High specific activity tritium labeling of vitamin D derivative RO275646

Article abstract of DOI:10.1002/jlcr.1014

The octahydroindenone intermediate was tritium labeled using T₂O isotope exchange labeling, and then elaborated to the vitamin D derivative RO275646. Though this method of labeling was expected to give the minor isomer, it was the shortest and

Full text of DOI:10.1002/jlcr.1014

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2005; 48: 1013–1023.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.1014
Research Article
High specific activity tritium labeling of vitamin D
derivative RO275646
Steve A. de Keczer*, Tim S. Lane, Tatyana Voronin and Mohammad R.
Masjedizadeh
Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
Summary
The octahydroindenone intermediate was tritium labeled using T2O isotope exchange
labeling, and then elaborated to the vitamin D derivative RO275646. Though this
method of labeling was expected to give the minor isomer, it was the shortest and most
convenient route to the high specific activity metabolically stable labeled sites. A total
of 34 mCi (from two separate runs) at 64 Ci/mmol of [3H]-RO275646 were prepared
by this method. Copyright # 2005 John Wiley & Sons, Ltd.
Key Words: tritium; T2O; exchange labeling; minor isomer; vitamin D
Introduction
Vitamin D derivatives have been shown to exhibit a much broader spectrum of
biological activity than originally thought. These steroid hormone analogs
have been used or have high potential for application as drugs in treating a
diverse range of human diseases such as rickets, renal osteodystrophy,
psoriasis, leukemia, breast cancer, prostate cancer, AIDS, and Alzheimer’s
disease.1 These and other bio-medically important applications have
stimulated our current research.
Biologically stable, high specific activity, tritium labeled vitamin D
derivative RO275646 was required for pharmaco-kinetics studies for an
osteoporosis program. Usually for DMPK studies 14C-labeled compounds are
prepared.2 However, in our case high specific activity was necessary because of
the low dosage requirement for this compound. Even if a double 14C-labeled
compound would have been used, the mass equivalent to the radioactive dose
*Correspondence to: S. A. de Keczer, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, USA.
E-mail: steve.de keczer@roche.com
Dedicated to the memory of Douglas Wren
Copyright # 2005 John Wiley & Sons, Ltd.
Received 7 September 2005
Revised 21 September 2005
Accepted 22 September 2005

1014
S. A. DE KECZER ET AL.
would have been higher than the toxic threshold. Thus, preparation of a
3H-labeled compound with the label at a non-exchangeable site was requested.
A variety of approaches for stable high specific activity tritium labeling of
vitamin D derivatives has appeared in the literature.y,2 Each of the potential
labeling sites was examined for ease of labeling, multiple labeling, existing
available routes, number of total steps required, number of hot steps required,
and expected liability/metabolic stability.
One of the sited approaches was an enolic exchange for deuterium (for a
different vitamin D derivative) that was potentially applicable to tritium
exchange.2 Although the amount of D2O is not of concern for deuterium
labeling, the amount of T2O must be kept to a minimum for practical and
safety considerations. Thus, this crucial step required further development. We
expected that the approach for tritium labeling would afford an isomer ratio
similar to that reported for proton exchange at analogous positions in the
vitamin D derivative. The reported synthesis3 gave 1:3.6 ratio of the desired
isomer. Necessary for this approach is a coupling step tethering the A-ring to
the carbonyl carbon of the C-ring. As precedent for this step, a Horner-Wittig
Olefination as applied to a non-labeled vitamin D synthesis 3 was sited. Also, a
similar attachment step had been developed by our Process Chemistry Group
for our particular compound, lending more support for this approach. Based
on the above evaluation, we chose tritium exchange labeling at the [5,5,3aa-3H]
positions (9,9,14a-3H in the vitamin D numbering system).1
Of particular advantage to tritium exchange labeling at this site was the high
specific activity attainable by the simultaneous enolic exchange of three
protons for three tritium atoms. Once these tritium atoms are locked in place
via Horner-Wittig Olefination they become chemically and metabolically
stable as exemplified by comparison with the corresponding 14C-labeled
vitamin D derivative in initial metabolic studies.4 Other advantages to labeling
at the [5,5,3aa-3H] positions were our experience with T2O exchange labeling,
the known synthesis of the required precursor, and the known relatively few
hot steps necessary for the remainder of the total synthesis.5,}
However, there were two major disadvantages to tritium exchange labeling
at this site. The first major disadvantage was the Curie amounts of T2O
required. Fortunately, we had ready access to the National Tritium Labeling
Facility (NTLF) at UC Berkeley where these amounts of activity could be
readily handled. The second major disadvantage was the adverse isomer ratio
y An excellent review with 540 references by Zhu and Okamura (reference 1) describes a number of
approaches for the synthesis and isotopic labeling of vitamin D derivatives.
} The synthesis of phosphine oxide 7 is in analogy to the chemistry developed by DeLuca (reference 9). The
Ring A precursor was technically developed by the process research group of Roche-Basel with further work
done at Roche Palo Alto.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

