Tritiation of the cannabinoid receptor antagonist SR144528 involving lithium aluminum tritide reduction; assessment of the kinetic isotope effect by 3H-NMR

Article abstract of DOI:10.1002/jlcr.952

The cannabinoid receptor antagonist SR144528 was synthesized by an approach that enabled the incorporation of high specific activity tritium label while circumventing the lability of the target compound to catalytic hydrogenation. Lithium aluminum tritide

Full text of DOI:10.1002/jlcr.952

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2005; 48: 589–596.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.952
Research Article
Tritiation of the cannabinoid receptor antagonist
SR144528 involving lithium aluminum tritide reduction;
3
assessment of the kinetic isotope effect by H-NMR
Herbert H. Seltzman*, Matthew C. Foster, Christopher D. Wyrick,
Jason P. Burgess and F. Ivy Carroll
Center for Organic and Medicinal Chemistry, Research Triangle Institute, P.O. Box
12194, Research Triangle Park, NC 27709, USA
Summary
The cannabinoid receptor antagonist SR144528 was synthesized by an approach that
enabled the incorporation of high specific activity tritium label while circumventing
the lability of the target compound to catalytic hydrogenation. Lithium aluminum
tritide of less than maximum specific activity was employed to introduce tritium,
resulting in an H/T incorporation indicative of no kinetic isotope effect for the
hydride/tritide reduction of a methyl benzoate. Copyright # 2005 John Wiley &
Sons, Ltd.
3
Key Words: SR144528; lithium aluminum tritide; KIE; H-NMR
Introduction
The pyrazole carboxamide SR144528 is a selective antagonist at the peripheral
cannabinoid CB2 receptor.1,2 The potential value of this compound to
researchers for pharmacological studies of the CB2 receptor3 and its potential
role in immune function prompted the synthesis of the tritiated compound.
Prior to tritium labelling, the compound was synthesized in unlabelled form by
two routes (Scheme 1) that differ in some detail from that reported by Sanofi.4
The first, followed our reported approach to the synthesis of the Sanofi central
cannabinoid CB1 receptor antagonist SR141716 employing a benzyl hydrazine
in a convergent synthesis.5,6 The second route, reserved the benzylation as the
last step in the synthesis for its application to the tritium labeling.
*Correspondence to: Herbert H. Seltzman, Center for Organic and Medicinal Chemistry, Research Triangle
Institute, P.O. Box 12194, Research Triangle Park, NC 27709, USA. E-mail: hhs@rti.org
Contract/grant sponsor: National Institutes on Drug Abuse; contract/grant number: N01DA-6-7054
Contract/grant sponsor: National Institutes of Health; contract/grant number: N01DA-6-7054
Copyright # 2005 John Wiley & Sons, Ltd.
Received 14 March 2005
Revised 16 March 2005
Accepted 17 March 2005

590
H. H. SELTZMAN ET AL.
Scheme 1.
The unlabeled SR144528 from the first route was characterized by 1H-NMR
and MS analyses. Additionally, the product exhibited a nuclear Overhauser
effect (nOe) interaction between the benzylic methylene and the 2- and
6-protons of the aryl chloride ring in MeOD. This through space interaction is
only possible with the desired 1-benzylic-5-aryl isomer, thus rigorously
identifying and distinguishing the major isomer from the minor 1-benzylic-3-
aryl isomer. The synthesis produced an 85:15 ratio of the 1,5:1,3 isomers which
were separated chromatographically and by recrystallization at the end of the
synthesis.
Tritium labeling by catalytic reduction of an unsaturated or further
halogenated analog of SR144528 is potentially problematic since cleavage of
the chloride or benzyl fragments could occur producing significant radio-
chemical by-products that could be difficult to remove. In support of this
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

