11C labelling of AG957 - A potential tyrphostin radiotracer for PET

Article abstract of DOI:10.1002/jlcr.546

Carbon-11 labelled 4-(N-2,5-dihydroxybenzyl)amino methyl benzoate (AG957), a potential radiotracer for imaging bcr-abl receptors was synthesized. [¹¹C]AG957 was prepared by labelling 4-aminobenzoic acid using [¹¹C]CH₃I, which affords the corresponding [¹¹C] methyl ester in excellent yields. Subsequent condensation of the amino group with 2,5-dihydroxy-benzaldehyde formed the respective Schiff base. Reduction of this compound with NaBH3CN gave [¹¹C]AG957 in overall decay corrected radiochemical yield of 65-75% (based on ¹¹CH3I) with an average specific radioactivity of 40 Bq/μmol (1.1 Ci/μmol). The total synthesis time from EOB including formulation was 45 min. At physiological pH, the compound was found to be sufficiently stable for in vitro and in vivo investigations. Copyright

Full text of DOI:10.1002/jlcr.546

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2002; 45: 157–165.
DOI: 10.1002/jlcr.546
Research Article
11C labelling of AG957}a potential
tyrphostin radiotracer for PET
Uwe Ackermann1,2,*, Henri J. Tochon-Danguy1,3, Kenneth Young1,
John I. Sachinidis1, J. Gordon Chan1 and Andrew M. Scott1,3,4
1 Centre for PET, Austin & Repatriation Medical Centre, Studley Road,
Heidelberg, VIC 3084, Australia
2 School of Chemistry, The University of Melbourne, Australia
3 Department of Medicine, The University of Melbourne, Australia
4 Ludwig Institute for Cancer Research, Australia
Summary
Carbon-11 labelled 4-(N-2,5-dihydroxybenzyl)amino methyl benzoate
(AG957), a potential radiotracer for imaging bcr–abl receptors was synthe-
sized. [11C]AG957 was prepared by labelling 4-aminobenzoicaicd using
[11C]CH3I, which affords the corresponding [11C] methyl ester in excellent
yields. Subsequent condensation of the amino group with 2,5-dihydroxy-
benzaldehyde formed the respective Schiff base. Reduction of this compound
with NaBH3CN gave [11C]AG957 in overall decay corrected radiochemical
yield of 65–75% (based on 11CH3I) with an average specific radioactivity of
40 GBq/mmol (1.1 Ci/mmol). The total synthesis time from EOB including
formulation was 45 min. At physiological pH, the compound was found to be
sufficiently stable for in vitro and in vivo investigations. Copyright # 2002
John Wiley & Sons, Ltd.
Key Words: PET; tyrphostins; AG957; chronic myeloid leukaemia; carbon-11
*Correspondence to: U. Ackermann, Centre for PET, Austin & Repatriation Medical Centre,
Studley Road, Heidelberg, VIC 3084, Australia.
Contract grant sponsor: Austin Hospital Medical Research Foundation
Copyright # 2002 John Wiley & Sons, Ltd.
Received 7 April 2001
Revised 15 October 2001
Accepted 27 October 2001