HIGH SPECIFIC ACTIVITY TRITIUM LABELING
1015
expected resulting at the isotope exchange step. The desired isomer was
expected to be the minor isomer.
Though at first glance this approach does not appear to be very elegant, in
reality the cost of labeling the major or the minor isomer is essentially the same
since a large excess of T2O is required in either case. The increased difficulty in
separating the minor isomer is far off-set by the fewer number of required
synthetic steps when compared to the other schemes that we evaluated.
Results and discussion
Deuterium experiments
The exchange reaction conditions for the conversion of 4 to the labeled
isomers 5a and 5b were worked out using D2O (Scheme 1).
Referring to Table 1, the best results obtained using D2O gave an average
incorporation of about 1.5 deuterium atoms per molecule, and an isomer
content of 20–25% desired 5a. The isomer content could be adjusted to about
45% desired (by reducing the reaction time, experiment 12), but then the
average incorporation of deuterium was only about 0.3 deuterium atoms per
HO
OH
+
HO
OH
D
D
D2O
NaOMe
D
D
H
D
D
O
O
O
5b
5a
4
Scheme 1. Deuterium exchange reactions
Table 1. Results from the deuterium exchange reactions
Substrate
D2O NaOMe
Expt. (mg) (umol) Solvent
(ul)
(mg) (umol) Time Temp Isomer % Isotope
Equiv.
(h)
(8C) desired
1
2
3
4
5
6
7
8
9
20
2
2
2
2
2
1
1
55
5.5
DMF
DMF
DMF
DMF
DMF
DMF
THF
1:1 THF:DMF
1:4 THF:DMF
DMF
200
10
10
10
10
10
1.25 0.16
1.25 0.16
6
12
12
12
10.9 200
14
4
1
4
1
4
3
3
4
80
RT
50
50
80
80
RT
RT
RT
RT
14
25
15
17
14
14
90
21
26
26
21
45
1.7
1.1
1.6
1.5
1.5
1.7
0
0.8
1.2
1.6
1.4
0.3
0.9
0.9
0.9
0.9
0.9
16
16
16
16
16
3
5.5
5.5
5.5
5.5
2.7
2.7
3
3.5 9.5
1.2 3.3
1.2 3.3
1
18
9
9
10
11
12
0.5
0.5
0.5
1
DMF
DMF
0.2 RT
0.02 RT
1
2.7
9
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

