Synthesis and radiopharmacological characterisation of a fluorine-18-labelled azadipeptide nitrile as a potential PET tracer for invivo imaging of cysteine cathepsins

CHEMMEDCHEM

FULL PAPERS

Visualising cathepsins in tumours: Cys-

teine cathepsins are key players in

tumour pathology. An azadipeptide ni-

trile with high affinity for cathepsins L,

S, B, and K was labelled with fluorine-18

and investigated for its pharmacokinetic

properties. PET imaging studies with

tumour-bearing mice indicate the

tumour accumulation of the probe and

the potential of tumour targeting for

this inhibitor class.

R. Lçser,* R. Bergmann, M. Frizler,

B. Mosch, L. Dombrowski, M. Kuchar,

J. Steinbach, M. Gꢀtschow, J. Pietzsch

&& – &&

Synthesis and Radiopharmacological

Characterisation of a Fluorine-18-

Labelled Azadipeptide Nitrile as

a Potential PET Tracer for in vivo

Imaging of Cysteine Cathepsins

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

1

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

DOI: 10.1002/cmdc.201300135

Synthesis and Radiopharmacological Characterisation of

a Fluorine-18-Labelled Azadipeptide Nitrile as a Potential

PET Tracer for in vivo Imaging of Cysteine Cathepsins

Reik Lçser,*^{[a, c]}Ralf Bergmann,^[a]Maxim Frizler,^[b]Birgit Mosch,^[a]Lilli Dombrowski,^[a]

Manuela Kuchar,^{[a, c]}Jçrg Steinbach,^{[a, c]}Michael Gꢁtschow,^[b]and Jens Pietzsch^{[a, c]}

A fluorinated cathepsin inhibitor based on the azadipeptide ni-

trile chemotype was prepared and selected for positron emis-

sion tomography (PET) tracer development owing to its high

affinity for the oncologically relevant cathepsins L, S, K and B.

Labelling with fluorine-18 was accomplished in an efficient and

reliable two-step, one-pot radiosynthesis by using 2-

[¹⁸F]fluoroethylnosylate as a prosthetic agent. The pharmacoki-

netic properties of the resulting radiotracer compound were

studied in vitro, ex vivo and in vivo in normal rats by radiome-

tabolite analysis and small-animal positron emission tomogra-

phy. These investigations revealed rapid conjugate formation

of the tracer with glutathione in the blood, which is associated

with slow blood clearance. The potential of the developed ¹⁸F-

labelled probe to image tumour-associated cathepsin activity

was investigated by dynamic small-animal PET imaging in

nude mice bearing tumours derived from the human NCI-H292

lung carcinoma cell line. Computational analysis of the ob-

tained image data indicated the time-dependent accumulation

of the radiotracer in the tumours. The expression of the target

enzymes in the tumours was confirmed by immunohistochem-

istry with specific antibodies. This indicates that azadipeptide

nitriles have the potential to target thiol-dependent cathepsins

in vivo despite their disadvantageous pharmacokinetics.

Introduction

The huge and diverse class of proteolytic enzymes accounts

for 2% of the human genome and plays a pivotal role in the

pathophysiology of tumours. Proteases of all mechanistic sub-

classes (Ser/Thr-, Cys-, Asp- and metalloproteases) have been

discussed to be involved in the extensive network of proteoly-

sis leading to tumour invasion and metastasis.^[1]A prominent

function in enabling tumour cells for metastasis has been as-

signed to matrix-metalloproteinases (MMPs).^[2]Beside these en-

zymes, the thiol-dependent cathepsins are attracting increas-

ing interest in tumour biology as key players in tumour-linked

proteolysis. Eleven different representatives of these papain-

like cysteine proteases are encoded in the human genome and

the expressed enzymes are localised inside the lysosome

under usual conditions.^[3]The most significant role of these

proteases in tumour progression has become evident for cath-

epsins L, B and S.^[4]Their particular function in tumour progres-

sion is mainly exerted by the extracellular action of the secret-

ed proteases on proteins that are important either for the in-

teraction of tumour cells with their microenvironment or for

the interaction of tumour cells with neighbouring tumour

cells.^[5]In detail, cathepsins are able to degrade the compo-

nents of both the basement membrane and the extracellular

matrix.^[6]Furthermore, they have the capability for proteolytic

activation of alternative proteases participating in this process,

such as MMP-1 and MMP-3, as well as the serine protease uro-

kinase plasminogen activator (uPA), from their respective zym-

ogens.^[6]The increased cathepsin activity in tumours does not

only originate from tumour cells but also from tumour-associ-

ated benign cells, such as macrophages and other immune

cells, endothelial cells and fibroblasts.^[5a]Tumour cells and en-

dothelial cells mainly contribute the cathepsins L and B, where-

as the cathepsin activity of tumour-associated immune cells is

dominated by cathepsin S. It has been shown in numerous

studies that increased cathepsin activity is correlated with the

invasive potential and, thus, the malignancy of tumour cells.^[7]

A correlation of elevated cathepsin B activity with malignancy

was evidenced for various tumour entities, such as melano-

ma,^[8]glioma^[9]and breast cancer.^[10]Cathepsin L has been

shown to be crucial for the invasive capacity of oncogenically

transformed fibroblasts^[11]and is involved in the metastasis of

melanoma cells.^[12]The connection of cathepsin S activity in tu-

mours to increased malignancy seems to be less evident than

that for the cathepsins B and L. Nevertheless, cathepsin S is im-

[a] Dr. R. Lçser, Dr. R. Bergmann, Dr. B. Mosch, L. Dombrowski, M. Kuchar,

Prof. Dr. J. Steinbach, Prof. Dr. J. Pietzsch

Institut fꢀr Radiopharmazeutische Krebsforschung

Helmholtz-Zentrum Dresden-Rossendorf

Bautzner Landstraße 400, 01328 Dresden (Germany)

E-mail: r.loeser@hzdr.de

[b] Dr. M. Frizler, Prof. Dr. M. Gꢀtschow

Pharmazeutisches Institut, Pharmazeutische Chemie I

Rheinische Friedrich-Wilhelms-Universitꢁt

An der Immenburg 4, 53121 Bonn (Germany)

[c] Dr. R. Lçser, M. Kuchar, Prof. Dr. J. Steinbach, Prof. Dr. J. Pietzsch

Fachrichtung Chemie und Lebensmittelchemie

Technische Universitꢁt Dresden

Bergstraße 66c, 01062 Dresden (Germany)

Supporting information for this article is available on the WWW under

http://dx.doi.org/10.1002/cmdc.201300135.

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

2

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

portant for tumour pathology. Concerning its prognostic value,

contradictory findings have been reported. In lung cancer, in-

creased activity of the enzyme was correlated with a better

prognosis,^[13]whereas recurrence-free survival was decreased

for patients with colorectal carcinoma overexpressing cathe-

psin S.^[14]In accordance with the latter result, the activity of

this protease was found to be lower in glioblastoma stem cells

than in their invasive progenitors.^[15]

109.8 min. Moreover, fluorine atoms can form highly stable co-

valent bonds with carbon atoms. For these reasons, fluorine-18

represents the radionuclide of choice to develop PET tracers

based on small molecules.^[28]The aim of this study was the

design and synthesis of a fluorine-containing cathepsin inhibi-

tor and its labelling with fluorine-18, as well as its characterisa-

tion in vivo.

Peptidic and non-peptidic compounds containing cyano

groups have been comprehensively explored as inhibitors for

thiol-dependent cathepsins in recent years.^[29]Among this class

of inhibitors, azadipeptide nitriles have proven advantageous

due to their high potency, combined with notable stability to-

wards degradation by serine proteases.^[30]Extensive structure–

activity relationships have been deduced for these com-

pounds^[31]and their efficiency for targeting papain-like cysteine

proteases in cellular systems has been established.^[32]There-

fore, this inhibitor chemotype was selected to develop a PET

tracer for visualisation of cysteine cathepsins in vivo.

The function of cathepsin K in neoplastic conditions is less

well understood than that of the enzymes mentioned above. It

is one of the most potent matrix-degrading enzymes known^[16]

and its expression mainly occurs in osteoclasts^[17]but is not re-

stricted to these cells. Cathepsin K could be detected in various

prostate cancers, with higher levels in metastatic than in pri-

mary tumour cells.^[18]Immunoreactivity against cathepsin K

was shown to be associated with the increased invasive poten-

tial of lung carcinoma.^[19]In addition, breast-tumour-associated

stromal fibroblasts were found to be cathepsin K positive and

able to stimulate the invasion of breast-tumour epithelial cells

in vitro.^[20]

Due to the evident importance of thiol-dependent cathe-

psins for tumour biology, several efforts have been undertaken

towards the development of molecular probes that are capa-

ble of imaging the activity of these enzymes in vivo but also of

unraveling their functions in cell-based experiments.^[21]

Cy5.5-conjugated methoxypoly(ethylene glycol)-grafted

poly-l-lysines were found to be useful agents for optical imag-

ing of cathepsin B in vivo. Multiple fluorophore units were

linked to the polymer backbone, which led to internal quench-

ing of the near-infrared fluorescence (NIRF). Unmodified lysines

within the polymer served as recognition sites for

cathepsin B-catalysed cleavage, which led to an en-

hanced fluorescence signal. This NIRF probe was

used for the imaging of human breast cancers im-

planted in nude mice and a correlation was found

with cathepsin B activity.^[22]Besides substrate-based

probes, fluorescently labelled inhibitors containing an

acyloxymethylketone substructure were successfully

developed for in vivo application.^[23]For application

in vivo, optical imaging has severe limitations in sen-

sitivity and penetration depth and is excelled by mo-

Results and Discussion

Synthesis of the non-radioactive reference compound and

enzyme inhibitory activity

The azadipeptide nitrile 1 was prepared as described recen-

tly.^[31a]Compound 1 was O-alkylated by using 1-bromo-2-fluo-

roethane (2), which in turn was prepared by a published

method.^[33]This furnished the fluorine-containing azadipeptide

nitrile 3 (Scheme 1), which served as a non-radioactive refer-

ence compound of the desired tracer.

Scheme 1. Synthesis of the fluorine-containing azadipeptide nitrile 3. Reagents and condi-

tions: a) DAST, CH₂Cl₂, ꢀ508C!RT; b) NaH, DMF. DAST=diethylaminosulfur trifluoride;

DMF=N,N-dimethylformamide.

lecular imaging based on radiolabelled probes in

these regards. Cathepsin-inhibiting peptide-derived

diazoketones and epoxides have been labelled with

iodine-125 and used for cell-based in vitro studies

but reports on their in vivo behaviour are missing.^[24]

Positron emission tomography (PET) represents the most fa-

vourable modality for functional imaging in terms of sensitivity,

quantitative image evaluation and penetration depth.^[25]For

example, successful attempts have already been made to

image MMPs with radiofluorinated inhibitors.^[26]The develop-

ment of cathepsin-targeting molecular probes labelled with

positron-emitting radionuclides has started only very recently

with copper-64-labelled acyloxymethylketones.^[27]Among the

various radionuclides available for PET imaging, fluorine-18 can

be considered as optimal because it has a high content of

positron emission (97%) and an intermediate half-life of

Azadipeptide nitriles have been found to be highly potent

slow-binding inhibitors of papain-like cysteine proteases.^[30,31]

The inhibitory activity of 1 towards the oncologically important

cathepsins L, S, B and K was reported to be in the sub-nano-

molar range (Table 1).^[31a]The inhibition constants, K_i, as well as

the rate constants, k_onand k_off, of the fluorine-containing azadi-

peptide nitrile 3 were determined in kinetic enzyme assays for

cathepsins L, S, B and K. Similarly to those of the parent com-

pound 1, the inhibition constants of the fluoroethyl derivative

3 were in the sub-nanomolar range, with the exception of

those for cathepsin B. The lower binding affinity of 3 relative

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

3

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Table 1. Inhibition of thiol-dependent cathepsins by the azadipeptide nitriles 1 and 3.^[a]

