Synthesis and amyloid binding properties of rhenium complexes: Preliminary progress toward a reagent for SPECT imaging of Alzheimer's disease brain

Article abstract of DOI:10.1021/jm990103w

The definitive diagnosis of Alzheimer's disease (AD) requires the detection of amyloid plaques in postmortem brain. Although the amount of fibrillar amyloid roughly correlates with the severity of symptoms at the time of death, the temporal relationship b

Full text of DOI:10.1021/jm990103w

J . Med. Chem. 1999, 42, 2805-2815
2805
Syn th esis a n d Am yloid Bin d in g P r op er ties of Rh en iu m Com p lexes: P r elim in a r y
P r ogr ess Tow a r d a Rea gen t for SP ECT Im a gin g of Alzh eim er ’s Disea se Br a in
,∇
,§
Weiguo Zhen,†,‡, Hogyu Han,‡, Magdalena Anguiano,†,‡ Cynthia A. Lemere,† Cheon-Gyu Cho,†,‡,| and
Peter T. Lansbury, J r.*,†
Center for Neurologic Diseases, Brigham and Women’s Hospital and Harvard Medical School, Harvard Institutes of Medicine,
Room 754, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, and Department of Chemistry, Massachusetts Institute of
Technology, Cambridge, Massachusetts 02139
Received March 8, 1999
The definitive diagnosis of Alzheimer’s disease (AD) requires the detection of amyloid plaques
in postmortem brain. Although the amount of fibrillar amyloid roughly correlates with the
severity of symptoms at the time of death, the temporal relationship between amyloid deposition,
neuronal loss, and cognitive decline is unclear. To elucidate this relationship, a noninvasive,
practical method for the quantitation of brain amyloid deposition is required. We describe herein
the initial stages of a strategy to accomplish this goal by single photon computed tomographic
imaging. The amyloid-binding dye Congo Red was modified to allow its conjugation to the
monoamine-monoamide bis(thiol) ligand. This ligand complexes technetium(V) in its neutral
oxo form. A biphenyl-containing building block was conjugated to the protected ligand, and
the product was coupled to the relevant aromatic compounds. Rhenium oxo complexes, which
are isosteric, but nonradioactive, analogues of the potential imaging agent technetium oxo
complexes, were synthesized. These complexes bound to Aâ amyloid fibrils produced in vitro
and stained amyloid plaques and vascular amyloid in AD brain sections.
In tr od u ction
genic cascade: (1) overexpression of APP is character-
istic of Down syndrome, and early-onset AD is a virtual
certainty in this population;8,9 (2) missense mutations
in APP cause early-onset AD;3,7 (3) mutations in the
presenilin proteins that also cause early-onset AD all
increase the expression of the variant Aâ42 that is
known to fibrillize more rapidly than Aâ40;7,10-12 (4) the
apoE variant encoded by the apoE4 allele, which confers
susceptibility to late-onset AD, is more permissive of Aâ
amyloid formation than the other apoE variants;13-16
and (5) transgenic mice that overexpress mutant APP
develop AD-like neuropathology.17,18 While these facts
strongly suggest that amyloid formation precedes neu-
rodegeneration, a direct proof is lacking.
Alzheimer’s disease (AD) is currently diagnosed based
on the clinical observation of cognitive decline, coupled
with the systematic elimination of other possible causes
of those symptoms.1,2 The confirmation of the clinical
diagnosis of “probable AD” can only be made by exami-
nation of the postmortem brain.1,3,4 The AD brain is
characterized by a loss of neurons in regions of the brain
responsible for learning and memory (e.g., hippocampus)
and by the appearance, in these regions, of two distinct
abnormal proteinaceous deposits: extracellular amyloid
plaques, which are characteristic of AD, and intracel-
lular neurofibrillary tangles (NFTs), which are found
in other neurodegenerative disorders.1-4 The amount of
amyloid deposits roughly correlates with the severity
of symptoms at the time of death;5 although synaptic
count, a more downstream marker, correlates more
closely.
Amyloid plaques comprise dystrophic neurites and
other altered astrocytes and microglia surrounding an
insoluble fibrillar core. AD amyloid fibrils comprise a
family of proteins known collectively as the amyloid
â-proteins (Aâ), predominantly two variants: Aâ40 and
Aâ42.6 Aâ is derived from the ubiquitously expressed
cell surface amyloid precursor protein (APP).3,4,7 Several
lines of circumstantial evidence suggest that Aâ amyloid
fibril formation is an initiating event in the AD patho-
We sought to elucidate the relationship between
amyloid formation and neurodegeneration by designing
amyloid probes that could be used to measure brain
amyloid noninvasively by single photon computed to-
mography (SPECT).19-26 Three approaches to the SPECT
imaging of amyloid fibrils, all involving protein probes,
have been reported. The amyloid-associated protein
serum amyloid P component (SAP), labeled with 123I,
accumulates at low levels in the cerebral cortex, possibly
in vessel walls, of patients with cerebral amyloidosis.19
Two other approaches have been discussed but have not
been reduced to practice. Iodinated Aâ1-40, which
binds AD amyloid plaque in tissue sections, can be
transported across the blood-brain barrier (BBB) by
conjugation to a protein that is actively transported.25
In addition, antibodies to Aâ have been proposed to be
useful imaging probes, although a method to deliver
these probes across the BBB has not been described.24
These approaches suffer from some or all of these
disadvantages: (1) 123I is not an ideal radioisotope for
SPECT applications, since it must be generated in a
† Brigham and Women’s Hospital and Harvard Medical School.
‡ Massachusetts Institute of Technology.
These two authors contributed equally to this work.
∇ Current address: Perkin-Elmer Biosystems, Foster City, CA
94404.
§ Current address: Department of Chemistry, Korea University,
1-Anamdong, Seoul 136-701, Korea.
| Current address: Department of Chemistry, Hangyang University,
17 Haengdang-Dong, Sungdong-Ku, Seoul 133-791, Korea.
10.1021/jm990103w CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/02/1999

2806 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15
Zhen et al.
Ch a r t 1. Congo Red (1) and a Technetium Complex of
Its Bipyridyl Analogue 2
fibrils in vitro with comparable affinity to CR and
similarly stained amyloid plaques in tissue sections from
AD brain.
Resu lts
Syn th esis of th e Bip h en yl Lin k er , Tr ia m in e 5.
Diazotization of 2-amino-4,4′-dinitrobiphenyl followed
by nucleophilic aromatic substitution afforded the iodide
3 (Scheme 1), which was subsequently converted to the
nitrile 4. Attempts to produce 4 via the bromide
analogue of 342 were less successful, since formation of
the bromide and its conversion to 4 proceeded in
significantly lower yields. Hydrogenation of the nitro
groups followed by reduction of the nitrile provided the
triamine 5 in 61% overall yield from the starting
monoamine. Ten grams of triamine 5 have been pro-
duced in comparable yield by this reaction sequence.
Con ju ga tion of th e MAMA′ Liga n d . The MAMA′
ligand 636,37,40,43 was alkylated with â-bromopropionate
to produce ester 7 in good yield (Scheme 2). Hydrolysis
of the ester, followed by treatment with dicyclohexyl-
carbodiimide (DCC), afforded the DCC ester, which was
converted to the N-hydroxysuccinimide (NHS) ester. The
NHS ester was added, without purification, to a solution
of 1 equiv of triamine 5 to afford compound 8 in ca. 78%
yield from 6. This strategy compares favorably with the
alternative method of MAMA′ ligation by alkylation, in
which the primary alkyl halide intermediate is suscep-
tible to elimination.28,39
Compound 8 was diazotized by a standard method
and separately coupled to two activated aromatic com-
pounds (Scheme 3). The optimal conditions for these two
couplings were not identical, due to the different solu-
bilities and pKa’s of the aromatic compounds. Treatment
of diazotized 8 with 4-amino-1-naphthalenesulfonic acid,
as sodium salt, provided the CR analogue 9 in 60% yield.
Alternatively, treatment with 1-hydroxy-2-naphthoic
acid, as sodium salt, afforded compound 11 in 60% yield.
This compound is a naphthyl congener of chrysamine
G (CG), a dye that has affinity for amyloid fibrils
comparable to that of CR.20,21,23 Coupling to o-hydroxy-
benzoic acid, to produce the CG analogue, was unsuc-
cessful. Since the hydroxy group ortho to the carboxylic
acid moiety present in 11 (and 14) will delocalize its
effective charge of the carboxylate via intramolecular
hydrogen bonding, these compounds should be less polar
than the CR analogues and may be more likely to cross
the BBB (see below). The CR analogue 9 was depro-
tected with silver nitrate23,44-48 and treated, without
purification, with ammonium pertechnate (99Tc), to
afford, after purification, the technetium-99 complex 10
in 88% yield. This procedure must be modified for
incorporation of technetium-99m; hence compound 11
was not converted to the corresponding technetium-99
complex. In addition, the isosteric rhenium oxo com-
plexes were synthesized by an alternative route.