TRITIUM-LABELED SR144528
591
concern, a sample of SR144528 that was stirred under a deuterium atmosphere
over 10% Pd/C in ethanol for 2 h consumed more than half of the material
through cleavage of either the chlorine, the benzyl moiety, or both the chlorine
and the benzyl moiety (HPLC-electrospray MS) (see Scheme 1).
While selectivity of the reduction of a bromo- or iodo-aryl analog of
SR144528 might prove fruitful, we sought an approach that would circumvent
the above hazards. This involved the benzylation of the des benzyl analog of
SR144528 (4) with tritium-labeled 4-methylbenzyl bromide 7 as shown in
route 2. Thus, treatment of the diketone 2 with hydrazine and subsequently
sodium hydroxide in refluxing ethanol gave the pyrazole acid 3 in 91% yield
and of suitable purity for use in the next step after simple extractive workup.
Conversion to the acid chloride (92%) and treatment with (-)-fenchylamine7–9
afforded the fenchylamide 4 in 76% yield and 97% purity (HPLC) by
extractive workup. A benzylation with commercially available 4-methylbenzyl
bromide mediated by NaH in toluene (1408C, 5 h) following the method used
for benzylating the methyl ester of 34 gave SR144528 in 47% yield comparable
(HPLC, TLC, NMR) to that produced via route 1. This approach also has the
advantage of affording a lesser percentage of the isomeric 3-aryl isomer (6%)
than did the benzyl hydrazine approach.
Applying route 2 to the synthesis of tritium-labeled SR144528 started with
the production of lithium aluminum tritide produced by the action of tritium
gas on n-butyllithium and subsequent treatment of the lithium tritide with
aluminum bromide as described.10 Addition of excess methyl 4-methylbenzo-
ate 5 to the lithium aluminum tritide afforded tritiated 4-methylbenzyl alcohol
6. Bromination with triphenylphosphine and carbon tetrbromide11,12 provided
the tritiated 4-methylbenzyl bromide 7 which was chromatographed and used
to benzylate 4 as described above. Chromatography afforded tritiated
SR144528 that was chromatographically identical (HPLC B-RAM, UV;
TLC-radioscan) to unlabeled SR144528. The specific activity was determined
by radioactivity counting and UV quantitation to be 32.5 Ci/mmol.
3
A proton decoupled H-NMR spectrum (533 mHz) of the product showed
major and minor singlets at 5.34 d and 5.36 d, respectively, as the only
resonances. This is indicative of tritiation only in the benzylic position as
3
anticipated. The proton-coupled H-NMR spectrum exhibited the 5.34 d
resonance as a singlet and the 5.36 d resonance as a doublet (J=16.4 Hz) (area
ratio 66:50) indicative of CT2 and CHT, respectively. This, in combination
with the 32.5 Ci/mmol specific activity indicates that [3H]SR144528 is a
32:49:19 mixture of CT2:CHT:CH2.y Using excess methyl 4-methylbenzoate
and complete consumption of the LiAl*H4 to give a specific activity of 32.5 Ci/
mmol indicates the percent tritium incorporation of the LiAl*H4 to be 56%
(32.5/57.5). A calculation for a random incorporation (i.e. no kinetic isotope
effect, KIE) of 56% tritide LiAl*H4 affords a 31:50:19 ratio.z Therefore,
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