1016
S. A. DE KECZER ET AL.
molecule. This may simply be a reflection of incomplete exchange and
incomplete isomer distribution at short times. In fact, there appears to be a
trend of higher ratio of desired isomer correlated to lower percent isotope
incorporation. The efficiency of deuterium incorporation and isomer distribu-
tion were measured using GC/MS of the dehydration product (the upper side
chain tertiary alcohols both dehydrate under GC conditions).
Referring to Table 1, the first experiment was performed at a relatively large
scale to provide material for analysis and as a comparison for the scaled-down
reactions. For example, comparing trial experiments 1, 5 and 6, the isomer
ratio and deuterium incorporation is relatively insensitive to the reaction time
(1, 4, and 14 h) and insensitive to an order of magnitude reduction in scale.
Comparison of experiments 3 and 4 with experiments 5 and 6 shows that the
isomer ratio and deuterium incorporation is also relatively insensitive to the
temperature (508 vs 808) and overall the reaction is quite reproducible at the
10 ml scale. Experiment 2 performed at ambient temperature affords an
increase in the desired isomer at the expense of deuterium incorporation. The
set of experiments 7–8 were aliquots from a stock solution to test THF as
solvent. The relatively high desired isomer ratio in 7 was presumed to be
simply a reflection of incomplete enolization and minimal (if any) deuterium
incorporation in the absence of DMF. Experiment 9 compared with 2 showed
that there was no advantage to incorporating THF as a co-solvent.
Experiments 10–11 show that a reaction time of 12 min at ambient
temperature is sufficient, while experiment 12 shows that 1 min at ambient
temperature is insufficient for reasonable deuterium incorporation. Deuterium
experiments where the ratio of deuterated water was decreased to give >4 M
concentrations of NaOD/D2O resulted in the formation of an impurity.
Tritium synthesis
The tritium exchange reaction was run at the Lawrence Berkeley National
Labs National Tritium Labeling Facility because of the large amount of
tritiated water used (ꢀ40 Ci). The tritium exchange reaction required the
transformation of tritium gas on a platinum oxide catalyst to tritiated water.
The tritiated water (ꢀ15 ml) was vacuum transferred onto 3.7 mg NaOMe to
give a ꢀ4 M solution of NaOT/T2O; then a DMF solution of 9.5 mg substrate
was injected. Based on deuterium experiments 10 and 11, the reaction was
presumed complete in 30 min at ambient temperature. Upon quenching the
reaction mixture with HOAc, the labile tritiums were removed by methanol
exchange affording 1.3 Ci of the crude mixed isomers. Isomer separation could
only be accomplished after the protection step. Thus, unlike the deuterium
experiments the desired isomer contents were based on the two steps. TLC-
radioscan showed only about 6% desired isomer.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

HIGH SPECIFIC ACTIVITY TRITIUM LABELING
1017
Tritium incorporation (ca. 550 Ci/mmol based on total activity/starting
mass) was better than expected as compared to the GC/MS deuterium
experiments. The mixed isomer products were TMS protected and the desired
isomer 6a was isolated by a combination of flash column chromatography and
preparative HPLC. Following procedures that were worked out using the
deuterium analog, the pure single isomer 6a was coupled to the anion of the
extended A-ring phosphene oxide 7 using Horner-Wittig Olefination (0.4 mmol
scale). The coupling product 8 was deprotected to afford 22 mCi [9,9,14a-3H]-
RO275646 9 in >99% purity by HPLC and having a specific activity of 64 Ci/
mmol. The measured specific activity of the final product 9 was determined by
HPLC/liquid scintillation corresponding to 2.2 tritium atoms incorporated per
molecule of the desired isomer of the final product. This was higher than
expected based on the highest (1.7 deuterium atoms per molecule of desired
isomer) GC/MS. This discrepancy may be explained because of the different
modes of analysis. The specific activity result for the final product 9 was
obtained by HPLC/liquid scintillation. The deuterium incorporation result
was obtained by GC/MS on intermediate 5a. The former method is presumed
to be more accurate than the latter. This is because the tritium is non-
exchangeable in the final product, while the deuterium is in an exchangeable
form in the intermediate, and is thought to partially exchange under GC
conditions as discussed above (Scheme 2).
Experimental
Proton spectra were run on a Bruker 300 MHz NMR spectrometer. LC/MS
data were recorded from an Agilent 1100 Series LC/MSD. Preparative HPLC
was run on a Beckman System 32 Karat Gold using a 125 Solvent Module and
a 166 Detector; activity was measured using an INUS System b-Ram. TLC
scans were run on a Bioscan System 200 Imaging Scanner; TLC plates (silica
1
gel 250 microns) were obtained from Analtech. H NMR (300 MHz) and MS
ES+ and GCMS were consistent with the each of the non-labeled products.
6-[(1R,3aR,4S,7aR)-4-(tert-Butyl-dimethyl-silanyloxy)-7a-methyl-octahy-
dro-inden-1-yl]-2,10-dimethyl-undecane-2,10-diol (2)
Dry THF (10 ml) was injected into a dry septum sealed 3-neck 50 ml flask
containing 994 mg of substrate 1 (1.95 mmol). Methyl magnesium bromide
(6.0 ml 3 M in ether) was added dropwise to the solution stirred at ꢁ308C
under nitrogen, and the reaction was allowed to slowly warm to ambient
temperature overnight. After cooling to 0 8C, the reaction was quenched by
slow addition of 15 ml saturated aqueous ammonium chloride. The product
was extracted three times with a total of 45 ml ethyl acetate. The organic phase
was dried with brine, and anhydrous sodium sulfate, then filtered and
concentrated under pump vacuum. Flash chromatography (silica gel: 40%
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

