Synthesis of [C2H3]methotrexate and [C2H3]7-hydroxymethotrexate

Article abstract of DOI:10.1002/jlcr.537

Stable label analogues of methotrexate (MTX) and 7-hydroxymethotrexate (7-OH-MTX) were required for use as internal standards for LC/MS quantitation. A minimum incorporation of stable isotopes to produce a mass increase of 3 in the non-glutamate derived p

Full text of DOI:10.1002/jlcr.537

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Labelled Cpd Radiopharm 2002; 45: 29–36.
DOI: 10.1002/jlcr.537
Research Article
Synthesis of [C2H3]methotrexate and
[C2H3]7-hydroxymethotrexate
Charles S. Elmore*, Dennis C. Dean, Yong Zhang, Carter Gibson,
Herb Jenkins, Allen N. Jones and David G. Melillo
Department of Drug Metabolism, Merck Research Laboratories, RY80R-104,
PO Box 2000, Rahway, NJ 07065, USA
Summary
Stable label analogues of methotrexate (MTX) and 7-hydroxymethotrexate
(7-OH-MTX) were required for use as internal standards for LC/MS
quantitation. A minimum incorporation of stable isotopes to produce a mass
increase of 3 in the non-glutamate derived portion of the molecule was
necessary for adequate MS detection. The commercial availability of des-
methyl MTX (aminopterin, 1) made methylation with C2H3I an attractive
option. Surprisingly, all attempted methylations of 1 and the dimethyl ester of
1 failed to provide a significant amount of the methylated aniline, apparently
due to attenuated reactivity of the secondary amine towards alkylation.
However, reductive amination of diacid 1 with C2H2O and NaB2H3CN gave
[C2H3]MTX in 52% yield. A previously reported method was utilized to
convert [C2H3]MTX to [C2H3]7-OH-MTX. Preparative HPLC purification of
[C2H3]7-OH-MTX resulted in extremely low recovery from the column; this
was resolved by switching to a column with few free silanols. Copyright #
2002 John Wiley & Sons, Ltd.
Key Words: methotrexate; 7-hydroxymethotrexate; rofecoxib; cyclooxygenase
II; reductive amination
*Correspondence to: C. S. Elmore, Department of Drug Metabolism, Merck Research
Laboratories, RY80R-104, PO Box 2000, Rahway, NJ 07065, USA.
E-mail: charles elmore@merck.com
Copyright # 2002 John Wiley & Sons, Ltd.
Received 27 August 2001
Revised 15 October 2001
Accepted 16 October 2001

30
C. S. ELMORE ET AL.
Introduction
Methotrexate (MTX) is a drug with several clinical applications
including treatment of rheumatoid arthritis1 which could lead to co-
administration of MTX and cyclooxygenase II (COX II) selective
inhibitors such as rofecoxib (Vioxx). Severe toxic effects have been
observed when MTX and several non-steroidal antiinflammatory drugs
(NSAIDs) are co-administered.2 Two hypotheses which attempt to
explain this observation are: (1) competition for renal clearance between
the NSAID and MTX and (2) cyclooxygenase I (COX I) induced
suppression of renal clearance;3 both hypotheses would lead to a longer
half-life for MTX. To probe whether rofecoxib displays the same drug–
drug interaction as the NSAIDS tested, a LC/MS internal standard of
MTX and its primary metabolite 7-hydroxymethotrexate (7-OH-MTX)
were required for which stable isotope labeled versions of each were
deemed ideal.
O
CO H
2
O
N
H
CO H
2
X
O
N
N
N
N
CH
3
N
X
SO Me
2
H N
2
NH
2
Rofecoxib
=
H
Methotrexate (MTX)
=
OH 7-Hydroxymethotrexate (7-OH-MTX)
Results and discussion
The syntheses of stable isotope labeled MTX with C-13 labels in the
pteridine ring has been reported by Cheung.4 However, the conversion
of commercially available aminopterin (1) to [C2H3]MTX was a less
complex route to the target and would provide a mass increase of 3
which was the minimum enrichment required for adequate discrimina-
tion from unlabeled drug by MS. The synthesis of 7-OH-MTX from
MTX has been reported by Dawson et al.;5 therefore, a successful
synthesis of [C2H3]MTX would provide access to both compounds.
Labeled N-methylation of the aniline nitrogen of aminopterin could be
conducted using C2H3I or by reductive amination with C2H2O/
NaB2H3CN (Scheme 1).
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

