Discovery of isoindoline and tetrahydroisoquinoline derivatives as potent, selective PPARδ agonists

Article abstract of DOI:10.1016/j.bmcl.2010.10.117

Small molecule isoindoline and tetrahydroisoquinoline derivatives have been identified as selective agonists of human peroxisome proliferator-activated receptor δ (PPARδ. Compound 18 demonstrated efficacy in a biomarker for increased fatty acid oxidation,

Full text of DOI:10.1016/j.bmcl.2010.10.117

Bioorganic & Medicinal Chemistry Letters 21 (2011) 492–496
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of isoindoline and tetrahydroisoquinoline derivatives as potent,
selective PPARd agonists
Christopher A. Luckhurst a, , Linda A. Stein a, Mark Furber a, Nicola Webb a, Marianne J. Ratcliffe b,
⇑
Gary Allenby b, Sara Botterell b, Wendy Tomlinson b, Barrie Martin a, Andrew Walding b
a Medicinal Chemistry, AstraZeneca R&D Charnwood, Loughborough, Leicestershire, LE11 5RH, UK
b Discovery Biosciences, AstraZeneca R&D Charnwood, Loughborough, Leicestershire, LE11 5RH, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Small molecule isoindoline and tetrahydroisoquinoline derivatives have been identified as selective
agonists of human peroxisome proliferator-activated receptor d (PPARd. Compound 18 demonstrated
efficacy in a biomarker for increased fatty acid oxidation, with upregulation of pyruvate dehydrogenase
kinase, isozyme 4 (PDK4) in human primary myotubes.
Received 5 August 2010
Accepted 21 October 2010
Available online 27 October 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
PPARd agonism
Isoindolines
Tetrahydroisoquinolines
The peroxisome proliferator-activated receptors (PPARs) are
ligand-activated transcription factors belonging to the nuclear
dyslipidemia model4,5 and also displays about 1000-fold selectivity
versus the PPAR and subtypes. The Bayer compound 2 is also
a
c
hormone receptor superfamily and comprise three subtypes, PPAR
(OMIM170998), PPAR (OMIM601487) and PPARd (OMIM600409).
Unlike PPAR and PPAR , PPARd is ubiquitously expressed, but
a
PPARd-selective and is less lipophilic than 1.
c
At the start of a hit identification programme, compound 2 was
seen as a good starting point. A number of pyrrolidine compounds
were found to display moderate PPARd activity (Fig. 2). Compound
potency was assessed using a human PPARd reporter assay.25 The
numbers in parentheses show level of efficacy. Compound 3
a
c
highest levels are found in tissues with high lipid metabolism such
as adipose tissue, skeletal muscle and intestine. PPARs have been
implicated in processes that regulate plasma levels of lipoproteins
and triglycerides. Subtype-selective PPAR agonists are regarded as
promising leads for novel drugs used for the treatment of metabolic
syndrome, diabetes, obesity, dyslipidemia (lipid metabolism disor-
ders) and atherosclerosis.1,2 In recent years, there has been signifi-
cant effort invested in the discovery and clinical development of
showed some activity at hPPARd (EC50 = 1.6 lM), perhaps indicat-
ing that the constraint imparted by the pyrrolidine acyl moiety
incorrectly positions the dichloroanilide moiety with respect to
the PPARd protein. An attempt to correct this by moving the amide
to the 3-position of the pyrrolidine ring in 4 was unsuccessful.
An additional ring constraint was introduced into these proline
compounds, and a small breakthrough came with racemic indoline
compound 5. This phenyl ring fusion provided a full agonist with
agonists for PPARa (fibrates) and PPARc (glitazones). There are no
known marketed drugs that have been identified as targeting
PPARd.3 In the past few years, a number of companies have identified
selective PPARd ligands and considerable structural and structure–
activity data have been published.4–23 Many of the reported PPAR
hPPARd EC50 = 0.025 lM (Fig. 3). Selectivity at PPARd over the other
PPAR subtypes was necessary in order to allow an understanding of
the in vivo effects of PPARd agonism, and also to minimise the risk
of toxicity that may arise as a result of agonism of the other PPAR
ligands (a, c and d) are considerably lipophilic, as shown by their
partition coefficient (c log P) values.24 Two selective PPARd agonists
are illustrated in Figure 1.
