Structure-activity relationships of 2-arylamido-5,7-dihydro-4H-thieno[2,3- c]pyran-3-carboxamide derivatives as cannabinoid receptor agonists and their analgesic action

Article abstract of DOI:10.1016/j.bmcl.2012.10.087

SAR studies were performed on a series of 2-arylamido-5,7-dihydro-4H- thieno[2,3-c]pyran-3-carboxamide derivatives as cannabinoid receptor agonists. Starting from a HTS hit both potency and selectivity could be improved. Modifications to the thiophene fusion and C-3 amides were studied. A representative compound 3t produced analgesia when dosed orally in inflammatory pain models of writhing and carrageenan-induced allodynia.

Full text of DOI:10.1016/j.bmcl.2012.10.087

Accepted Manuscript
Structure-activity relationships of 2-arylamido-5,7-dihydro-4H-thieno[2,3-
c]pyran-3-carboxamide derivatives as cannabinoid receptor agonists and their
analgesic action
Yithachu Thur, Amit Bhalerao, Zaki Munshi, Nisha Pansare, Klaus Mann,
Guido Hanauer, Hans-Peter Kley, Sandra Nappe, Cornelia Weiss-Haljiti,
Claude Ostermann, Christof Zitt, Michaela Schaefer, Dibyendu Mondal, Afsar
Ali Siddiki, Velavan Armugam, Vinod Gudaghe, Mahendra Gupta, Pramila
Rayudu, Frank M. Dautzenberg, Koushik Das Sarma
PII:
S0960-894X(12)01378-9
DOI:
http://dx.doi.org/10.1016/j.bmcl.2012.10.087
BMCL 19710
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
14 August 2012
5 October 2012
18 October 2012
Please cite this article as: Thur, Y., Bhalerao, A., Munshi, Z., Pansare, N., Mann, K., Hanauer, G., Kley, H-P., Nappe,
S., Weiss-Haljiti, C., Ostermann, C., Zitt, C., Schaefer, M., Mondal, D., Siddiki, A.A., Armugam, V., Gudaghe, V.,
Gupta, M., Rayudu, P., Dautzenberg, F.M., Sarma, K.D., Structure-activity relationships of 2-arylamido-5,7-
dihydro-4H-thieno[2,3-c]pyran-3-carboxamide derivatives as cannabinoid receptor agonists and their analgesic
action, Bioorganic & Medicinal Chemistry Letters (2012), doi: http://dx.doi.org/10.1016/j.bmcl.2012.10.087
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Structure-activity relationships of 2-arylamido-5,7-dihydro-4H-thieno[2,3-c]pyran-3-
carboxamide derivatives as cannabinoid receptor agonists and their analgesic action
Yithachu Thur,a Amit Bhalerao,a Zaki Munshi,a Nisha Pansare,a Klaus Mann,b Guido
Hanauer,b Hans-Peter Kley,b Sandra Nappe,b Cornelia Weiss-Haljiti,b Claude Ostermann,b
Christof Zitt,b Michaela Schaefer,b Dibyendu Mondal,a Afsar Ali Siddiki,a Velavan
Armugam,a Vinod Gudaghe,a Mahendra Gupta,a Pramila Rayudu,a Frank M. Dautzenberg,b,*
Koushik Das Sarma,a,c,*
a Global Discovery, Nycomed Pharma Pvt. Ltd, Plot number 29-30-31, Suren Road, Andheri
(East), Mumbai 400 093, Maharashtra, India
b Global Discovery, Nycomed GmbH, Byk-Gulden-Str 2, 78467 Konstanz, Germany
* Corresponding authors. Tel: +91 120-4093300; e-mail: frankdautzenberg@yahoo.de;
kdsarma@gmail.com
c Present address: Jubilant Chemsys Ltd. B-34, Sector-58, NOIDA 201 301, Uttar Pradesh,
India
Cannabinoid receptors have been shown to play an important role in the modulation of pain
transmission. The two identified receptors, CB1 and CB2 belong to the large superfamily of
G protein coupled receptors (GPCR),1 and are mainly coupled to the inhibitory G protein Gi,
which upon receptor activation results in reduction of intracellular cyclic AMP (cAMP)
levels.2 Up until recently, distribution pattern of CB1 and CB2 receptors were thought to be
distinctly separate, with a predominant expression of CB1 receptors in the central nervous
system and an exclusive peripheral expression pattern of CB2 receptors.3 However, this
picture has eroded in the recent years and evidence for central expression of CB2 receptors
has accumulated.4
The analgesic activity of CB1 receptors is limited due to its profound side effect
profile, such as the development of addiction.5 Thus, attention has been shifted to the
development of CB2 receptor agonists for the treatment of chronic pain. However, due to the
CB1-mediated side effect potential, CB2-selectivity is regarded as the most important feature
of such agonists. Several selective CB2 agonists have been discovered and the list is growing
rapidly.6 As a part of our research program aimed at discovering novel selective CB2 agonists
1