Syn th esis of th e Rh en iu m Oxo Com p lexes. The
direct conversion of compounds 9 and 11 to the analo-
gous rhenium complexes was not possible because the
conditions for Re complexation reduced the diazo bonds
in the product. Therefore, complexation of Re to 8 was
accomplished prior to diazo coupling. The major product
12 was a single isomer, presumed to be syn with respect
to the Re oxo bond and the linker because: (1) all known
cyclotron; (2) proteins are susceptible to proteolysis; (3)
most proteins do not cross the BBB; (4) proteins
penetrate tissue poorly; and (5) proteins, especially
antibodies, can be difficult to mass produce.
We chose to take a straightforward approach to the
design of a low-molecular-weight, nonprotein amyloid
probe by redesigning Congo Red (1, CR; Chart 1), a
molecule that has been used for many decades to stain
amyloid plaques in AD brain tissue sections. Congo Red
stains amyloid in tissue to produce birefringent staining,
indicative of long-range order of binding sites. Analysis
of the fibril-binding properties of several analogues of
CR established that the central biphenyl moiety could
be modified at the 2 and 2′ positions without signifi-
cantly diminishing the amyloid affinity.22 Furthermore,
introduction of a metal-binding bipyridyl unit and
chelation of 99Tc did not reduce amyloid affinity.23
However, the resultant Tc(bipyridyl) complex 2 (Chart
1) has two disadvantages relevant to its use as an
imaging agent. First, the mode of technetium complex-
ation results in a positive charge on the metal, which
may impede or prevent transport across the BBB.21,27-31
Second, there is no simple way to make this complex in
high yield and purity on a very small scale, a require-
ment for this probe to be useful clinically. To solve both
of these problems, we sought to employ the monoamine-
monoamide bis(thiol) (MAMA′) ligand, which chelates
technetium(V) as a neutral oxo complex.28,32-40 A com-
pound comprising a Tc(MAMA′) oxo complex conjugated
to a high-affinity ligand for the dopamine transporter
crosses the BBB in a rhesus monkey and can be
shown, by SPECT imaging, to accumulate in the stria-
tum.28,39,41,42 In addition, the Tc(MAMA′) complex is
simple to prepare from readily available sources of
99mTc.28,36,37,39,40 Finally, nonradioactive rhenium(MA-
MA′) oxo complexes can be simply prepared and used
for biochemical studies that do not require radioactivity,
since they are isosteric and isoelectronic with the Tc
complexes.33,36,37,40
We report herein the synthesis of one technetium-99
oxo complex and two rhenium oxo complexes that are
models of potential SPECT imaging agents for AD
amyloid. The rhenium complexes bound to amyloid

Progress Toward a SPECT Imaging Reagent of AD Brain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15 2807
Sch em e 1. Production of Triamine 5a
a
Reagents and conditions: (a) H2SO4/KI, NaNO2, 5 °C; (b) CuCN, DMSO, ∆; (c) Pd/C, MeOH, 25 °C; (d) LiAlH4, THF, ∆.
Sch em e 2. Production of Compound 8a
a
Reagents and conditions: (a) methyl 3-bromopropionate, KHCO3/K2CO3, CH3CN, ∆; (b) LiOH, 1:1 MeOH/H2O, 25 °C; (c) NHS, DCC,
THF, 25 °C; (d) 5/THF, 25 °C.
examples of Re complexes of this type are predominantly
Th e Rh en iu m Oxo Com p lexes 13 a n d 14 Ha ve
Com p a r a ble Affin ity for Aâ40 Am yloid F ibr ils a s
Con go Red . CR, 13, and 14 showed saturable binding
to amyloid fibrils formed in vitro from synthetic Aâ40.23
Since compound 14 is insoluble at high salt, binding
studies were done at pH 7.4, in 10 mM phosphate buffer,
with and without an added 100 mM NaCl. The intrinsic
affinity of CR for Aâ40 fibrils was insensitive to salt;
Kd (apparent dissociation constant) was measured to be
1.5 ( 0.005 µM at 110 mM salt and 1.1 ( 0.4 µM at 10
mM salt (Figure 1). However, the stoichiometry of
binding site saturation was salt-sensitive; more sites on
the Aâ40 fibril were available at high salt (Bmax is 2.6
( 0.14 mol of CR bound/mol of Aâ40 vs 0.60 ( 0.05 at
low salt). The salt effect suggests that fibril-CR inter-
actions are not predominantly ionic in nature (in
contrast to one model for binding20), but interactions
between bound dye molecules may be. Thus ionic
shielding of sulfonate groups projecting away from the
fibril surface may allow closer approach of like-charged
dye molecules and a higher stoichiometry of binding.
The rhenium complex 13 bound with comparable affin-
ity as CR at high salt (Kd ) 1.1 ( 0.2 µM), but with
lower stoichiometry (Bmax ) 0.44 ( 0.07). In contrast,
at low salt, compound 14 bound slightly more avidly to
Aâ40 amyloid fibrils than did CR (Kd ) 0.83 ( 0.04 µM),
with a higher stoichiometry (Bmax ) 1.4 ( 0.05).
1
syn,28,40 and (2) the diagnostic H NMR resonances of
the protons R to the tertiary amine in the anti complex
1
linker are not seen in the H NMR spectrum of 12.28,49
The infrared absorption frequency of the Re-oxo bond
in 12 was not informative in this regard, since the
experimental absorption of 967 cm-1 is midway between
published absorptions of 955 cm-1 for a model syn
complex and 980 cm-1 for a model anti complex.40
Rhenium complex 12 (Scheme 4) was successfully dia-
zotized and coupled to both 4-amino-1-naphthalene-
sulfonic acid and 1-hydroxy-2-naphthoic acid, both as
sodium salt, in moderate yield, to afford the Re com-
plexes 13 and 14, respectively.
Th e Rh en iu m Oxo Com p lex 14 Is Rela tively
Hyd r op h obic. The partitioning of a compound between
aqueous buffer and octanol roughly correlates with its
ability to cross the BBB by passive transport; that is,
compounds that prefer octanol are more likely to
cross.20,30,50 This property is expressed as the P value,
where P is the concentration in octanol phase/concen-
tration in aqueous phase. It has been proposed that a
P value between 7.8 and 316 (log P between 0.9 and
2.5) is optimal for BBB crossing.30 We measured a P
value of 0.16 ( 0.02 for CR. This value compares to a P
value of 0.323 ( 0.015 for CR using a different aqueous
buffer (the same paper reports a P value of 3.69 ( 0.21
for CG).20 The Re complex based on CR, 13, had a P
value of 0.12 ( 0.04, comparable to that of CR. Com-
pound 14, loosely based on the CG structure, had a P
value of 4.6 ( 0.9 (log P ) 0.70), indicative of its
significantly greater hydrophobicity. It is expected that
the 99mTc analogues of 13 and 14 will have similar
polarities as the Re complexes.
Sta in in g of AD Br a in Tissu e w ith Rh en iu m Oxo
Com p lexes 13 a n d 14. Rhenium complexes 13 and 14
were utilized to stain AD brain sections, following a
standard CR staining protocol for paraffin sections (see
Experimental Section). Serial sections were stained with
CR, Re complex 13, and Re complex 14. Each complex
produced birefringence, yellow/green in the case of CR

2808 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15
Zhen et al.
Sch em e 3. Production of Bistrityl Precursors and Technetium Complex 10a
a
Reagents and conditions: (a) NaNO2, THF/H2O/HCl, 5 °C, 2 min; (b) 4-amino-1-naphthalenesulfonic acid sodium salt, NaOAc/Na2CO3/
THF, 5 °C; (c) AgNO3, 1:1 MeOH/H2O, 25 °C; (d) DTT, 1:1 THF/H2O, then Na2CO3, 25 °C; (e) NH4TcO4, NaOH/EtOH, 75 °C; (f) 1-hydroxy-
2-naphthoic acid, Na2CO3/THF, 5 °C.
and 13 and yellow/white in the case of 14, indicative of
ordered long-range binding (Figure 2). However, a 2.5-
fold higher concentration of 14 had to be utilized in
order to observe birefringent staining over background.