subtypes. Gratifyingly, 5 displayed excellent selectivity over PPAR
and PPAR . The base pKa of this compound was measured at 5.6
and the acid moiety at 3.1.
a
The GSK compound, GW501516, 1, (Fig. 1) has provided under-
standing of the pharmacological role of PPARd, the first selective
PPARd agonist to have entered clinical trials, and has demonstrated
efficacy in animal models. Compound 1 has been reported to mark-
edly increase serum HDLc levels whilst reducing LDLc in a primate
c
The synthesis of compound 5 is illustrated in Scheme 1. Efforts to
form the acid chloride from acid 6 (using Vilsmeier reagent or oxalyl
chloride) then treatment with the aniline failed to provide the de-
sired amide. Attempts with coupling reagents (HOBt, EDCI or uroni-
um reagents) to effect the same transformation were also
unsuccessful. The acid 6 was converted to methyl ester 7. Formation
⇑
Corresponding author.
E-mail address: chris.luckhurst@astrazeneca.com (C.A. Luckhurst).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.10.117

C. A. Luckhurst et al. / Bioorg. Med. Chem. Lett. 21 (2011) 492–496
493
Cl
O
O
O
O
O
OH
N
H
OH
S
Cl
N
S
F3C
N
GW501516 (1)
Bayer Compound 33 (2)
hPPARδ 0.002 μM (full agonist)
hPPARδ 0.003 μM (full agonist)
>1000-fold selectivity
clogP 3.4
>1000-fold selectivity vs. PPARα and PPARγ
clogP 5.8
Figure 1. Two selective PPARd agonists.
O
O
O
O
OH
O
Cl
OH
N
N
NH
N
H
O
Cl
Cl
Cl
3, hPPARδ EC50= 1.6 μM (1.0)
clogP 1.9
4, PPARδ EC50 = 1.0 μM (0.8)
clogP 1.6
Figure 2. Early pyrrolidine derivatives with moderate PPARd activity.
O
column chromatography, amide 8 was obtained in 31% yield. Later,
with the aniline and 4-chloroaniline examples, warming to room
temperature prior to quenching versus warming to À40 °C afforded
better conversion of the corresponding amides, in 67% and 98%
yields, respectively. However, warming to room temperature did
not greatly improve the yield of 8. Finally, deprotection of the carba-
mate protecting group and reductive alkylation of 9 with the lithium
salt of 2-(4-formyl-2-methylphenoxy) acetic acid provided 5 in 44%
yield. All compounds gave satisfactory spectral and analytical
data.27
O
OH
N
HN
O
(rac)-5
Cl
PPARδ 0.025 μM (full agonist)
PPARα >30 μM
PPARγ >30 μM
clogP= 5.0; logD = 2.6
Cl
The effect of the position of the substituent in the anilide moi-
ety on PPARd potency was explored, and the results highlight the
difficulty of replacing the 2,4-dichloro substitution on the aryl ring
whilst maintaining PPARd potency (Fig. 4). The reduction in po-
tency on going from 2,4-dichloro- (5) to 2,5-dichloro-substituted
anilide (10) illustrates the effect of subtle structural modifications
without changing bulk lipophilicity. The structure–activity rela-
tionships are summarized in Fig. 4.
Figure 3. First indoline compound 5.
of the dilithium anilide of 2,4-dichloroaniline was effected with
2 equiv of n-butyllithium at 0 °C in THF, followed by the addition
of methyl ester 7 at À78 °C.26 Little conversion was seen at this tem-
perature, so the mixture was warmed to À40 °C, then after 4 h was
quenched with saturated aqueous ammonium chloride. After flash
N
b
Boc
N
N
O
a
Boc
Boc
O
HN
O
O
HO
Cl
7
8
6
Cl
O
O
NH
O
OH
d
c
N
HN
HN
Cl
O
Cl
9
5
Cl
Cl
Scheme 1. Synthesis of isoindoline compound 5. Reagents and conditions: (a) TMS diazomethane, AcOH, MeOH, PhMe, rt, 2 h (99%); (b) 2,4-dichloroaniline, n-BuLi (2 equiv),
THF, À78 °C to À40 °C, 4 h (31%); (c) 15% v/v TFA/DCM, rt, 2 h (100%); (d) lithium 2-(4-formyl-2-methylphenoxy)acetate, NaBH3CN, AcOH, MeOH, rt, 16 h (44%).