for the treatment of pain, we performed a high-throughput screening (HTS) campaign and
improved the identified hits for their selectivity versus CB1 receptors.
Figure 1. HTS hit
A HTS screening of ~300,000 compounds consisting of commercially available as
well as proprietary compounds resulted in 1 (Figure 1) as the hit from one of the priority
clusters. As this is a commercially available screening compound, the initial effort was driven
towards designing novel compounds while retaining the key structural features.
A series of analogues were synthesized using the Gewald reaction to generate
corresponding 2-amino-3-carboethoxy thiophene.7 Acylation with a suitable acid chloride
generated the C-2 amide. Saponification, followed by amide formation using standard
coupling, installed the C-3 carboxamide (Scheme 1).
Scheme 1. Synthesis route for synthesis of compounds studied (D)
2

A simple modification on the cyclohexyl ring fused to thiophene resulted in
compound 2a (Table 1).8 The single point substitution of –CH2– by –O– resulted 1log loss of
activity in both hCB2 and hCB1 receptors.9 However, the selectivity remained similar for
both 1 and 2a and both of these compounds were found to be full agonists. Moving the Cl to
3 position (2d) resulted in a 6 fold drop in activity whereas, corresponding 4-Cl derivative
(2e) did not show any measurable activity up to 10 µM. Addition of one more Cl had a
detrimental effect on the activity. A 3 fold drop in activity was observed for 2,6-Cl2
compound (2b). For dichloro compound like 2c, the loss of activity was over 40 fold. SAR
data therefore strongly supported a requirement for substitution at the 2-position for CB2
activity. Also, multiple chloro substitutions of the aryl ring did not have any positive impact
on the activity.
As a next step, we decided to ascertain the roles of various substituents on the aryl
ring on activity and selectivity. As shown in table 1, substituting the Cl in 2a with Me (2f)
resulted in a mild loss of activity in both hCB2 and hCB1 receptors. A stronger electron
donating group –OMe resulted in a complete loss of activity in 2h. The incremental activity
improvement in difluoromethyl 2i and trifluoromethyl 2j established the importance of
maintaining a threshold electron deficiency on the aryl ring. The selectivity was also higher
(>90) in 2j. The importance of 2-substitution was again proven by over 7 fold loss of activity
in meta substituted 2k with a concomitant loss in selectivity. As expected the activity was
further improved with a stronger electron withdrawing group in the 2-position (2l) which
showed 50 fold selectivity over hCB1 activity. As seen in 2a and 2b, there was a slight loss in
activity in 2m (~3 fold) when a fluoro group was added to the 6-position of 2l. However, the
selectivity against hCB1 remained similar for both 2l and 2m. The activity and selectivity of
isopropyl 2g was found to be similar to 2l. A plausible explanation could be that the inductive
effect of iPr would be weaker than the resonance effect of OMe (2h) that will contribute to
less charge delocalization in the aryl ring for 2g as compared to 2h. Another plausible
explanation could point towards the necessity of having an optimal steric bulk at the 2-
position. The high micromolar activity of 2n and lower activity of 2f (with a lower inductive
contribution) supported this hypothesis.
Table 1. Screening of substitutions on the aryl ring
3

Compound # R1
R2
EC50 (hCB2) (nM/%) EC50 (hCB1) (nM/%) Selectivity
2-Cl
H
227 (112)
3270 (85)
4639 (78)
>10000
14
2a
2b
2c
2d
2e
2f
695
(76)
2-Cl
6-Cl
7
2-Cl
4-Cl 9354 (108)
NA
NA
NA
12
3-Cl
H
H
H
H
H
H
H
H
H
6-F
H
1421 (98)
>10000
>10000
4-Cl
>10000
2-Me
2-iPr
646 (105)
7452 (97)
1500 (105)
>10000
35
(101)
42
2g
2h
2i
2-OMe
2-OCF2
2-OCF3
3-OCF3
2-CF3
2-CF3
2-Ph
>10000
NA
>20
>90
>10
50
488
(96)
>10000
103 (106)
>10000
2j
749
(77)
>10000
2k
2l
35
(102)
1753 (84)
4538 (56)
>10000
95
(100)
48
2m
2n
7892 (63)
NA
The screening of substitutions of the benzamide ring resulted in modest improvement
of activity and selectivity. To further improve the biological profile we next initiated the
optimization of the morpholine amide (3-C, Table 2). Replacing the amide with a pivaloyl
4