1018
S. A. DE KECZER ET AL.
CO Et
2
EtO C
2
HO
OH HO
H SiF
OH
MeMgBr
2
6
H
H
O 2
H
OH
Et O/THF
2
TBDMS O
TBDMS
3
H O/THF/ACN
2
1
PDC
CH Cl
2
2
HO
OH
T O
2
NaOMe
HO
OH
+ T
T
T
T
T
T
5b
O
O
H
O
5a
4
TMSCl
Imidazole
DMF
TMSO
OTMS
T
+
T
T
T
T
T
6a
O
O
6b
P(O)Ph
2
n-BuLi
THF
OTBDMS
TBDMSO
7
OTMS
TMSO
HO
OH
T
T
T
T
T
T
Bu NF
4
THF
OTBDMS
OH
TDBMSO
HO
3
8
9
H-RO275646
Scheme 2. Preparation of [3H]-RO275646
ethyl acetate/hexane) afforded 720 mg (1.49 mmol, 76% yield) of compound 2
as a brittle white foam.
6-((1R,3aR,4S,7aR)-4-Hydroxy-7a-methyl-octahydro-inden-1-yl)-2,10-dimethyl-
undecane-2,10-diol (3)
A 700 mg (1.45 mmol) portion of product 2 was rinsed with 4 ml THF and 4 ml
acetonitrile into a 30 ml Teflon bottle. A 25% aqueous solution of fluorosilicic
acid (5 ml) was added and the reaction was stirred at ambient temperature for
69 h, and then at 37 8 for 20 h until nearly all of compound 2 had been
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

