The development of improved syntheses of PPARγ-sparing, insulin sensitizing thiazolidinedione-ketones

Article abstract of DOI:10.1016/j.tetlet.2019.07.022

Ketones 2 (MSDC-0160) and 3 (MSDC-0602) had been selected for clinical development, however their initial syntheses were considered suboptimal for application deep into clinical trials. Difficulties ranging from the nature of the starting material, alcohol oxidation problems, epoxide opening regioisomeric issues, and endgame ketone redox problems had been encountered. Direct ketone introduction/maintenance was desired for maximum efficiency and convergence was found to be critically dependent upon the acidity of the nucleophilic species (13, 18) and the use of pre- or post-alkylative oximino-ether/oxime protection (vide infra). Improvements in overall yield for the syntheses of 2 (MSDC-0160) and 3 (MSDC-0602) from 20% (2) and 31% (3) respectively, to 44% (2) and 59% (3) were realized.

Full text of DOI:10.1016/j.tetlet.2019.07.022

Tetrahedron Letters xxx (xxxx) xxx
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
The development of improved syntheses of PPAR
sensitizing thiazolidinedione-ketones
c-sparing, insulin
Steven P. Tanis a, , Jerry R. Colca b,c, Timothy T. Parker d, Gerald D. Artman III d,1, Scott D. Larsen e,
⇑
Robert C. Gadwood d, James R. Zeller f
a SPTanis PharmaChem Consulting LLC, Carlsbad, CA 92011, United States
b Metabolic Solutions Development Co., Kalamazoo, MI 49007, United States
c Cirius Therapeutics, 12651 High Bluff Drive, Suite 150, San Diego, CA 92130, United States
d Kalexsyn, Kalamazoo, MI 49008, United States
e Vahlteich Medicinal Chemistry Core, University of Michigan, Ann Arbor, MI 48109, United States
f Jim Zeller Consulting, Scottsdale AZ 85259, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Ketones 2 (MSDC-0160) and 3 (MSDC-0602) had been selected for clinical development, however their
initial syntheses were considered suboptimal for application deep into clinical trials. Difficulties ranging
from the nature of the starting material, alcohol oxidation problems, epoxide opening regioisomeric
issues, and endgame ketone redox problems had been encountered. Direct ketone introduction/mainte-
nance was desired for maximum efficiency and convergence was found to be critically dependent upon
the acidity of the nucleophilic species (13, 18) and the use of pre- or post-alkylative oximino-ether/oxime
protection (vide infra). Improvements in overall yield for the syntheses of 2 (MSDC-0160) and 3 (MSDC-
0602) from 20% (2) and 31% (3) respectively, to 44% (2) and 59% (3) were realized.
Received 13 May 2019
Revised 28 June 2019
Accepted 8 July 2019
Available online xxxx
Ó 2019 Elsevier Ltd. All rights reserved.
Introduction
structure that would maintain the interaction with the recently
identified MPC and minimize interaction with the nuclear receptor.
Novel thiazolidinediones (TZDs) are now undergoing clinical
development in multiple therapeutic areas based on their ability
to reduce insulin resistance and metabolic dysfunction [1]. These
new analogs have been designed based on the developing under-
standing that pharmacology against metabolic dysfunction is
mediated by attenuation of pyruvate uptake through the mito-
chondrial pyruvate carrier (MPC); this recent understanding has
opened up an alternate route into new insulin sensitizers [1,2].
We have summarized the medicinal chemistry approach to limit
Several approaches were undertaken including introduction of
polar substituents, but the successful effort involved introduction
of a ketone functionality [3] and in the progressions of two clinical
candidates shown in Fig. 1. Both of these clinical candidates have
been shown to be stereoselectively converted to alcohol
metabolites that also spare the direct activation of PPARc [1,3,5].
Compound 1 (MSDC-0160) is progressing through clinical develop-
ment as a potential treatment of neurodegeneration [6–8] and
compound 2 (MSDC-0602) is currently in clinical development as
a potential treatment for non-alcoholic steatohepatitis (NASH)
[9,10].