This may simply be a reflection of the shifted absorption
of the induced birefringence, which makes it difficult
to distinguish from background. Fibrillar Aâ amyloids,
both in blood vessel walls and in plaques (plaques were
located by immunostaining with an anti-Aâ antibody),
were detected by each compound. This pattern of
staining (i.e., lesion specificity) was consistent for all
three dyes; however, the intensity of birefringence for
CR and 13 was brighter than for 14. For all three
compounds, vascular amyloid was labeled more in-
tensely than plaque amyloid, possibly reflecting a
greater density of Aâ fibrils in vascular amyloid (Figure
2).
have major repercussions for the understanding of the
pathogenesis of AD and also for its diagnosis and
treatment. First, fibrillar amyloid quantitation would
allow for a confirmation of the clinical diagnosis of
“probable AD”, which is based on the observation of
dementia and the systematic elimination of other pos-
sible causes. Second, it may allow for the detection of
early disease in individuals who have not yet developed
symptoms. These individuals, who could be targeted for
imaging based on their “AD susceptibility profile” (e.g.,
apoE genotype, family history, incidence of head trauma,
etc.), are likely to be more easily treated than symp-
tomatic individuals, who have extensive neurodegen-
eration. Finally, such a method would be valuable both
for clinical trials, to monitor the effect of treatment with
amyloid inhibitors, and eventually for monitoring the
dose of such a drug in a clinical setting.51
Discu ssion
SPECT imaging is the ideal method for the quanti-
tation of brain amyloid, since presence and amount of
fibrillar amyloid, not its high-resolution anatomical
The quantitation of brain amyloid by a simple,
inexpensive, and noninvasive imaging method could

Progress Toward a SPECT Imaging Reagent of AD Brain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15 2809
Sch em e 4. Production of Rhenium Complexes 13 and 14a
a
Reagents and conditions: (a) MeOH/THF/HCl, SnCl2/NaReO4, ∆; (b) 4-amino-1-naphthalenesulfonic acid sodium salt, NaOAc/Na2CO3/
THF, 5 °C; (c) 1-hydroxy-2-naphthoic acid, Na2CO3/THF, 5 °C.
localization, is the critical issue. An ideal amyloid
SPECT probe would incorporate 99mTc, the optimal
radioisotope for SPECT.39,41 In addition, such a probe
would: (1) utilize a practical method of technetium
complexation such that the complex could be prepared
in a radiopharmacy immediately prior to its use; (2) be
easily accessible in large quantities at high purity; (3)
demonstrate a high practical affinity for amyloid, that
is, be able to distinguish AD tissue from age-matched
normal brain that typically contains little or no fibrillar
amyloid; (4) be physiologically stable; (5) cross the BBB
in sufficient quantities to allow realization of point (3);
and (6) be nontoxic. This manuscript addresses points
(1) and (2). The amyloid probes described herein utilize
a practical method of Tc complexation identical to that
of the brain imaging agent Technepine.39 In addition,
the probe precursors can be prepared in large quantities
by a simple sequence of reactions. The practical issues
of 99mTc incorporation, which should ideally be ac-
complished in a simple final step at the imaging site
(this rules out the route used herein for production of
Re complexes 13 and 14), remain to be solved.
Regarding point (3), the in vitro affinity of probes 13
and 14 for Aâ40 amyloid fibrils is comparable to that of
CR and CG, in the high-nanomolar/low-micromolar
range.20,21,23 The relationship between this value and
the effective in vivo affinity is complex. The latter
involves, in addition to the intrinsic affinity, the stoi-
chiometry of binding (Bmax herein), the deliverable local
concentration, the number of binding sites in the target
tissue, and the kinetics of BBB crossing in both direc-
tions. Practical affinity can therefore only be optimized
by iterative experimentation in a live organism. Nev-
ertheless, in vitro affinity of these probes can be
optimized by screening of analogue libraries. The syn-
thetic route outlined above is adaptable to a parallel
synthesis mode, whereby libraries of analogues, with
different aromatic termini flanking the biphenyl core,
for example, could be produced and screened for tight
binding and specificity. The rhenium complexes 13 and
14 stain AD brain sections in a manner similar to CR.
It remains to be demonstrated that the specific lesion
binding that is selectively visualized by birefringence
(Figure 2) represents a sufficient portion of the total
binding (i.e., birefringent + nonspecific) such that a
difference between normal brain and AD brain can be
measured by SPECT. It is encouraging that the analo-
gous azo dye [14C]CG binds to AD brain tissue in
significantly greater amount than to normal brain
tissue.21
Points (4), (5), and (6) are not addressed in these
studies. However, there are published reports regarding
analogous compounds that are relevant to each
point. These complexes, in contrast to protein-imaging
agents,19,24,25 are expected to be physiologically stable.
The reductive cleavage of diazo bonds in analogous dyes,
including CR, proceeds at low levels in biological
media.52 However, [14C]CG, recovered from mouse brain
and liver 1 h after intravenous injection, was >95%
radiochemically pure as judged by HPLC.21 This result
suggests that the physiological stability of these com-
plexes may be adequate for SPECT studies.
The ability of these complexes to cross the BBB in
quantities sufficient to realize point (3) is difficult to

2810 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15
Zhen et al.
F igu r e 1. Standard binding curves are shown in the left column, with Scatchard double-reciprocal plots in the right column.
Panels A and B show a comparison of CR and compound 13 at high salt (110 mM, see text). Panels C and D show a comparison
of CR and compound 14 at low salt (10 mM).
predict. The ability of a compound to cross the BBB is
correlated to its partition coefficient between octanol
and aqueous buffer (P value).30,42,50 However, it must
be emphasized that this correlation is not linear, and
these in vitro measurements cannot accurately predict
the outcome of an in vivo study. The Re complex 14 has
a P value of 4.6, which is very close to the optimal P
values recommended for brain imaging.30 Low-molecu-
lar-weight compounds cross the BBB more efficiently;
the BBB-crossing ability seems to correlate inversely
with the square route of the MW,50 at least for com-
pounds under 600 Da. There have been few studies of
compounds in the MW range of these complexes (the
Tc analogue of 14 has a MW of ca. 970), although
peptide analogues with a MW over 1000 Da have been
shown to cross.19,25 By way of comparison, compounds
with similar P values to that of 14 have been shown to
cross the BBB at levels that may be sufficient for
SPECT imaging (e.g., CG: P ) 3.7, MW ) 48221).
Furthermore, several neutral Tc oxo complexes that
were designed to label the dopamine transporter have
been show to effectively cross the BBB and label the
striatum in rhesus monkeys (Technepine: MW ) 608,
P is unreported28,32) and in humans (TRODAT: MW )
520, P ) 22740,41). Although 10 is not likely to efficiently
cross the BBB, it may be suitable for SPECT detection
of systemic amyloid in patients suffering from systemic
amyloid diseases such as B-cell myeloma or familial
amyloidotic polyneuropathy.53 In these populations,
there is a clinical need to identify sites of amyloid
deposition.
Finally, the toxicity of the technetium complexes
discussed herein in the concentration range utilized for
SPECT imaging is unlikely to be significant. Typical
doses for SPECT imaging in humans are in the range
of 10 µg/kg.40 The LD50 of CR is reported to be 190 mg/
kg in rats.27 It has been reported that CG has no
behavioral or toxic effects in mice for periods up to 72 h
at doses of 100 mg/kg.21 Finally CR is negative in the
standard Ames test.52
In conclusion, the 99mTc analogues of 13 and, espe-
cially, 14 have the potential to be useful in imaging
amyloid in AD brain. The synthetic challenges posed
by practical issues surrounding Tc-99m complexation
remain to be solved. In addition, in vivo studies are
required in order to determine whether any of these
probes cross the BBB in adequate concentrations and
whether the ratio of specific to nonspecific binding
allows the distinction of AD brain tissue from brain
tissue of an age-matched, cognitively normal individual.
However, the chemistry reported herein provides a
critical first step toward this important goal.
Exp er im en ta l Section
Gen er a l. 1H NMR spectra were recorded using a Varian
XL-300 NMR spectrometer (300 MHz). Chemical shifts are

Progress Toward a SPECT Imaging Reagent of AD Brain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15 2811
2-Iod o-4,4′-d in itr obip h en yl, 3. To a 500-mL three-neck
round-bottom flask charged with 15.6 g (60.3 mmol) of
2-amino-4,4′-dinitrobiphenyl was added 100 mL of concen-
trated sulfuric acid. After the mixture became homogeneous,
the reaction was cooled to 0 °C. In a separate 100-mL flask
containing 5.60 g (81.2 mmol) of sodium nitrite was slowly
added 50 mL of sulfuric acid over 30 min at 5 °C. The resultant
solution was then dropwise added to the amine solution with
cooling. Manual agitation of the reaction mixture was neces-
sary. Next, 150 mL of 85% phosphoric acid was added over 1
h via an addition funnel, during which time the temperature
of the reaction mixture was maintained below 10 °C. The
mixture was agitated periodically to ensure homogeneity.
Conversion to the diazonium salt was monitored by adding a
drop of the reaction mixture to 20 mL of water. When this
solution turned light-yellow and no insoluble material was
formed, the reaction was complete. The reaction mixture was
then allowed to warm to room temperature over 2-3 h and
added to 1000 mL of ice water with stirring. After 30 min,
excess urea (12.0 g) was slowly added to consume excess
nitrous acid, and the reaction mixture was stirred for 30 min.