158
U. ACKERMANN ET AL.
Introduction
Chronic myeloid leukemia (CML) is a cancer of one of the two main
types of white blood cells, the neutrophils. CML is a disease that
progresses slowly and is difficult to detect in its early stages.
A useful prognostic classification in CML is of considerable
importance in order for patients to receive the appropriate therapeutic
procedure. So far, clinical criteria such as spleen size, hematocrit and
platelet count have been used to classify patients into high-risk and low-
risk groups. However, by employing molecular criteria, it may be
possible to define low-risk patients with increased accuracy.
The Philadelphia (Ph) chromosome is the most prominent cytogenic
abnormality in CML. The existence of this chromosome is due to a
translocation, involving the bcr-gene on chromosome 22 and the
abl-gene on chromosome 9.1 The expression of the tyrosine kinase
p210bcr–abl, a 210 kDa fusion protein, is the result of this translocation.
Although the exact function of this protein in the cell is not known, it
has been demonstrated that p210bcr–abl induces a myeloproliferate
disorder in mice, which is similar to CML. Tyrphostins are inhibitors of
this enzyme and have been synthesized as potential drugs.2,3 The
chemical structure of the tyrphostin AG957 (1) is shown in Figure 1.
This compound is of particular interest because of its small size, which
enables the molecule to permeate cells,4 and because of its selectivity for
p210bcr–abl over the normally expressed p140c-abl
.
AG957 has previously been radiolabelled with H and 14C to collect
pharmacological data.5 In this study, we have labelled AG957 with the
PET isotope 11C in order to be evaluated as a potential agent for clinical
PET studies of chronic myeloid leukaemia.6
3
Results and discussion
Two different syntheticpathways leading to [ 11C]AG957 have been
explored. As outlined in Scheme 1, we have initially synthesised 4-(N-2,
Figure 1. Chemical structure of AG957
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

11C LABELLING OF AG957
159
Scheme 1. Synthesis and radiolabelling of AG957 (first pathway). Reagents and
conditions: (i) methanol, reflux 9 h; (ii) NaBH3CN, ethanol/water 1 : 1, 1 h, RT;
(iii) [11C]CH3I, BnNMe3OH, 708C, 5 min
5-dihydroxybenzyl)amino benzoicaicd ( 2), assuming that this com-
pound would be a suitable desmethyl precursor for labelling with
[11C]CH3I. The procedure for the synthesis of 2 was based on a
published method for the preparation of structurally similar com-
pounds7 and involves a condensation reaction between 2,5-dihydroxy-
benzaldehyde (3) and 4-amino benzoicacid ( 4) to form the Schiff base 5.
The imino function is subsequently reduced using NaBH3CN to afford
the desired precursor 2. This reaction sequence is a ‘one pot’ reaction
and does not involve the isolation of the unstable imine 5.
However, the labelling of 2 with [11C]CH3I in the presence of benzyl
trimethylammonium hydroxide as base produces a complex mixture of
radiolabelled products, from which only small amounts of [11C]AG957
could be isolated. The formation of many different radioactive
compounds is most likely due to the instability of AG957 in basic
medium as well as the labelling of other functional groups in the
desmethyl precursor 2. The use of different bases such as NaH or
NaHCO3 did not improve the yield of [11C]AG957.
As outlined in Scheme 2, we have therefore decided to change our
synthesis strategy and performed the labelling of 4-amino benzoicacid
(4) as the first synthesis step.8 This gave the corresponding [11C] methyl
ester 6 in 85–95% decay corrected radiochemical yield, without the
formation of any unwanted N-methylated side products. Crude 6 was
subsequently reacted with 2,5-dihydroxybenzaldehyde (3) to give the
radiolabelled imine 7, which was then reduced with NaBH3CN to form
the desired [11C] 1.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

160
U. ACKERMANN ET AL.
Scheme 2. Synthesis and radiolabelling of AG957 (second pathway). Reagents
and conditions: (i) [11C]CH3I, BnNMe3OH, DMF, 5 min, 708C; (ii) DMF,
TosOH, 5 min, RT; (iii) NaBH3CN, RT, 30 s
This route involves two additional synthesis steps after the labelling
reaction and is only possible because of the very fast formation of the
imine 7 in the presence of p-toluene sulphonic acid and because of the
instantaneous reduction of 7 with NaBH3CN. Without p-toluene
sulphonic acid as catalyst the condensation reaction is too slow to be
utilized in PET radiochemistry.
The imine 7 is an unstable compound and cannot be isolated. It
undergoes retro-cleavage upon injection into the HPLC system and only
starting compounds 3 and 6 can therefore be detected.
The HPLC chromatogram of the crude reaction mixture after
addition of 2,5-dihydroxy benzaldehyde and subsequent reduction with
NaBH3CN is shown in Figure 2. The upper trace represents radio-
activity, the bottom trace shows UV absorbance.
The radioactive peak at 9.6 min co-elutes with an authentic sample of
AG957, which we have synthesized following a published procedure.7
The radioactive peak at 7.5 min corresponds to the methyl ester 6.
Because of the instability of 7 we could not establish, whether the
presence of 6 in the crude reaction mixture was due to incomplete
reduction of 7 or incomplete condensation of 3 and 6 from the previous
synthesis step.
[11C]AG957 was isolated by semi-preparative HPLC and formulated
in 10% ethanol and 0.01 mg/ml of ascorbic acid. The synthesis time
without formulation from EOB was 40 min, formulation took an
additional 5 min. The formulation was found to be stable at room
temperature over a period of >2 h.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