SYNTHESIS OF [C2H3] METHOTREXATE
31
O
CO H
2
N
H
CO H
2
N
H
N
N
N
N
2
2
H N
2
NH
1
[C H ]MTX
2
3
[C H ]7-OH-MTX
3
+
2
C H O
2
2
or
C H I
3
+
2
NaB H CN
3
Scheme 1. Retrosynthetic analysis of [C2H3]7-OH-MTX and [C2H3]MTX
O
CO H
2
O
CO H
2
N
H
CO H
2
N
H
CO H
2
N
H
N
C
N
N
2
2
N
N
1. C H O, HCl,
2
2
2
2
N
N
NaB H CN, H O
H
3
3
2
N
N
2. Preparative HPLC
43%
2
H N
2
NH
[C H ]MTX
3
H N
2
NH
2
2
1
Scheme 2. Synthesis of [C2H3]MTX
Reaction of aminopterin (1) with C2H3I was first explored.
Unfortunately, repeated attempts under several different reaction
conditions always produced the desired methylated product in less
than 10% yield. Several other major by-products were formed which
had mass spectra consistent with mono-methylation at an alternative
site and dimethylation.
Reductive amination of aminopterin (1) with C2H2O and NaB2H3CN
2
in H2O gave a 52% yield of [C2H3]MTX along with a 35% impurity
which had the same M+H/Z ratio by LC/MS as [C2H3]MTX (Scheme 2).
Sparging the reaction mixture with N2 prior to reaction and maintaining
the pH of the solution between 6 and 7 as the reaction progressed were
critical to the success of the reaction. A portion of the reaction mixture
was purified by preparative HPLC to give [C2H3]MTX in moderate
yield and 99% purity.
For the preparation of labeled 7-OH-MTX, crude [C2H3]MTX was
esterified using boron trifluoride in methanol,5 and the reaction mixture
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

32
C. S. ELMORE ET AL.
purified by column chromatography to give diester 2 in 75% yield and
88% purity. Esterification of 1 followed by reductive amination also
provided 2, but the reaction was heterogenous and had considerable
run-to-run variability. As reported previously,5 reaction of 2 with diethyl
cyanophosphonate gave nitrile 3, which was converted to [C2H3]7-OH-
MTX by reaction with aqueous sodium hydroxide (Scheme 3).
Precipitation of 7-OH-MTX from the reaction solution by adjusting
the pH to 3.5 with AcOH provided material with 93% purity, which was
not sufficient for our purposes. An initial probe of the preparative
HPLC purification (Zorbax RX C8, MeCN–0.1% TFA) resulted in low
recovery; this result was confirmed with unlabeled reference 7-OH-
MTX, which gave a 15% recovery after preparative HPLC. Surpris-
ingly, attempted isolation of the reference compound by SepPak (ODS-
AM C18) resulted in a further 50% loss of material. The loss of material
on the ODS-AM C18 SepPak and the Zorbax C8 column could be the
result of strong binding of 7-OH-MTX to uncapped silanol sites on the
reversed-phase packing. Analysis of several other columns showed
improved recovery from Hamilton PRP-1 polymer-based column and
the highly capped Waters Xterra RP18 column (60–90%). Use of a
polymer-based SepPak (Waters OasisTMHLB) resulted in quantitative
recovery of reference 7-OH-MTX. This information was then applied to
the purification of [C2H3]7-OH-MTX by iterative preparative HPLC on
O
CO Me
2
O
CO Me
2
N
H
CO Me
2
N
H
CO Me
2
CN
N
N
N
N
N
N
2
N
C H
(EtO) P(O)CN
3
2
2
N
N
C H
3
N
NMP, H O, 80 ºC
2
H N
2
NH
2
H N
2
NH
2
2
3
O
CO H
2
N
H
CO H
2
OH
1. NaOH, H O, THF
2
N
N
N
2. Preparative HPLC
2
N
C H
3
N
11% from 2
H N
2
NH
2
2
[C H ]7-OH-MTX
3
Scheme 3. Synthesis of [C2H3]7-OH-MTX from diester 2
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