494
C. A. Luckhurst et al. / Bioorg. Med. Chem. Lett. 21 (2011) 492–496
O
O
OH
N
Xn
N
H
O
Aryl replacements:
Aryl substitution:
O
HO
N
N
N
5: 2,4-dichloro 0.025 μM (1.1) clogP 5.0
10: 2,5-dichloro 0.32 μM (1.0) clogP 5.0
11: 4-chloro
12: deschloro
2.5 μM (0.7) clogP 4.6
>60 μM (0.2) clogP 4.1
13: IA
clogP 3.1
14: IA
15: IA
16: >20 μM (>0.5)
clogP 4.1
clogP 3.7
clogP 2.8
Figure 4. Summary of structure–PPARd activity relationships for the isoindoline amide compounds, with level of agonism shown in parentheses.
O
O
OH
N
H
N
Cl
O
Cl
Figure 6. (S)-Configuration of compound 5.
isoindoline moiety in our compounds (Fig. 5). Racemic compound
5 was co-crystallised with PPARd and binding within the pocket
determined to 3.00 Å resolution. The two pictures in Figure 5 show
the binding pocket of PPARd containing compound 5 in two orien-
tations rotated by 90°. The experimentally observed electron den-
sity map clearly shows that only the (S)-configuration of
compound 5 (Fig. 6) is bound to the active site. The ligand 5 forms
five specific hydrogen bonds, namely to the side chain atoms of
Thr289, His323, His449, and Tyr473.
An improvement in PPARd potency was achieved through mod-
ification of the cyclic constraint. On changing the isoindoline scaf-
fold to a tetrahydroisoquinoline ring (17), PPARd potency increased
to 10 nM, although with some reduction in selectivity versus
PPARa and PPARd (Fig. 7).
More interestingly, the regioisomeric tetrahydroisoquinoline
pair of enantiomers 18 and 19 were identified as more potent
and PPARd-selective ligands (Fig. 8). These were prepared similarly
to the method described for 5 from the homochiral tetrahydroiso-
quinoline acids. Compound (R)-19 was found to be more active
with a PPARd potency of 0.5 nM, 10-fold more potent than its enan-
tiomer (S)-18. Both compounds 18 and 19 demonstrated excellent
PPAR subtype selectivity. For example, compound 19 showed only
weak activity at PPAR
more, the efficacy ratios against known PPAR
for these subtypes were low, 0.4 and 0.08, respectively. In short,
compound 19 was over 2400-fold selective versus PPAR and
1500-fold versus PPAR
Incubation of 18 with rat hepatocytes (1.5 h incubation at
M) revealed that the majority of the metabolism of the parent
a
(1.25
l
M) and PPAR
c
(0.8
l
M). Further-
Figure 5. Binding pocket of PPARd containing compound 5 in two orientations
rotated by 90°.
a
and PPAR
c
agonists
a
c.
To further understand the binding mode of compounds with the
PPARd protein, X-ray crystallization studies were undertaken with
the racemate 5 (PDB code 3OZ0).
10
l
compound occurs by hydroxylation of the tetrahydroisoquinoline
portion, partial oxidation (-2H) of the same moiety, and some
amide hydrolysis. The compounds show no cardiac QT liability
in vitro and exhibit excellent glutathione stability, chemical and
plasma stability.
The PPAR subtypes show significant structural differences be-
tween the ligand-binding domains in the PPARa, PPARc and PPARd
proteins. The PPAR and PPAR active sites contain a larger acid
a
c
binding domain, whilst PPARd contains a selectivity pocket consid-
erably larger than that of the other two subtypes, occupied by the

C. A. Luckhurst et al. / Bioorg. Med. Chem. Lett. 21 (2011) 492–496
495
O
O
O
O
OH
OH
N
O
N
HN
Cl
HN
O
Cl
Cl
Cl
(rac)-5
(rac)-17
PPARδ 0.025 μM (1.1)
PPARα >30 μM
PPARδ 0.010 μM (0.9)
PPARα 6.3 μM
PPARγ >30 μM
clogP= 5.0; logD = 2.6
PPARγ 2.7 μM
clogP= 5.5; logD = 2.6
Figure 7. Isoindoline compound 5 and tetrahydroisoquinoline 17.