Compd

Cathepsin L

Cathepsin S

Cathepsin B

Cathepsin K

K_i

[nm]

k_on

k_off

[10^ꢀ3s^ꢀ1

K_i

[nm]

k_on

k_off

[10^ꢀ3s^ꢀ1

K_i

[nm]

k_on

k_off

[10^ꢀ3s^ꢀ1

K_i

[nm]

k_on

k_off

[10^ꢀ3s^ꢀ1

]

[10³m^ꢀ1s^ꢀ1

]

[10³m^ꢀ1s^ꢀ1

]

[10³m^ꢀ1s^ꢀ1

]

[10³m^ꢀ1s^ꢀ1

]

1^[b]

3

0.36ꢁ0.03 4200ꢁ100 1.5ꢁ0.1

0.86ꢁ0.02 500ꢁ30

0.43ꢁ0.03 0.38ꢁ0.03

79ꢁ22 0.030ꢁ0.009 0.16ꢁ0.01 320ꢁ20 0.051ꢁ0.005

190ꢁ10 0.46ꢁ0.03 0.17ꢁ0.01 210ꢁ50 0.036ꢁ0.009

0.73ꢁ0.06 930ꢁ100 0.68ꢁ0.09 0.79ꢁ0.06 560ꢁ100 0.44ꢁ0.09 2.4ꢁ0.1

[a] Data represent the meanꢁSEM; experiments were performed in duplicate with at least five different inhibitor concentrations. [b] Values taken from Ref. [31a].

to 1 towards cathepsin B can be attributed to a notably in-

creased first-order rate constant, k_off, to describe the decay of

the enzyme–inhibitor complex, whereas the second-order rate

constant for the association to the complex, k_on, was slightly

higher than that of the parent compound 1. Owing to its high

affinity for the oncologically important thiol-dependent cathe-

psins, 3 was used as a basis to develop a PET tracer targeting

these enzymes.

manner, which indicated the decomposition of [¹⁸F]6a under

these reaction conditions. The degradation of the ¹⁸F-fluoroe-

thylation agent was even more pronounced in DMF at 908C

and only traces of [¹⁸F]6a and [¹⁸F]3 could be detected at

1508C with this solvent.

The isolation of the labelled azadipeptide nitrile [¹⁸F]3 was

attempted for the most favourable conditions (acetonitrile as

the solvent, 1158C, 20 min reaction time) by semi-preparative

reversed-phase (RP) HPLC. This resulted in [¹⁸F]3 of poor radio-

chemical purity (RCP) because the separation from the remain-

ing [¹⁸F]6a proved difficult. The low labelling yields and the dif-

ficult separation of [¹⁸F]3 from [¹⁸F]6a led to the use of alterna-

tive 2-[¹⁸F]fluoroethylbenzenesulfonates being envisaged. The

selection of the benzenesulfonate leaving groups was inspired

by the study of Musachio et al.^[35]The required ethylene glycol-

1,2-dibenzenesulfonates 6b and 6c were prepared by follow-

ing published procedures (see scheme S1 in the Supporting In-

formation).

Synthesis of labelling precursors and radiolabelling

In analogy to the synthesis of the non-radioactive reference 3,

¹⁸F-fluoroethylation of 1 appeared to be a promising strategy

to afford the radiolabelled azadipeptide nitrile [¹⁸F]3. As radio-

labelling by ¹⁸F-fluoroethylation is commonly performed by

employing 2-[¹⁸F]fluoroethyltosylate,^[34]initial efforts towards

the preparation of [¹⁸F]3 focused on the conversion of 1 with

this alkylating agent. A two-step, one-pot strategy was fol-

lowed, which consisted of the reaction of commercially avail-

able ethylene glycol-1,2-ditosylate (5a) with [¹⁸F]fluoride to

form 2-[¹⁸F]fluoroethyltosylate ([¹⁸F]6a), which was converted

without isolation, with 1 used in a slight excess over 5a

(Scheme 2). Analysis by radio-TLC indicated the formation of

[¹⁸F]3 in acetonitrile at 1008C. However, the labelling yields

The reaction of ethylene glycol-1,2-dinosylate (5b) with

[¹⁸F]fluoride was optimised with regard to the amount of pre-

cursor and reaction time. The highest labelling yields of [¹⁸F]6b

were achieved with 3 and 5 mg of 5b in acetonitrile. Conver-

sion of [¹⁸F]6b with 1 resulted in the formation of [¹⁸F]3 in an

analytical radiochemical yield (determined by radio-TLC; see

figure S2 in the Supporting Infor-

mation) of 74% after 10 min in

acetonitrile at 1158C (Table 2).

More reproducible results could

be obtained when 5 mg of 5b

were used instead of 3 mg. Radi-

ofluorination of ethylene glycol-

1,2-bis(3,4-dibromobenzenesul-

fonate) (5c) proceeded with

a lower analytical radiochemical

yield than those with 5a and 5b

and the results for ¹⁸F-fluoroe-

thylation with [¹⁸F]6c were even

less beneficial than with the cor-

responding tosylate (Table 2).

Besides the two-step, one-pot

Scheme 2. Two-step, one-pot radiosynthesis of [¹⁸F]3. Reagents and conditions: a) [¹⁸F]KF, K₂CO₃, K222, CH₃CN,

908C, 5 min; b) reaction solution from (a), 1m Cs₂CO₃/H₂O, 1158C, 10 min. K222=Kryptofix 222.

were low and a considerable amount of the ¹⁸F activity re-

mained in the intermediate 2-[¹⁸F]fluoroethyltosylate ([¹⁸F]6a;

see table S1 in the Supporting Information). An increase in the

temperature to 1158C resulted in higher yields of the desired

product, which, however, did not exceed 26%. The amount of

[¹⁸F]6a continuously decreased with the reaction time. This de-

cline was not accompanied by an increase in [¹⁸F]3 in the same

approach for ¹⁸F-labelling of 3, direct radiofluorination was also

considered. Hence, the corresponding tosylate and nosylate

precursors 7a and 7b were prepared by treating 1 with an

excess of 5a and 5b, respectively, as described for other com-

pounds (Scheme 3).^[36]The precursors 7a and 7b were treated

with [¹⁸F]fluoride as detailed in the Experimental Section.

Under favourable conditions (3 mg of precursor, 5 min reaction

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

4

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

chemical purities were greater than 98% and the specific activ-

ities ranged from 2.9–30.4 GBqmmol^ꢀ1(see figure S3 and S4 in

the Supporting Information for the chromatograms).

Immunohistochemical detection of cathepsins in xenograft-

ed NCI-H292 tumours

The overexpression of cathepsins L and B in lung carcinoma

has been reported in several studies.^[37]Therefore, we decided

to use tumour xenografts derived from the human pulmonary

mucoepidermoid carcinoma cell line NCI-H292 in nude mice as

a model to investigate the potential of fluorine-18-labelled aza-

dipeptide nitrile [¹⁸F]3 for imaging the cathepsin activity in tu-

mours.

The expression of cathepsins L, S, B and K was analysed in

selected mouse and rat organs and in tumour xenografts by

using immunohistochemical staining. Representative images of

mouse specimens and tumour xenografts are shown in

Figure 1 and the corresponding negative controls (staining

without primary antibodies) can be found in the Supporting

Information (figure S5). In tumour xenografts, all four cathe-

psins were detected with higher intensity for cathepsins L and

K than for cathepsins S and B. The skeletal muscle was chosen

as an internal negative control. It is of note that cathepsin L ex-

pression was detectable in mouse but not in rat muscle. Cathe-

psins S, B and K showed a signal in neither mouse nor rat

muscle. In the liver, all cathepsins could be detected, although

cathepsins B and K had very low levels in the mouse liver. The

spleen showed strong expression of all cathepsins, with the ex-

ception of cathepsin B in the rat, for which no specific signal

was detectable (data not shown in detail). A likewise intensive

cathepsin staining was observed in the kidney, with the excep-

tion of weak signals for cathepsin B in the mouse and cathe-

psin K in the rat. Taken together, we detected considerable ex-

Scheme 3. Synthesis of direct-labelling precursors 7a and 7b and one-step

radiolabelling of [¹⁸F]3. Reagents and conditions: a) 5a or 5b, respectively,

K₂CO₃, CH₃CN, reflux; b) [¹⁸F]KF, K₂CO₃, K222, CH₃CN, 908C.

Table 2. Yields and conditions for the radiosynthesis of [¹⁸F]3 by ¹⁸F-fluo-

roethylation of

[¹⁸F]fluoroethyl-3,4-dibromobenzenesulfonate ([¹⁸F]6c).

1

with 2-[¹⁸F]fluoroethylnosylate ([¹⁸F]6b) and 2-

Radiofluorination

precursor

t_reaction[min]

Analytical radiochemical

yield [%]^[a]

[¹⁸F]6b/c

[¹⁸F]3

5b

5c

5

10

20

5

10

20

31

6

0

7

3

58

74

49

30

29

22

0

[a] Determined by radio-TLC.

time), [¹⁸F]3 was formed from

the nosylate 7b with an analyti-

cal radiochemical yield of 12%,

whereas no product formation

could be observed with the cor-

responding tosylate 7a (see

table S2 and figure S2 in the

Supporting Information). There-

fore, it was concluded that direct

labelling is less advantageous

than ¹⁸F-fluoroethylation and the

preparative radiosynthesis of

[¹⁸F]3 was optimised for the two-

step, one-pot procedure with 2-

[¹⁸F]fluoroethylnosylate ([¹⁸F]6b)

as the alkylating agent. This led

to a reliable procedure for the

preparation of [¹⁸F]3 with an

average radiochemical yield after

isolation of (26.3ꢁ4.6)% (n=

13), with a synthesis time of

140 min by starting with drying

Figure 1. Immunohistochemical detection of cathepsins in NCI-H292 tumour xenografts and mouse organs (skele-

tal muscle, liver, spleen, kidney). Tissue pieces were paraffin-embedded, sliced into 5 mm sections and stained as

described in the Experimental Section. Red colour indicates cathepsin staining; blue colour displays cell nuclei

of the [¹⁸F]fluoride. The radio- (counterstaining with Mayer’s Hematoxylin). Bar=100 mm; cath=cathepsin.

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

5

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

pression of cathepsins L, S, B and K in rodent liver, spleen and

kidney samples, as well as in the tumour xenografts. Thus, the

NCI-H292 tumour xenografts were chosen as a promising

model for radiopharmacological

blood-cell membrane seems to be facilitated by aquaporins.^[39]

As cyanamides are known to react spontaneously with

thiols,^[40]it was assumed that the glutathione (GSH) conjugate

characterisation of [¹⁸F]3 as

a tumour-imaging agent.

Radiopharmacological charac-

terisation in vitro, ex vivo and

in vivo

LogP and in vitro stability

A logP value of 2.3 was deter-

mined for [¹⁸F]3 by the shaking-

flask method.^[38]The radiotracer

will neither be protonated nor

deprotonated in the physiologi-

cal pH range. Probe [¹⁸F]3 was

proven to be stable in phos-

phate-buffered saline (PBS) over

6 h at pH 7.4 (see figure S6 in

the Supporting Information). In-

cubation of [¹⁸F]3 in rat blood

plasma and fresh whole blood

was carried out to model the be-

haviour of the compound in bio-

logical media without the inter-

action with the endothelium and

the tissues in the organism. To

a minor extent, [¹⁸F]3 was con-

verted into a compound repre-

sented by a second peak with

a lower retention time upon in-

cubation in rat blood plasma.

Analysis of the compound’s fate

in whole rat blood by radio-

HPLC indicated its rapid transfor-

mation into more hydrophilic

metabolites (see chromatogram

for 0 min in Figure 2). Further-

more, determination of the dis-

tribution of the ¹⁸F activity over

the blood components revealed

that 94% were accumulated im-

mediately in the erythrocytes

(Table 3). However, there are no

cathepsins located in red blood

cells; therefore, the accumula-

tion in the erythrocytes must be

driven by other mechanisms.