592
H. H. SELTZMAN ET AL.
comparing the observed ratio (based on 3H-NMR integration) to the
calculated ratio suggests a negligible KIE for the lithium aluminum tritide/
hydride reduction of the methyl benzoate.}
y3H-NMR peak
CT2
CTH/CHT
CH2
3H peak area
C*H2 mole ratio
C*H2 mol%
66
50
0
0
33
50
33/83
39.8
50/83
60.2
32.5 Ci/mmol=
(57.5)39.8 + (28.8)60.2 + (0)X;
X=23.8
39.8 +
39.8 +
32
60.2 +
60.2 +
49
X
total mmol=123.8 =
mol%C*H2 =C*H2/123.8 =
23.8
19
zH/T Combinations
CTT
CTH
CHT
CHH
Probability of incorporation
(from 56% LIAL*H4)
Molar ratio (calc)
56 ꢀ 55
56 ꢀ 44
44 ꢀ 56
44 ꢀ 43
31
25
25
19
Conclusions
A synthesis of tritium-labeled SR144528 was developed to circumvent the
lability of the target compound to hydrogenolysis. The synthetic approach
involved tritiation via lithium aluminum tritide reduction of a methyl
benzoate. The 56% labeled lithium aluminum tritide that was prepared
resulted in an incorporation that was determined by 3H-NMR to be the same
as a random introduction of hydride and tritide. This is consistent with no
kinetic isotope effect in the LiAl*H4 reduction of the methyl benzoate.
Experimental
The preparation of lithium aluminum tritide was conducted on an IN/US
Systems Inc. Tri-Sorber Model TS-1000 tritiation manifold. 3H-NMR spectra
1
were acquired on a Bruker AMX-500 operating at 500.13 MHz for H and
3
533.4 MHz for H. Spectra were acquired at ambient temperature using a
} The hydride is suspected to have derived from trace water in the production of lithium tritide. Trace water,
possibly introduced with TMEDA, would react with n-BuLi to generate LiOH which in turn would react
with LiT to generate HT. Alternatively, LiT could react with water if it is generated faster than water is
consumed by n-BuLi. The HT would then reduce n-BuLi and form LiH and LiT which are subsequently
converted to lithium aluminum tritide/hydride thus accounting for the production of mixed incorporation
[3H]SR144528.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

TRITIUM-LABELED SR144528
593
1
5 mm 3H probe with H decoupling coil. Spectra were obtained as a 16 K free
1
induction decay over a spectral width or 5550 Hz. A H-coupled 3H spectrum
was acquired in 4000 transients with a 1 s interpulse delay using a pulse
1
1
sequence that gated the H decoupler off during acquisition. A H-decoupled
3H spectrum was acquired in 10240 transients with a 1 s interpulse delay. Both
spectra were zero filled to 32 K prior to multiplication by a 0.1 Hz Lorentzian
function and fourier transformation.
Ethyl 4-(4-chloro-3-methylphenyl)-2,4-dioxobutanoate (2)
Potassium t-butoxide (5.0 ml, 5.0 mmol, 1.0 M in THF) and anhydrous ethyl
ether (150 ml) under dry nitrogen at ꢁ788C was treated with lithium
diisopropylamide (25.2 ml, 50.4 mmol, 2.0 M in heptane:THF:ethyl-benzene).
After 15 min, 4-chloro-3-methylacetophenone 1 (8.50 g, 50.4 mmol) in 40 ml
anhydrous ether was added dropwise with stirring over 25 min. After stirring
for 30 min, the resulting orange-yellow mixture was treated with diethyl
oxalate (6.85 ml, 50.4 mmol) and allowed to warm to ambient temperature.
After the resulting orange solution was allowed to stand for 3 days it was
extracted with 1 N HCl (3 ꢀ 200 ml) and the ether layer was dried over
Na2SO4. The volatiles were removed in vacuo and subsequent high vacuum
afforded 14.2 g of an orange oil. Chromatography on silica gel (10:1) eluting
with 30% EtOAc:hexanes and combining fractions that were homogeneous by
TLC (SiO2, 30% EtOAc:hexanes, PMA/Ce+4) yielded 9.21 g (68%) of the title
1
compound as an orange waxy solid. H-NMR (250 MHz, CDCl3) d 7.92 (dd,
1H, J=1.8, 26.8 Hz, Ar-5-H), 7.76 (dd, 1 H, J=2.2, 8.4 Hz, Ar-2-H), 7.42 (dd,
1H, J=8.2, 25.7 Hz, Ar-6-H), 7.03 (s, 1H, C3-H of enolate), 4.41 (q, 2H,
J=7.14 Hz, CH2-CH3), 2.46 (s, 3H, Ar-CH3), 1.42 (t, 3H, J=7.14 Hz, CH2-
CH3). MS C13H13ClO4 268 (M+), 195, 153, 125, 84, 69.
5 -(4-Chloro-3-methylphenyl)-1-H-pyrazole-3-carboxylic Acid (3)
Ethyl 4-(4-chloro-3-methylphenyl)-2,4-dioxobutanoate 2 (4.0 g, 14.9 mmol)
was added to hydrazine monohydrate (0.746 g, 14.9 mmol) dissolved in 75 ml
absolute ethanol and heated at reflux under nitrogen with stirring for 1.5 h.
Solid NaOH (1.79 g, 44.8 mmol) was added to the solution affording a turbid
mixture that was heated at reflux a further 1.5 h. The volatiles were removed in
vacuo to afford an orange-white solid which was suspended in water and
washed with ether (4 ꢀ 100 ml). The aqueous suspension was acidified with
HCl to pH 2 to afford a yellow-white precipitate which was extracted with
ether (6 ꢀ 100 ml), combined and dried over Na2SO4 to give 3.18 g (91%) of
the title compound after filtration and removal of volatiles. This material was
1
used in the next step without further purification. H-NMR (250 MHz, CDCl3/
DMSO-d6) d 7.75 (br d, 1H, J=31 Hz, Ar-5-H), 7.56 (d, 1H, J=8.4 Hz, Ar-2-H),
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

