18F-barbiturates are PET tracers with diagnostic potential in Alzheimer's disease

Article abstract of DOI:10.1039/c2cc38443d

Three fluoro-barbiturates were synthesised, showing in vivo sedative efficacy. One of them, [¹⁸F]1a, was synthesised in radiofluorinated form. PET/CT Imaging with [¹⁸F]1a identified β-amyloid over-expressing transgenic mice (βA mice) compared to wild type and tau lines. The fluorescent barbiturate 9 was able to label βA plaques in brain sections of βA mice, and co-localise with a fluorescent Zn(ii) indicator.

Full text of DOI:10.1039/c2cc38443d

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
18F-barbiturates are PET tracers with diagnostic
potential in Alzheimer’s disease†
Elisa Calamai,a Sergio Dall’Angelo,a David Koss,a Juozas Domarkas,a
Timothy J. McCarthy,b Marco Mingarelli,a Gernot Riedel,a Lutz F. Schweiger,a
Andy Welch,a Bettina Platt*a and Matteo Zanda*a
Cite this: Chem. Commun., 2013,
49, 792
Received 18th July 2012,
Accepted 4th December 2012
DOI: 10.1039/c2cc38443d
www.rsc.org/chemcomm
Three fluoro-barbiturates were synthesised, showing in vivo sedative and associate with aggregated proteins; (4) barbiturates have strong
efficacy. One of them, [18F]1a, was synthesised in radiofluorinated form. capacity to form supra-molecular structures with synthetic recep-
PET/CT Imaging with [18F]1a identified b-amyloid over-expressing tors, biopolymers, etc.11 therefore they may also be able to bind to
transgenic mice (bA mice) compared to wild type and tau lines. The amyloid peptides and fibrils, or other protein aggregates (e.g. tau,
fluorescent barbiturate 9 was able to label bA plaques in brain sections synuclein, TDP43), key hallmarks of neurodegenerative diseases.
of bA mice, and co-localise with a fluorescent Zn(II) indicator.
Here we describe (1) the synthesis of fluoro-barbiturates 1a–c,
(2) the radio-synthesis of [18F]1a, (3) its in vivo use as a PET
Barbiturates are central nervous system (CNS) depressants acting as tracer in b-amyloid (bA)-expressing transgenic (TG) mice, and
positive allosteric modulators of the GABAA receptor.1 Their binding (4) histological evidence on the localisation of the fluorescent
prolongs the opening of the GABAA chloride channel causing a barbiturate 9 in Zn(II)-containing bA aggregates.
hyperpolarization of the neuronal membrane that produces the well
Three intermediate o-hydroxy-barbiturates 5a–c (Scheme 1) were
known sedative effect.2 Barbiturates, benzodiazepines and the endo- synthesised via condensation between urea and the corresponding
genous neurotransmitter GABA (g-aminobutyric acid) bind to dis- 2,2-disubstituted malonates 3a–c. Next, three fluoro-barbiturates
tinct sites on the GABAA receptor. Barbiturates have never been used (1a–c) and the parent tosyl (Ts) derivatives (6a–c) were synthesised
as radiotracers for brain PET imaging, although [18F]-barbiturate- from 5a–c, demonstrating the scope of the method.
type matrix metalloproteinase inhibitors have been investigated in
mice, showing low brain uptake.3 [11C]-barbiturates have also been
reported,4 but no use for PET imaging was described. Radio-
iodinated vinyl barbiturates for SPECT imaging showed good
capacity to cross intact blood–brain–barrier (BBB).5 In contrast,
radiolabelled benzodiazepines like 11C-flumazenil, have been used
with little success as PET tracers in Alzheimer’s disease (AD).6
We hypothesised that [18F]-barbiturates7 might be useful tracers
for the PET imaging of CNS disorders, especially proteinopathies
such as AD, for the reasons below: (1) barbiturates generally feature
excellent BBB crossing ability; (2) barbiturates receptors, i.e. the
GABAA and other ligand-gated ion channels,8 play a key role in CNS
disorders, suggesting that [18F]-barbiturates may provide informa-
tion on brain function; (3) barbiturates are effective chelators9 of
metal cations, such as Zn(II) and Cu(II), which are found in elevated
concentrations in neurodegenerative disorders, including AD,10
a Kosterlitz Centre for Therapeutics, Institute of Medical Sciences and J. Mallard
Scottish PET Centre, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD,
Scotland, UK. E-mail: m.zanda@abdn.ac.uk, b.platt@abdn.ac.uk
b Precision Medicine, Pfizer WRD, Groton, CT, USA
Scheme 1 Synthesis of the intermediate o-OH-barbiturates 5a–c, F-barbiturates
1a–c and tracer precursors 6a–c. Reagents and conditions: (a) NaH, dry DMF, 0 1C to
r.t. (b) Urea, NaH, dry DMF, 0 1C to r.t. (c) H2, Pd(OH)2/C, MeOH, r.t. (d) Tosyl chloride,
dry pyridine, dry CHCl3, 0 1C to 60 1C (to r.t. for 6a). (e) DAST, dry DME, À78 1C to r.t.