As ketones 2 and 3 progressed into development we examined
the fitness of the initial synthetic processes and found them want-
ing. Scheme 1 depicts the original synthesis of 2 which utilized
pioglitazone as the starting material [11]. Pioglitazone 1 was oxi-
dized (MCPBA) to afford the related N-oxide (4, 96%) which was
smoothly rearranged and hydrolyzed to give the racemic hydroxyl
compound 5 (74%), setting the stage for a simple benzylic alcohol
oxidation to provide the target ketone 2. That oxidation proved
to be somewhat problematic with a wide variety of standard oxi-
dation conditions [12a] leading to low yields and the production
of considerable impurities. The modified DMSO/P2O5 oxidation
reported by Taber [11,12b] resulted in a good yield (88%) of 2.
direct binding to the nuclear receptor PPARc [3], which is known
to drive dose-limiting side effects of the first generation TZDs.
These efforts took advantage of the fact that marketed drug piogli-
tazone 1 (ActosÒ, Fig. 1) was known to be effective as anti-diabetic
drug both in terms of treating the symptoms of diabetes and, more
importantly, in reducing adverse cardiovascular outcomes;
however, the use of pioglitazone is still limited by its additional
effect of directly activating PPARc, which drives the dose-limiting
side effects of volume expansion, edema, and bone loss [4]. We rea-
soned that it made sense to look for modest changes in a proven
⇑
Corresponding author.
E-mail address: sptanis@sbcglobal.net (S.P. Tanis).
Current address: Celgene, Summit, NJ 07901, United States.
1
https://doi.org/10.1016/j.tetlet.2019.07.022
0040-4039/Ó 2019 Elsevier Ltd. All rights reserved.
Please cite this article as: S. P. Tanis, J. R. Colca, T. T. Parker et al., The development of improved syntheses of PPAR
lidinedione-ketones, Tetrahedron Letters, https://doi.org/10.1016/j.tetlet.2019.07.022
c-sparing, insulin sensitizing thiazo-

2
S.P. Tanis et al. / Tetrahedron Letters xxx (xxxx) xxx
O
O
O
NH
NH
S
NH
S
S
N
O
N
O
H3CO
O
O
O
O
O
O
Pioglitazone (Actos®) 1
2 (MSDC-0160)
3 (MSDC-0602)
Fig. 1. Structures of Pioglitazone and Clinical Candidates 2 and 3.
O
O
b , c
a
NH
S
NH
O
S
N
O
N
O
O
O
1
4
O
O
d
NH
NH
S
S
N
O
N
O
O
O
OH
O
5
2 (MSDC-0160)
Scheme 1. (a) MCPBA, CH2Cl2, 96%; (b) TFAA, CH2Cl2,
D; (c) aq. NaHCO3, THF, 74%; (d) DMSO, P2O5 Et3N, CH2Cl2, 88%.
An approach, starting from the pyridyl epoxide 6, was the first
in the Scheme 3 approach wherein the ketone was directly
introduced.
synthesis to afford large quantities of 2 without starting from
pioglitazone 1 (Scheme 2) [13]. Treatment of the epoxide with
the sodium salt of 4-hydroxybenzaldehyde (phase transfer
water-toluene, PEG 4000) gave (60%) a 1:4 mixture of 7 and 8. Kno-
evenagel condensation with 2,4-thiazolidinedione afforded a mix-
ture of 9 and the Knoevenagel product of 7, leading to alcohol 9
(56%) after trituration with methanol. Alcohol 9 was transformed
to ketone 2 upon cobalt mediated reduction [11] followed by an
application of the Taber [11,12b] oxidation conditions. An overall
yield of ca. 20% was realized, however the epoxide 6 is not readily
available and its synthesis requires 4 steps from 2-methyl-5-ethyl
pyridine.