To this mixture was added an aqueous solution of potassium
iodide (22.68 g in 150 mL, 129.4 mmol). The reaction mixture
was then heated to 70 °C. The reaction mixture was extracted
with CH2Cl2, dried over MgSO4, and concentrated. The crude
product was purified by silica gel chromatography (1:1 hexane:
CH2Cl2) to afford the dinitroiodide product 3 (19.66 g, 53.72
1
mmol, 89% yield) as a light-yellow solid. H NMR (CDCl3): δ
8.82 (d, J ) 2.1 Hz, 1H), 8.35 (d, J ) 8.7 Hz, 2H), 8.30 (dd, J
) 2.1, 8.8 Hz, 1H), 7.55 (d, J ) 8.7 Hz, 2H), 7.48 (d, J ) 8.4
Hz, 1H). 13C NMR (CDCl3): δ 150.4, 148.1, 147.8, 147.5,
134.50, 129.9, 128.2, 124.2, 123.6, 123.2. HRMS: EI parent
ion, C12H7N2O4I, 369.9455; calcd, 369.9450. Anal. (exptl.
(calcd)): C 39.07 (38.94), H 1.90 (1.91), N 7.54 (7.57).
F igu r e 2. Staining of serial occipital sections from the brain
of a 69-year-old AD patient with Aâ antibody (top row), CR
(1) (second row), and rhenium complexes 14 (third row) and
13 (fourth row). a-d: Amyloid plaques (large arrowhead) and
blood vessel walls (small arrowheads) in adjacent sections by
each staining method (a, Aâ antibody R1282; b, CR (1); c, 14;
d, 13). The horizontal and vertical axes in a-d represent 700
and 470 µm, respectively. e-h: Amyloid in the wall of a blood
vessel in adjacent sections of occipital cortex of the same
patient (e, R1282; f, 1; g, 14; h, 13). The horizontal and vertical
axes in e-h represent 350 and 235 µm, respectively. (Figure
was reduced to 55% of original for reproduction.)
2-Cya n o-4,4′-d in itr obip h en yl, 4. A solution of the iodide
3 (10.0 g, 27.3 mmol) and CuCN (3.30 g, 36.8 mmol) in 60 mL
of anhydrous dimethyl sulfoxide was heated (180 °C) under
an argon atmosphere. After 2 h, the reaction mixture was
added to a solution of aqueous ammonium chloride (200 mL)
at 0 °C and filtered through a Buchner funnel. The insoluble
product was partitioned into methylene chloride and water.
The separated organic layer was dried over MgSO4 and
concentrated. The crude product was purified by silica gel
chromatography (1:1 hexane:CH2Cl2 to 30%:69%:1% hexane:
CH2Cl2:MeOH) to afford the nitrile 4 (6.60 g, 24.5 mmol, 90%
1
from 3) as a light-yellow solid. H NMR (CDCl3): δ 8.69 (d, J
) 2.4 Hz, 1H), 8.56 (dd, J ) 8.4, 2.4 Hz, 1H), 8.42 (d, J ) 8.4
Hz, 2H), 7.78 (dd, J ) 8.7, 1.8 Hz, 3H). 13C NMR (CDCl3): δ
148.7, 148.5, 147.5, 141.9, 131.3, 129.8, 128.9, 127.6, 124.9,
115.8, 113.0. HRMS: EI, C13H7N3O4, 269.0435; calcd, 269.0436.
Anal. (exptl (calcd)): C 58.06 (57.98), H 2.58 (2.62), N 15.58
(15.61).
reported in parts per million downfield from tetramethylsilane,
and coupling constants are in hertz (Hz). Mass spectra were
recorded using fast atom bombardment (FAB) techniques (by
Andrew Rhomberg of the MIT Mass Spectrometry Facility;
supported by NIH Grant RR00317 to K. Biemann) or matrix-
assisted laser desorption ionization (MALDI) mass spectrom-
etry by Quality Controlled Biochemicals, Inc. (Hopkinton, MA).
High-resolution mass spectra were measured at the MIT
Department of Chemistry Spectroscopy Laboratory under
either electron impact (EI) or FAB ionization conditions.
Several compounds, notably 10 and 14, did not produce parent
ions under EI or FAB conditions; therefore, these were
analyzed by a low-resolution MALDI technique. Preparative
HPLC was performed on a Waters Prep LC 4000 system using
a C18 reversed-phase column (22 × 250 mm, 15-20-µm particle
size, TP silica; Vydac, Southboro, MA) with H2O and MeOH
as eluents at a flow rate of 15 mL/min. UV spectra were
measured on a Hewlett-Packard 8452A diode array spectro-
photometer. Flash chromatography was carried out using silica
gel 60 (230-400; mesh EM Science). Thin-layer chromatog-
raphy (TLC) was performed on silica gel 60 F254 precoated
plates (250 µm; EM Science). All chemicals for synthesis were
purchased from Aldrich unless otherwise specified. CR was
purchased from Fluka and purified using preparative HPLC
before use. All compounds reported herein were judged to be
> 95% pure by TLC, HPLC, and 1H NMR. Therefore, recrys-
tallization for the sole purpose of melting point determination
was not attempted.
Tr ia m in e 5. A solution of nitrile 4 (3.0 g, 11 mmol) in
methanol (150 mL) was shaken with 10% Pd on C (500 mg)
under H2 (55 psi) for 24 h using a Parr apparatus. The reaction
mixture was filtered through Celite and concentrated to afford
the diaminonitrile as a yellow solid (2.2 g, 11 mmol, 96%), Rf
) 0.54 (7:93 MeOH:CH2Cl2). The crude product was dissolved
in tetrahydrofuran (5 mL) and added dropwise to a suspension
of LiAlH4 (3.0 g, 79 mmol) in THF (90.0 mL), and the reaction
mixture was heated at reflux under argon. After 8 h, the
reaction mixture was cooled to room temperature and quenched
by sequential addition of H2O and 5% NaOH. The mixture was
filtered to remove insoluble aluminum salts, and the filtrate
was extracted with CH2Cl2, dried over potassium carbonate,
filtered, and concentrated in vacuo to give the triamine 5 as a
tan oil. The oil was further purified by redissolving in 200 mL
of pH 1-2 water and extracted with CH2Cl2. The acidic
aqueous solution was adjusted to pH 10-11 and exhaustively
extracted with CH2Cl2 to give a light-yellow solid after
1
evaporation (1.7 g, 8.1 mmol, 76% from 4). H NMR (DMSO-
d6): δ 6.97 (d, J ) 8.3 Hz, 2H), 6.81 (d, J ) 8.1 Hz, 1H), 6.73
(d, J ) 1.8 Hz, 1H), 6.60 (d, J ) 8.2 Hz, 2H), 6.47 (dd, J ) 8.1,
2.2 Hz, 1H), 4.94 (bs, 4H), 4.20 (bs, 2H), 3.57 (bs, 2H). 13C NMR

2812 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15
Zhen et al.
(DMSO-d6): δ 147.6, 146.2, 132.8, 132.4, 132.0, 131.0, 127.8,
116.3, 116.0, 115.2, 67.1. MALDI MS for C13H15N3 [M + H]+:
213. HRMS: EI, 213.1254; calcd, 213.1265.
(m, 12H), 7.16-7.07 (m, 18H), 4.48 (s, 2H), 2.97 (t, J ) 6.8
Hz, 2H), 2.86 (s, 2H), 2.62 (t, J ) 5.9 Hz, 2H), 2.34-2.25 (m,
8H). 13C NMR (CDCl3): δ 174.1, 173.4, 159.0, 153.9, 153.5,
146.3, 146.1, 145.8, 143.1, 142.8, 138.0, 132.4, 132.1, 131.1,
130.6, 130.4, 130.3, 129.9, 128.9, 128.4, 127.7, 126.5, 126.0,
124.5, 124.4, 124.1, 123.5, 123.2, 68.0, 67.9, 58.8, 54.3, 51.9,
42.6 39.4, 34.9, 32.8, 31.1. MALDI MS for C80H70N9O8S4 [M +
H]-: 1413; calcd, 1413.43; C80H69N9O8S4Na [M + Na]-, 1436.
HRMS FAB (M + Na)-: 1434.4028; calcd, 1434.4049.
Com p ou n d 7. To a solution of MAMA′36,37,40 (6; 5.36 g, 7.92
mmol) in acetonitrile (150 mL) were added methyl 3-bro-
mopropionate (4.0 mL, 58 mmol), KHCO3 (500 mg), and K2CO3
(500 mg). The reaction mixture was refluxed under argon at
80 °C for 12 h. After cooling to room temperature and filtering,
the solution was concentrated. The crude product was purified
by flash chromatography (hexane, then 1:3 ethyl acetate/
hexane) to give compound 7 as a pale-yellow solid (3.9 g, 5.2
mmol, 65% from 6; 87% based on recovered starting material),
Com p ou n d 10. To a solution of 9 (9.0 mg, 6.4 µmol) in
MeOH/H2O (1:1, 3 mL) was added 0.1 M aqueous silver nitrate
(330 µL). The dark silver mercaptide derived from 9 precipi-
tated immediately. After 5 min, the precipitate was collected
by centrifugation and resuspended in THF/H2O (1:1, 3 mL).