HIGH SPECIFIC ACTIVITY TRITIUM LABELING
1019
deprotected. The reaction was quenched with 10 ml saturated aqueous sodium
bicarbonate, and then extracted three times with 20 ml portions of ethyl
acetate. The combined organics were dried with an equal volume of brine, and
then with anhydrous sodium sulfate. The solution was filtered and
concentrated to a clear, colorless glass. Flash chromatography (silica gel;
50% ethyl acetate/hexane) followed by the concentration of pure product
fractions afforded 399 mg (1.08 mmol, 74% yield) of triol 3.
(1R,3aR,7aR)-1-[5-Hydroxy-1-(4-hydroxy-4-methyl-pentyl)-5-methyl-hexyl]-
7a-methyl-octahydro-inden-4-one (4)
Triol 3 (93.8 mg, 0.255 mmol) was dissolved in 2.5 ml dichloromethane.
Pyridinium dichromate (PDC, 389 mg, 1.01 mmol) was added and the deep
brown solution was stirred at ambient temperature overnight. An additional
200 mg of PDC and 100 mg dry Celite were added and the reaction was stirred
for a second night until the reaction was about 75% complete. The salts were
precipitated by dilution with 25 ml ether. The mixture was filtered, and the
filtrate was concentrated to dryness. Product 4 was isolated by flash
chromatography (silica gel; 2.5% MeOH/CH2Cl2) affording 66.7 mg white
foamy semi-solid (0.18 mmol, 71% yield).
(1R,3aR,7aR)-[3a,5,5-3H]-1-[5-hydroxy-1-(4-hydroxy-4-methyl-pentyl)-5-methyl-
hexyl]-7a-methyl-octahydro-inden-4-one (5a)
Cold preparation. All glassware was dried. The reaction apparatus consisted
of a ‘wishbone’ equipped with three stopcocks arranged so that the apparatus
and each of two 5 ml septum side-arm flasks could be isolated. To the flask on
the left was added 200 mg PtO2, and to the flask on the right was added ca.
3.7 mg NaOMe under nitrogen. Both flasks (each containing a micro-spin bar)
were evacuated. Separately, the substrate 4 (9.5 mg, 0.026 mmol) was dissolved
in 0.5 ml freshly degassed and distilled DMF.
Hot work. The tritium gas (ꢀ40 Ci) was transferred onto the PtO2 catalyst at
ambient temperature. After the flash, the T2O generated was cooled to
ꢁ198 8C, and the excess tritium gas was removed. The T2O was then vacuum
transferred on to the NaOMe. The mixture was warmed to ambient
temperature, the vacuum was released with nitrogen and the NaOMe was
dissolved in the T2O by means of the internal micro spin bar. A 0.5 ml DMF
solution of the substrate 4 was injected all at once, followed by a 0.25 ml DMF
rinse. After stirring for 30 min, the reaction was neutralized by the injection of
50 ml 1:3 v/v HOAc:DMF. After stirring for 10 min, volatiles were removed by
vacuum transfer, and chased with three 0.5 ml portions of MeOH to exchange
off the tritium from the hydroxyl groups, and once with toluene to azeotrope
off the excess HOAc. The residue was redissolved in 2 ml ether, and filtered
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

1020
S. A. DE KECZER ET AL.
(0.2 m Teflon filter, to remove NaOAc) with two 2 ml ether rinses into a 25 ml
flask. The solution was concentrated to dryness by N2 evaporation, and the
residue was redissolved in freshly distilled THF. A 50 mCi portion was
removed for tritium NMR and the solution was assayed for activity (1.3 Ci
total) and transferred to two ampoules in a total of 2.7 ml THF each and
3
sealed under nitrogen. H NMR (300 MHz) d 1.95 (2T, m), d 2.22 (1T, m).
(1R,3aR,7aR)-1-[5-(tert-Butyl-dimethyl-silanyloxy)-1-(4-dimethylsilany-
loxy-4-methyl-pentyl)-5-methyl-hexyl]-[3a,5,5-3H]-7a-methyl-octahydro-inden-4-
one (6a)
The 2.7 ml THF solution of mixed isomers was transferred to a dry 15 ml #7
O-ring stopcock flask containing a micro-spin bar. The THF was removed by
vacuum transfer, and the residue was chased with 1 ml dry toluene. A 0.5 ml
DMF solution of TMSCl (0.13 mmol) and imidazole (0.33 mmol) from a
freshly prepared bulk solution was added and the mixture was stirred for 17 h
at ambient temperature. The DMF was removed by vacuum transfer and the
residue was triturated with 0.5 ml hexane. The hexane solution was applied
directly to an 8 ml SiO2 disposable flash column pre-equilibrated with 2.5%
EtOAc/hexane+0.25% TEA, rinsing with two 0.25 ml portions of hexane and
two 0.5 ml portions of eluent. The hexane insoluble residue was dissolved in
1.0 ml DMF. Chromatography of the hexane soluble portion, combining all
product containing fractions, purified the isomers and enriched the isomer
ratio to ꢀ20% correct isomer. A second re-purification gave about a ꢀ50 mCi
1:1 mixture of isomers. The pure desired isomer 6a (26 mCi) was isolated by
preparative HPLC (4.6 ꢂ 250 mm Zorbax RX-Silica; 0.1% TEA in pet ether,
2 ml/min; dry b-Ram cell; compound 6b 8 min, compound 6a 12.5 min).
(1R,3aR,7aR)-4-[2-[(3R,5R)-3,5-Bis-(tert-butyl-dimethyl-silanyloxy)-cyclo-
hexylidene]-eth-(Z)-ylidene]-[3a,5,5-3H]-7a-methyl-1-[5-methyl-1-(4-methyl-
4-trimethylsilanyloxy-pentyl)-5-trimethylsilanyloxy-hexyl]-octahydro-indene (8)
The reaction was run on a vacuum line using a #7 O-ring ‘Tribone’ manifold
containing four stopcocks. The stopcock at the top was used to isolate the
manifold from a Firestone valve (that allows for vacuum or atmospheric
nitrogen pressure), and stopcocks at each of three arms. One arm was used for
the attachment of a 5 ml pear shaped septum side-arm reaction flask
containing a spin bar. The second arm was used for the attachment of a
10 ml pear shaped septum side-arm phosphene oxide anion generation flask
containing a spin bar. The third arm was used for attachment of either a
distillate receiver or a THF/LAH reservoir. The system was dried by pump
overnight. From this point on all products were treated as air and light
sensitive. A 0.5 ml toluene solution of 18.5 mCi (ca. 0.4 mmol) of the labeled
bis-TMS ketone substrate 6a was transferred to the reaction flask. The toluene
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