group resulted in loss of activity for both 3a (35 fold against hCB2 as compared to 2l) and 3b
(35 fold against hCB2 as compared to 2j). The free carboxylic acid 3c did not show any
measurable activity up to 10 µM. Interestingly, there was a dramatic improvement in activity
(>1log) for all compounds (3d-f) with piperidine amides at the amide terminal. The
selectivity of 3d increased 4-fold over 2l. Amides with cyclopropylmethyl amine, 3g and 3h
retained higher hCB2 activity (3-9 fold) compared to 2j and 2m, but 3h only experienced a
50% improvement in selectivity.
The terminal oxygen of the morpholine amide could also serve as a potential H-bond
acceptor. We decided to study the role of a flexible H-bond acceptor instead of the
constrained morpholine amide to ascertain if a H-bonding interaction could influence the
selectivity. In the initial attempt with 2-methoxyethylamine, there was a significant increase
in selectivity (525 fold) for 3i as compared to 3d and 2l. It was evident that this rise in
selectivity was due to a loss of hCB1 activity. Similarly for 3k there was a significant
decrease in hCB1 activity (as compared to 2j, 3e and 3g) which attributed to 385 fold
selectivity. For 3j, the selectivity was over 100 fold due to increase in both hCB1 and hCB2
components. The same trend was retained for the bulkier and more constrained 1-methoxy-2-
methylpropan-2-amine. However, there was a slight loss of selectivity due to increase in
hCB2 potency for 3l compared to 3i. There was no significant change in activity and
selectivity for 3m as compared to 3j. For the constrained 3-methoxyazetidine, the selectivity
increased again for 3o and 3q (as compared to 3l and 3n respectively) due to increase in
hCB2 potency. Thus, 3o and 3q showed 387 and 275 fold selectivity over hCB1 respectively.
Interestingly, this same trend was not observed for 3p which showed a 1log drop in both
hCB2 and hCB1 receptor activity compared to 3j and 3m resulting in only a 37 fold
selectivity.10
These ongoing studies indicated that with an aryl amide at the C-2 of the thiophene
ring, a terminal hydrogen bond acceptor at the C-3 amide was necessary for attaining high
activity and selectivity for the hCB2 receptor. The difluoromethyl group has been reported to
function both as H-bond donor (isostere of –OH) and H-bond acceptor (isostere of –O-).
These interactions can result in a modified binding to a receptor.11 As a next step, we decided
to study the role of gem-difluorides at the amide terminal. With 3,3-difluoroazetidine, there
was a significant increase in hCB2 potency for 3r-t. However, the selectivity for 3r and 3t
was not as high as in 3i and 3k respectively due to a concomitant increase in hCB1 potency.
5

Nevertheless, both of these compounds retained over 100 fold selectivity. There was a further
drop in selectivity for 3u and 3w (57 and 71 fold respectively). This was due to mainly an
increase in hCB1 activity. The higher selectivity in 3s was a result of a higher increase in
hCB2 potency as compared to hCB1. However, this trend was not retained with 3v which
showed a modest selectivity (38 fold). Overall, gem-difluorides were found to retain activity
of the compounds studied, while selectivity was sensitive to ring size.
Table 2. Screening of modifications on the amide part of the carboxylic acid end
Compound R1
#
R2
R3
EC50
(hCB2)
EC50
(hCB1)
Selectivity
(nM/%) (nM/%)
2-CF3
H
H
H
H
-tBu
-tBu
-OH
897
(88)
>10000
>10000
>10000
11
3a
3b
3c
3d
2-OCF3
2-Cl
939
(102)
11
>10000
NA
204
2-CF3
3.0
607
(99)
(93)
2-OCF3
2-CF3
H
9.5
(101)
925
(85)
98
71
3e
3f
6-F
14
(103)
1000
(94)
2-OCF3
2-CF3
H
35
(94)
519
(88)
15
3g
3h
3i
6-F
H
11
(112)
812
(90)
74
2-CF3
17
(93)
8710
(45)
525
2-OCF3
H
96
>10000
>104
3j
(100)
6