The solution was treated with 0.1 M dithiothreitol (DTT; 660
µL) for 5 min followed by 10% Na2CO3 (80 µL), and the
supernatant was collected by centrifugation. To a 10-mL vial
containing 25 mM NH4[TcO4] (600 µL; New England Nuclear)
and 0.01 N NaOH (4.5 mL, pH 12) were added the supernatant
and 1 M Na2S2O4 (30 µL) in 0.01 N NaOH added sequentially
at room temperature. The reaction mixture was heated at ca.
75 °C for 30 min, cooled to room temperature, and purified by
flash chromatography using C18 corasil (37-50 µm, Waters, 2
× 4 cm) (washed with H2O, then eluted with MeOH). Purifica-
tion by preparative HPLC (0-5 min 0% MeOH, 5-10 min 10%
MeOH, 10-25 min 10-100% MeOH gradient, Rv ) 285-345
mL) afforded a red solid 10 (7.4 mg, 7.1 µmol, 88% from 9).
Specific activity ) 1.26 mCi/mmol. UV (10 mM Na2HPO4, pH
7.4): λmax 482 nm (ꢀ ) 2.40 × 104 cm-1‚M-1), 328 nm (ꢀ ) 2.86
× 104 cm-1‚M-1). 1H NMR (CD3OD): δ 8.77 (d, J ) 9.1 Hz,
2H), 8.62 (s, 1H), 8.60 (s, 1H), 8.30 (d, J ) 7.7 Hz, 2H), 8.02-
7.90 (m, 4H), 7.67-7.47 (m, 7H), 4.56-4.48 (m, 4H), 3.54-
3.47 (m, 2H), 3.13-2.40 (m, 10H). IR (film) 1651, 960 (TcdO)
cm-1. MALDI MS for C42H38N9O9S4Tc [MH]+: 1042.0; calcd,
1042.20.
Com p ou n d 11. To a solution of 8 (50 mg, 0.53 mmol) in
THF (5 mL), H2O (2.0 mL), and 10% HCl (140 µL) at 5 °C was
added NaNO2 (8.0 mg, 0.12 mmol) in H2O (40 µL). After
stirring at 5 °C for 2 min, the resulting yellow solution was
added dropwise to a solution of 2.0 mL of 1-hydroxy-2-
naphthoic acid (45 mg, 0.24 mmol, in 1:1 THF/0.5 M Na2CO3)
at 5 °C. A distinct color change from yellow to darker brown
was immediately observed. After stirring at 5 °C for 1 h, the
solution was warmed to room temperature and acidified to pH
4-5 by addition of 10% HCl. This solution was extracted by
EtOAc and dried over Na2SO4. Purification by flash chroma-
tography (CH2Cl2, then 1:9, 3:7 MeOH:CH2Cl2) followed by
preparative HPLC (0-5 min 10% MeOH, 5-20 min 10-100%
MeOH gradient, 19-21 min 100% MeOH, Rv ) 285-315 mL)
afforded 11 as a golden-yellow solid (42 mg, 0.32 mmol, 60%
from 8), Rf ) 0.50 (25:75 MeOH:CH2Cl2). UV (10 mM Na2HPO4,
pH 7.4): λmax 424 nm (ꢀ ) 2.70 × 104 cm-1‚M-1). 1H NMR
(CD3OD): δ 8.90 (dd, J ) 8.1, 7.8 Hz, 2H), 8.56 (s, 2H), 8.40
(d, J ) 8.1 Hz, 2H), 7.99-7.07 (m, 4H), 7.50-7.66 (m, 7H),
7.22-7.29 (m, 12H), 7.03-7.14 (m, 18H), 4.50 (s, 2H), 2.91 (t,
J ) 6.9 Hz, 2H), 2.82 (s, 2H), 2.56-2.64 (m, 2H), 2.16-2.34
(m, 8H). 13C NMR (CDCl3): δ 176.2, 176.1, 174.1, 173.3, 165.8,
165.7, 154.3, 153.9, 146.1, 146.2, 143.6, 143.3, 139.7, 138.4,
136.3, 132.3, 131.3, 130.8, 130.0, 129.0, 128.8, 127.9, 127.2,
127.1, 126.5, 124.9, 124.8, 124.2, 123.9, 123.3, 123.2, 115.9,
115.6, 113.0, 112.9, 110.0, 68.0, 67.9, 58.8, 54.4, 52.0, 42.7 39.4,
34.9, 32.7, 31.1. MALDI MS for C82H69N7O8S2 [MH]+: 1343;
calcd, 1343.46. HRMS FAB (MH)+: 1343.4649; calcd, 1343.4649.
1
Rf ) 0.35 (1:3 EtOAc:hexane). H NMR (CDCl3): δ 7.39 (t, J
) 7.2 Hz, 12H), 7.17-7.26 (m, 18H), 3.53 (s, 3H), 3.05 (q, J )
7.0, 12 Hz, 2H), 2.85 (s, 2H), 2.60 (t, J ) 7.0 Hz, 2H), 2.26-
2.40 (m, 8H). 13C NMR (CDCl3): δ 172.3, 170.5, 144.7, 144.6,
129.5, 127.8, 126.7, 126.6, 66.8, 66.6, 58.2, 53.3, 51.6, 49.8, 37.9,
31.9, 31.7, 29.6. MALDI MS for C48H48N2O3S2Na [M + Na]+:
786.28; calcd, 787.30. Anal. (exptl (calcd)): C 75.03 (75.36), H
6.64 (6.33), N 3.66 (3.66).
Com p ou n d 8. To a solution of 7 (3.1 g, 4.0 mmol) in MeOH
(30 mL), H2O (15 mL), and THF (30 mL) was added LiOH‚
H2O (336 mg, 8.00 mmol). The resulting solution was stirred
for 3 h at room temperature. Organic solvent was removed in
vacuo, and the resultant aqueous solution was acidified to pH
4 with 10% HCl and extracted with ethyl acetate. The
combined ethyl acetate fractions were dried over MgSO4 and
evaporated to give the acid as a white solid (3.0 g, Rf ) 0.71,
7:1:92 MeOH:acetic acid:CH2Cl2). 1H NMR (DMSO-d6): δ 7.67
(t, J ) 5.8 Hz, 1H), 7.31-7.18 (m, 30H), 2.95 (q, J ) 6.7 Hz,
2H), 2.79 (s, 2H), 2.51-2.46 (m, 2H), 2.30-2.16 (m, 8H).
To a solution of the crude acid (3.0 g, 4.0 mmol) and
N-hydroxysuccinimide (460 mg, 4.00 mmol) in THF (60 mL)
was added 1,3-dicyclohexylcarbodiimide (860 mg, 2.00 mmol)
in THF (20 mL). After stirring at room temperature for 2 h,
the solution was filtered into a flask containing the triamine
5 (810 mg, 3.80 mmol) in THF (20 mL). After 2 h at room
temperature, the reaction mixture was concentrated and the
crude product was purified by flash chromatography (CH2Cl2,
then 3:97 MeOH:CH2Cl2) to give compound 8 as a yellow solid
(3.5 g, 3.6 mmol, 90% from 7, Rf ) 0.49, 7:1:92 MeOH:acetic
acid:CH2Cl2). 1H NMR (DMSO-d6): δ 8.09 (t, J ) 5.0 Hz, 1H),
7.71 (t, J ) 5.4 Hz, 1H), 7.19-7.35 (m, 30H), 6.89 (d, J ) 8.4
Hz, 2H), 6.82 (d, J ) 8.7 Hz, 1H), 6.55 (d, J ) 8.3 Hz, 2H),
6.47-6.49 (m, 2H), 4.99 (bs, 2H), 4.95 (bs, 2H), 4.04 (d, J )
6.4 Hz, 2H), 2.91-2.95 (m, 2H), 2.76 (s, 2H), 2.31-2.35 (m,
2H), 2.13-2.22 (m, 8H). 13C NMR (CDCl3): δ 170.8, 170.6,
145.6, 145.2, 144.7, 136.2, 131.8, 131.3, 130.9, 130.1, 129.5,
127.9, 126.7, 126.6, 114.9, 114.2, 66.9, 66.7, 58.0, 53.5, 50.2,
49.1, 41.7, 38.0, 33.9, 29.5. MALDI MS for C60H59N5O2S2Na
[M + Na]+: 968.01; calcd, 968.40. HRMS FAB (M + H)+:
946.4182; calcd, 946.4188.