SYNTHESIS OF [C2H3] METHOTREXATE
33
a Waters Xterra RP18 column followed by isolation using a Waters
OasisTMHLB SepPak, and after lyophilization, provided 21 mg of
[C2H3]7-OH-MTX as a yellow solid (11% yield from 2, 50% recovery).
HPLC analysis showed the compound to be 99% pure and LC/MS
2
2
showed 93% H3, 7% H2.
Experimental
General: Aminopterin, methotrexate, and 7-hydroxymethotrexate were
2
2
obtained from Sigma. NaB2H3CN (96% H), 20% C2H2O in H2O
2
2
2
2
2
(98% H), 20% HCl in H2O (99.5% H), 40% NaO2 H in H2O
2
2
2
(99.9% H) were obtained from Aldrich, and H2O (99.9% H) was
obtained from Isotec. Anhydrous solvents were obtained from Aldrich
(
and were dried over 4 A molecular sieves for at least 24 h prior to use.
Analytical HPLC was performed using a Shimadzu HPLC system with
LC-10ATVP pumps, SPD-10AVP UV detector, CTO-10ASVP column
oven heated to 308C, and a SCL-10A controller. 1H NMR spectra were
recorded on a Varian U-400 spectrometer and are referenced to the
2
residual solvent peak (2.5 for H5-DMSO). LC/MS data was acquired
using an HP MSD-100 with electrospray ionization. The final products
were identified by HPLC comparison with the commercially available
material using either method A (12% MeCN–0.1% trifluoroacetic acid
on a Zorbax RX C-18 column); method B (15% MeCN–0.1%
trifluoroacetic acid on a Zorbax RX C-18 column); method C (5–40%
MeCN–0.1% trifluoacetic acid over 30 min, Waters Xterra C-8); or
method D (20:10:70 MeCN:MeOH:0.1% HClO4, Discovery Amide RP-
C16). All HPLC analyses were conducted with the column heated to
308C and concluded with a 10 min wash of 100% MeCN. Purity is
reported as UV area % (wavelength).
[C2H3]Methotrexate: A suspension of 1.0 g (2.1 mmol) of aminopterin
in 100 ml of 2H2O was sparged with N2 for 30 min at rt, and the pH was
adjusted to 9 with 40% NaO2 H in 2H2O to give a homogenous solution.
2
The pH of the reaction mixture was adjusted to 6.5 with 1 M HCl in
2
2H2O, and 3.36 ml (21.0 mmol) of a 20% solution of C2H2O in H2O
was added. The reaction mixture was sparged for 10 min with N2, and
208 mg (3.16 mmol) of NaB2H3CN was then added. The pH was closely
monitored for the next 1 h and was kept in the range of 6–7 by addition
2
of 40% NaO2H in H2O. HPLC assay (method A) at 1 h showed no
starting material remaining so the pH of the solution was adjusted to 3.5
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

34
C. S. ELMORE ET AL.
with aqueous HCl and the yellow solid was isolated by filtration to give
971 mg (58%) with a 60% purity (254 nm). Purification of 88 mg was
effected by preparative HPLC (21.2 ꢀ 250 mm Zorbax RX C-18, 9%
MeCN–0.1% TFA, 20 ml/min) in four batches. The fractions contain-
ing product were combined, and the pH of the solution adjusted to 3.6
with aqueous NaOH. The MeCN was removed at reduced pressure and
the solution was lyophilized to give 1.1 g of a yellow solid. The solid was
dissolved in a minimum of water and was passed through a Waters C-18
SepPak (300 mg, 3 cm3). The SepPak was rinsed with 5 ml of water, and
the product was eluted with MeOH. The organic solution was dried
(MgSO4), filtered and concentrated to dryness. The solid was then taken
up in 2 ml of water and the pH of the solution was adjusted to 3.6. After
14 h at 48C the solution was filtered to give 39 mg of [C2H3]MTX as a
yellow solid (analysis by HPLC (method D)) 99.7% (216 nm), 98.7
(242 nm), 99.0 (306 nm). MS (M+H+/Z, abundance): 458 (92.3%), 457
1
2
(7.1%) H NMR (400 MHz, H6-DMSO) d 8.67 (s, 1H), 7.71(d, 2H,
J ¼ 8:8 Hz), 6.78 (d, 2H, J ¼ 8:8 Hz), 4.82 (s, 2H), 4.32 (m, 1H), 2.29 (t,
2H, J ¼ 7:2 Hz), 2.02 (m, 1H), 1.90 (m, 1H).
Dimethyl ester of [C2H3]methotrexate (2): A suspension of 875 mg
(1.91 mmol) of [C2H3]MTX (60% purity (254 nm)) in 19 ml of MeOH
.
was stirred at rt as 2.10 g (14.7 mmol) of BF3 OEt2 was added in two
portions, and the reaction mixture was stirred overnight. HPLC analysis
(method B) showed no remaining diacid so the mixture was diluted in
300 ml of CHCl3 and washed with 3 ꢀ 100 ml of NaHCO3 and 60 ml of
saturated aqueous NaCl. The organicsolution was dried (MgSO ),
4
filtered, and concentrated to afford 1.69 g of an orange oil. A portion of
the sample (1.5 g) was purified by flash column chromatography on
silica (5% MeOH–CHCl3) to give 660 mg of a light yellow solid (88.5%
1
2
purity at 254 nm, 75% yield). H NMR (400 MHz, H6-DMSO) d 8.54
(s, 1H), 7.68 (d, 2H, J ¼ 8:9 Hz), 6.79 (d, 2H, J ¼ 8:9 Hz), 4.75 (s, 2H),
4.36 (m, 1H), 3.57 (s, 3H), 3.54 (s, 3H), 2.39 (t, 2H, J ¼ 7:4), 2.04 (m,
1H), 1.94 (m, 1H).
Dimethyl ester of [C2H3]7-cyanomethotrexate (3): The following
procedure is a modification of that reported by Dawson.5 A solution
of 200 mg (0.41 mmol, 89% UV purity at 254 nm) of diester 2 in 2.9 ml
of N-methylpyrrolidinone was sparged for 10 min with N2 then 1.21 g
(12.0 mmol) of triethylamine, 1.0 ml (6.2 mmol) of diethyl cyanopho-
sphonate (DEPC), and 530 ml of water were added sequentially to the
solution at 08C. The mixture was stirred for 10 min at 08C and 2.5 h at
808C. HPLC (method C) showed a 2:1 ratio of 2:3 so 1.21 g (12.0 mmol)
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