Several drug molecules are

known to accumulate in the er-

ythrocytes by reversible binding

Figure 2. a) Structures of conjugates of [¹⁸F]3 with different endogenous cysteine derivatives. b) Radio-HPLC chro-

matograms of [¹⁸F]3, [¹⁸F]3–GSH, [¹⁸F]3–AcCys, [¹⁸F]3–Cys (obtained by incubation of [¹⁸F]3 with the corresponding

cysteine derivatives) and metabolites of [¹⁸F]3 detected in rat blood, urine and intestine (0 min: blood in vitro;

60 min p.i.: blood, urine and intestine ex vivo, respectively).

to proteins or to undergo trans-

formation inside these cells.

Their penetration of the red-

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

6

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Table 3. ¹⁸F activity (% of blood activity) in the erythrocytes in vitro and

ex vivo.

in vitro

ex vivo

time [min]

3

30

60

3

30

60

number of values

25th percentile

median

3

2

5

93.0

94.0

96.0

81.0

82.5

84.0

65.0

69.5

74.0

18.0

20.0

59.0

30.5

34.0

65.5

44.0

54.0

65.5

75th percentile

of [¹⁸F]3 was formed and resulted in temporary trapping inside

the erythrocytes. To prove this, the radiotracer was incubated

with 10 mm glutathione in PBS, a concentration that is similar

to the glutathione level in erythrocytes.^[41]Analysis by radio-

HPLC confirmed the identity of the observed metabolite in rat

blood with the product formed upon incubation with gluta-

thione (Figure 2). To suppress this early metabolism of the radi-

otracer, attempts to lower the blood glutathione level were

undertaken. Buthionine sulfoximine (BSO), a mechanism-based

inhibitor of g-glutamylcysteine synthetase,^[42]the enzyme cata-

lysing the first step of glutathione biosynthesis, has been used

to systemically decrease the glutathione level in vivo.^[43]With

this aim, samples of rat blood were incubated with varying

doses of BSO. Even in the presence of the highest dose of

3.75 mm, the metabolism of [¹⁸F]3 was not influenced. To se-

quester residual amounts of glutathione, blood samples were

additionally incubated with varying doses of etacrynic acid, an

antidiuretic agent, the pharmacological action of which relies

on Michael adduct formation with cysteine.^[44]Even the pres-

ence of both BSO and etacrynic acid, at concentrations of 3.53

and 2.94 mm, respectively, did not prevent the transformation

of [¹⁸F]3 into the glutathione conjugate. Diethyl maleate,

a stronger Michael acceptor than etacrynic acid, was also em-

ployed to sequester the glutathione of the blood.^[45]Incubation

of blood samples with 2% of this agent resulted in significant

suppression of the radiotracer conversion (see figure S7 in the

Supporting Information). This provides further evidence for the

major metabolism of [¹⁸F]3 by reaction with glutathione. To

prove this mechanism, erythrocyte ghosts were prepared and

incubated with [¹⁸F]3, whereupon no conversion of the radio-

tracer into the conjugate could be observed (see figure S8 in

the Supporting Information). The conjugation of [¹⁸F]3 was

also observed with cysteine to a minor extent in rat blood

(Figure 2). The comparison of the ¹⁸F-activity accumulation in

the erythrocytes in vitro and in vivo revealed a fast uptake in

the erythrocytes under both conditions (Table 3).

Figure 3. Biodistribution of [¹⁸F]3 in Wistar rats at 5 (&) and 60 min (&)

after a single intravenous injection. a) Activity amount in selected organs

(data are meanꢁSD in %ID for 16 animals). b) Activity concentration (data

are meanꢁSD in SUV_biofor 16 animals).

Our data indicate that the majority of the ¹⁸F activity after in-

travenous injection was eliminated through the kidneys. An

amount of 41%ID was calculated for the urine after one hour.

A minor part is eliminated through the hepatobiliary route.

About 18%ID are accumulated in the liver and intestine after

one hour. The ratio of approximately 2:1 between renal and

hepatobiliary elimination can be explained by the conjugate

formation between [¹⁸F]3 and glutathione leading to a consid-

erably increased hydrophilicity, which favours renal filtration

but also specific transport in the liver (see below).^[46]Another

large fraction of the ¹⁸F activity is distributed in the whole or-

ganism (carcass), which might be explained by the high activi-

ty in the blood even after one hour. The clearance of the ¹⁸F

activity from the blood was relatively slow with a half-life of

39 min (Figure 4a); however, the metabolite-corrected [¹⁸F]3

clearance shows an even lower half-life of 114 min. An increase

of the ¹⁸F-activity concentration over time was only observed

in the kidneys and resulted in an SUV of 4.98ꢁ0.53 at 60 min

p.i. The accumulation mechanism was not investigated in

detail. The lung was a further organ with an SUV significantly

greater than 1 after 60 min p.i., which probably can be attrib-

uted to the slow blood clearance of the ¹⁸F activity.

In vivo studies

Biodistribution and in vivo stability

The biodistribution of [¹⁸F]3 was studied in normal rats after 5

and 60 min post injection (p.i.). The ¹⁸F-activity distribution

data are shown both as a percentage of injected dose (%ID) in

selected organs (Figure 3a) and as standardised uptake values

(SUV_bio; Figure 3b).

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

7

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Figure 5. ¹⁸F activity in the erythrocytes of arterial blood for five rats (&

meanꢁSEM) and five mice (* each data point: 1 mouse). The activity in

the erythrocytes is expressed as % of activity in the whole blood sample.

activity in the red blood cells seems to be the reason for the

low biological availability.

As mentioned above, ¹⁸F-activity uptake in the erythrocytes

in vitro is very fast and high (median 94.0%; decreased to

69.5% after one hour). Contrastingly, in vivo, the initial uptake

after 3 min was only 20% (median) and increased to 54% after

60 min. This difference seems to be a result of the different re-

action volume, as well as the resulting higher dilution in vivo

and the continuous elimination of the radiolabelled compound

and its metabolites from the plasma, which results in an in-

creasing activity ratio in the erythrocytes to the plasma. A simi-

lar behaviour of [¹⁸F]3 was also observed in the blood of mice

(Figure 5).

Figure 4. Ex vivo pharmacokinetics of the ¹⁸F activity derived from com-

pound [¹⁸F]3 distribution in the blood. a) ¹⁸F-activity clearance in arterial

blood of rats (meanꢁSEM in %ID per mL of blood for 4 animals): t_1/

₂=39.3 min. b) Fraction of ¹⁸F metabolites of the total activity in the arterial

blood plasma.

The in vivo stability of [¹⁸F]3 was investigated in rat arterial

blood samples. The overall blood volume of rats allows for col-

lection of blood samples in amounts sufficient for radio-HPLC.

The samples were collected at 1, 3, 5, 10, 20, 30 and 60 min p.i.

Some of the radio-HPLC chromatograms are shown in Figure 2.

Even though [¹⁸F]3 was synthesised with a very high RCP value

of (98.5ꢁ0.3)% (meanꢁSEM for 23 syntheses), a small part

was immediately hydrolysed to the free acid after dissolving

[¹⁸F]3 in the injection electrolyte solution (E-153). The remain-

ing RCP value of [¹⁸F]3 was (84ꢁ3.5)% (meanꢁSEM). The next

step, incubation of [¹⁸F]3 with the first rat blood sample in vitro

(designated as 0 s), revealed the immediate formation of me-

tabolites. In the blood sample collected at one minute after in-

jection, only 56% of the blood activity remained as original

[¹⁸F]3 (see table S3 in the Supporting Information). The main

metabolites were [¹⁸F]3–GSH and [¹⁸F]3–Cys. The [¹⁸F]3–GSH

was located in the erythrocytes, in which the ¹⁸F activity in-

creased from 30% of the total blood activity at 1 min to 55%

after 60 min (Figure 5). Notably, despite vibrant conjugate for-

mation, around 35% of original [¹⁸F]3 was still be detectable in

the rat blood after one hour, as is evident from Figure 4b. Er-

ythrocytes represent by far the most abundant cellular compo-

nents of the blood, so the localisation of the radioactive me-

tabolites inside these cells has great influence on the biodistri-

bution and tissue-specific accumulation. An in vivo transport of

glutathione conjugates inside rat erythrocytes and their elimi-

nation through the liver and bile into the intestine has also

been reported for other xenobiotics.^[46,47]This retention of the

The urine and the content of the intestine were investigated

by radio-HPLC 60 min after injection of [¹⁸F]3 into the rat.[¹⁸F]3

was also incubated with the respective biogenic thiols and, by

comparison, [¹⁸F]3–GSH and [¹⁸F]3–Cys were identified as the

main metabolites in the intestine. In contrast to this finding,

the conjugate of [¹⁸F]3 and N-acetylcysteine (AcCys; mercaptu-

ric acid) appeared in considerable amounts in the urine, in ad-

dition to [¹⁸F]3–Cys (Figure 2). In both the urine and the intes-

tine, only negligible amounts of original [¹⁸F]3 were detectable.

The formation of the conjugates derived from Cys and AcCys

can be explained by peptidolytic processing of [¹⁸F]3–GSH into

[¹⁸F]3–Cys followed by acetylation to form [¹⁸F]3–AcCys mainly

occurring in the liver but also in the epithelial cells of the prox-

imal tubulus.^[46]

Small-animal PET studies

The biodistribution and biokinetics of [¹⁸F]3 were also studied

with small-animal PET in rats, NMRI mice and NMRI nude mice

(nu/nu) bearing NCI-H292 tumour xenografts. The resulting

images are depicted in Figure 6 (normal rat and normal

mouse) and Figure 7 (tumour mouse) and confirm the data ob-

tained for the ex vivo biodistribution of the compound. A large

proportion of the activity was eliminated through the kidneys

into the urine and through the hepatobiliary route into the in-

testine. The dynamics and metabolism of [¹⁸F]3 in mice seem

to be faster than in rats (Figure 8). Hepatobiliary elimination of

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

8

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Figure 6. Representative small-animal PET images (transversal, coronal, sagittal planes and maximum intensity projection (MIP)) of [¹⁸F]3 in a Wistar rat and

NMRI mouse after 1, 5, 60 and 120 min p.i., respectively.

the compound was more pro-

nounced in mice; about 40%ID

(calculated from the PET data)

was in the gall bladder and the

intestine after one hour. In the

rat, only 3%ID (calculated from

the PET data) was in the intes-

tine at the same time point (Fig-

ure 8d). The blood clearance, as

calculated from the PET data,

was also substantially faster in

the mice (half-life of 11.7 min;

95% confidence interval: 6.7–

45.4 min) than in the rats

(30.9 min; 95% confidence inter-

val: 19.9–69.6 min). The faster

hepatobiliary elimination rates

observed in mice follow the gen-

eral trend that excretion of xeno-

biotics proceeds faster in mice

than in rats.^[48]

Even after 2 h, the ¹⁸F-activity

concentration in the blood of

the mice was high enough to

visualise parts of the vascular

system. After rescaling of the

Figure 7. Maximum intensity projections of a representative PET study with a NCI-H292 tumour-bearing NMRI nu/

nu mouse at 1, 5, 60 and 120 min after a single intravenous injection of [¹⁸F]3 (upper images are scaled to the

images, the tumour was clearly

visible too (Figure 7). However,

maximum in the field of view; lower images are manually scaled; circle indicates the region of the tumour).