† Electronic supplementary information (ESI) available: Experimental
procedures for the synthesis, copies of NMR spectra of all the new compounds,
biological tests and imaging. See DOI: 10.1039/c2cc38443d
c
792 Chem. Commun., 2013, 49, 792--794
This journal is The Royal Society of Chemistry 2013

View Article Online
Communication
ChemComm
Scheme 2 [18F]-Hexethal radiosynthesis. Reagents and conditions:
2.2.2-Kryptofixs, dry CH3CN, 100 1C, 15 min.
[
18F]KF,
The partition coefficient (log P) of 1a–c was measured via HPLC
(Scheme 1 and ESI†). These values fall into the optimum range for
brain uptake, which is comprised between 0 and 4. Water soluble
Na(I) salts of 1a–c were prepared by treatment of 1a–c with NaOH in
THF/EtOH. Then their sedative action, and thus brain bioavailability,
was confirmed in three adult C57BL/6 mice injected intraperitoneally
(i.p.) with 50 mg kgÀ1 doses of each Na(I) salt. In all cases, first symp-
toms of sleepiness and drowsiness occurred 5–7 min post-injection,
and animals resumed normal motor activity after B40–60 min, in
line with the previously observed sedative effect of Hexethal.12
Compound 1a, a fluorinated analogue of the sedative barbiturate
Hexethal, was then synthesised in radiolabelled version. The
radiosynthesis of [18F]1a was achieved from 6a by nucleophilic
substitution of the Ts group with the [18F]KF 2.2.2-Kryptofixs
complex (Scheme 2). Tracer identity was confirmed by compari-
son of its HPLC profile with that of the ‘‘cold’’ compound 1a
(Fig. S1, ESI†). The tracer [18F]1a was obtained in radiochemical
purity Z 99% by HPLC purification, and a 17 Æ 3% (n = 7)
decay-corrected radiochemical yield (RCY) in 90 Æ 10 min.‡
The in vivo imaging potential of the barbiturate tracer [18F]1a
was assessed by dynamic PET/CT brain scans in pre-clinical
models of AD performed on four aged (>12 months) wild-type
(WT) mice, and five aged (>12 months) bA over-expressing mice
(APP/PS1), which develop a high density of amyloid pathology
including plaques from 6 months of age.13 Additionally, four
TG Tau mice (PLB2-Tau),14 which show tau histopathology
from 6 months of age, were utilised to determine specificity
for bA over-expression. Scans were acquired from 0 to 60 min
after intravenous (i.v.) injection (Fig. 1 and 2; Fig. S8, ESI†).
For all subjects, peak activity occurred at B7.5 min post injec-
tion (Fig. 2), the maximum brain uptake was between 3 and 5% ID
per g (injected dose per gram). Decline in brain radioactivity
content was quite slow, and after 60 min (Fig. 1 and 2; Fig. S8,
Fig. 1 Brain PET-CT scans from a WT mouse with [18F]1a at various times post
injection. Images were converted to units of % ID per g and a colour scale
running from 0–10% ID per g was used (bottom left). The bottom right image
shows the brain activity in isolation, and images orientation.
Fig. 2 Time-course of global [18F]1a activity, normalized to peak values, in WT
(n = 4), TG bA (n = 5) and Tau mice (n = 4). Significant interaction and genotype
effect were observed from t = 135 sec (dotted line).
ESI†) a significant amount of tracer was still present in the brain. ID per g, APP/PS1: 0.85% ID per g; p o 0.05), while WT and PLB2-Tau
Neither total activity over the first scan minute nor peak activity mice did not differ (p > 0.05). The low variability obtained in all
differed among WT, APP/PS1 and Tau-PLB2 mice (P > 0.05). Thus, groups suggests highly reproducible results, with activity tightly
differences in tracer retention time courses and brain area activities grouped particularly in the WT group (Fig. S6, ESI† for details).