SYNTHESIS OF [C2H3] METHOTREXATE
35
of triethylamine, 1.0 ml (6.2 mmol) of DEPC, and 530 ml of water were
added and the heating was continued. After another 1.5 h, HPLC
analysis (method C) showed a 1:11 ratio of 2:3; therefore, 100 ml of
EtOAcwas added and the solution was washed with 50 ml of NaHCO 3
and 2 ꢀ 50 ml of saturated NaCl. The solution was dried (Na2SO4),
filtered and concentrated to give 229 mg of a brown oil (55.8% purity at
254 nm).
[C2H3]7-Hydroxymethotrexate: The following procedure is a modifica-
tion of that reported by Dawson.5 A solution of [C2H3]7-cyanometho-
trexate in 20 ml of THF and 5 ml of water was cooled to 08C and 2.4 ml
of 1 M NaOH was added. The reaction mixture was stirred for 3 h at
which time HPLC analysis (method C) showed no remaining starting
material. The THF was removed at reduced pressure, and the pH of the
solution was adjusted to 6.5 with HCl. The remaining water was
removed at reduced pressure (no heating) to give 690 mg of a red oil.
Purification was effected in 15 batches using a Gilson Automated
Preparative HPLC system (19 ꢀ 300 mm Waters Xterra RP18, 11%
MeCN–0.1% TFA, 20 ml/min, sample loaded in 10% MeCN–1%
TFA). The fractions containing product were combined, and the MeCN
was removed at reduced pressure. The solution was loaded onto a
Waters OasisTMHLB SepPak (500 mg, 6 ml) and was washed with water
and MeOH, then the product eluted with 50% aqueous MeCN. The
solution was lyophilized to give 21 mg (11% yield from 2) of a yellow
solid [HPLC (method E) purity 97.8% (202 nm), 98.0 (220 nm), 96.9
(242 nm), 98.9 (306 nm)]. MS (M+H+/Z, abundance): 474 90.3%, 473
1
(9.7%) H NMR (2H6-DMSO) d 7.69 (d, 2H, J ¼ 8:5 Hz), 6.79 (d, 2H,
J ¼ 8:5 Hz), 4.54 (s, 2H), 4.33 (dd, 1H, J ¼ 5:0, 9.5 Hz), 2.31 (t, 2H,
J ¼ 7:5 Hz), 2.03 (m, 1H), 1.91 (m, 1H).
References
1. Bertino JR, Kamen B, Romanini A. Chapter 60: folate antagonists. In
Cancer Medicine (4th edition), Holland JF, Frei III E, Blast RC, Kufe DW,
Morton DL, Weichselbaum RR (eds.). Williams and Wilkins: Baltimore,
1997; 907–922.
2. Singh RR, Malaviya AN, Pandey JN, Guleria, JS. Lancet 1986; 327: 1390;
Badr MZ, Chen TS. Toxicology 1985; 34: 333–340; Daly H, Boyle J,
Roberts C, Scott G. Lancet 1986; 327: 557; Daly HM, Scott GL, Boyle J,
Roberts CJC. Br J Dermatology 1986; 114: 733–735; Thyss A, Milano G,
Kubar J, Namer M, Schneider M. Lancet 1986; 327: 256–258.
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36

36
C. S. ELMORE ET AL.
3. Huang KC, Wenczak BA, Liu YK. Cancer Res 1979; 39: 4843–4848.
4. Cheung HTA, Smal M, Chau DD. Heterocycles 1987; 25: 507–514; Cheung
HTA, Gray PG. J Labelled Compd Radiopharm 1984; 21: 471–483.
5. Dawson MI, O’Krongly D, Hobbs PD, Barrueco JR, Sirotnak FM.
J Pharm Sci 1987; 76: 635–638.
Copyright # 2002 John Wiley & Sons, Ltd.
J Labelled Cpd Radiopharm 2002; 45: 29–36