Scheme 3 outlines our initial, more direct approach to ketone 3
[3]. Direct ketone introduction was accomplished by an alkylation
of the readily available 3-methoxy-phenacyl bromide 10 to give 11
(62%). Knoevenagel condensation afforded 12 (76%), which, as
expected, suffered double reduction when exposed to the conju-
gate reduction conditions to give 13 (85%). Benzylic alcohol oxida-
tion in this non-pyridyl molecule was more readily accomplished
with IBX to provide 3 (77%). An overall yield of 31% was realized
The syntheses presented in Schemes 2 (for 2) and Scheme 3 (for
3) had been utilized to provide clinical supplies to date. As we
anticipated possible transition into Phase 3 clinical development,
we felt that these routes presented issues which would make fur-
ther scale-up difficult. The direct ketone introduction in Scheme 3
is noteworthy, however Scheme 3 does not deal with the undesired
reduction of an installed ketone and a subsequent re-oxidation,
resulting in a modest 31% overall yield. Improved syntheses of 2
and 3 must be higher in yield, they should be more convergent,
and direct ketone introduction would be desirable as long as we
avoid ketone red-ox in the end game. Success in these endeavors
would enable facile production of needed clinical supplies.
More concise syntheses of 2 and 3 might result if a reduced 5-
(4-hydroxybenzyl)-2,4-thiazolidinedione [14] right-hand half moi-
ety 14 could be directly alkylated with a phenacylhalide such as
the commercially available 3-methoxy-phenacyl bromide 10 of
Scheme 3, leading directly to 3, or 2-bromo-1-(5-ethyl-2-pyridi-
nyl)-ethanone 15, not reported before these studies, which would
afford 2 as shown in Scheme 4.
CHO
a
O
+
CHO
N
N
N
O
O
OH
OH
6
7
8
O
O
d , e
b , c
NH
NH
S
S
N
O
N
O
O
O
OH
O
9
2 (MSDC-0160)
Scheme 2. (a) NaOH, 4-HO-PhCHO, PhCH3, PEG 4000 60% 7:8 1:4; (b) 2,4-thiazolidinedione, pyrrolidine, CH3OH,
D; (c) separate, 56%; (d) CoCl2-6H2O, dimethylglyoxime,
NaBH4, THF, aq. NaOH; (e) P2O5, DMSO, Et3N, CH2Cl2 75%
Please cite this article as: S. P. Tanis, J. R. Colca, T. T. Parker et al., The development of improved syntheses of PPAR
lidinedione-ketones, Tetrahedron Letters, https://doi.org/10.1016/j.tetlet.2019.07.022
c-sparing, insulin sensitizing thiazo-

S.P. Tanis et al. / Tetrahedron Letters xxx (xxxx) xxx
3
Scheme 3. (a) 4-HO-benzaldehyde, K2CO3, acetone 62%; (b) 2,4-thiazolidinedione, piperidine, abs. EtOH,
85%; (d) IBX, EtOAc, , 77%
D
, 76%; (c) CoCl2-6H2O, 2,20-bipyridine, 1 N aq. NaOH, THF, NaBH4
D
Scheme 4. An idealized convergent approach to 2 and 3.
position of the more reactive 15 observed. When the more acidic
phenol 19 [14,19] (c-pKa = 9.14, calculated in water) was
employed in this reaction (Scheme 6, panel B), a good yield (85%)
of 12 (Scheme 3) was obtained and a poor yield of 20 in the pyridyl
series (0–20%), again with extensive decomposition of 15, was
realized.
Scheme 5. (a) i. n-BuLi, PhCH3, ꢀ40 °C, ii. N,N-dimethylacetamide, 80%; (b) PdCl2-
DPPF, Et2Zn, dioxane, 50 °C, 88%; (c) Br2, HBr-HOAc, 80%.
Clearly, the acidity of the phenol had a major impact on the
alkylation in the phenyl series.
The required 2-bromo-1-(5-ethyl-2-pyridinyl)-ethanone 15
was prepared as shown in Scheme 5 [15]. 2,5-Dibromopyridine
16 is selectively lithiated (n-BuLi) at the 2 position and reacted
with dimethylacetamide to afford 17 (80%) [16]. The 5-position
ethyl is then installed by a palladium mediated reaction with Et2-
Zn [17] to give 18 and then bromination (Br2, HBr/HOAc) provides
the HBr salt of the desired pyridyl bromoketone 15 (80%) [18].