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

9

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Figure 8. Biokinetics of [¹⁸F]3 in rats (&) and mice (*). The time–activity curves were derived from the regions of interest over a) the vena cava, b) the liver,

c) the kidneys, and d) the intestine. Curves a)–c) are meansꢁSEM for five mice and eight rats. The time–activity curve of the rat intestine could be drawn

only from one rat, because of the limited field of view; the elimination into the intestine of the rat is shown in table S5 in the Supporting Information.

the pharmacokinetics of [¹⁸F]3 was found to be complex; after

an increase of the tumour uptake to an SUV of 0.86

(2.5%IDg^ꢀ1), the uptake decreases over the time without

reaching a steady state (Figure 9). This reversibility was also re-

flected in the kinetic PET analyses, in which the reversible two-

tissue compartment model successfully described the tumour

accumulation of [¹⁸F]3. This model is defined by the following

rate constants: K₁(apparent influx rate constant), k₂(efflux rate

constant), k₃(local uptake rate constant) and k₄(rate constant

of local release), the fractional blood volume (FBV), and the

delay of tumour-tracer uptake in comparison with blood time

activity in the vena cava and the aorta (see figure S9 in the

Supporting Information for a schematic representation).^[49]The

¹⁸F-activity values were corrected for metabolites. The follow-

20 min before administration of [¹⁸F]3 did not block the

tumour uptake of the radiotracer (see figure S10 in the Sup-

porting Information). This does not necessarily indicate a ca-

thepsin-independent mechanism for tumour accumulation.

Similarly to azadipeptide nitrile 3, the peptide aldehyde leu-

peptin interacts covalently with the active-site cysteine residue

of cathepsins.^[50]We assume that leupeptin, similarly to com-

pound 3, reacts spontaneously with other endogenous thiols

and, therefore, protected [¹⁸F]3 from nonspecific interactions

with thiols. In consequence, more unmetabolised [¹⁸F]3 should

be available for tumour uptake. Such a phenomenon has also

been observed with other radiotracers.^[51]

ing values were calculated: K₁=0.034 min^ꢀ1mL^ꢀ1

,

k₂=

Conclusions

0.483 min^ꢀ1, k₃=0.082 min^ꢀ1, k₄=0.026 min^ꢀ1, FBV=0.012. The

fractional blood volume (1.2%) was relatively low; this seems

to be a result of the artificially high interstitial pressure that is

often observed in subcutaneously xenotransplanted tumours.

The volume of distribution (V_T), which is equal to the ratio of

tissue and plasma concentrations at true equilibrium, was

0.292 mLcm^ꢀ3. The low FBV might be caused by limited vascu-

larisation. The tumour to muscle ratio increased up to 10 after

two hours, which indicated specific interactions of [¹⁸F]3 in the

tumour (Figure 9). As the presence of all four cathepsins that

were investigated for inhibition by azadipeptide nitrile 3 could

be confirmed immunohistochemically in the neoplastic tissue,

a cathepsin-mediated tumour uptake of [¹⁸F]3 appears to be

obvious. Injection of leupeptin (10 mgkg^ꢀ1body weight)

The radiosynthesis of a fluorine-18-labelled PET tracer targeting

cathepsins was established for the first time on the basis of

compound 3, a fluoroethylated inhibitor of the azadipeptide

nitrile chemotype. Among different approaches tried, a two-

step, one-pot procedure with 2-[¹⁸F]fluoroethylnosylate was re-

vealed to be the most efficient one for achieving sufficient and

reproducible radiochemical yields in the radiosynthesis of

[¹⁸F]3.

The radiopharmacological characterisation of the tracer com-

pound provided insight into the pharmacokinetic properties of

azadipeptide nitriles. Their inherent thiol reactivity leads to

a slow blood clearance, which limits their suitability as radio-

tracers for PET imaging.

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&10

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

Figure 9. [¹⁸F]3 kinetics in NCI-H292 tumour-bearing mice. The time–activity curves represent a) tumour, b) blood, c) tumour-to-muscle and d) tumour-to-

blood ratios. The values are meansꢁSEM for nine animals.

Dynamic PET studies with [¹⁸F]3 in NMRI nu/nu mice bearing

by using Merck silica gel (mesh size 230–400 ASTM) with solvent

mixtures as specified for the particular compounds. TLC was per-

tumours derived from the human lung carcinoma cell line NCI-

formed on Merck silica gel F-254 aluminium plates with visualisa-

H292 indicated a reversible tumour accumulation of the tracer.

tion under UV (254 nm). Radio-TLC was conducted in the same

The kinetics of this process could be well described with a re-

way with petroleum ether/EtOAc (1:2) as the eluent. Radioactive

versible two-tissue compartment model. Immunohistochemical

spots were visualised by using the Fujix Bas2000 TR radioluminog-

analysis of the tumour sections for the presence cathepsins L,

raphy system. The identity of the radioactive spots was proven

S, K and B confirmed the expression of the target enzymes. In

combination, these results reveal strong evidence that [¹⁸F]3 is

able to visualise tumour-associated cathepsins in vivo. The cor-

relation of tumour uptake and tumour-associated cathepsin ac-

tivity will be elaborated in further studies, together with the in-

vestigation of the radiotracer in other tumour models.

with the corresponding non-radioactive reference compounds that

were spotted with radioactivity after visualisation under UV.

Analytical radio-HPLC was performed on a Luna C₁₈5 mm column

(Phenomenex, 250ꢂ4.6 mm) by using 50% CH₃CN/H₂O containing

0.1% trifluoroacetic acid as an isocratic eluent pumped by a Merck

Hitachi L7100 gradient pump at a flow rate of 1 mLmin^ꢀ1through

a Jasco DG2080 4-line degasser with UV detection by a Merck Hita-

chi L7450 diode array detector and g detection by a Raytest GABI

detector. The system was operated by the D-700 HSM software

with a Merck Hitachi D7000 interface. Semi-preparative radio-HPLC

was done on a Nucleodur C₁₈Isis 5 mm column (Macherey–Nagel,

250ꢂ10 mm) by using 50% CH₃CN/H₂O containing 0.1% trifluoro-

acetic acid as an isocratic eluent (degassed before use by ultrasoni-

cation) pumped by a Jasco PU1580 pump at a flow rate of

4 mLmin^ꢀ1with UV detection at 254 nm by a Jasco UV1575 detec-

tor. Sample injection was accomplished in an automated manner

by using the equipment of a Nuclear Interface radiofluorination

module that also contained the g detector.

Experimental Section

Syntheses of reference compounds and precursors

General remarks: All commercial reagents and solvents were used

without further purification unless otherwise specified. NMR spec-

tra were recorded on a Varian Unity 400 MHz and on a Bruker

Avance 500 MHz spectrometer. NMR chemical shifts were refer-

enced with the residual solvent resonances relative to tetramethyl-

silane (TMS). The IR spectrum was recorded on a Bruker Tensor 27

FTIR spectrometer. Mass spectra were obtained on a Quattro/LC

mass spectrometer (Micromass) by electrospray ionisation. LC-DAD

chromatograms and ESI-MS spectra were recorded on an Agilent

1100 HPLC system with an Applied Biosystems API-2000 mass

spectrometer. Preparative column chromatography was carried out

2-Bromo-1-fluoroethane (2): 2-Bromoethanol (1.8 mL, 25 mmol)

was added to a solution of DAST (2.7 mL, 20 mmol) in diglyme

(15 mL) at ꢀ508C. Cooling was continued for 30 min and the reac-

tion mixture was stirred at room temperature for 14 h. The reaction

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&

11

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

mixture was distilled at atmospheric pressure to obtain 2 (0.51 g,

Ethylene glycol-1,2-bis(3,4-dibromobenzenesulfonate) (5c): Po-

tassium trimethylsilanolate (1.92 g, 14.96 mmol) was added as

a solid in portions to a solution of ethylene glycol (0.19 g,

2.99 mmol) and 4 (2.5 g, 7.48 mmol) in THF (6 mL) under ice cool-

ing. The reaction mixture was stirred for 2 h in the ice bath and

subsequently diluted with ice-cold H₂O (30 mL) and CH₂Cl₂(20 mL).

The aqueous layer was extracted with CH₂Cl₂(2ꢂ20 mL) and the

combined organic phases were washed with brine (20 mL), dried

over Na₂SO₄and evaporated in vacuo. The obtained crude product

was recrystallised from petroleum ether/CHCl₃to yield 5c (0.51 g,

26%) as white needles; mp: 167–1688C (lit.:^[35]1688C); ¹H NMR

1

16%) as a colourless liquid; H NMR (400 MHz, CDCl₃): d=3.55 (dt,

³J_H,F=20.1 Hz, ³J=5.9 Hz, 2H, CH₂Br), 4.67 ppm (dt, ²J_H,F=46.7 Hz,

³J=5.9 Hz, 2H, FCH₂); ¹⁹F NMR (376 MHz, CDCl₃): d=ꢀ212.4 ppm

(m).

N-(Benzyloxycarbonyl)-(O-(2-fluoroethyl)tyrosine)-methylazala-

nine-nitrile (3): Compound 1^[31a](1.28 g, 3.35 mmol) was dissolved

in dry DMF (20 mL) and cooled to ꢀ108C. The resulting solution

was treated with NaH (0.14 g (60% in mineral oil), 3.50 mmol) and

stirred at ꢀ108C for 30 min. Compound 2 (0.83 g (77% purity),

5.03 mmol) was added. The solution was slowly warmed to room

temperature and stirred for 19 h. The solvent was removed under

reduced pressure and the resulting residue was suspended in H₂O

and extracted with EtOAc (3ꢂ30 mL). The combined organic layers

were washed with H₂O (30 mL) and brine (30 mL), then dried over

Na₂SO₄. The crude product was purified by column chromatogra-

phy with EtOAc/petroleum ether (1:1) as the eluent to obtain 3 as

3

4

(400 MHz, CDCl₃): d=4.31 (s, 4H, 2ꢂCH₂), 7.64 (dd, J=8.4 Hz, J=

3

2.2 Hz, 2H, H-6, H-6’), 7.83 (d, J=8.4 Hz, 2H, H-5, H-5’), 8.09 ppm

(d, ⁴J=2.2 Hz, 2H, H-2, H-2’); MS (ESI, positive): m/z=681 [M+

Na]⁺; elemental analysis: calcd for C₈H₁₀Br₄O₆S₂: C 25.56, H 1.53, S

9.75; found: C 25.61, H 1.42, S 9.90.

2-Fluoroethyltosylate (6a): Tosyl chloride (2.23 g, 11.7 mmol), trie-

thylamine (1.62 mL, 11.7 mmol) and 4-dimethylaminopyridine

(0.11 g, 0.12 mmol) were successively added to a solution of 2-fluo-

roethanol (0.45 mL, 7.8 mmol) in dry CH₂Cl₂(17 mL) under ice cool-

ing. After it had been stirred for 2 h at room temperature, the reac-

tion mixture was diluted with CH₂Cl₂(20 mL) and 1m HCl (20 mL)

was added. The organic layer was separated and the aqueous

phase was extracted with CH₂Cl₂(2ꢂ20 mL). The combined organic

layers were washed with sat. NaHCO₃and brine (20 mL each),

dried over Na₂SO₄and evaporated to dryness. The obtained resi-

due was purified by column chromatography on silica gel (petrole-

1

colourless oil (0.53 g, 37%); [a]²_D⁰= +22.4 (c=0.77, CHCl₃); H NMR

(500 MHz, [D₆]DMSO, mixture of rotamers): d=2.73–2.86 (m, 2H,

3

CHCH₂), 2.98, 3.08, 3.12, 3.22 (4ꢂs, 6H, 2ꢂNCH₃), 4.19 (dt, J_H,F

=

30.0 Hz, ³J=3.5 Hz, 2H, OCH₂CH₂F), 4.66–4.77 (m, 3H, OCH₂CH₂F,

NHCHCO), 4.95 (s, 2H, CH₂O), 6.88–7.34 (m, 9H, H_arom), 7.78_a,

7.87_bppm (brs_a, d_b, J=7.9 Hz, 1H, NHCHCO); ¹³C NMR (125 MHz,

3

[D₆]DMSO, mixture of rotamers): d=30.28, 30.59, 35.65, 36.04,

40.48, 40.88 (2ꢂNCH₃, CHCH₂), 52.89, 53.34 (NHCHCO), 65.62

2

1

(CH₂O), 67.13 (d, J_C,F=18.8 Hz, OCH₂CH₂F), 82.28 (d, J_C,F=165.6 Hz,

OCH₂CH₂F), 114.21, 114.57, 127.74, 127.91, 128.42, 129.85, 130.28,

130.50, 136.91 (CN, C_arom), 156.28 (OCON), 157.12 (C_aromO),

173.50 ppm (NHCHCO); FTIR (KBr): n˜ =2222 cm^ꢀ1(CꢂN); LC-

MS(ESI) (90% H₂O to 100% MeOH in 20 min, then 100% MeOH for

30 min; DAD: 219.7–300.7 nm): t_R=17.23, 96% purity, m/z=429.5

[M+H]⁺.

um ether/EtOAc, 4:1) to yield 6a (1.12 g, 66%) as a yellowish oil;

3

¹H NMR (400 MHz, CDCl₃): d=2.46 (s, 3H, CH₃), 4.20 (dt, J_H,F

=

27.0 Hz, ³J=4.2 Hz, 2H, CH₂O), 4.50 (dt, ²J_H,F=47.0 Hz, ³J=4.1 Hz,

2H, FCH₂), 7.36 (d, ³J=8.0 Hz, 2H, H-3’, H-5’), 7.81 ppm (d, ³J=

8.4 Hz, 2H, H-2’, H-6’); MS (ESI, positive): m/z=241 [M+Na]⁺; ele-

mental analysis: calcd for C₉H₁₁FO₃S: C 49.53, H 5.00, S 14.69;

found: C 49.39, H 5.11, S 14.63.