Com p ou n d 9. To a solution of 8 (50 mg, 0.053 mmol) in
THF (5 mL), H2O (2 mL), and 10% HCl (140 µL) at 5 °C was
added a solution of NaNO2 (8.0 mg, 0.12 mmol) in H2O (40
µL). After stirring at 5 °C for 2 min, the resulting yellow
solution was added dropwise to a solution of 4-amino-1-
naphthalenesulfonic acid sodium salt (61 mg, 0.25 mmol),
sodium acetate trihydrate (108 mg, 0.790 mmol), and Na2CO3
(10 mg, 0.094 mmol) in H2O (1 mL) at 5 °C. A distinct color
change from yellow to red was immediately observed. After
the mixture stirred at 5 °C for 3 min, THF (1 mL) was added.
After 1 h at 5 °C the mixture was concentrated, and the crude
product was purified by flash chromatography (CH2Cl2, then
1:9, 3:7 MeOH:CH2Cl2) followed by preparative HPLC (0-5
min 10% MeOH, 5-20 min 10-100% MeOH gradient, 20-25
min 100% MeOH, Rv ) 285-345 mL) to afford 9 as a red solid
(45 mg, 0. 032 mmol, 60% from 8, Rf ) 0.29, 25:75 MeOH:
CH2Cl2). UV (10 mM Na2HPO4, pH 7.4): λmax 486 nm (ꢀ ) 2.90
Com p ou n d 12. To a boiling solution of 8 (133 mg, 0.141
mmol) in 16 mL of 7:1 CH3OH:THF was added a solution of
tin(II) chloride (60 mg, 0.30 mmol, in 400 uL of 0.1 M HCl),
followed immediately by a solution of sodium perrhenate (80
mg, 0.30 mmol, in 400 µL of distilled water). Refluxing was
continued for 16 h, after which the solution was filtered and
concentrated. Flash column chromatography of this brown
solid on silica gel was performed, eluting with 1-5% methanol
in dichloromethane. The major isomer, presumed to be the syn
isomer (a minor isomer, presumed to be the anti isomer,28
constituted ca. 5% of the amount of the major isomer and was
1
× 104 cm-1‚M-1), 322 nm (ꢀ ) 3.00 × 104 cm-1‚M-1). H NMR
(CD3OD): δ 8.81 (d, J ) 8.1 Hz, 2H), 8.65 (d, J ) 3.0 Hz, 1H),
8.29 (d, J ) 7.9 Hz, 1H), 8.28 (d, J ) 7.8 Hz, 1H), 7.96 (d, J )
2.0 Hz, 1H), 7.91 (d, J ) 8.1 Hz, 3H), 7.66 (t, J ) 7.9 Hz, 2H),
7.53 (t, J ) 8.4 Hz, 4H), 7.46 (d, J ) 7.5 Hz, 1H), 7.30-7.25

Progress Toward a SPECT Imaging Reagent of AD Brain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15 2813
not characterized), was isolated as a pale-purple solid (56 mg,
100 mM NaCl) to reach final compound concentrations of 42,
34, and 27 µM for CR, 96, 60, and 26 µM for 13, and 100, 65,
and 34 µM for 14. It is important to note that, in the case of
13 and 14, the solution at this point contained ca. 6-25%
methanol (although no difference in measured P resulted). This
solution was added to an Eppendorf tube containing an equal
volume of 1-octanol. The mixture was vortexed for 1 min,
allowed to equilibrate for 30 min at 25 °C, and centrifuged for
5 min at 14 000 rpm (Eppendorf microcentrifuge 5415C) to
facilitate the formation of two clear phases. The phases were
separated, and the concentration of CR, 13, or 14 in each phase
was determined by UV-vis spectroscopy, according to prede-
termined standards (λmax of CR in buffer 480 nm (ꢀ ) 19 880
cm-1‚M-1) and in octanol 508 nm (ꢀ ) 26 600 cm-1‚M-1); 13,
480 nm in buffer (ꢀ ) 24 400 cm-1‚M-1) and 492 nm in octanol
(ꢀ ) 36 000 cm-1‚M-1); 14, 420 nm in buffer (ꢀ ) 21 800
cm-1‚M-1) and 426 nm in octanol (ꢀ ) 31 200 cm-1‚M-1). The
partition coefficient was expressed as the compound concen-
tration in the octanol layer divided by its concentration in the
aqueous layer.20,29-31 Determinations were based on averages
of three parallel tubes at each of the three concentrations.
0.085 mmol, 60%), Rf ) 0.50 (2% methanol in CH2Cl2). IR (KBr
disk): 1640, 967 (RedO) cm-1 1H NMR (DMSO-d6): δ 8.29
.
(t, J ) 4.5 Hz, 1H), 6.87 (d, J ) 8.7 Hz, 2H), 6.79 (d, J ) 8.4
Hz, 1H), 6.52 (d, J ) 7.8 Hz, 3H), 6.45 (dd, J ) 8.1, 2.1 Hz,
1H), 5.10 (bs, 4H), 4.95 (d, J ) 16.8 Hz, 1H), 4.41 (qt, J ) 5.4,
12.0 Hz, 1H), 4.20 (d, J ) 16.8 Hz, 1H), 3.96-4.15 (m, 4H),
3.70-3.85 (m, 1H), 3.42-3.47 (dd, J ) 1.8, 3.0 Hz, 1H), 3.18-
3.23 (dd, J ) 1.8, 3.0 Hz, 1H), 2.88-3.08 (m, 3H), 2.61-2.83
(m, 2H), 1.68 (dt, J ) 4.5, 12.6 Hz, 1H). 13C NMR (DMSO-d6):
δ 187.7, 168.8, 147.2, 146.9, 135.7, 130.4, 129.6, 128.4, 113.62,
113.59, 112.9, 65.6, 63.6, 59.3, 59.1, 46.5, 40.9, 37.7, 30.4.
MALDI MS for C22H29N5O3S2Re [MH]+: 661; calcd, 662.12;
C22H28N5O3S2ReNa [M + Na]+, 683; calcd, 684.11. HRMS FAB
(M + H)+: 662.1279; calcd, 662.1269.
Com p ou n d 13. To a solution of 12 (36 mg, 0.055 mmol) in
THF (4 mL), H2O (2 mL), and 10% HCl (140 µL) at 5 °C was
added a solution of NaNO2 (8.0 mg, 0.12 mmol) in H2O (40
µL). After stirring at 5 °C for 2 min, the resulting yellow
solution was added dropwise to a solution of 4-amino-1-
naphthalenesulfonic acid sodium salt (61 mg, 0.25 mmol),
sodium acetate trihydrate (108 mg, 0.790 mmol), and Na2CO3
(10 mg, 0.094 mmol) in H2O (1 mL) and 0.2 mL THF at 5 °C.
A distinct color change from yellow to red was immediately
observed. After 1 h at 5 °C and another hour at room
temperature, the crude product was purified by preparative
HPLC (0-15 min 10-100% MeOH gradient in water, 20-25
min 100% MeOH, Rv ) 150-225 mL) and lyophilized to afford
13 as a red solid (31 mg, 0.026 mmol, 50% from 12). UV (10
mM Na2HPO4, pH 7.4): λmax 486 nm (ꢀ ) 2.60 × 104 cm-1‚M-1),
324 nm (ꢀ ) 2.86 × 104 cm-1‚M-1). 1H NMR (DMSO-d6): δ
8.94 (d, J ) 8.4 Hz, 2H), 8.80 (t, J ) 5.5 Hz, 1H), 8.55 (d, J )
1.5 Hz, 1H), 8.42 (d, J ) 8.7 Hz, 2H), 8.14 (d, J ) 1.5 Hz, 1H),
8.10 (d, J ) 8.4 Hz, 2H), 8.00 (dd, J ) 8.4, 1.5 Hz, 2H), 7.72-
7.79 (m, 4H), 7.54-7.62 (m, 3H), 5.10 (d, J ) 16.8 Hz, 1H),
4.53 (d, J ) 4.8 Hz, 2H), 4.46 (qt, J ) 5.7, 12.0 Hz, 1H), 4.30
(d, J ) 16.8 Hz, 1H), 4.15-4.22 (m, 1H), 4.03 (dd, J ) 4.9, 3.9
Hz, 1H), 3.83-3.91 (m, 1H), 3.32-3.60 (m, 2H), 2.77-3.15 (m,
5H), 1.68 (dt, J ) 4.5, 12.6 Hz, 1H). IR (KBr disk): 1640, 967
(RedO) cm-1. MALDI MS for C42H38N9O9S4ReNa2 [M - Na]-:
1152. HRMS FAB [M - Na]-: 1150.1138; calcd, 1150.1131.
F ibr il Bin d in g stu d ies. Stock solutions (see above) of CR,
13, and 14 were diluted ca. 40-fold into buffer (pH 7.4, 10 mM
phosphate buffer, 100 or 0 mM NaCl), and to 180 µL of the
resultant solution was added of 20 µL of 50 µM (total [Aâ40])
fibril suspension. Final compound concentrations ranged
between 2 and 18 µM, corresponding to methanol concentra-
tions of e3 vol %, in the case of 13 and 14. The suspension
was mixed by vortexing for ca. 10 s and allowed to equilibrate
at 25 °C for 1 h. Fibril-bound probe was separated by
sedimentation (Eppendorf microcentrifuge 5415 C, 14 000 rpm
for 4 min). In the concentration range employed, there was
no significant loss of free probe from the soluble phase in the
absence of added Aâ40 fibrils. The concentration of the probe
remaining in the supernatant solution (“free” probe) was then
measured by UV-vis spectroscopy.54 All compounds showed
saturable binding, indicating the absence of nonspecific bind-
ing over these concentration ranges. Three incubations were
run in parallel, and the combined data was used to calculate
Kd and Bmax by standard methods.55 These values were
averaged with those obtained from a second set of three
parallel incubations to produce the values ((standard devia-
tions) reported herein.