(Fig. 2; Fig. S4 and S5, ESI†), are genuine and not caused by genotype-
Significant differences between WT and TG groups were further
dependent differences in brain uptake. The [18F]1a time-course confirmed by brain area analyses (Fig. S4, ESI†). Furthermore, whole
activity, normalized to peak values, in WT and TG mice revealed a body scans performed with one animal from each group (Fig. S7,
significant genotype effect (p o 0.05) from t = 135 s onwards (Fig. 2) ESI†), showed (1) no significant differences between the three
and an interaction (2-way ANOVA, group  time: F(32,160) = 2.4; p o animals and (2) uptake into the liver followed by accumulation in
0.001), confirming retention and washout differences between groups the bladder, which is the proposed metabolism of barbiturates. To
over time. Planned paired comparison indicated this to be due to assess the in vivo stability of [18F]1a, cardiac blood samples were
differences between WT and APP/PS1 mice, while the time course drawn at the end of the scans (ca. 1 hour post injection) from two WT,
between WT and PLB2-Tau was very similar (p = 0.7). Accordingly, the two APP/PS1 and two PLB2-Tau mice (Fig. S2, ESI†) and subjected
end of the scan activity (60 min post injection) detected in all four WT to HPLC analysis, which suggested that activity of [18F]1a is at least
mice was lower than that in the five APP/PS1 mice (means WT: 0.69% 65% of the total activity in the plasma (main peak, Rt = 6.6 min).
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 792--794 793

View Article Online
ChemComm
Communication
over-expressing mice. This is consistent with [18F]1a binding to bA
aggregates, which was confirmed by an in vitro affinity assay. However,
the in vitro binding affinity of 1a for artificial bA aggregates was
relatively low (bA tracers like PIB, Florbetapir and Florbetaben have bA
binding affinities in the low nM range),16 therefore specific binding in
mice in vivo would normally not be expected. As there is little pre-
cedent for obtaining a specific signal in TG mice with a radiotracer of
high nM affinity for the target (and consequently, of low binding site
density), we hypothesised that [18F]1a may bind to Zn(II), and possibly
also to other metal cations (such as iron and copper) which are
present in bA plaques.10 This hypothesis was confirmed by showing
that the fluorescent analogue 9 targets bA plaques formed in the brain
of TG APP/PS1 mice, and co-localises with the pooled Zn(II) within bA
plaques. Conversely, GABAA receptors are very unlikely to play a
significant role since (1) their expression is high in the cerebellum
and low in the striatum, thus not matching the tracer profile obtained
here (Fig. S5, ESI†); (2) isoflurane, which is also a GABAA ligand,17 was
used as anaesthetic for the in vivo PET imaging experiments.
Fig. 3 Triple staining of TG APP/PS1 mouse brain sections with bA antibody 6E10
(a), TSQ (b) and 9 (FLC-BARB) (c). Co-localisation is evident from merged images
(d). White bar: 50 mm. Arrows highlight multiple plaque labelling in images
acquired with a 10Â objective.
Notes and references
‡ The specific activity of [18F]1a could not be reliably measured because
A smaller peak accounting for ca. 31% of the whole radioactivity
(Rt = 3.5 min), which may be a metabolite, was detected. No
metabolism differences were detected between WT and TG mice.
Evidence for binding of compound 1a to bA was first sought using
in vitro matured (as confirmed by transmission electron micro-
scopy, Fig. S9, ESI†) recombinant bA. Aggregates were used to
validate the affinity of the barbiturate 1a for bA in a competitive
in vitro Congo Red (CR) binding assay. Absorbance of CR is
dependent upon binding to bA (measured at 530 nm), and was
reduced in a concentration dependent manner by 1a (Fig. S9b,
ESI†). The inhibitory profile indicated binding between the barbi-
turate and bA thus occluding CR interaction, with a KD of 123 nM.
Further studies were carried out using the fluorescent barbiturate
9, a fluorescein-analogue of 1a having the same C5-n-hexyl and ethyl
substituents (see ESI† for the synthesis of 9). Following treatment with
9, the tissue showed some labelling throughout the brain section, but
particularly intense fluorescence was detected in plaques and bA
aggregates (Fig. S9c, ESI†). Such labelling was neither seen following
incubation of brain sections from TG bA mice with the fluorophore
alone (Fig. S9c, ESI†), nor in wild-type brain sections (data not shown).
In separate labelling experiments, the specificity of labelling with 9 for
Zn(II) accumulations in bA plaque was confirmed via triple staining
with the bA specific antibody 6E10 as well TSQ, a fluorescent zinc
indicator (Fig. 3a–c).15 The antibody staining for bA in TG mouse
brain sections co-localised with the fluorescence of 9 (Fig. 3d), con-
firming the successful labelling of bA plaques formed within the brain
of TG bA mice, as well as with the pooled zinc within such plaques.
of its very low UV absorbance.
1 J. A. Richter and J. R. Holtman Jr, Prog. Neurobiol., 1982, 18, 275–319.
2 C. D’Hulst, J. R. Atack and R. F. Cooy, Drug Discovery Today, 2009, 14,
866–875.