594
H. H. SELTZMAN ET AL.
7.31 (dd, 1H, J=8.0, 21 Hz, Ar-6-H), 7.04 (s, 1H, pyrazole 4-H), 2.41 (s, 3H,
Ar-Me).
N-[(1S)-endo-1,3,3-trimethylbicyclo[2.2.1]heptan-2-yl]-5-(4-chloro-3-methyl-
phenyl)-1-H-pyrazole-3-carboxamide ð4Þ
5-(4-Chloro-3-methylphenyl)-1-H-pyrazole-3-carboxylic acid
3
(3.18 g,
13.5 mmol) was suspended in 200 ml dry toluene under dry nitrogen and
treated dropwise with distilled thionyl chloride (3.21 g, 27.0 mmol) with
stirring at ambient temperature. The reaction was heated at reflux for 1.25 h
and the volatiles removed in vacuo to afford 3.15 g (92%) of the acid chloride
as a yellow solid which was used directly in the next step.
The acid chloride (3.15 g, 12.4 mmol) suspended in 400 ml dry CH2Cl2 was
treated with (-)-fenchylamine7,9 (2.85 g, 18.6 mmol) in 5 ml dry CH2Cl2
followed by triethylamine (3.76 g, 37.2 mmol) added dropwise over 2 min with
stirring under a nitrogen atmosphere. At 1 h TLC (SiO2, 25% EtOAc:hexanes,
UV, phosphomolybdic acid/ceric sulfate detection) showed the reaction to be
complete. The reaction was extracted with saturated aqueous Na2CO3
(2 ꢀ 200 ml), 1 N HCl (2 ꢀ 200 ml), and saturated aqueous Na2CO3
(2 ꢀ 200 ml). Drying over Na2SO4, filtration and evaporation in vacuo yielded
3.48 g (76%) of the title compound as a tan solid that was 97% pure by HPLC
1
(Waters C18, 75% CH3CN:H2O, 280 nm detection). H-NMR (250 MHz,
MeOD) d 7.71 (br d, 1H, J=21 Hz, Ar-5-H), 7.53 (d, 1H, J=8.2 Hz, Ar-2-H),
7.40 (dd, 1H, J=8.2, 14.0 Hz, Ar-6-H), 7.06 (br s, 1H, pyrazole 4-H), 3.78
(s, 1H, aH), 2.42 (s, 3H, Ar-Me), 1.80–1.23 (m, 7H, fenchyl CH2 and CH), 1.16
(s, 3H, fenchyl Me), 1.09 (s, 3H, fenchyl Me0), 0.88 (s, 3H, fenchyl Me00).
[C3H2OH]-4-methylbenzyl Alcohol ð6Þ
To LiAlT4 (prepared from 0.15 mmol carrier free tritium gas, 0.24 mmol
n-BuLi, 0.24 mol TMEDA, and 0.06 mmol ALBr3 as reported10), an excess of
methyl 4-methylbenzoate 5 in dry benzene (36 mg/ml) (0.5 ml, 0.12 mmol) was
added and the reaction mixture was stirred for 1 h. Anhydrous ether (0.5 ml)
was added affording a homogeneous solution and stirred for 15 min. The
reaction was quenched with 0.1 ml water, 0.1 ml 15% NaOH, and 0.1 ml water
and stored in 2 ml ethanol at ꢁ708C overnight. Evaporation of the ethanol
(2 ꢀ ) in vacuo without going to dryness and partitioning the concentrate
between water and ether (3 ꢀ ) was followed by drying the combined organic
layers over Na2SO4. This was concentrated and stored in toluene to afford
485 mCi (9%) of product that coelutes with authentic unlabeled 4-
methylbenzyl alcohol on TLC radioscan (SiO2, 25% ether-hexane). Significant
product losses that were experienced were likely due to volatility and
incomplete extraction that can be addressed by dilution of the quenched
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