O
O
O
O
OH
Cl
OH
N
O
N
O
Cl
N
H
N
H
Cl
Cl
(S)-18
(R)-19
PPARδ 0.005 μM (0.9)
PPARα 2.5 μM (0.8)
PPARγ >20 μM (>0.2)
clogP = 5.4; logD = 2.4
PPARδ 0.0005 μM (1.0)
PPARα 1.25 μM (0.4)
PPARγ 0.80 μM (0.1)
clogP = 5.4
Figure 8. Tetrahydroisoquinoline enantiomers 18 and 19.
(Fig. 9). This assay was configured to measure upregulation of
PDK4 mRNA, a modulator of glucose and fatty acid metabolism.
For compound 18, potency in this assay (EC50 = 15 nM) was consis-
tent with potency in the hPPARd reporter assay (EC50 = 5 nM). Sim-
ilarly, compound (rac)-17 displayed a potency of 70 nM in the
biological effect assay, and 10 nM in the reporter assay.
In spite of the very high potency and good selectivity seen with
compounds 18 and 19, the principal objective for this series is to
lower lipophilicity. Some work has been initiated to this end,
including incorporation of one or more nitrogen atoms into the
dichlorophenyl and/or the tetrahydroisoquinoline rings (20 and
21), but the examples in Figure 10 highlight the difficulty with this
approach. The exploration of options for reducing lipophilicity
whilst retaining PPARd activity and selectivity are the main thrust
of ongoing efforts.
Figure 9. Dose–response curve for upregulation of PDK4 mRNA with compounds
(rac)-17 and (S)-18. d GW501516, j compound (rac)-17, N compound (S)-18.
Activity in a relevant human biological effect assay (primary
human myotubes)28 has been determined for two of the key com-
pounds, with GW501516 (1) included as a reference compound
In summary, a series of potent, selective small molecule PPARd
agonists have been discovered. Compound 18 demonstrated effi-
cacy in upregulation of PDK4 in human primary myotubes.
O
O
N
O
O
N
OH
Cl
OH
N
O
N
O
Cl
N
H
N
H
Cl
Cl
(rac)-20
(rac)-21
PPARδ 0.5 μM (1.0)
PPARδ 2.0 μM (1.3)
clogP = 3.9
clogP = 3.9
Figure 10. Aza-tetrahydroisoquinoline compounds 20 and 21.

496
C. A. Luckhurst et al. / Bioorg. Med. Chem. Lett. 21 (2011) 492–496
18. Zhang, R.; Wang, A.; Deangelis, A.; Pelton, P.; Xu, J.; Zhu, P.; Zhou, L.; Demarest,
Acknowledgements
K. T.; Murray, W. V.; Kuo, G.-H. Bioorg. Med. Chem. Lett. 2007, 17, 3855.
19. Kasuga, J.; Oyama, T.; Nakagome, I.; Makishima, M.; Hirono, S.; Morikawa, K.;
Hashimoto, Y.; Miyachi, H. ChemMedChem 2008, 3, 1662.
20. Kasuga, J.; Nakagome, I.; Aoyama, A.; Sako, K.; Ishizawa, M.; Ogura, M.;
Makishima, M.; Hirono, S.; Hashimoto, Y.; Miyachi, H. Bioorg. Med. Chem. 2007,
15, 5177.
The authors gratefully acknowledge the support of Stefan
Steinbacher, Grazyna Seifert, and Martin Augustin of Proteros in
obtaining the PPARd crystal structure with compound 5.
21. Keil, S.; Wendler, W.; Glien, M.; Goerlitzer, J.; Chandross, K.; Mcgarry, D. G.;
Merrill, J.; Bernardelli, P.; Ronan, B.; Terrier, C. WO2005097786, Chem. Abstr.
2008, 148, 472054.
References and notes
22. Fujieda, H.; Usui, S.; Suzuki, T.; Nakagawa, H.; Ogura, M.; Makishima, M.;
Miyata, N. Bioorg. Med. Chem. Lett. 2007, 17, 4351.