11C LABELLING OF AG957
161
Figure 2. HPLC chromatogram of crude [11C] 1
Figure 3. Oxidation of AG957
This reaction sequence enabled us to produce [11C] 1 in high
radiochemical yields of 65–75% (decay corrected for [11C]CH3I) and
high average specific radioactivity of 40 GBq/mmol (1.1 Ci/mmol). High
specific activity should, however, not be required for future biological
studies, because of the high density of the target protein.
Previous studies5 on the stability of AG957 have shown that the
molecule decomposes at room temperature when exposed to air and
light. However, despite their importance for future metabolite studies,
no decomposition products have thus far been identified. Our stability
tests on the non-formulated radioligand revealed that AG957 undergoes
oxidation to its corresponding quinone 8 if exposed to atmospheric
oxygen, as shown in Figure 3.
The oxidation of 1 in the presence of atmospheric oxygen is pH
dependent and is faster in basic medium than under acidic conditions,
but does not occur within 2 h in the presence of ascorbic acid. In order
to ensure reproducibility of our results the addition of ascorbic acid to
the final formulation was therefore necessary.
For independent quality control, as well as for future metabolite
studies, it was necessary to chemically synthesize and characterize this
quinone. We achieved this by reacting AG957 with an aqueous solution
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

162
U. ACKERMANN ET AL.
of iodine and extraction of the quinone with ethyl acetate. After
characterization by NMR and accurate mass spectroscopy, methyl 4-
{[(3,6-dioxocyclohexa-1,4-dien-1-yl)methyl]amino}benzoate (8) was
used as HPLC standard for independent quality control in order to
ensure that oxidation of the formulated radioligand had not occurred.
Experimental
General
[11C]CO2 was produced by the 14N(p, a)11C nuclear reaction using a
target gas that consisted of 98.03% nitrogen and 1.97% oxygen. A
10 MeV proton beam was generated using the IBA Cyclone 10/5
cyclotron at the Austin & Repatriation Medical Centre.9 Typical
irradiation parameters were 40 mA/h for 30 min, which produced 22.2–
25.9 GBq (600–700 mCi) of [11C]CO2. Except for LiAlH4, all chemicals
and solvents were purchased from Sigma-Aldrich LiAlH4 (1 M in THF)
was obtained in 1 ml vials from ABX advanced biochemical com-
pounds.
A cold standard of AG957 was synthesized according to a literature
procedure.7 The synthesized compound showed NMR data, which were
consistent with the published values.
The labelling procedure including semi-preparative HPLC purifica-
tion, formulation and sterile filtration was carried out using an in-house
built automated system.
Semi-preparative HPLC was performed using a Shimadzu LC-10AS
isocratic pump equipped with a 1.2 ml injection loop, a reversed phase
column (Exsil C-18, 250 mm, I.D. 10 mm), a Shimadzu SPD-6AV UV
.
detector (254 nm) and a Geiger–Muller tube as radiodetector.
Independent quality control was performed by HPLC, using an
Alltech Alltima C-18 reverse phase column (C-18, 53 mm ꢀ 7 mm, 20 ml
injection loop), with 40 : 60 acetonitrile:water as mobile phase at a flow
rate of 1 ml/min. Detection of chemical compounds was achieved with a
Shimadzu SPD-6AV UV detector (254 nm) and a Berthold LB506 C-1
monitor for radiochemical detection.
4-(N-2,5-dihydroxybenzyl)amino benzoic acid ð2Þ
%
A solution of 4-amino benzoicaidc (0.41 g, 3 mmol) and 2,5-
dihydroxybenzaldehyde (0.41 g, 3 mmol) in 20 ml of methanol were
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