Com p ou n d 14. To a solution of 12 (36 mg, 0.55 mmol) in
THF (4 mL), H2O (2.0 mL), and 10% HCl (140 µL) at 5 °C was
added NaNO2 (8.0 mg, 0.12 mmol) in H2O (40 µL). After
stirring at 5 °C for 2 min, the resulting yellow solution was
added dropwise a solution of 1-hydroxy-2-naphthoic acid (45
mg, 0.24 mmol, in 2 mL of THF, 1.40 mL of 0.5 M Na2CO3) at
5 °C. A distinct color change from yellow to darker brown was
immediately observed. After stirring at 5 °C for 1 h, the
solution was warmed to room temperature. This crude solution
was purified by preparative HPLC (0-15 min 10-100% MeOH
gradient with water, 20-25 min 100% MeOH, Rv ) 150-225
mL) and lyophilized to afford 14 as a brown solid (30 mg, 0.026
mmol, 52% from 12). UV (10 mM Na2HPO4, pH 7.4): λmax 422
Sta in in g of AD Br a in Section s. Autopsied brain from a
69-year-old AD patient was dissected to obtain blocks of
cerebral cortex, and these were immersed in 10% buffered
formalin for 2 h. The tissue was dehydrated in graded ethanol
solutions, cleared in xylene, and embedded in paraffin. Eight
micron serial sections were cut, mounted on microscopic glass
slides, allowed to dry overnight at room temperature, and
baked for 1 h at 60 °C. Sections were deparaffinized in
Histoclear (National Diagnostics, Atlanta, GA) and rehydrated
to water prior to a 30-min incubation in one of three stains:
1% CR (10 mg of dye in 1.0 mL of buffer or solvent, 15 µM),
1% 13 in 50% DMSO in water, or 2.5% 14 (37.5 µM) in 50%
DMSO in water. The sections were dipped once in a saturated
lithium carbonate solution and once in 80% aqueous ethanol
to differentiate staining. After washing in water, the sections
were dehydrated, cleared, and coverslipped. Visualization of
the birefringent staining was achieved using polarizing filters
in an Olympus microscope. Adjacent sections were stained
immunohistochemically with the polyclonal Aâ antibody R1282
(see ref 56 for details).
1
nm (ꢀ ) 2.81 × 104 cm-1‚M-1). H NMR (DMSO-d6): δ 8.92-
9.03 (m, 2H), 8.75 (d, J ) 8.7 Hz, 2H), 8.72 (t, J ) 5.5 Hz,
2H), 8.45 (d, J ) 8.4 Hz, 2H), 8.32 (d, J ) 4.2 Hz, 1H), 8.10 (d,
J ) 8.4 Hz, 2H), 8.05 (dd, J ) 8.4, 2.7 Hz, 1H), 7.48-7.70 (m,
7H), 5.00 (d, J ) 16.8 Hz, 1H), 4.44 (d, J ) 5.5 Hz, 1H), 4.37
(qt, J ) 5.4, 12.0 Hz, 1H), 4.21 (d, J ) 16.8 Hz, 1H), 4.03-
4.13 (m, 1H), 3.95 (dd, J ) 3.8, 4.8 Hz, 1H), 3.73-3.82 (m,
1H), 3.24-3.50 (m, 2H), 2.68-3.08 (m, 5H),1.68 (dt, J ) 4.5,
12.6 Hz, 1H). IR (KBr disk): 1640, 967 (RedO) cm-1. MALDI
MS for C44H39N7O9S2Re [MH]+: 1058; calcd, 1059.16.
Deter m in a tion of P a r tition Coefficien ts. A stock solu-
tion of CR in buffer (pH 7.4, 10 mM phosphate, 100 mM NaCl)
was made ([CR] ) 1 mg/mL). 13 and 14 are insoluble in this
buffer, so stock solutions in methanol were used (ca. 0.5 mg/
mL). All three stock solutions were filtered through a MILLEX-
FG Millipore filter (0.2 µm, 13 mm) before use. The measure-
ment of partition coefficient was made at three different
concentrations for each compound; no concentration depen-
dence was observed. Stock solutions of CR and compounds 13
and 14 were diluted into buffer (pH 7.4, 10 mM phosphate,
Ack n ow led gm en t. We thank Dr. Ashfaq Mahmood
for helpful advice during the course of this work and
during the preparation of this manuscript and Drs. Alun
J ones and Dennis Selkoe for comments on the manu-
script. The work was partially supported by the Na-
tional Institutes of Health (Grant AG08470).

2814 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15
Zhen et al.
(28) Meltzer, P. C.; et al. A Technetium-99m SPECT Imaging Agent
Which Targets the Dopamine Transporter in Primate Brain. J .
Med. Chem. 1997, 40, 1835-1844.
(29) Kung, H. F.; Molnar, M.; Billings, J .; Wicks, R.; Blau, M.
Synthesis and Biodistribution of Neutral Lipid-Soluble Tc-99m
Complexes that cross the Blood-Brain Barrier. J . Nucl. Med.
1984, 25, 326-332.
(30) Dischino, D. D.; Welch, M. J .; Kilbourn, M. R.; Raichle, M. E.
Relationship Between Lipophilicity and Brain Extraction of
C-11-Labeled Radiopharmaceuticals. J . Nucl. Med. 1983, 24,
1030-1038.
(31) Neirinckx, R. D.; et al. Technetium-99m d,l-HM-PAO: A New
Radiopharmaceutical for SPECT Imaging of regional Cerebral
Blood Perfusion. J . Nucl. Med. 1987, 28, 191-202.
(32) Abrams, M. J .; Murrer, B. A. Metal compounds in therapy and
diagnosis. Science 1993, 261, 725-730.
(33) Francesconi, L. C.; et al. Synthesis and characterization of
neutral MVO (M ) Tc, Re) amine-thiol complexes containing a
pendant phenylpiperidine group. Inorg. Chem. 1993, 32, 3114-
3124.
(34) Davison, A.; Sohn, M.; Orvig, C.; J ones, A. G.; LaTegola, M. R.
A tetradentate ligand designed specifically to coordinate tech-
netium. J . Nucl. Med. 1979, 20, 641.
(35) Davison, A.; J ones, A. G.; Orvig, C.; Sohh, M. A new class of
oxotechnetium (+5) chelate complexes containing a TcON2S2
core. Inorg. Chem. 1981, 20, 1629-1632.
(36) DiZio, J . P.; Fiaschi, R.; Davison, A.; J ones, A.; Katzenellenbogen,
J . A. Progestins-rhenium complexes: metal-labeled steroids with
receptor binding activity, potential receptor-directed agents for
diagnostic imaging therapy. Bioconjugate Chem. 1991, 2, 353-
366.
Refer en ces
(1) McKhann, G., et al. Clinical diagnosis of Alzheimer’s disease:
a report of the NINCDS-ADRDA Work Group, Department of
Health and Human Services Task Force on Alzheimer’s Disease.
Neurology 1984, 34, 939-944.
(2) Weiner, M. F. Alzheimer’s disease: diagnosis and treatment.
Harvard Rev. Psychiatry 1997, 4, 306-316.
(3) Selkoe, D. J . Amyloid â-protein and the genetics of Alzheimer’s
disease. J . Biol. Chem. 1996, 271, 18295-18298.
(4) Yankner, B. A. Mechanisms of neuronal degeneration in Alzhe-
imer’s disease. Neuron 1996, 16, 921-932.
(5) Cummings, B. J .; Pike, C. J .; Shankle, R.; Cotman, C. W.
â-Amyloid deposition and other measures of neuropathology
predict cognitive status in Alzheimer’s disease. Neurobiol. Aging
1996, 17, 921-933.
(6) Lansbury, P. T., J r. A Reductionist view of Alzheimer’s disease.
Acc. Chem. Res. 1996, 29, 317-321.
(7) Selkoe, D. Alzheimer’s disease: genotypes, phenotype, and
treatments. Science 1997, 275, 630-631.
(8) Teller, J . K.; et al. Presence of soluble amyloid beta-peptide
precedes amyloid plaque formation in Down’s syndrome. Nature
Med. 1996, 2, 93-95.
(9) Lemere, C. A.; et al. Sequence of deposition of heterogeneous
amyloid â-peptides and Apo E in Down syndrome: Implications
for initial events in amyloid plaque formation. Neurobiol. Disease
1996, 3, 16-32.