HIGH SPECIFIC ACTIVITY TRITIUM LABELING
1021
was removed by vacuum transfer to a distillate receiver flask. Two 0.5 ml
portions of dry toluene were used to rinse in the remainder of the substrate
into the reaction flask, followed by removal of volatiles by vacuum transfer.
The distillate receiver flask was replaced by the THF/LAH reservoir flask.
About 0.5 ml THF was distilled into the reaction flask containing the semi-
solid substrate residue. The residue was dissolved by use of an internal spin
bar, and cooled to ꢁ72 8C. The phosphine oxide was dried by twice vacuum
transfer of volatiles from 1 ml toluene solutions of the phosphine oxide 7
(18.8 mg, 0.033 mmol, containing 4% water) contained in the anion generation
flask. The residue was redissolved in about 1 ml THF distilled into the flask.
The anion (deep orange-red) was generated by adding 16 ml of 1.6 N n-BuLi/
hexane (0.026 mmol) to the stirred ꢁ72 8C solution. With both flasks at
ꢁ72 8C, the anion was transferred dropwise (by syringe injection of dry
nitrogen into the sealed anion flask) by canula to the reaction flask until an
orange color remained (9 drops), a further 9 drops was then added. The
titration procedure was continued at 30–60 min increments until the orange
color remained in the reaction flask for about 30 min (about 2.5 h total and
about 15 drops total more anion added). The reaction was quenched by canula
transfer from the reaction flask to a well stirred ambient temperature 2 ml 2 M
pH 7 sodium phosphate buffer solution. After rinsing in with distilled THF
(5 ml), the aqueous phase was extracted three times with hexane (2 ml ea.). The
combined hexane/THF phase was dried with brine and anhydrous sodium
sulfate, filtered, and concentrated. The residue was redissolved in 0.5 ml
hexane and applied to 8 ml silica gel pre-equilibrated with the eluent (1.25%
EtOAc/hexane+0.125% TEA) in a 10 ml disposable column. Following flash
chromatography, the product containing fractions were combined and
concentrated to dryness and re-dissolved in hexane affording a nearly
quantitative recovery of compound 8 (TLC: Silica gel: 5% EtOAc/
hexane+0.5% TEA, Rf 0.8).
(1R,3R)-5-{2-[(1R,3aR,7aR)-[3a,5,5-3H]-1-[5-hydroxy-1-(4-hydroxy-4-
methyl-pentyl)-5-methyl-hexyl]-7a-methyl-octahydro-inden-(4Z)-ylidene]-
ethylidene}-cyclohexane-1,3-diol (9)
3
A portion (90%) of the hexane solution of H-labeled tetra-silyl substrate 8
from the previous step was transferred to a 10 ml pear shaped flask and the
volatiles were removed by vacuum transfer. The residue was treated with a
large excess (1 ml) of tetra-butyl ammonium fluoride 1 M in THF. The
solution (wrapped in foil to exclude light) was stirred for 15 h, and then
partitioned between 5 ml ea. water and EtOAc. The aqueous phase was
extracted twice with 5 ml portions of EtOAc, and the combined organic phase
was dried with brine and anhydrous Na2SO4. After filtration, the filtrate was
concentrated to dryness and immediately re-dissolved under nitrogen in 0.5 ml
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