The direct coupling of phenacyl halides with 5-(4-hydroxyben-
zyl)-2,4-thiazolidinedione 14 (c-pKa = 6.38 (TZD) and 9.92 (phe-
nol), calculated in water) [14,19] will require the proper selection
As shown in Scheme 6 (panel A and B), the reactivity of the pyr-
idyl-phenacyl bromide precluded a successful alkylation. As the
reactions shown in Scheme 6 are conducted in DMSO, a pKa com-
parison of the phenol nucleophiles employed (14 and 19) and the
electrophiles utilized (10 and 15) will assist in our analysis of the
Scheme 6 data. 4-Methylphenol has a reported DMSO pKa = 18.9
[20a] suggesting the pKa of 14 in DMSO should be approximately
18.9, with no direct DMSO comparator for 19, the
D value for the
water calculated pKa value (ca. 0.8) will be applied to suggest a
DMSO pKa = 18.1 for 19. The DMSO pKa values for 10 and 15 were
suggested by the DMSO pKa of 3-methoxy-acetophenone (24.5)
[20b], 2-acetyl-pyridine (DMSO pKa = 23.6) [20c], and 2-fluoro-1-
phenylethanone (DMSO pKa = 21.7) [20c]. Assuming that the
reduction in measured pKa from 3-methoxy-acetophenone (24.5)
to 2-fluoro-1-phenylethanone (DMSO pKa = 21.7), ca. 3 pK units,
can be applied to 10, a suggested pKa = 21.7 results. Similarly, the
2-acetyl-pyridine (DMSO pKa = 23.6) conversion to 15 leads to a
projected DMSO pKa = 20.6. These calculated and suggested pKa
values suggest proton transfer and decomposition as an issue for
15 (DMSO pKa = 20.6) with phenols 14 (DMSO pKa = 18.9) and 19
(DMSO pKa = 18.1), while the larger spread in DMSO pKa values
for 19 (DMSO pKa = 18.1) and 10 (DMSO pKa = 21.7) allows a
successful alkylation leading to 12 in 85% yield.
of reaction conditions in order to promote O- vs N-alkylation. A
brief base survey, using the commercially available 10, was per-
formed to set the conditions for the chemistry of Scheme 6. As
illustrated, by the graphic in panel A, (Scheme 6) 14 was exposed
to a variety of base/solvent conditions: 2 eq. K2CO3 in DMSO (N-
alkylation), 2 eq. KOH in EtOH (decomposition, N-alkylation),
phase transfer conditions (NaOH, n-Bu4NBr, PhCH3; decomposi-
tion), and 2 eq. of KOtBu in DMSO (15–30% 3). These results sug-
gested that the stronger base system KOtBu/DMSO, to insure
complete deprotonation of 14, was needed for the chemistry of
Scheme 6. When 15 free base was employed as the electrophile
in Path A, Scheme 6, with KOtBu in DMSO as the base/solvent sys-
tem, poor 0–20% yields of 2 were observed with extensive decom-
Please cite this article as: S. P. Tanis, J. R. Colca, T. T. Parker et al., The development of improved syntheses of PPAR
lidinedione-ketones, Tetrahedron Letters, https://doi.org/10.1016/j.tetlet.2019.07.022
c-sparing, insulin sensitizing thiazo-

4
S.P. Tanis et al. / Tetrahedron Letters xxx (xxxx) xxx
Scheme 6. More convergent synthetic studies. Impact of nucleophile acidity.
[22] and 6 N aq. HCl at 75 °C was found to efficiently cleave the
methoxime, giving 2 in 70% yield, 44% overall from 21.
The success (Scheme 6 – Panel B) of the coupling of 19 with
3-methoxy-phenacyl bromide 10 to give 12 suggested that a
post-alkylative transitory carbonyl protection might facilitate a
more convergent synthesis of 3. Toward that end (Scheme 8) we
elected to examine the utility of a standard oxime as the blocking
function. As alluded to in Scheme 6, Panel B, the alkylation of 19
with 2-bromo-1-(3-methoxyphenyl)-ethanone 10 proceeds
smoothly (KOtBu) to give 12 in 85% yield. The oxime blocking
group is easily introduced (HONH2-HCl, THF, DMF) to afford 23
(93%) and this function remains intact during the cobalt mediated
conjugate reduction, leading to 24 in 90% yield. Acidic hydrolysis
then affords 3 (83%), with an overall yield of 59% (vs 31% in
Scheme 3) for the more convergent Scheme 8 approach.