23. Zhang, R.; DeAngelis, A.; Wang, A.; Sieber-McMaster, E.; Li, X.; Russell, R.;
Pelton, P.; Xu, J.; Zhu, P.; Zhou, L.; Demarest, K.; Murray, W. V.; Kuo, G. Bioorg.
Med. Chem. 2009, 19, 1101.
1. Linsel-Nitscke, P.; Tall, A. R. Nat. Rev. Drug Disc. 2005, 4, 193.
2. Willson, T. M.; Brown, P. J.; Sternbach, D. D.; Henke, B. R. J. Med. Chem. 2000, 43,
527.
3. Willson, T. M.; Wahli, W. Curr. Opin. Chem. Biol. 1997, 1, 235.
4. Sznaidman, M. L.; Haffner, C. D.; Maloney, P. R.; Fivush, A.; Chao, E.; Goreham,
D.; Sierra, M. L.; LeGrumelec, C.; Xu, H. E.; Montana, V. G.; Lambert, M. H.;
Willson, T. M.; Oliver, W. R., Jr.; Sternbach, D. D. Bioorg. Med. Chem. Lett. 2003,
13, 1517.
5. Oliver, W. R., Jr.; Shenk, J. L.; Snaith, M. R.; Russell, C. S.; Plunket, K. D.; Bodkin,
N. L.; Lewis, M. C.; Winegar, D. A.; Sznaidman, M. L.; Lambert, M. H.; Xu, H. E.;
Sternbach, D. D.; Kliewer, S. A.; Hansen, B. C.; Willson, T. M. Proc. Natl. Acad. Sci.
U.S.A. 2001, 98, 5306.
6. Weigand, S.; Bischoff, H.; Dittrich-Wengenroth, E.; Heckroth, H.; Lang, D.;
Vaupel, A.; Woltering, M. Bioorg. Med. Chem. Lett. 2005, 15, 4619.
7. Adams, A. D.; Yuen, W.; Hu, Z.; Santini, C.; Jones, A. B.; MacNaul, K. L.; Berger, J.
P.; Doebber, T. W.; Moller, D. E. Bioorg. Med. Chem. Lett. 2003, 13, 931.
8. Bischoff, H.; Dittrich-Wengenroth, E.; Heckroth, H.; Vaupel, A.; Woltering, M.;
Weigand, S. WO2003035603; Chem. Abstr. 2003, 138, 353734.
9. Urbahns, K.; Woltering, M.; Nikolic, S.; Pernerstorfer, J.; Hinzen, B.; Dittrich-
Wengenroth, E.; Bischoff, H.; Hirth-Dietrich, C.; Lustig, K. WO2002028821;
Chem. Abstr. 2002, 136, 294653.
10. Pettersson, I.; Ebdrup, S.; Havranek, M.; Pihera, P.; Korinek, M.; Mogensen, J. P.;
Jeppesen, C. B.; Johansson, E.; Sauerberg, P. Bioorg. Med. Chem. Lett. 2007, 17,
4625.
11. Havranek, M.; Sauerberg, P.; Mogensen, J. P.; Kratina, P.; Jeppesen, C. B.;
Pettersson, I.; Pihera, P. Bioorg. Med. Chem. Lett. 2007, 17, 4144.
12. Kuo, G.-H.; Zhang, Y.; Shen, L.; Lu, S.; Demarest, K. T.; Peiton, P. US2007244094;
Chem. Abstr. 2007, 147, 469249.
13. Epple, R.; Russo, R.; Azimioara, M.; Cow, C.; Xie, Y.; Wang, X.; Wityak, J.;
Karanewsky, D.; Gerken, A.; Iskandar, M.; Saez, E.; Seidel, H. M.; Tian, S. Bioorg.
Med. Chem. Lett. 2006, 16, 4376.
14. Epple, R.; Azimioara, M.; Russo, R.; Xie, Y.; Wang, X.; Cow, C.; Wityak, J.;
Karanewsky, D.; Bursulaya, B.; Kreusch, A.; Tuntland, T.; Gerken, A.; Iskandar,
M.; Saez, E.; Seidel, H. M.; Tian, S. Bioorg. Med. Chem. Lett. 2006, 16, 5488.