TRITIUM-LABELED SR144528
595
reaction with ether and drying over Na2SO4 to eliminate ethanol evaporation
and partitioning.
[C3H2Br]-4-methylbenzyl bromide ð7Þ
Tritiated 4-methylbenzyl alcohol (389 mCi) 6 was freed from most of the
solvent and transferred with two 0.25 ml dry CH2Cl2 washes, to CBr4 (10 mg,
0.030 mmol) under dry nitrogen. A 1.24 M solution of triphenyl phosphine in
dry CH2Cl2 (25 ml, 0.030 mmol) was added and stirred at ambient temperature
for 1 h. Three further additions of CBr4 (25 ml, 0.030 mmol) and triphenyl
phosphine (25 ml, 0.030 mmol) in CH2Cl2 were required to consume all the
starting 4-methylbenzyl alcohol (GC analysis; DB-17, 1008C, fid). This was
due to consumption of reagent by residual ethanol and possibly TMEDA. The
reaction mixture was chromatographed twice (once to remove problematic
toluene which causes coelution of 7 and ester) on silica gel (0.27 g, 0.5 g in
pentane) eluted with pentane. Concentrating the product containing fractions
by bulb-to-bulb rotary evaporation at aspirator vacuum and ambient
temperature afforded 141 mCi (36%) of radiochemically pure product (TLC-
radioscan; SiO2, hexane) that was free of ester (GC). The material was stored
in 10 ml of pentane.
N-[(1S)-endo-1,3,3-trimethylbicyclo[2.2.1]heptan-2-yl]-5-(4-chloro-3-methyl-
phenyl)-1-([C3H2]-4-methylbenzyl)-pyrazole-3-carboxamide, tritium-labeled
SR144528
Desbenzyl SR144528 4 (3.16 mg, 0.0086 mmol) in 0.15 ml dry toluene and 60%
NaH dispersion in mineral oil (1.23 mg, 0.0513 mmol) in 0.20 ml dry toluene
was stirred in a capped 1 ml reacti-vial in a 658C oil bath for 1 h under dry N2.
Tritium-labeled 4-methylbenzyl bromide 7 (141 mCi, 0.0039 mmol) was
exchanged into ꢂ0.1 ml dry toluene and transferred to the cooled sodium
salt with two 75 ml toluene washes. The reaction was stirred in a 1408C oil bath
for 20 h by which time only modest progress was observed by HPLC (Waters
C-18, 8 mm ꢀ 10 cm ꢀ 10 m Resolve RCM column, 85% CH3CN:H2O, UV,
bRAM) and by TLC radioscan (SiO2, pentane). A further addition of NaH
(0.6 mg, 0.026 mmol) in 0.1 ml toluene was made and the reaction was heated
an additional 10 h by which time the [3H]-4-methylbenzyl bromide was now
only ꢂ11% of the total radioactivity (HPLC bRAM). Work up by
quenching\partitioning between saturated brine and ether (3 ꢀ ), drying the
combined organic layers over Na2SO4 with CH2Cl2 added, filtration and
concentration under a stream of nitrogen afforded 68 mCi of predominantly
target radioligand. Chromatography on a silica gel column (0.5 g in hexane)
was conducted eluting with a step gradient of 3–10% EtOAc-hexane in 1 ml
steps, followed by EtOAc elution while collecting 0.5 ml fractions. Fractions
15, 16 (each 92% by HPLC bRAM) and the EtOAc (74%) contained the
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596