2-CF3
2-CF3
2-OCF3
2-CF3
6-F
H
24
(96)
9360
(58)
385
3k
3l
41
(93)
>10000
>10000
>10000
>243
>100
>135
H
101
(81)
3m
3n
6-F
74
(94)
2-CF3
2-OCF3
2-CF3
2-CF3
H
15
(98)
5960
(53)
387
37
3o
3p
3q
3r
H
38
(100)
1416
(45)
6-F
H
24
(100)
6611
(55)
275
118
2.5
296
(106)
(84)
2-OCF3
2-CF3
H
4.7
(104)
1974
(104)
420
122
57
3s
3t
6-F
H
6.5
(96)
798
(83)
2-CF3
1.4
80
3u
(99)
(90)
2-OCF3
2-CF3
H
3.3
(96)
125
(79)
38
71
3v
6-F
3.6
251
3w
(97)
(94)
With clarity in the choice of amines for the C-3 amide position, we wanted to evaluate
the possibilities of non-aryl acids at the C-2 amine terminal (Table 3). As shown by 4a-d,
there was a modest activity and selectivity for 4a and 4c. However, activity against both
receptors dropped by up to 1log when a pyran ring was used as in 4b and 4d. Among all the
substrates studied, 3,3-dimethylbutanoic acid resulted in compounds that were of single digit
nanomolar potency (4j and 4m). However, selectivity was modest in these compounds (63
and 24 fold respectively). A better selectivity profile, however, was achieved with 2-
(piperidin-1-yl)acetic acid for 4k (>178 fold) and 4n (126 fold). However, a reduction in ring
7

size [2-(pyrrolidin-1-yl)acetic acid] in 4l and 4o resulted in a loss of hCB2 potency by
approximately 1log with a concomitant loss in selectivity. Interestingly, gem-difluoride C-3
amides were found to result in higher potency for compounds in this series as seen previously
for C-2 aryl amides. It is also noteworthy that a basic nitrogen at C-2 amides (4f, i, k, l, n, o)
did not provide any additional advantage over compounds without any basic nitrogen.
Clearly, hydrophobic interaction around the C-2 amides dominated the binding of these
compounds to the CB receptors. Overall, a suitably substituted benzamide was preferred
compared to the aliphatic C-2 amides.
Table 3. Screening of non-aryl substituents
Compound X
#
R1
R2
EC50
(hCB2) (hCB1)
(nM/%) (nM/%)
EC50
Selectivity
CH2
23
(94)
1905
(90)
84
4a
4b
4c
4d
4e
4f
O
489
(101)
>10000 >20
CH2
O
71
(81)
8030
(61)
113
313
(84)
>10000 >32
>10000 >49
>10000 >65
CH2
CH2
201
(101)
152
(98)
CH2
344
>10000 >29
4g
(88)
8

CH2
CH2
47
(110)
3972
(63)
85
4h
4i
219
>10000 >45
(92)
CH2
CH2
5.4
(99)
336
(91)
63
4j
56
(103)
>10000 >178
>10000 >25
4k
CH2
394
4l
(89)
CH2
CH2
4.7
(82)
113
(86)
24
4m
4n
26
(104)
3217
(100)
126
CH2
209
>10000 >48
4o
(87)
In the next step we decided to optimize the alicyclic ring fused to the thiophene ring
(Table 4). The cyclohexyl HTS hit (1) was found to be more active against both hCB1/2 as
compared to pyran 2a. A cyclopentyl system (5a-b) retained similar activity against hCB2 as
compared to 2a but agonist efficacy was reduced (52 and 72% for 5a and 5b, respectively).
However, there was an increase in selectivity (~5 fold for 5a and ~7 fold for 5b) due to
absence of hCB1 activity up to 10 µM. With a gem-dimethyl cyclopentyl ring (5c-f), there
was a significant improvement in activity for unconstrained C-3 amides, especially for 5c, for
both hCB1 and 2. As a result, the selectivity was found to be for these compoundsmodest.
With 3,3-dimethylbutanoic acid in place of 2-chlorobenzoic acid, a similar trend was
observed with various C-2 amides. Even with the 4-methyl furan, the activity and selectivity
of 5h was similar to 4j. Although high CB2 potency was achievable with 5-membered ring
fusions no significant improvement in selectivity against hCB1 was possible.
9