(10) Lemere, C. A.; et al. The E280A presenilin 1 Alzheimer mutation
produces increased Aâ42 deposition and severe cerebellar pa-
thology. Nature Med. 1996, 2, 1146-1148.
(11) Scheuner, D.; et al. Secreted amyloid â-protein similar to that
in the senile plaques of Alzheimer’s disease is increased in vivo
by the presenilin 1 and 2 and APP mutations linked to familial
Alzheimer’s disease. Nature Med. 1996, 2, 864-870.
(12) J arrett, J . T.; Berger, E. P.; Lansbury, P. T., J r. The carboxy
terminus of the beta amyloid protein is critical for the seeding
of amyloid formation: implications for the pathogenesis of
Alzheimer’s disease. Biochemistry 1993, 32, 4693-4697.
(13) Greenberg, S. M.; Rebeck, G. W.; Vonsattel, J . P. G.; Gomez-
Isla, T.; Hyman, B. T. Apolipoprotein E ꢀ4 and cerebral hemor-
rhage associated with amyloid angiopathy. Ann. Neurol. 1995,
38, 254-259.
(14) Gearing, M.; Mori, H.; Mirra, S. S. Aâ peptide length and
apolipoprotein E genotype in Alzheimer’s disease. Ann. Neurol.
1996, 39, 395-399.
(15) Hyman, B. T.; et al. Quantitative analysis of senile plaques in
Alzheimer’s disease: observation of log-normal size distribution
and molecular epidemiology of differences associated with apo-
lipoprotein E genotype and trisomy 21 (Down syndrome). Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 3586-3590.
(16) Nicoll, J . A. R.; Roberts, G. W.; Graham, D. I. Apolipoprotein E
E4 allele is associated with deposition of amyloid â-protein
following head injury. Nature Med. 1995, 1, 135-137.
(17) Games, D.; et al. Alzheimer-type neuropathology in transgenic
mice overexpressing V717F â-amyloid precursor protein. Nature
1995, 373, 523-527.
(18) Hsiao, K.; et al. Correlative memory deficits, Aâ elevation, and
amyloid plaques in transgenic mice. Science 1996, 274, 99-102.
(19) Bornebroek, M.; et al. Potential for imaging cerebral amyloid
deposits using. 123I-labeled serum amyloid P component and
SPET. Nucl. Med. Commun. 1996, 17, 929-933.
(20) Klunk, W. E.; Debnath, M. L.; Pettegrew, J . W. Development of
small molecule probes for the beta-amyloid protein of Alzhe-
imer’s disease. Neurobiol. Aging 1994, 15, 691-698.
(21) Klunk, W. E.; Debnath, M. L.; Pettegrew, J . W. Chrysamine-G
Binding to Alzheimer and Control Brain: Autopsy Study of a
New Amyloid Probe. Neurobiol. Aging 1995, 541-548.
(22) Ashburn, T. T.; Han, H.; McGuinness, B. F.; Lansbury, P. T.
Amyloid probes based on Congo Red distinguish between fibrils
comprising different peptides. Chem. Biol. 1996, 3, 351-358.
(23) Han, H.; Cho, C.-G.; Lansbury, P. T., J r. Technetium complexes
for the quantitation of brain amyloid. J . Am. Chem. Soc. 1996,
118, 4506-4507.
(24) Majocha, R.; Marotta, C. A.; Zain, S. Patent #5,231,000, 1993.
(25) Saito, Y.; Buciak, J .; Yang, J .; Pardridge, W. M. Vector-mediated
delivery of 125I-Labeled â-amyloid peptide Aâ1-40 through the
blood-brain barrier and binding to Alzheimer’s disease amyloid
of the Aâ1-40/vector complex. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 10227-10231.
(37) DiZio, J . P.; et al. Technetium and rhenium labeled progestins:
synthesis, receptor binding and in vivo distribution of an 11â-
substituted progestin labeled with technetium-99 and rhenium-
186. J . Nucl. Med. 1992, 33, 558-569.
(38) Gustavson, L. M.; Rao, T. N.; J ones, D. S.; Fritzberg, A. R.
Synthesis of a new class of Tc chelating agents: monoamine-
monoamide (MAMA) ligands. Tetrahedron Lett. 1991, 32, 5485-
5488.
(39) Madras, B. K.; et al. Technepine: A high-affinity 99mtechnetium
probe to label the dopamine transporter in brain by SPECT
imaging. Synapse 1996, 22, 239-246.
(40) O’Neil, J . P.; Wilson, S. R.; Katzenellenbogen, J . A. Preparation
and structural characterization of monoamine-monoamide bis(thi-
ol) oxo complexes of technetium (V) and rhenium (V). Inorg.
Chem. 1994, 33, 319-323.
(41) Kung, H. F.; et al. Imaging of dopamine transporters in humans
with technetium-99m TRODAT-1. Eur. J . Nucl. Med. 1996, 23,
1527-1530.
(42) Meegalla, S. K.; et al. Synthesis and Characterization of Tech-
netium-99m-Labeled Tropanes as Dopamine Transporter-Imag-
ing Agents. J . Med. Chem. 1997, 40, 9-17.
(43) Wolfe, J . A. Ph.D. Thesis, MIT, Cambridge, MA 02139, 1991.
(44) Connolly, B. A. Chemical synthesis of oligonucleotides containing
a free sulphydryl group and subsequent attachment of thiol
specific probes. Nucleic Acids Res. 1985, 13, 4485-4502.
(45) Mag, M.; Luking, S.; Engels, J . W. Synthesis and selective
cleavage of an oligonucleotide containing a bridged internucle-
otide 5′-phosphorothioate linkage. Nucleic Acids Res. 1991, 19,
1437-1441.
(46) Tomiyama, T.; et al. Rifampicin prevents the aggregation and
neurotoxicity of amyloid beta protein in vitro. Biochem. Biophys.
Res. Commun. 1994, 204, 76-83.
(47) Tomiyama, T.; et al. Racemization of Asp23 residue affects the
aggregation properties of Alzheimer amyloid â protein analogues.
J . Biol. Chem. 1994, 269, 10205-10208.
(48) Soto, C.; Castano, E. M.; Kumar, R. A.; Beavis, R. C.; Frangione,
B. Fibrillogenesis of synthetic amyloid-beta peptides is depend-
ent on their initial secondary structure. Neurosci. Lett. 1995,
200, 105-108.
(49) Lever, S. Z.; Baidoo, K. E.; Mahmood, A. Structural proof of syn/
anti isomerism in N-alkylated diamine dithiol complexes of
technetium. Inorg. Chim. Acta 1990, 176, 183-184.
(50) Levin, V. Relationship of Octanol/Water Partition Coefficient and
Molecular Weight to Rat Brain Capillary Permeability. J . Med.
Chem. 1980, 23, 682-684.
(51) Lansbury, P. T. Inhibition of amyloid formation: a strategy to
delay the onset of Alzheimer’s disease. Curr. Opin. Chem. Biol.
1997, 1, 260-267.
(52) Reid, T. M.; Morton, K. C.; Wang, C. Y.; King, C. M. Conversion
of Congo red and 2-azoxyfluorene to mutagens following in vitro
reduction by whole-cell rat cecal bacteria. Mutat. Res. 1983, 117,
105-112.
(26) Hom, R. K.; Katzenellenbogen, J . A. Technetium-99m-labeled
receptor-specific small-molecule radiopharmaceuticals: recent
developments and encouraging results. Nucl. Med. Biol. 1997,
24, 485-498.
(27) Tubis, M.; Blahd, W. H.; Nordyke, R. A. The preparation and
use of radioiodinated Congo Red in detecting amyloidosis. J . Am.
Pharm. Assoc. 1960, 49, 422-425.
(53) Kelly, J . W.; Lansbury, P. T., J r. A chemical approach to
elucidate the mechanism of transthyretin and â-protein amyloid
fibril formation. Amyloid: Int. J . Exp. Clin. Invest. 1994, 1, 186-
205.

Progress Toward a SPECT Imaging Reagent of AD Brain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 15 2815
(54) Klunk, W. E.; Pettegrew, J . W.; Abraham, D. J . Quantitative
evaluation of congo red binding to amyloid-like proteins with a
beta-pleated sheet conformation. J . Histochem. Cytochem. 1989,
37, 1273-1281.
(55) Bennett, J . J .; Yamamura, H. I. Neurotransmitter Hormone, or
Drug Receptor Binding Methods. In Neurotransmitter Receptor
Binding; Yamamura, H. I., et al., Eds.; Raven Press: New York,
1985; pp 61-89.
(56) Lemere, C. A.; et al. The lysosomal cysteine protease, cathepsin
S, is increased in Alzheimer’s disease and Down syndrome
brain: An immunocytochemical study. Am. J . Pathol. 1995, 146,
848-860.
J M990103W