Ethylene glycol dinosylate (5b): Nosyl chloride (10.64 g, 48 mmol)

as a solid was added to a solution of ethylene glycol (1.12 mL,

20 mmol) in CHCl₃(20 mL) and tetrahydrofuran (THF; 10 mL). Pyri-

dine (6.46 mL, 80 mmol) was added dropwise to the resulting mix-

ture under ice cooling. After the reaction mixture had stirred for

4 h at room temperature, the precipitate was collected by filtration

and recrystallised from acetone (100 mL). The crystallised solid was

collected by filtration, washed with a small amount of ice-cold ace-

tone, petroleum ether and diethyl ether and recrystallised again

from acetone to yield 5b (3.38 g, 38%) as slightly yellow crystals;

mp: 1858C (lit.:^[52]1868C); ¹H NMR (400 MHz, [D₆]DMSO): d=4.38

2-Fluoroethylnosylate (6b): A solution of nosyl chloride (0.42 g,

1.9 mmol) in THF (3 mL) and potassium trimethylsilanolate (0.87 g,

6.8 mmol) as a solid were added to a solution of 2-fluoroethanol

(0.078 mL, 1.36 mmol) in THF (2 mL) under ice cooling. The result-

ing brownish slurry was stirred for 2 h at 48C. H₂O (30 mL) was

added and the mixture extracted with CH₂Cl₂(3ꢂ20 mL). The com-

bined organic layers were washed with brine (20 mL), dried over

Na₂SO₄and evaporated to dryness. The obtained residue was puri-

fied by column chromatography on silica gel (petroleum ether/

EtOAc, 4:1) to yield 6b (0.20 g, 59%) as an off-white solid; mp:

116–1198C (lit.:^[35]118 8C); ¹H NMR (400 MHz, CDCl₃): d=4.41 (dt,

³J_H,F=27.3 Hz, ³J=4.20 Hz, 2H, CH₂O), 4.60 (dt, ²J_H,F=47.0 Hz, ³J=

3

(s, 4H, 2ꢂCH₂), 8.13 (d, J=9.0 Hz, 4H, 2’-H, 6’-H), 8.44 ppm (d, J=

8.9 Hz, 4H, 3’-H, 5’-H); ¹³C NMR (101 MHz, [D₆]DMSO): d=69.6

(CH₂), 125.6 (C-3’, C-5’), 130.1 (C-2’, C-6’), 140.7 (C-1’), 151.8 ppm (C-

4’); elemental analysis: calcd for C₁₄H₁₂N₂O₁₀S₂: C 38.80, H 2.80, N

6.48, S 14.83; found: C 38.85, H 2.72, N 6.57, S 15.03.

3

4.1 Hz, 2H, FCH₂), 8.14 (d, J=9.0 Hz, 2H, H-2’, H-6’), 8.42 ppm (d,

³J=9.0 Hz, 2H, H-3’, H-5’); MS (ESI, positive): m/z=457 [M+

5CH₃CN+H]⁺.

3,4-Dibromobenzenesulfonyl chloride (4): Chlorosulfonic acid

(8 mL, 120.1 mmol) was added dropwise to a solution of 1,2-dibro-

mobenzene (1.8 mL, 15.1 mmol) in CHCl₃(40 mL) at ꢀ108C. The re-

action mixture was stirred for 24 h at room temperature and

heated at reflux for 1 h, before being cooled and poured onto

crushed ice (50 g). The organic layer was separated and the aque-

ous phase was extracted with CH₂Cl₂. The combined organic layers

were washed with sat. NaHCO₃, H₂O (2ꢂ20 mL) and brine (20 mL)

and dried over Na₂SO₄. The solvent was removed in vacuo and the

remaining residue was purified by vacuum distillation (0.038 mbar,

1208C) to obtain 4 (2.96 g, 77%) as a yellowish waxy solid; mp:

2-Fluoroethyl-3,4-dibromobenzenesulfonate (6c): Potassium tri-

methylsilanolate (0.87 g, 6.8 mmol) was added as a solid in por-

tions to a solution of 2-fluoroethanol (0.08 mL, 1.4 mmol) and 4

(0.57 g, 1.7 mmol) in THF (2 mL) under ice cooling. The reaction

mixture was stirred for 2 h in the ice bath and subsequently dilut-

ed with ice-cold H₂O (30 mL) and CH₂Cl₂(20 mL). The aqueous

layer was extracted with CH₂Cl₂(2ꢂ20 mL) and the combined or-

ganic phases were washed with sat. NaHCO₃(20 mL) and brine

(20 mL), dried over Na₂SO₄and evaporated in vacuo to yield 6c

(0.28 g, 57%) as a yellow solid; mp: 67–698C (lit.:^[35]68–708C);

¹H NMR (400 MHz, CDCl₃): d=4.35 (dt, ³J_H,F=27.1 Hz, ³J=4.0 Hz,

1

3

38–408C (lit.:^[53]348C); H NMR (400 MHz, CDCl₃): d=7.82 (dd, J=

4

3

2

3

8.5 Hz, J=2.3 Hz, 1H, H-6), 7.89 (d, J=8.5 Hz, 1H, H-5), 8.26 ppm

2H, H-7a, H-7b), 4.61 (dt, J_H,F=47.1 Hz, J=4.1 Hz, 2H, H-8a, H-8b),

(d, ⁴J=2.2 Hz, 1H, H-2).

7.70 (dd, J=8.4 Hz, J=2.2 Hz, 1H, H-6), 7.83 (d, J=8.4 Hz, 1H, H-

3

4

3

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&12

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

5), 8.16 ppm (d, ⁴J=2.1 Hz, 1H, H-2); ¹³C NMR (101 MHz, CDCl₃):

d=69.25 (C-7), 81.38 (C-8), 126.26, 127.53, 132.15, 132.96, 134.75,

136.26 ppm (C_arom); elemental analysis: calcd for C₈H₇Br₂FO₃S: C

26.54, H 1.95, S 8.86; found: C 27.13, H 1.91, S 8.98.

distillation with CH₃CN in a stream of nitrogen gas at 908C with

stirring. In total, 4ꢂ2 mL of CH₃CN were added and the remaining

residue was dissolved in the corresponding reaction medium.

Preparation of 2-[¹⁸F]fluoroethylbenzenesulfonates: A solution of

the appropriate ethylene dibenzenesulfonate (5a–c) in CH₃CN

(1 mL) was added to the [¹⁸F]KF-containing residue obtained by

azeotropic distillation and brought to the reaction temperature of

908C in the screw-cap-sealed vial. To optimise the reaction with

regard to the amount of precursor and time, 10 mL aliquots of the

reaction were withdrawn after 5, 10 and 15 min, diluted with

CH₃CN (90 mL) and analysed by radio-TLC.

(S)-2-(4-(2-(Benzyloxycarbonylamino)-3-(2-cyano-1,2-dimethylhy-

drazinyl)-3-oxopropyl)phenoxy)ethyl-4-methylbenzenesulfonate

(7a): Ethylene glycol-1,2-ditosylate (5a; 0.35 g, 0.94 mmol) and po-

tassium carbonate (0.04 mg, 0.3 mmol) were added as solids to

a solution of 1 (0.2 g, 0.52 mmol) in CH₃CN (6 mL). The resulting

mixture was heated at reflux for 6 h. After cooling, the reaction

mixture was acidified with 1m HCl (20 mL) and extracted with

EtOAc (3ꢂ30 mL). The combined organic layers were washed with

brine and dried over Na₂SO₄. The crude product obtained after

evaporation of the solvent was purified by column chromatogra-

phy (CH₂Cl₂/MeOH, 150:1) to yield 7a (0.11 g, 37%) as a colourless

Conditions for ¹⁸F-fluoroethylations: A solution of 1 (1.5-fold

excess related to the corresponding ethylene dibenzenesulfonate)

in CH₃CN (or DMF; 100 mL) was added to the solution containing

the 2-[¹⁸F]fluoroethylbenzenesulfonate, followed by a 1m solution

of Cs₂CO₃in H₂O (volume corresponding to an amount equimolar

to 1). The conditions were specified in the Results and Discussion

section. When the ¹⁸F-fluoroethylation was performed in DMF, the

solution of [¹⁸F]6a in CH₃CN obtained after reaction with

[¹⁸F]fluoride was evaporated in the nitrogen stream and the ob-

tained residue was redissolved in DMF (1 mL).

1

waxy solid; mp: 40–448C; H NMR (400 MHz, CDCl₃, mixture of ro-

tamers, only the signals of the major rotamer are listed): d=2.45

(s, 3H, Ts-CH₃), 2.83 (dd, ²J=14.1 Hz, ³J=8.1 Hz, 1H, C_bHH), 3.05

2

3

(dd, J=14.0 Hz, J=5.5 Hz, 1H, C_bHH), 3.20 (s, 3H, NCH₃), 3.24 (s,

3H, NCH₃), 4.09–4.14 (m, 2H, CH₂OTs), 4.31–4.38 (m, 2H, PhOCH₂),

3

4.95–5.18 (m, 3H, PhCH₂O, C_aH), 5.26 (d, J=8.4 Hz, 1H, NH), 6.72

3

(d, J=8.4 Hz, 2H, PhO-oH), 7.06 (d, J=8.4 Hz, 2H, PhO-mH), 7.26–

3

7.40 (m, 7H, Ph, Ts-H-3’, Ts-H-5’), 7.82 ppm (d, J=8.3 Hz, 2H, Ts-H-

Conversion of the direct precursors 7a and 7b with

[¹⁸F]fluoride: A solution containing 7a or 7b (1, 3 or 5 mg) in

CH₃CN (1 mL) was added to the [¹⁸F]KF-containing residue ob-

tained by azeotropic distillation and brought to the reaction tem-

perature of 908C. Aliquots (10 mL) were withdrawn after 5, 10 and

15 min, diluted with CH₃CN (90 mL each) and analysed by radio-

TLC.

2’, Ts-H-6’); ¹³C NMR (101 MHz, CDCl₃): d=21.81 (C-24), 30.61, 37.78,

41.43 (C-10, C-11, C-13), 52.17 (C-8), 65.62, 67.33, 68.23 (C-5, C-18,

C-19), 113.56 (C-12), 115.10 (C-16, C-16’), 128.18, 128.39, 128.68,

129.99, 130.10, 130.44, 130.76, 136.11, 145.06, 145.41 (C_arom), 156.07,

157.49 (C-17, C-6), 173.40 ppm (C-9); elemental analysis: calcd for

C₂₉H₃₂N₄O₇S: C 59.99, H 5.55, N 9.65, S 5.54; found: C 59.38, H 5.55,

N 9.19, S 5.59.