11C LABELLING OF AG957
163
refluxed for 16 h. The reaction was cooled to room temperature and
NaBH3CN (0.22 g, 3.5 mmol) was added. After stirring at room
temperature for 2 h the reaction mixture was hydrolysed with 100 ml
of 0.1 M HCl and extracted with ethyl acetate. The organic layer was
dried over MgSO4 and the solvent evaporated to give a light yellow oil
which was then triturated with CHCl3 and filtered to give 0.2 g (26%
yield) of 2 as a light yellow solid.
1H-NMR (300 MHz, DMSO-d6) d 7.62 (2H, d, J 11.8 Hz), 6.84 (1H,
m), 6.61 (1H, d, J 11.2 Hz), 6.53 (2H, d, J 11.8 Hz), 6.42 (1H, dd, J 11.2,
4 Hz), 4.15 (2H, s).
Accurate Mass (ESI) calc (M+H) 260.0917. Found 260.0919.
[11C] Methyl iodide
[11C]CO2 was purged from the target, trapped in a cryogenic loop using
liquid nitrogen and then passed through a vessel, containing a 1 M
solution of LiAlH4 in THF (150 ml). After completion of the transfer,
the THF was evaporated and non-stabilized HI (300 ml, 55–58%) was
added. The vessel was heated up to 1258C to form [11C] methyl iodide.
[11C]AG957 (first pathway)
[11C] methyl iodide was distilled in a stream of nitrogen gas into a
reaction vial containing 4-(N-2,5-dihydroxybenzyl)amino benzoicacid
(2) (7 mg, 27 mmol) and benzyl trimethylammonium hydroxide (6.5 ml of
a 40 wt% solution in water, 16 mmol) in 300 ml of DMF. After complete
transfer of the methyl iodide, the vial was kept at 708C for 5 min. For
injection into the HPLC, 800 ml of mobile phase was added and the
crude product was then purified by semi-preparative HPLC using the
above-mentioned set-up with 70% 0.1 M ammonium formate/30%
acetonitrile as mobile phase at a flow rate of 4 ml/min. Only 185 MBq
(5 mCi) of [11C]AG957 could be isolated.
[11C]AG957 (second pathway)
[11C] methyl iodide was distilled in a stream of nitrogen gas into a
reaction vial containing 4-amino benzoic acid (3.5 mg, 25 mmol) and
benzyl trimethylammonium hydroxide (6.5 ml of a 40 wt% solution in
water, 16 mmol) in 300 ml of DMF. After complete transfer of the
methyl iodide, the vial was kept at 708C for 5 min. A solution of
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