596
H. H. SELTZMAN ET AL.
product with a close eluting shoulder (44.4 mCi total, 31%). The product
containing fractions were chromatographed on the above HPLC analytical
system in batches of up to 19 mCi with no loss of resolution to afford
20.3 mCi of the 599% radiochemically pure (HPLC) title compound. A two-
dimensional TLC-radioscan (SiO2, 20% EtOAc-hexane), however, showed
the presence of 16% of a slightly more polar radioactive impurity (or
decomposition product). The product was then chromatographed on silica gel
eluting with 10% EtOAc-hexane to afford 3.86 mCi of radiochemically pure
product that was chromatographically identical to the unlabeled compound by
HPLC and TLC-radioscan. The specific activity was determined by counting
and UV quantitation (e239=24 000 l/mol cm in 85% CH3CN:H2O) to be
32.5 Ci/mmol. The material was stored in 85% CH3CN-water mobile phase.
Acknowledgements
This work was supported by Federal funds from the National Institutes on
Drug Abuse, National Institutes of Health, Department of Health and Human
Services, under Contract No. N01DA-6-7054.
References
1. Barth F, Rinaldi-Carmona M, Millan J, Derocq J-M, Bouaboula M, Casellas P,
Congy C, Oustric D, Sarran M, Calandra B, Portier M, Shire D, Breliere J-C, Le
Fur G. In Internation Cannabinoid Research Society. ICRS: Stone Mountain, GA,
1997; 11.
2. Rinaldi-Carmona M, Barth F, Millan J, Derocq J-M, Casellas P, Congy C,
Oustric D, Sarran M, Bouaboula M, Calandra B, Portier M, Shire D, Breliere
J-C, Le Fur G. J Pharmacol Exp Ther 1998; 284: 644–650.
3. Munro S, Thomas KL, Abu-Shaar M. Nature (London) 1993; 365: 61–65.
4. Barth F, Casellas P, Millan J, Oustric D, Rinaldi M, Sarran M. Sanofi Recherche,
1997. patent no. W097/21682
5. Seltzman HH, Carroll FI, Burgess JP, Wyrick CD, Burch DF. J Chem Soc, Chem
Commun 1995; 15: 1549–1550.
6. Seltzman HH, Carroll FI, Burgess JP, Wyrick CD, Burch DF. J Label Compd
Radiopharm 2002; 44: 1–12.
7. Suchocki JA, May EL, Martin TJ, George C, Martin BR. J Med Chem 1991; 34:
1003–1010.
8. Ingersoll AW, Brown JH, Kim CK, Beauchamp WD, Jennings G. J Am Chem
Soc 1936; 58: 1808–1811.
9. Ingersoll AW, DeWitt HD. J Am Chem Soc 1951; 73: 3360–3362.
10. Andres H, Morimoto H, Williams PG. J Chem Soc, Chem Commun 1990; 8:
627–628.
11. Kocienski PJ, Cernigliaro G, Feldstein G. J Org Chem 1977; 42: 353–355.
12. Hooz J, Gilani SSH. Can J Chem 1968; 46: 86–87.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 589–596