Preparation of [¹⁸F]3 by the optimised procedure: A solution of

5b (5 mg, 11.6 mmol) in CH₃CN (1 mL) was added to the [¹⁸F]KF-

containing residue obtained by azeotropic distillation and brought

to the reaction temperature of 908C for 5 min. The reaction solu-

tion was cooled to 48C. A solution of 1 (6.65 mg, 15.5 mmol) in

CH₃CN (100 mL) was added, followed by 1m Cs₂CO₃in H₂O

(15.5 mL). The resulting mixture was heated at 1158C for 10 min

and subsequently cooled to 48C. The solution was diluted with

0.1% CF₃COOH/H₂O (1 mL) and filtered. The filtrate was transferred

into a hot cell and purified by semi-preparative radio-HPLC. The

peak representing the product was collected and immediately di-

luted with H₂O (30 mL). The resulting solution was subjected to

solid-phase extraction by using a 500 mg LiChrolut RP-18E (40–

63 mm) cartridge. The cartridge was washed with H₂O (3 mL) and

the product was eluted with 2ꢂ0.5 mL of EtOH. The product solu-

tion was completely evaporated and the obtained residue was re-

dissolved in EtOH (55 mL). A 5 mL aliquot of the final product solu-

tion was withdrawn, diluted with CH₃CN (50 mL) and used for quali-

ty control by radio-HPLC and radio-TLC.

(S)-2-(4-(2-(Benzyloxycarbonylamino)-3-(2-cyano-1,2-dimethylhy-

drazinyl)-3-oxopropyl)phenoxy)ethyl-4-nitrobenzenesulfonate

(7b): The procedure described for the synthesis of 7a was fol-

lowed, with the exception that 5b (0.40 g, 0.94 mmol) was used in-

stead of 5a and the reaction time was decreased to 2 h. Com-

pound 7b (0.10 g, 31%) was obtained as a yellow solid; mp: 74–

1

798C; H NMR (400 MHz, CDCl₃, mixture of rotamers, only the sig-

2

3

nals of the major rotamer are listed): d=2.84 (dd, J=14.0 Hz, J=

2

3

7.7 Hz, 1H, C_bHH), 3.03 (dd, J=14.0 Hz, J=5.7 Hz, 1H, C_bHH), 3.19

(s, 3H, NCH₃), 3.23 (s, 3H, NCH₃), 4.12–4.18 (m, 2H, CH₂ONs), 4.46–

3

4.52 (m, 2H, PhOCH₂), 4.96–5.18 (m, 3H, PhCH₂O, C_aH), 5.27 (d, J=

8.8 Hz, 1H, NH), 6.65 (d, ³J=8.5 Hz, 2H, PhO-oH), 7.05 (d, ³J=

3

8.5 Hz, 2H, PhO-mH), 7.38–7.27 (m, 5H, Ph), 8.11 (d, J=8.8 Hz, 2H,

Ns-H-2’, Ns-H-6’), 8.35 ppm (d, ³J=8.5 Hz, 2H, Ns-H-3’, Ns-H-5’);

¹³C NMR (101 MHz, CDCl₃): d=30.57, 37.83, 41.42 (C-10, C-11, C-13),

52.10 (C-8), 65.40, 67.23, 69.60 (C-5, C-18, C-19), 113.51 (C-12),

114.98 (C-16, C-16’), 124.55, 128.16, 128.41, 128.62, 128.69, 129.42,

130.52, 130.82 (C_arom), 141.95, 150.91, 156.06, 157.12 (C-20, C-23, C-

17, C-6), 173.35 ppm (C-9); elemental analysis: calcd for

C₂₈H₂₉N₅O₉S: C 54.98, H 4.78, N 11.45, S 5.24; found: C 53.06, H

5.04, N 10.77, S 4.85.

Enzyme inhibition assays

Cathepsin assays were performed as described previously.^[31a]

Radiosynthesis

Immunohistochemistry of tumour and organ sections

Processing of [¹⁸F]fluoride for radiosynthesis: No-carrier-added

aqueous [¹⁸F]HF was produced in a IBA CYCLONE 18/9 cyclotron

by irradiation of [¹⁸O]H₂O through the ¹⁸O(p,n)¹⁸F nuclear reaction.

The aqueous [¹⁸F]fluoride (500–3000 MBq) was adsorbed on an

anion-exchange cartridge (QMA plus, Waters) and eluted with a so-

lution (1 mL) of 26.6 mm Kryptofix 222 and 13.4 mm K₂CO₃in 14%

H₂O/CH₃CN into a conical glass vial equipped with a screw cap and

rubber septum. Removal of H₂O was accomplished by azeotropic

Tumour and organ pieces were fixed in 4% phosphate-buffered

paraformaldehyde, paraffin-embedded, cut into 5 mm sections and

mounted onto SuperFrostPlus object slides. Tissue sections were

rehydrated and pretreated in 500 mm tris(hydroxymethyl)amino-

methane (Tris)/10 mm ethylenediaminetetraacetate (EDTA; pH 9.0)

or 10 mm citrate buffer (pH 6.0, staining of cathepsin S) heated for

20 min by microwave. After blocking unspecific binding sites with

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&13

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

a solution containing 0.3% milk powder, 0.1% gelatin and 2%

bovine serum albumin in Tris-buffered saline containing 0.05%

Tween 20 (TBS-0.05% Tween 20; pH 7.4), the primary antibody was

incubated at 48C overnight. Cathepsins were visualised by using

antibodies from Abcam (cathepsin L: mouse monoclonal #ab6314,

dilution 1:100; cathepsin S: goat polyclonal #ab18822, dilution

1:40; cathepsin B: rabbit polyclonal #ab33538, dilution 1:200; cath-

epsin K: rabbit polyclonal #ab19027, dilution 1:100; typical vol-

umes were 50–100 mL each). After washing with TBS-0.05%

Tween 20 (pH 7.4; 3ꢂ10 min), primary antibodies were visualised

with the corresponding biotinylated secondary antibodies goat

anti-rabbit, goat anti-mouse or donkey anti-goat (Dianova, dilution

1:200, 1–2 h incubation time), followed by incubation with ExtrAvi-

din peroxidase (Sigma–Aldrich) and the AEC substrate kit (BD Phar-

mingen). Finally, tumour sections were counterstained by using

Mayer’s Hematoxylin and embedded in glycerol. Slides were ana-

lysed by using an AxioImager.A1 microscope and the appropriate

AxioVision software package (Carl Zeiss MicroImaging).

The metabolite fraction in the blood was calculated as a percentage

of the parent compound in the blood from the HPLC. The extracts

from other tissues, urine or intestine content were prepared by ul-

trasound homogenisation as a 10% solution in PBS, centrifugation

(58C, 70000 gꢂmin) and subsequent precipitation by 60% acetoni-

trile. The supernatant was analysed by the described radio-HPLC.

Experiments to suppress conjugate formation between [¹⁸F]3 and

glutathione were performed with whole rat blood as noted above.

Solutions of 10 mm buthionine sulfoximine (BSO; purchased from

Sigma, diastereomeric mixture with varying chirality at the sulfur

atom) in PBS and 50 mm etacrynic acid in EtOH were freshly pre-

pared. Blood samples (each 400 mL) were treated with BSO (80, 160

or 240 mL) and incubated for 2 h (final concentrations of BSO: 1.67,

2.56 and 3.75 mm, respectively). [¹⁸F]3 was added and the samples

were analysed by radio-HPLC as described above. In a further ex-

periment, etacrynic acid (40 mL) was added to each of the blood

samples in addition to BSO (final concentrations of etacrynic acid:

3.85, 3.33 and 2.94 mm, respectively). Independently, blood sam-

ples were treated with diethylmaleate (DEM) in amounts that were

2% of the sample volume and incubated for 1 h before addition of

[¹⁸F]3.

Chemical stability, distribution coefficient, in vitro stability,

and cell uptake experiments

Chemical stability: [¹⁸F]3 (10–20 MBq) dissolved in EtOH was dilut-

ed 1:10 or 1:20 in different buffered aqueous media and incubated

at 378C in a thermomixer. Aliquots of 20 mL were withdrawn after

0, 1, 2, 3, 4 and 5 h, diluted with CH₃CN (100 mL) and spotted on

a TLC plate (10 mL). In addition, non-radioactive 3 was spotted. The

TLC plate was developed in petroleum ether/EtOAc (2:1) and vi-

sualised under UV as well as by radioluminography, as described

above. The aqueous media used were 0.1m sodium phosphate

(pH 6.0), 10ꢂPBS (pH 7.4), and 10ꢂTBS (pH 8.0).

In vivo experiments

Animal experiments were carried out according to the guidelines

of the German Regulations for Animal Welfare. The protocol was

approved by the local Ethical Committee for Animal Experiments.

Besides the in vivo stability studies, biodistribution and small-

animal PET experiments were also performed in male Wistar rats.

For the in vivo stability studies, rats (aged 7–9 weeks, (183ꢁ56) g

(mean body weight ꢁ standard deviation (SD)), n=19) were anaes-

thetised with desflurane (initially 9%, then 6% in 20% oxygen/air).

The guide value for breathing frequency was 65 breathsmin^ꢀ1. Ani-

mals were put in the supine position and placed on a heating pad

to maintain body temperature. The spontaneously breathing rats

were treated with 100 unitskg^ꢀ1heparin (Heparin-Natrium 25000-

ratiopharm, Ratiopharm, Germany) by subcutaneous injection to

prevent blood clotting on intravascular catheters. After local anaes-

thesia by injection of 1% Lignocaine (Xylocitin loc, Mibe, Jena, Ger-

many) into the right groin, a catheter was introduced into the

right femoral artery (0.8 mm Umbilical Vessel Catheter, Tyco Health-

care, Tullamore, Ireland) for blood samples for metabolite analysis,

for gas analysis and for arterial blood pressure measurements;

a second needle (35 G) catheter in one tail vein was used for ad-

ministration of the [¹⁸F]3.

Water–octanol distribution coefficient: [¹⁸F]3 (~5 MBq) dissolved

in 10% EtOH/H₂O (100 mL) was added to a mixture of n-octanol

(5 mL) and H₂O (5 mL) in a shaking flask. After vigorous shaking,

the separated aqueous phase was discarded and H₂O (5 mL) was

added to the shaking flask. After shaking again, 3ꢂ100 mL of each

layer were transferred to scintillation vials, each containing H₂O

(1 mL), and the activities (A) were measured in a g counter. The

values were averaged and the logP value was calculated according

to logP=log(A(organic phase)/A(aqueous phase)).

In vitro and in vivo stability: In order to investigate in vitro stabili-

ty, [¹⁸F]3 was incubated at 378C with venous blood obtained from

male Wistar rats. The time points of analysis corresponded to

those used in the studies of in vivo stability. The latter were carried

out on arterial blood samples that were collected from the right

femoral artery at 1, 3, 5, 10, 20, 30 and 60 min after intravenous ad-

ministration of [¹⁸F]3.

From rat heparin blood, erythrocyte ghosts were also prepared.

Heparin blood was centrifuged for 5000 gꢂmin at 48C and the

plasma and white cells were discarded. Packed erythrocytes were

washed three times with isotonic chloride solution and, each time,

the buffy coat was carefully removed. Red blood cells (1 mL) were

added to deionised water (3 mL) at 48C and gently vortexed. At

the end of the hypotonic hemolysis, the isotonicity was restored

with isotonic sodium chloride (0.5 mL) and the erythrocytes were

thus allowed to restore during 30 min. The restored cells were

then washed three times with PBS and used within 4 h.

Blood cells were separated by centrifugation (58C, 5 min,

8000 rpm) and plasma proteins were precipitated by using 60%

acetonitrile and subsequent centrifugation (58C, 5 min, 8000 rpm).