With the failure to achieve any improvement of selectivity for various 5-membered
ring systems, we decided to look further into the pyran/cyclohexyl ring system. A gem-
difluoro could easily be inserted at the 6 position of the ring using a standard chemistry. The
gem-difluoride group has lipophilicity between CH2 and O and thus could provide useful
insight into the present SAR studies.11 The first set of compounds (5i-5k) showed good in
vitro potency but without any improvement in selectivity due to hCB1 activity. In fact, 5i was
almost equipotent against both receptors. The selectivity observed for 5j was in the range of
20 fold. Though there was a significant increase in selectivity for 5l (>145 fold), the hCB2
potency was modest.
Table 4. Modifications on the ring fused to thiophene
Compoun A
d #
R1
R2
EC50
(hCB2)
EC50
(hCB1)
Selectivity
(nM/%) (nM/%)
137
(52)
>10000 >72
5a
5b
5c
106
(76)
>10000 >94
0.55
16
30
(116)
(86)
129
(96)
>10000 >77
5d
5e
6.4
243
38
(101)
(96)
10

5.9
(100)
368
(98)
62
42
36
5f
40
(91)
1656
(88)
5g
5h
6.0
217
(105)
(94)
49
81
2
5i
(98)
(95)
46
(105)
958
(97)
21
8
5j
5k
5l
18
(99)
139
(93)
63
>10000 >145
(96)
Selected agonists from the C-2 benzamides (Table 2) were further evaluated in
receptor binding assays (Table 5).12 In order to show physical interaction between the novel
synthesized compounds and CB1 and CB2 receptors, a binding assay was employed. In
general, the functional assay displayed higher activity of all compounds when compared to
their binding affinity for both receptors. This observation is not uncommon when using
recombinant systems, in which a large receptor reserve exists and the activation of only a
small fraction of receptors resulted in full activation of the second messenger cascade. We
correlated activity of functional responses with the binding affinity of all tested compounds in
house and observed a consistent rightward shift by a factor ~20-fold for almost all
compounds in both receptor systems (data not shown).
11

The same set of molecules were further subjected to in vitro metabolic stability testing
in mouse and rat liver microsomes. Both microsomal preparations were selected because the
initial in vivo screening model, acetic-acid induced writhing, was conducted in female mice
whereas carrageenan-induced inflammatory pain readouts were performed in rats. As can be
seen in Table 5, the compounds studied suffered from low metabolic stability in microsomal
preparations, which predict low exposure levels in vivo. Interestingly, compound 3t showed a
substantially higher stability in mouse microsomes than in the rat.13
Table 5. Binding affinity, h/rCB2/1 and microsomal stability data of key compounds
Compound Ki (hCB2) Ki (hCB1) Ki (rCB2) Ki (rCB1) t1/2
t1/2
#
(nM)
241
347
999
904
899
81.5
240
444
(nM)
(nM)
1610
2130
2540
1,700
3,110
594
(nM)
(MLM)
(min.)
(RLM)
(min.)
11,700
19,400
>30,000
>30,000
>30,000
7,190
15,100
22,600
>30,000
>30,000
>30,000
>30,000
>30,000
13,500
15.5
10.0
19.8
21.0
37.3
37.4
25.5
51.3
5.28
3.52
17.8
13.6
50.4
8.0
3d
3e
3i
3k
3o
3r
3s
3t
>30,000
11,100
4,620
630
6.4
14.2
As a preliminary in vivo screen preferred compounds were tested for their efficacy in
inhibiting acetic acid induced writhing.14 Compounds were tested in a screening mode at 30
mg/kg p.o. in female mice. The strongest inhibition at this dose was observed for compound
3t (38+3% inhibition) with a slightly lower activity for 3d (32+2% inhibition). Compounds
3o, and 3r showed only very low level of response ( 15%) and the rest did not show any
measurable response.15 Due to its strongest activity in the writhing model, compound 3t was
selected for further testing in the more elaborate carrageenan-induced inflammatory pain
model.
12