1022
S. A. DE KECZER ET AL.
EtOAc. This solution was applied to a 10 ml disposable flash column
containing 8 ml silica gel in EtOAc. The product was eluted with 50 ml of
EtOAc, then 50 ml 2% i-PrOH/EtOAc, and finally, 50 ml 5% i-PrOH/EtOAc.
Product containing fractions were combined affording 12.5 mCi of 9. A second
run afforded 22 mCi of 9.
Conclusion
The method that we have developed for T2O exchange labeling was based on
D2O exchange labeling experiments. This method is quite robust based on the
reproducibility of the tritiation run and trial deuterium incorporation
experiments. The reactions run in DMF were relatively insensitive to time
(12 min–14 h), temperature (22–80 8C), and scale (3–55 mmol substrate) as long
as 2.5 equiv. of NaOMe and at least 0.5 ml of D2O/mmol of substrate was used.
The method was readily translated to T2O exchange labeling at scale.
Following procedures that were worked out using the deuterium analog, the
pure single isomer 6a was coupled to the anion of the extended A-ring
phosphene oxide 7 using Horner-Wittig Olefination (0.4 mmol scale). The
coupling product 8 was deprotected to afford 34 mCi [9,9,14a-3H]-RO275646 9
(from two runs) in >99% purity by HPLC and having a specific activity of
64 Ci/mmol.
Acknowledgements
We wish to acknowledge Kate Comstock from the Mass Spectroscopy group
for her expert analytical support, and the UC Berkeley National Tritium
Labelling Facility (NTLF) and especially Hiromi Morimoto and Dr Phil
Williams for sharing their facility and for their expertise in the preparation and
incorporation of T2O.
References
1. Zhu G-D, Okamura WH. Chem Rev 1995; 95: 1877–1952.
2. Hartung T, Hilpert H, Mohr P. Synthesis of the carbon-14 labelled antipsoriatic
vitamin D3 derivative RO 65-2299. In Synthesis and Applications of Isotopically
Labelled Compounds, Pleiss U, Voges R (ed). vol. 7. Wiley: New York, 2000;
189–192.
3. Napoli JL, Fivizzani MA, Hamstra AJ, Schnoes HK, DeLuca HF. Anal Biochem
1979; 96: 481–488.
4. Perlman K, Schnoes HK, Tanaka Y, DeLuca HF, Kobayashi Y, Taguchi T.
Biochemistry 1984; 23: 5041–5048.
5. Yamada S, Schnoes HK, DeLuca HF. Anal Biochem 1978; 85: 34–41.
6. DeLuca HF, Weller M, Blunt JW, Neville PF. Arch Biochem Biophys 1968; 124:
122–128.
7. Hoeger CA, Okamura WH. J Am Chem Soc 1985; 107: 268–270.
8. Neville PF, DeLuca HF. Biochemistry 1966; 5: 2201–2207.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023

HIGH SPECIFIC ACTIVITY TRITIUM LABELING
1023
9. Holick SA, Holick MF, Frommer JE, Henley JW, Lenz JA. Biochemistry 1980;
97: 1031–1037.
10. Holick MF, Tavela TE, Holick SA, Schnoes HK, DeLuca HF, Gallagher BM.
J Biol Chem 1976; 251: 1020–1024.
11. Okamura WH, Hoeger CA, Miller KJ, Reischl W. J Am Chem Soc 1988; 110:
973–974.
12. Perlmam KL, Swenson RE, Paaren HE, DeLuca HF. Tetrahedron Lett 1991; 32:
7663–7666.
13. Ahene AB, Chang FE, Masjedizadeh MR, Wu SY, de Keczer SA, Tarnowski T.
Poster presented at AAPS Meeting in Salt Lake City, October 26–30, 2003.
14. Hilpert H, Wirz B. Roche Internal Report, 2000.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 1013–1023