Scheme 7. (a) i. CH3ONH2-HCl, EtOH, ii. aq. NaHCO3, 85%; (b) i. 14, 2 eq. KOtBu,
DMSO, ii. 21, 74%; (c) 6 N aq. HCl, pyruvic acid,
D, 70%.
These outcomes render the direct Scheme 6, Panel A approach
non-operative as depicted. As an alternative we considered the
potential utility of a pre-alkylation protection/stabilization in the
pyridyl 15) series to prepare 2, and a post-alkylation protection/
deprotection (Scheme 5, Panel B), avoiding the Scheme 3 type
red-ox protocol to achieve a synthesis of 3.
Conclusion
Convergent and efficient syntheses of the clinical candidates 2
(MSDC-0160) and 3 (MSDC-0602) were realized after careful con-
sideration/pairing of the c-pKa of the thiazolidinedione right hand
half entities 14 and 19 with either a methoxime-blocked pyridyl
phenacyl halide 21 (paired with 14 leading to 2) or a phenacyl
halide 10 (paired with 19 leading to 3). Judicious use of pre-alkyla-
tive or post-alkylative ketone protection led to improved syntheses
of 2 (20%–44%) and 3 (31%–59%), potentially setting the stage for
more facile syntheses of 2 and 3 as clinical development progresses
toward Phase 3.
Prior utilization of
a
phenacyl bromide protected as an
O-methyl-oxime in
a
phenol alkylation [21] suggested the
potential of that blocking group for the present issue. In the event,
bromide 15 (Scheme 7) was reacted with CH3ONH2-HCl in ethanol
to provide the relatively stable methoxime 21 (85%) after
neutralization. The reaction of 21 with the dianion of 14, in DMSO,
proceeded smoothly to give 22, a protected variant of 2, in 74%
yield, a dramatic improvement over the results presented in
Scheme 6. After some experimentation, a mixture of pyruvic acid
Scheme 8. (a) i. KOtBu, DMSO, ii.10, 85%; (b) HONH2-HCl, THF, DMF, 93%; (c) CoCl2-6H2O, 2,20bipyridyl, NaBH4, 90%; (d) 6 N aq. HCl, THF,
D
, 83%.
Please cite this article as: S. P. Tanis, J. R. Colca, T. T. Parker et al., The development of improved syntheses of PPAR
lidinedione-ketones, Tetrahedron Letters, https://doi.org/10.1016/j.tetlet.2019.07.022
c-sparing, insulin sensitizing thiazo-

S.P. Tanis et al. / Tetrahedron Letters xxx (xxxx) xxx
5
[10] J.R. Colca, W.G. McDonald, K.S. McCommis, B.N. Finck, Hepatol. Commun. 1
(2017) 193–197.
Declaration of Competing Interest
[11] S.P. Tanis, T.T. Parker, J.R. Colca, R.M. Fisher, R.F. Kletzien, J. Med. Chem. 39
(1996) 5053–5063.
[12] (a) For a review of oxidation conditions which have been run on process scale
(ꢁ100g) see: S. Caron, R.W. Dugger, S.G. Ruggeri, J.A. Ragan, D.H.B. Ripin Chem.
Rev. 106 (2006) 2943–2989;
JRC has an interest in Metabolic Solutions Development Co. and
in Cirius Therapeutics.
(b) D.F. Taber, J.C. Amedio Jr., K.-Y. Jung, J. Org. Chem. 52 (1987) 5621–5622.
[13] D.E. Carpenter, R.J. Imbordino, M.T. Maloney, J.A. Moselein, M.R. Reeder, A.
Scott, Org. Proc. Res. Dev. 6 (2002) 721–728.
[14] T. Storr, D. Mitchell, P. Buglyo, K.H. Thompson, V.G. Yuen, J.H. McNeill, C. Orvig,
Bioconj. Chem. 14 (2003) 212–221.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.tetlet.2019.07.022.