The supernatant was analysed by radio-HPLC. The radio-HPLC

system (Agilent 1100 series) applied for metabolite analysis was

equipped with UV detection (254 nm) and an external radiochemi-

cal detector (RAMONA, Raytest GmbH, Straubenhardt, Germany).

Analysis was performed on a Zorbax C₁₈300SB (250ꢂ9.4 mm;

4 mm) column with an eluent system comprising C (H₂O with 0.1%

trifluoroacetic acid (TFA)) and D (acetonitrile with 0.1% TFA) with

a gradient of 5 min 95% C, 10 min to 95% D and 5 min at 95% D

at a flow rate of 3 mLmin^ꢀ1. HPLC were performed on [¹⁸F]3 added

to a rat blood sample, on arterial blood samples from 1 to 60 min

after injections and on a urine sample from 60 min after injection.

For biodistribution experiments, two groups of four rats (aged 7 to

9 weeks; mean body weightꢁSD: (186ꢁ11) g) for each time point

were intravenously injected into a tail vein with [¹⁸F]3 (~0.3 MBq)

dissolved in E-153 electrolyte solution (0.5 mL). The specific activity

ranged between 5 and 30 GBqmmol^ꢀ1at the time of injection. Ani-

mals were sacrificed at 5 and 60 min post injection. Blood and the

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&14

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

major organs were collected, weighed and counted in a Wallac

WIZARD automatic g counter (Perkin–Elmer, Germany). The radio-

activity of the tissue samples was decay corrected and calibrated

by comparing the counts in tissue with the counts in aliquots of

the injected tracer that had been measured in the g counter at the

same time. The activity amount in the selected tissues and organs

was expressed as percent injected dose (%ID). The activity concen-

tration in the biodistribution measurements was calculated as

SUV_bio(g/g, SUV=(activity per g of tissue)/(injected activity/body

weight)) and expressed as the meanꢁSD for each group of sixteen

animals.

tivity only minimally changed over the time of measurement. The

ROI data and TAC were further analysed by using R (R is available

as Free Software under the terms of the Free Software Founda-

tion’s GNU General Public License in source code form) and espe-

cially developed program packages (Jçrg van den Hoff, Helmholtz-

Zentrum Dresden-Rossendorf, Dresden, Germany). The standar-

dised uptake values (SUV=(activity per mL of tissue)/(injected ac-

tivity/body weight), gmL^ꢀ1)) were calculated in the ROIs.

The standard reversible two-tissue compartment model^[49]was

fitted to the mean tumour time–activity curve of nine animals,

with the mean blood time–activity derived input curve of nine ani-

mals. In the parameter estimation, the fractional blood volume and

tracer delay at the tumour were included. The fitting was per-

formed by the nonlinear “compfit” algorithm implemented in R by

J. van den Hoff.

Additionally, small-animal PET experiments were performed in

normal NMRI mice and NCI-H292 xenotransplanted NMRI nu/nu

mice. With this aim, female NMRI (nu/nu) mice (10–14 weeks old)

were purchased from the specific pathogen-free breeding facility

of the Experimental Centre of the Medical Faculty Carl Gustav

Carus, University of Technology, Dresden. For the generation of

subcutaneous tumours, NCI-H292 cells (ATCC CRL-1848) were used.

Tumour cells were harvested, washed in PBS and transferred to

0.9% sodium chloride. Only single-cell suspensions (5ꢂ10⁶cells

per 100 mL) with at least 90% viability (trypan blue exclusion) were

used for injection. Animals were anaesthetised by using desflurane

(10% v/v in 30% oxygen/air) and tumour cells were subcutaneous-

ly injected into the right hind leg. Tumour size was monitored

three times a week by calliper measurements and the tumour

volume was calculated by using the formula V=p/6ꢂ(tumour

length ꢂ tumour width ꢂ tumour width). Animals entered the

imaging study 15–20 days after tumour-cell injection with tumour

volumes of 100–400 mm³.

Statistical analyses

Data are expressed as meanꢁSD. Values were compared by using

an unpaired Student’s t-test with Welch’s correction and an F-test

to compare the variances (GraphPad Prism 5.02 for Windows,

GraphPad Software, San Diego, CA). The non-parametric Wilcoxon

signed rank test and the D’Agostino–Pearson normality test were

used for statistical evaluation for some of the data. A P value of

<0.05 was considered significant and is indicated by an asterisk.

Acknowledgements

For small-animal PET experiments, anaesthetised, spontaneously

breathing animals were allowed to stabilise for 10 min after prepa-

ration. A 10 min transmission scan was recorded during this time

for each subject by using a rotating point source of ⁵⁷Co (microPET

P4; Siemens preclinical solutions, Knoxville, TN) or a whole-body

CT was recorded (NanoScan PET/CT; Mediso Budapest, Hungary).

The transmission scans were used to correct the emission scan for

g-ray attenuation caused by body tissues and supporting struc-

tures; they were also used to demarcate the body field for image

registration. The activity of the injection solution was measured in

a well counter (Isomed 2000, Dresden, Germany) cross-calibrated

to the small-animal PET scanners. The PET acquisition of 60- or

120 min emission scans was started and the infusion of the [¹⁸F]3

into a tail vein was initiated with a delay of 30 s. At the end of the

experiment, the animals were deeply anaesthetised and sacrificed

by an intravenous injection of potassium chloride.

The dedicated assistance of Andrea Suhr, Natalie Thieme, and

Regina Herrlich in the animal experiments is gratefully acknowl-

edged. We cordially appreciate the expert support of Catharina

Heinig, Aline Morgenegg, and Mareike Barth in the immunohisto-

chemical analyses, as well as that of Stephan Preusche and his

team for providing [¹⁸F]fluoride. Further thanks are due to Dr.

Torsten Knieß for radiochemical advice. This work is part of a re-

search initiative within the Helmholtz-Portfoliothema “Technolo-

gie und Medizin—Multimodale Bildgebung zur Aufklꢁrung des In-

vivo-Verhaltens von polymeren Biomaterialien”.

Keywords: cathepsins

· fluorine · glutathione · imaging

agents · positron emission tomography · radiochemistry

[1] a) J. M. Rothberg, M. Sameni, K. Moin, B. F. Sloane, Biochim. Biophys.

Acta Proteins Proteomics 2012 , 1824 , 12 3 –1 32; b) S. D. Mason, J. A.

Joyce, Trends Cell Biol. 2011 , 21 , 22 8 –2 37.

[2] E. I. Deryugina, J. P. Quigley, Cancer Metastasis Rev. 2006 , 25 , 9 –3 4.

[3] F. Lecaille, J. Kaleta, D. Brçmme, Chem. Rev. 2002 , 102 , 445 9 –4 488.

[4] J. A. Joyce, A. Baruch, K. Chehade, N. Meyer-Morse, E. Giraudo, F. Y. Tsai,

D. C. Greenbaum, J. H. Hager, M. Bogyo, D. Hanahan, Cancer Cell 2004 ,

5 , 44 3 –4 53.

[5] a) T. Nomura, N. Katunuma, J. Med. Invest. 2005 , 52 , 1 –9; b) M. M. Mo-

hamed, B. F. Sloane, Nat. Rev. Cancer 2006 , 6 , 76 4 –7 75; c) T. Reinheckel,

C. Peters, A. Krꢁger, B. Turk, O. Vasiljeva, Front. Pharmacol. 2012, 3, 133.

[6] V. Gocheva, J. A. Joyce, Cell Cycle 2007 , 6 , 6 0 –6 4.

[7] a) J. Kos, T. T. Lah, Oncol. Rep. 1998, 5, 1349–1361; b) J. Kos, B. Werle, T.

Lah, N. Brunner, Int. J. Biol. Markers 2000, 15, 84–89; c) I. Berdowska,

Clin. Chim. Acta 2004 , 342 , 4 1 –6 9.

Data acquisition was performed in 3D list mode. Emission data

were collected continuously. The list mode data were sorted with

32 or 38 frames (15ꢂ10 s, 5ꢂ30 s, 5ꢂ60 s, 4ꢂ300 s, 3ꢂ600 s or

9ꢂ600 s). The data were decay, scatter and attenuation corrected.

The frames were reconstructed by Ordered Subset Expectation

Maximisation applied to 3D sinograms (OSEM3D). The voxel size

was 0.06ꢂ0.06ꢂ0.12 cm and the resolution in the centre of the

field of view was equal to or lower than 1.8 mm. The image

volume data were converted into the Siemens ECAT7 format for

further processing. The image files were then processed by using

the ROVER software (ABX GmbH, Radeberg, Germany). Masks for

defining three-dimensional regions of interest (ROIs) were set and

the ROIs were defined by thresholding; ROI time–activity curves

(TAC) were derived for the subsequent data analysis. The input

curves were calculated from the ROIs over the vena cava by using

a recovery of 26% and the ROIs over the tumours. No especially

partial volume correction was applied because the background ac-

[8] a) E. Frçhlich, B. Schlagenhauff, M. Mçhrle, E. Weber, C. Klessen, G. Rass-

ner, Cancer 2001, 91, 972–982; b) P. Matarrese, B. Ascione, L. Ciarlo, R.

Vona, C. Leonetti, M. Scarsella, A. M. Mileo, C. Catricala, M. G. Paggi, W.

Malorni, Mol. Cancer 2010 , 9 , 207.

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&15

&

ÞÞ

These are not the final page numbers!

CHEMMEDCHEM

FULL PAPERS

www.chemmedchem.org

[9] a) B. Gole, M. B. D. Alonso, V. Dolenc, T. Lah, Pathol. Oncol. Res. 2009 , 15 ,

71 1 –7 23; b) C. Colin, B. Voutsinos-Porche, I. Nanni, F. Fina, P. Metellus,

D. Intagliata, N. Baeza, C. Bouvier, C. Delfino, A. Loundou, O. Chinot, T.

Lah, J. Kos, P. M. Martin, L. Ouafik, D. Figarella-Branger, Acta Neuropa-

thol. 2009 , 118 , 74 5 –7 54.

[29] a) M. Frizler, M. Stirnberg, M. T. Sisay, M. Gꢁtschow, Curr. Top. Med.

Chem. 2010, 10, 294–322; b) F. F. Fleming, L. Yao, P. C. Ravikumar, L.

Funk, B. C. Shook, J. Med. Chem. 2010 , 53 , 790 2 –7 917; c) V. Turk, V.

Stoka, O. Vasiljeva, M. Renko, T. Sun, B. Turk, D. Turk, Biochim. Biophys.

Acta Proteins Proteomics 2012 , 1824 , 6 8 –8 8.

[10] a) B. F. Sloane, S. Q. Yan, I. Podgorski, B. E. Linebaugh, M. L. Cher, J. X.

Mai, D. Cavallo-Medved, M. Sameni, J. Dosescu, K. Moin, Semin. Cancer

Biol. 2005 , 15 , 14 9 –1 57; b) C. J. Watson, P. A. Kreuzaler, J. Mammary

Gland Biol. Neoplasia 2009 , 14 , 17 1 –1 7 9; c) L. Sevenich, U. Schurigt, K.

Sachse, M. Gajda, F. Werner, S. Mꢁller, O. Vasiljeva, A. Schwinde, N.

Klemm, J. Deussing, C. Peters, T. Reinheckel, Proc. Natl. Acad. Sci. USA

2010 , 107 , 249 7 –2 502; d) L. Sevenich, F. Werner, M. Gajda, U. Schurigt,

C. Sieber, S. Mꢁller, M. Follo, C. Peters, T. Reinheckel, Oncogene 2011 , 30 ,

5 4 –6 4; e) N. P. Withana, G. Blum, M. Sameni, C. Slaney, A. Anbalagan,

M. B. Olive, B. N. Bidwell, L. Edgington, L. Wang, K. Moin, B. F. Sloane,

R. L. Anderson, M. S. Bogyo, B. S. Parker, Cancer Res. 2012 , 72 , 119 9 –

1209.