15. Epple, R.; Russo, R.; Azimioara, M.; Xie, Y. WO2005113519; Chem. Abstr. 2005,
144, 22911.
24. c log P 4.3 is a predicted octanol/water partition coefficient from Daylight
Chemical Information Systems and BioByte Corp.
25. HEK293 cells are stably transfected with a construct comprising the ligand-
binding domain of human PPARd fused to the GAL4 DNA binding domain. It has
been confirmed that the primary PPARd screening assay contains the most
common sequence of PPARD (NM_006238). The cell also stably expresses a
luciferase gene under the control of a GAL4 promoter. Hence, compounds
which bind and activate the PPARd/GAL4 fusion promote luciferase production,
which is quantified by measuring luminescence generated from a luciferase
substrate. GW501516, 1, was used as reference compound in every assay to
ensure assay stability, and the intrinsic activity of this compound is set to 100%,
against which all project compounds are scaled.
26. Ooi, T.; Tayama, E.; Yamada, M.; Maruoka, K. Synlett 1999, 6, 729.
27. Selected spectroscopic information: Compound (rac)-5: 1H NMR (400 MHz,
DMSO) d 10.06 (s, 1H), 8.25 (d, J = 8.9 Hz, 1H), 7.67 (d, J = 2.4 Hz, 1H), 7.42 (dd,
J = 8.9, 2.4 Hz, 1H), 7.33 (d, J = 7.3 Hz, 1H), 7.31–7.23 (m, 4H), 7.16 (d, J = 8.3 Hz,
1H), 6.67 (d, J = 8.4 Hz, 1H), 4.68 (s, 1H), 4.28 (s, 2H), 4.28 (dd, J = 13.6, 2.4 Hz,
1H), 4.04 (d, J = 12.6 Hz, 1H), 3.89 (dd, J = 13.7, 2.1 Hz, 1H), 3.80 (d, J = 12.7 Hz,
1H), 2.12 (s, 3H); acid OH resonance absent. m/z [M+H]+ = 485, 487.
Compound (rac)-17: 1H NMR: (400 MHz, DMSO) d 10.15 (s, 1H), 8.14 (d,
J = 8.7 Hz, 1H), 7.68 (d, J = 2.3 Hz, 1H), 7.42–7.39 (m, 2H), 7.21–7.14 (m, 4H),
7.11 (dd, J = 8.3, 1.9 Hz, 1H), 6.70 (d, J = 8.5 Hz, 1H), 4.41 (s, 2H), 4.38 (s, 1H),
3.77 (d, J = 13.1 Hz, 1H), 3.61 (d, J = 13.1 Hz, 1H), 3.15–3.08 (m, 1H), 2.83–2.78
(m, 2H), 2.61–2.54 (m, 1H), 2.13 (s, 3H); acid OH resonance absent. m/z
[M+H]+ = 499, 501.
Compound (S)-18 (trifluoroacetate salt): 1H NMR (400 MHz, DMSO, 90 °C) d
9.95 (s, 1H), 8.02–7.93 (m, 1H), 7.57 (d, J = 2.4 Hz, 1H), 7.35 (dd, J = 8.8, 2.4 Hz,
1H), 7.26–7.09 (m, 6H), 6.81 (d, J = 8.3 Hz, 1H), 4.62 (s, 2H), 4.04 (d, J = 14.6 Hz,
2H), 3.95–3.83 (m, 3H; co-incident NH+ br s, 1H), 3.25–3.08 (m, 2H), 2.17 (s,
3H); acid OH resonance absent. m/z [M+H]+ = 499, 501.
28. Human skeletal muscle myoblast cells were isolated by enzymatic digestion
from the vastus lateralis. The myoblasts were differentiated for 7 days in 2%
horse serum. Cells were treated with compound for 24 h before measuring fold
induction of PDK4 mRNA.
16. Epple, R.; Azimioara, M.; Russo, R.; Bursulaya, B.; Tian, S.; Gerken, A.; Iskandar,
M. Bioorg. Med. Chem. Lett. 2006, 16, 2969.
17. Deangelis, A.; Demarest, K. T.; Kuo, G.-H.; Pelton, P.; Wang, A.; Zhang, R.
US2006058393; Chem. Abstr. 2006, 144, 311792.