164
U. ACKERMANN ET AL.
2,5-dihydroxybenzaldehyde (8 mg, 58 mmol) and p-toluene sulphonic
acid monohydrate (40 mg, 210 mmol) in 300 ml of DMF was then added
and the vial was left at room temperature for 5 min.
Sodium cyanoborohydride in tetrahydrofuran (15 ml of a 1 M
solution, 15 mmol) was added to the reaction vial containing the red
solution of imine 7. The red colour disappeared immediately upon
addition of the cyanoborohydride and the mixture was kept at room
temperature for 30 s. For injection into the HPLC, 600 ml of mobile
phase was added and the crude product was then purified by semi-
preparative HPLC using the above-mentioned set-up with 70% 0.1 M
ammonium formate/30% acetonitrile as mobile phase at a flow rate of
4 ml/min.
The radioactive peak at approximately 12 min was collected and the
radioligand subsequently formulated in 10% ethanol and 0.01 mg/ml of
ascorbic acid using the SepPak method.10 In the final step, the
formulation was filtered through a sterilized 0.22 mm non-vented
Millipore GS filter unit. With this method, we produced a decay
corrected radiochemical yield of 65–75% (2.8–3.1 GBq, 75–84 mCi) of
[11C]AG957 based on [11C]CH3I. Synthesis time from EOB was 40 min.
Formulation using the SepPak method took an additional 5 min.
Methyl 4-{[(3,6-dioxocyclohexa-1,4-dien-1-yl)methyl]amino}benzoate
To a solution of AG957 (273 mg, 1 mmol) in 10 ml of acetonitrile was
added a solution of iodine (50 mg/ml, 5 ml) in water. Then, 50 ml of
water was added and the aqueous phase extracted with EtOAc. The
organicphase was dried over MgSO and the solvent removed to form
4
195 mg (70% yield) of red crystals.
1H-NMR (300 MHz, CDCl3) d 7.88 (2H, d, J 11.8 Hz), 6.82–6.69 (3H,
m), 6.60 (2H, d, J 11.8 Hz), 4.28 (2H, d, J 2.2 Hz), 3.85 (3H, s).
Accurate Mass (ESI) calc (M+H) 272.0917. Found 272.0920.
Conclusions
In conclusion, [11C]AG957 was synthesized in high yields via a three
step synthesis starting from commercially available compounds.
In vivo studies are currently being conducted to determine the
potential of this radioligand as a PET radiotracer.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165

11C LABELLING OF AG957
165
Acknowledgements
This work was supported, in part, by the Austin Hospital Medical
Research Foundation. The authors wish to acknowledge the scientific
! !
staff at the PET Centre of the Service Hospitalier FredericJoliot in
Orsay, Paris for helpful discussions. U. Ackermann is particularly
"
grateful to Prof. Bernard Maziere who has kindly organized a working
! !
visit to the Centre and to FredericDoll e for introducing him to the
!
various labelling techniques.
The authors would also like to acknowledge Prof. A. Levitzki for
providing an authenticsample of AG957.
References
1. Heisterkamp N, Stephenson JR, Groffen J, Hansen PF, de Klein A,
Bartram CR, Grosveld G. Nature: 1983; 306: 239.
2. Gazit A, Yaish P, Gilon C, Levitzki A. J Med Chem 1989; 32: 2344.
3. Gazit A, Osherov N, Posner I, Yaish P, Poradosu E, Gilon C, Levitzki A.
J Med Chem 1991; 34: 1896.
4. Anafi M, Gazit A, Gilon C, Ben-Neriah Y, Levitzki A. J Biol Chem 1992;
267: 4518.
5. Taylor GF, Hristova-Kazmierski MK, Ley FR, Kepler JA. J Labelled Cpd
Radiopharm 1996; 38(2): 187.
6. Ackermann U, Tochon-Danguy HJ, Young K, Sachinidis JI, Chan JG,
Scott AM. J Labelled Cpd Radiopharm 2001; 44 (Suppl. 1): S307.
7. Levitzki A, Gazit A, Nor Harim J, Ben-Neriah Y. Appl WO 94 (26): 260.
!
8. Dolle F, Hinnen F, Valette H, Fuseau C, Duval R, Peglion J-L, Crouzel C.
!
Bioorg & Med Chem 1997; 5(4): 749.
9. Tochon-Danguy HJ, Sachinidis JI, Egan GF, Chan JG, Berlangieri SU,
McKay WJ, Scott AM. Australas Phys Eng Sci Med 1999; 22(4): 137.
10. Lemaire C, Plenevaux A, Del Fiore G, Aerts J, Luxen A. J Labelled Cpd
Radiopharm 1995; 37: 723.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 157–165