[30] R. Lçser, M. Frizler, K. Schilling, M. Gꢁtschow, Angew. Chem. 2008 , 120 ,

440 3 –4 406; Angew. Chem. Int. Ed. 2008 , 47 , 433 1 –4 334.

[31] a) M. Frizler, F. Lohr, N. Furtmann, J. Klꢃs, M. Gꢁtschow, J. Med. Chem.

2011 , 54 , 39 6 –4 00; b) M. Frizler, F. Lohr, M. Lꢁlsdorff, M. Gꢁtschow,

Chem. Eur. J. 2011 , 17 , 1141 9 –1 142 3; c) M. Frizler, J. Schmitz, A.-C.

Schulz-Fincke, M. Gꢁtschow, J. Med. Chem. 2012 , 55 , 598 2 –5 986; d) M.

Frizler, M. D. Mertens, M. Gꢁtschow, Bioorg. Med. Chem. Lett. 2012 , 22 ,

771 5 –7 718; e) P. A. Ottersbach, J. Schmitz, G. Schnakenburg, M. Gꢁt-

schow, Org. Lett. 2013 , 15 , 44 8 –4 51; f) X. F. Ren, H. W. Li, X. Fang, Y. Wu,

L. Wang, S. Zou, Org. Biomol. Chem. 2013 , 11 , 114 3 –1 148.

[32] a) R. Lçser, J. Gut, P. J. Rosenthal, M. Frizler, M. Gꢁtschow, K. T. Andrews,

Bioorg. Med. Chem. Lett. 2010 , 20 , 25 2 –2 55; b) Y. H. Loh, H. B. Shi, M. Y.

Hu, S. Q. Yao, Chem. Commun. 2010 , 46 , 840 7 –8 409; c) P. Y. Yang, M.

Wang, L. Li, H. Wu, C. Y. He, S. Q. Yao, Chem. Eur. J. 2012 , 18 , 652 8 –

6541.

[11] K. Ravanko, K. Jꢃrvinen, J. Helin, N. Kalkkinen, E. Hçlttꢃ, Cancer Res.

2004 , 64 , 883 1 –8 838.

[12] N. Rousselet, L. Mills, D. Jean, C. Tellez, M. Bar-Eli, R. Frade, Cancer Res.

2004 , 64 , 14 6 –1 51.

[33] W. J. Middleton, J. Org. Chem. 1975 , 40 , 57 4 –5 78.

[13] J. Kos, A. Sekirnik, G. Kopitar, N. Cimerman, K. Kayser, A. Stremmer, W.

Fiehn, B. Werle, Br. J. Cancer 2001 , 85 , 119 3 –1 200.

[14] J. A. Gormley, S. M. Hegarty, A. O’Grady, M. R. Stevenson, R. E. Burden,

H. L. Barrett, C. J. Scott, J. A. Johnston, R. H. Wilson, E. W. Kay, P. G. John-

ston, S. A. Olwill, Br. J. Cancer 2011 , 105 , 148 7 –1 494.

[15] S. Y. Ardebili, I. Zajc, B. Gole, B. Campos, C. Herold-Mende, S. Drmota,

T. T. Lah, Radiol. Oncol. 2011, 45, 102–115.

[16] P. Garnero, O. Borel, I. Byrjalsen, M. Ferreras, F. H. Drake, M. S. McQue-

ney, N. T. Foged, P. D. Delmas, J. M. Delaisse, J. Biol. Chem. 1998 , 273 ,

3234 7 –3 2352.

[17] D. Brçmme, K. Okamoto, B. B. Wang, S. Biroc, J. Biol. Chem. 1996, 271,

2126–2132.

[18] K. D. Brubaker, R. L. Vessella, L. D. True, R. Thomas, E. Corey, J. Bone

Miner. Res. 2003 , 18 , 22 2 –2 30.

[19] I. Rapa, M. Volante, S. Cappia, R. Rosas, G. V. Scagliotti, M. Papotti, Am. J.

Clin. Pathol. 2006 , 125 , 84 7 –8 54.

[34] a) W. Wadsak, L. K. Mien, D. E. Ettlinger, H. Eidherr, D. Haeusler, K. M. Sin-

delar, B. K. Keppler, R. Dudczak, K. Kletter, M. Mitterhauser, Nucl. Med.

Biol. 2007 , 34 , 101 9 –1 02 8; b) M. R. Zhang, K. Suzuki, Curr. Top. Med.

Chem. 2007, 7, 1817–1828.

[35] J. L. Musachio, J. Shah, V. W. Pike, J. Labelled Compd. Radiopharm. 2005 ,

48 , 73 5 –7 47.

[36] O. Prante, R. Tietze, C. Hocke, S. Lçber, H. Hꢁbner, T. Kuwert, P. Gmeiner,

J. Med. Chem. 2008 , 51 , 180 0 –1 810.

[37] a) H.-H. Heidtmann, U. Salge, K. Havemann, H. Kirschke, B. Wiederand-

ers, Oncol. Res. 1993, 5, 441–451; b) B. Werle, W. Ebert, W. Klein, E.

Spiess, Biol. Chem. Hoppe-Seyler 1995 , 376 , 15 7 –1 64; c) P. Ledakis, W. T.

Tester, N. Rosenberg, D. Romero-Fischmann, I. Daskal, T. T. Lah, Clin.

Cancer Res. 1996, 2, 561–568; d) E. Krepela, Neoplasma 2001, 48, 332–

349; e) B. Werle, M. Kotzsch, T. T. Lah, J. Kos, D. Gabrijelcic-Geiger, E.

Spiess, J. Schirren, W. Ebert, W. Fiehn, T. Luther, V. Magdolen, M. Schmitt,

N. Harbeck, Anticancer Res. 2004, 24, 4147–4161.

[20] C. G. Kleer, N. Bloushtain-Qimron, Y. H. Chen, D. Carrasco, M. Hu, J. Yao,

S. K. Kraeft, L. C. Collins, M. S. Sabel, P. Argani, R. Gelman, S. J. Schnitt,

I. E. Krop, K. Polyak, Clin. Cancer Res. 2008 , 14 , 535 7 –5 367.

[21] a) B. F. Sloane, M. Sameni, I. Podgorski, D. Cavallo-Medved, K. Moin,

Annu. Rev. Pharmacol. Toxicol. 2006 , 46 , 30 1 –3 15; b) G. Blum, Curr. Opin.

Drug Discovery Dev. 2008, 11, 708–716; c) T. E. McCann, N. Kosaka, B.

Turkbey, M. Mitsunaga, P. L. Choyke, H. Kobayashi, NMR Biomed. 2011,

24, 561–568; d) S. Serim, U. Haedke, S. H. L. Verhelst, ChemMedChem

2012 , 7 , 114 6 –1 159.

[38] a) A. A. Wilson, L. Jin, A. Garcia, J. N. DaSilva, S. Houle, Appl. Radiat. Isot.

2001 , 54 , 20 3 –2 08; b) R. N. Waterhouse, Mol. Imaging Biol. 2003 , 5 ,

37 6 –3 89.

[39] P. H. Hinderling, Pharmacol. Rev. 1997, 49, 279–295.

[40] a) K. N. Dalby, W. P. Jencks, J. Chem. Soc. Perkin Trans. 2 1997 , 155 5 –

1563; b) P. A. MacFaul, A. D. Morley, J. J. Crawford, Bioorg. Med. Chem.

Lett. 2009 , 19 , 113 6 –1 138.

[41] A. Pastore, G. Federici, E. Bertini, F. Piemonte, Clin. Chim. Acta 2003 , 333 ,

1 9 –3 9.

[22] C. Bremer, C. H. Tung, A. Bogdanov, R. Weissleder, Radiology 2002 , 222 ,

81 4 –8 18.

[23] G. Blum, G. von Degenfeld, M. J. Merchant, H. M. Blau, M. Bogyo, Nat.

Chem. Biol. 2007 , 3 , 66 8 –6 77.

[24] a) R. W. Mason, D. Wilcox, P. Wikstrom, E. N. Shaw, Biochem. J. 1989, 257,

125–129; b) M. Bogyo, S. Verhelst, V. Bellingard-Dubouchaud, S. Toba,

D. Greenbaum, Chem. Biol. 2000 , 7 , 2 7 –3 8.

[42] O. W. Griffith, A. Meister, J. Biol. Chem. 1979, 254, 7558–7560.

[43] O. W. Griffith, J. Biol. Chem. 1982, 257, 13704–13712.

[44] H. C. Palfrey, S. Leung, Am. J. Physiol. 1993, 264, C1270–C1277.

[45] A. Gupta, S. Seifert, R. Syhre, M. Scheunemann, P. Brust, B. Johannsen,

Radiochim. Acta 2001 , 89 , 4 3 –4 9.

[46] J. N. Commandeur, G. J. Stijntjes, N. P. Vermeulen, Pharmacol. Rev. 1995,

47, 271–330.

[25] T. Quillard, K. Croce, F. A. Jaffer, R. Weissleder, P. Libby, Thromb. Haemo-

stasis 2011 , 105 , 82 8 –8 36.

[47] K. L. Davison, G. L. Larsen, Xenobiotica 1993 , 23 , 29 7 –3 05.

[48] J. H. Lin, Drug Metab. Dispos. 1995, 23, 1008–1021.

[26] a) H. J. Breyholz, S. Wagner, A. Faust, B. Riemann, C. Holtke, S. Hermann,

O. Schober, M. Schꢃfers, K. Kopka, ChemMedChem 2010 , 5 , 77 7 –7 89;

b) U. auf dem Keller, C. L. Bellac, Y. Li, Y. Lou, P. F. Lange, R. Ting, C.

Harwig, R. Kappelhoff, S. Dedhar, M. J. Adam, T. J. Ruth, F. Benard, D. M.

Perrin, C. M. Overall, Cancer Res. 2010 , 70 , 756 2 –7 56 9; c) V. Hugenberg,

H. J. Breyholz, B. Riemann, S. Hermann, O. Schober, M. Schꢃfers, U.

Gangadharmath, V. Mocharla, H. Kolb, J. Walsh, W. Zhang, K. Kopka, S.

Wagner, J. Med. Chem. 2012 , 55 , 471 4 –4 727.

[49] J. van den Hoff, Amino Acids 2005 , 29 , 34 1 –3 53.

[50] H.-H. Otto, T. Schirmeister, Chem. Rev. 1997 , 97 , 13 3 –1 72.

[51] R. C. Eastman, R. E. Carson, K. A. Jacobsen, S. Yechial, M. A. Channing,

B. B. Dunn, J. D. Bacher, E. Baas, E. Jones, K. L. Kirk, M. A. Lesniak, J.

Roth, Diabetes 1992 , 41 , 85 5 –8 60.

[52] J. Schnekenburger, Arch. Pharm. 1969 , 302 , 82 2 –8 27.

[53] Z. P. Li, J. van Lier, C. C. Leznoff, Can. J. Chem. 1999 , 77 , 13 8 –1 45.

[27] G. Ren, G. Blum, M. Verdoes, H. Liu, S. Syed, L. E. Edgington, O. Ghey-

sens, Z. Miao, H. Jiang, S. S. Gambhir, M. Bogyo, Z. Cheng, PLoS One

2011 , 6 , e28029.

[28] S. M. Ametamey, M. Honer, P. A. Schubiger, Chem. Rev. 2008 , 108 , 150 1 –

1516.

Received: March 27, 2013

Revised: May 22, 2013

Published online on && &&, 0000

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemMedChem 0000, 00, 1 – 16

&16

&

ÞÞ

These are not the final page numbers!

Article Doi

Synthesis and radiopharmacological characterisation of a fluorine-18-labelled azadipeptide nitrile as a potential PET tracer for invivo imaging of cysteine cathepsins

DOI: 10.1002/cmdc.201300135

Source and publish data:

Authors:

Article abstract of DOI:10.1002/cmdc.201300135

Full text of DOI:10.1002/cmdc.201300135

Products guided by the article

R&D Labs maybe for 113426-12-3

Relevant to this article

Hot Product