3 D. Schrigten, H. J. Breyholz, S. Wagner, S. Hermann, O. Schober,
¨
M. Schafers, G. Haufe and K. Kopka, J. Med. Chem., 2012, 55, 223–232.
¨
4 A. Gee and B. Långstrom, Appl. Radiat. Isot., 1991, 42, 1195–1198.
5 (a) P. C. Srivastava, A. P. Callahan, E. B. Cunningham and
F. F. Knapp Jr, J. Med. Chem., 1983, 26, 742–746; (b) P. A. Mason
and B. Law, J. Labelled Compd. Radiopharm., 1982, 19, 357–364.
6 (a) M. Ohyama, M. Senda, K. Ishiwata, S. Kitamura, M. Mishina,
K. Ishii, H. Toyama, K. Oda and Y. Katayama, Ann. Nucl. Med., 1999,
13, 309–315; (b) S. Venneti, B. J. Lopresti and C. A. Wiley, Prog.
Neurobiol., 2006, 80, 308–322.
7 N. Moussier, L. Bruche, F. Viani and M. Zanda, Curr. Org. Chem.,
2003, 7, 1071–1080.
8 T. Ito, T. Suzuki, S. E. Wellman and I. K. Ho, Life Sci., 1996, 59, 169–195.
9 (a) A. Tochowicz, K. Maskos, R. Huber, V. D. Oltenfreiter, A. Yiotakis,
M. Zanda, W. Bode and P. Goettig, J. Mol. Biol., 2007, 371, 989–1006;
(b) M. S. Masoud, S. S. Haggag and E. A. Khalil, Nucleosides, Nucleo-
tides Nucleic Acids, 2006, 25, 73–87; (c) Y. Xiong, C. He, T. C. An,
C. H. Cha and X. H. Zhu, Transition. Met. Chem., 2003, 28, 69–73.
10 (a) B. Platt, J. Alzheimer’s Dis., 2006, 10, 203–213; (b) L. M. Sayre, G. Perry,
P. L. R. Harris, Y. Liu, K. A. Schubert and M. A. Smith, J. Neurochem.,
2000, 74, 270–278; (c) C. L. Masters, R. Cappai, K. J. Barnham and
V. L. Villemagne, J. Neurochem., 2006, 97, 1700–1725; (d) A. Rauk, Chem.
´
Soc. Rev., 2009, 38, 2698–2715; (e) M. P. Cuajungco and K. Y. Faget, Brain
Res. Rev., 2003, 41, 44–56; ( f ) D. Beauchemin and R. Kisilevsky, Anal.
Chem., 1998, 70, 1026–1029; (g) S. A. James, I. Volitakis, P. A. Adlard,
J. A. Duce, C. L. Masters, R. A. Cherny and A. I. Bush, Free Radical
Biol. Med., 2012, 52, 298–302.
11 S. J. Yagai, J. Photochem. Photobiol., C, 2006, 7, 164–182.
12 T. C. Butler, J. Pharmacol. Exp. Ther., 1950, 100, 219–226.
13 A. Jyoti, A. Plano, G. Riedel and B. Platt, J. Alzheimer’s Dis., 2010, 22,
873–887.
14 B. Platt, B. Drever, D. Koss, S. Stoppelkamp, A. Jyoti, A. Plano, A. Utan,
G. Merrick, D. Ryan, V. Melis, H. Wan, M. Mingarelli, E. Porcu,
L. Scrocchi, A. Welch and G. Riedel, PLoS One, 2011, 6, e27068.
´
15 A. Cote, M. Chiasson, M. R. Peralta 3rd, K. Lafortune, L. Pellegrini
and K. Toth, J. Physiol., 2005, 566, 821–837.
16 (a) A. Manook, B. H. Yousefi, A. Willuweit, S. Platzer, S. Reder, A. Voss,
M. Huisman, M. Settles, F. Neff, J. Velden, M. Schoor, H. Von der Kammer,
H. J. Wester, M. Schwaiger, G. Henriksen and A. Drzezga, PLoS One, 2012,
7, e31310; (b) G. Poisnel, M. Dhilly, O. Moustie, J. Delamare, A. Abbas,
D. Guilloteau and L. Barre, Neurobiol. Aging, 2012, 33, 2561–2571.
17 F. Jia, M. Yue, D. Chandra, G. E. Homanics, P. A. Goldstein and
N. L. Harrison, J. Pharmacol. Exp. Ther., 2008, 324, 1127–1135.
In conclusion, PET imaging with the tracer [18F]1a showed repro-
ducible brain uptake and clearance in three mice genotypes. Signifi-
cantly different [18F]1a retention modes identified bA (but not tau)
c
794 Chem. Commun., 2013, 49, 792--794
This journal is The Royal Society of Chemistry 2013