroid based shell-less chorioallantoic mem-
brane 3-D model in bioengineering research, J. Biomed. Sci. Eng. 3 (2010) 20–26.
[52] A. Ivascu, M. Kubbies, Rapid generation of single-tumor spheroids for high-through-
put cell function and toxicity analysis, J. Biomol. Screen. 11 (2006) 922–932.
[33] Y. Berger, L. Ingrassia, R. Neier, L. Juillerat-Jeanneret, Evaluation of dipeptide-deriv-
atives of 5-aminolevulinic acid as precursors for photosensitizers in photodynamic
therapy, Bioorg. Med. Chem. 11 (2003) 1343–1351.
[34] F. Giuntini, L. Bourre, A.J. MacRobert, M. Wilson, I.M. Eggleston, Improved peptide
prodrugs of 5-ALA for PDT: rationalization of cellular accumulation and