[15] S.P. Tanis, T. Parker, R.C. Gadwood, J.R. Zeller, G.D. Artman III, S.D. Larsen, A
process for the preparation of thiazolidinedione derivatives for use as PPAR
receptor modulators WO2012021467.
c
References
[16] B. Abarca, R. Ballesteros, R. Ballesteros-Garrido, F. Colobert, F.R. Leroux,
Tetrahedron 64 (2008) 3794–3801.
[17] B. Pelcman, A. Sanin, P. Nilsson, T. Groth, H. Kromann, Preparation of pyrazole-
3-carboxamide derivatives useful in the treatment of inflammation
WO2007051981.
[18] K. Nakao, R.W. Stevens, K. Kawamura, C. Uchida, H. Koike, S. Caron, 2-3-
Substituted indole compounds as cox-2 inhibitors WO9935130.
[19] Calculated using ACD Labs Structure Designer V12.0.
[20] (a) F.G. Bordwell, R.J. McCallum, W.N. Olmstead, J. Org. Chem. 49 (1984)
1424–1427;
[1] J.R. Colca, S.P. Tanis, W.G. McDonald, R.F. Kletzien, Expert Opin. Investig. Drugs
23 (2014) 1–7.
[2] J.R. Colca, Expert Opin. Drug Discov. 2015 (10) (2015) 1259–1270.
[3] S.P. Tanis, J.R. Colca, T.T. Parker, G.D. Artman III, S.D. Larsen, W.G. McDonald, R.
C. Gadwood, R.F. Kletzien, J.B. Zeller, P.H. Lee, W.J. Adams, Bioorg. Med. Chem.
26 (2018) 5870–5884.
[4] R.A. DeFronzo, S. Inzucchi, M. Abdul-Ghani, S.E. Nissen, Diabet. Vasc. Dis. Res.
(2019) 1–11.
[5] W.J. Adams, H.M. Dotzauer, T. VanNoord, J.R. Colca, 17th North American
Regional ISSX Meeting 16-20 October 2011, Abstract – p307.
[6] J.R. Colca, J.T. VanderLugt, W.J. Adams, A. Shashlo, W.G. McDonald, J. Liang, R.
Zhou, D.G. Orloff, Clin. Pharmacol. Ther. 93 (2013) 352–359.
[7] R.C. Shah, D.C. Matthews, R.D. Andrews, A.W. Capuano, D.A. Fleischman, J.T.
VanderLugt, J.R. Colca, Curr. Alzheimer Res. 11 (2014) 564–573.
[8] E. Quansah, W. Peelaerts, J.W. Langston, D.K. Simon, J. Colca, P. Brundin, Mol.
Neurodegener. 13 (2018) 28/1-28/12.
(b) F.G. Bordwell, F.J. Cornforth, J. Org. Chem. 43 (1978) 1763–1768;
(c) F.G. Bordwell, Acc. Chem. Res. 21 (1988) 456–463.
[21] G.T. Kim, Hee O. Han, J.S. Koh, Seung H. Kim, K.-H. Kim, H.-K. Chung, Ok K. Park,
S.K. Jeon, S.H. Woo, H.J. Yim, G.-C. Hur, S.W. Hong, K.D. Koo, C.-S. Lee, M.S.
Chang, S.-Y. Koo, H.W. Joo, J.U. Song, S.P. Moon, Song-mi-a, Novel compounds
as agonist for PPAR gamma and PPAR alpha, method for preparation of the
same, and pharmaceutical composition containing the same WO2007058504.
[22] E.B. Hershberg, J. Org. Chem. 13 (1948) 542–546.
[9] J.R. Colca, W.G. McDonald, W.J. Adams, Expert Opin. Investig. Drugs 27 (2018)
631–636.
Please cite this article as: S. P. Tanis, J. R. Colca, T. T. Parker et al., The development of improved syntheses of PPAR
lidinedione-ketones, Tetrahedron Letters, https://doi.org/10.1016/j.tetlet.2019.07.022
c-sparing, insulin sensitizing thiazo-