In vivo PK profiling indicated low exposure of 3t after oral administration in rats.
Interestingly, however, after oral dosing the exposure levels were stable for at least four hours
after administration (Figure 2). Increasing doses (10, 30, 100 mg/kg) of compound 3t were
dosed orally to rats, 90 min. after subcutaneous injection of carrageenan.16 Effects on paw
withdrawal thresholds were measured 30 min. and 90 min. after compound dosing. While at
10 mg/kg compound 3t did not show an effect on withdrawal thresholds in the inflamed paw,
strong increases in the threshold force were observed at 30 and 100 mg/kg, indicating an
analgesic activity of 3t (Figure 3A). This analgesic activity was fully retained 90 min. after
compound dosing (Figure 3B, 10 and 30 mg/kg).17
Route Dose
(mg/kg)
AUC
(mg.h/l)
Cl/kg
(l/h/kg) (h)
Tmax t1/2
Vss/kg
(l/kg)
F (%)
(h)
i.v.
1.0
3.0
0.165
6.12
0.5
3.61
8
(AUC
p.o.
1.0
8.9
last)
Figure 2. Pharmacokinetic profile of 3t in rats
13

Figure 3. Effects of compound 3t in carrageenan-induced allodynia. Increasing doses of
compound 3t (10, 30, 100 mg/kg) were administered orally after carrageenan injection. Paw
withdrawal thresholds in the non-inflamed (black bars) and the inflamed (red bars) paw were
measured using the electronic von Frey filament system. Animals in 10 and 30 mg/kg dose
14

groups were tested at 30 and 90 min. post administration. The 100 mg/kg group was tested at
30 min. post administration.
In conclusion, we have studied the SAR of 2-arylamido-5,7-dihydro-4H-thieno[2,3-
c]pyran-3-carboxamide derivatives as potential cannabinoid receptor agonists. Starting from
an HTS hit, structural features required for potency and selectivity were analyzed by
incremental modification. In general, substituting the cyclohexyl ring fused to the thiophene
ring in the HTS hit 1 by a pyran ring was well tolerated. The screening of aryl rings at the C-
2 amide terminal highlighted the importance of 2-substitution on the phenyl ring. For
selectivity an amine was required with a terminal H-bond acceptor potential at the
carboxyamido end (3-C). An inverse effect of ring size on selectivity against CB1 was
established (difluoroazetidine vs. difluoropiperidine). Thus, the best selectivities were
obtained with 2-methoxyethylamine, 1,1-dimethyl-2-methoxyethylamine and followed by 2-
methoxyethylamine and 3,3-difluoroazetidine. Non-aryl C-2 amide did not show any further
improvement in both potency and selectivity. Similar trends were observed with other rings
fused to thiophene. No improvement in selectivity was observed with gem-difluoride at the 6
position of the ring fused to thiophene.
Some of the compounds were studied for their binding affinity towards both human
and rat CB receptors. Although shifted 20 folds, the results were in agreement with the trend
observed in the functional assay. Compounds 3d and 3t showed a moderate response in the
acetic acid induced writhing model. Compound 3t was further profiled in the carrageenan
induced allodynia model whereby it showed analgesic effects by oral administration which
were fully retained even 1.5 h after dosing.
We are in the process of further optimizing these compounds to improve their
selectivity and pharmacokinetic profile. The results will be published in due course.
Acknowledgements
The authors wish to thank Sham Nikam and Karin Rimvall for their support. Help from
Steffen Weinbrenner, Arijit Das, Avdhoot Velankar, Amit Das, Dilipkumar Yadav, Cornelia
Auriga, Sigrid Gabriel and Srivani Palukuri at various time points is gratefully
acknowledged. The authors also thank the reviewers for their valuable feedback.
15

References
1 Bosier, B., Muccioli, G. G., Hermans, E., Lambert, D. M. Biochem. Pharmacol. 2010, 80, 1.
2 Slipetz, D. M., O'Neill, G. P., Favreau, L., Dufresne, C., Gallant, M., Gareau, Y., Guay, D.,
Labelle, M., Metters, K. M. Mol. Pharmacol. 1995, 48, 352.
3 Drysdale, A.J.; Platt, B. Curr. Med. Chem. 2003, 10, 2719.
4 (a) Onaivi, E. S., Ishiguro, H., Gong, J. P., Patel, S., Perchuk, A., Meozzi, P. A., Myers, L.,
Mora, Z., Tagliaferro, P., Gardner, E., Brusco, A., Akinshola, B. E., Liu, Q. R., Hope, B.,
Iwasaki, S., Arinami, T., Teasenfitz, L., Uhl, G. R. Ann. N. Y. Acad. Sci. 2006, 1074, 514. (b)
Stella, N. Glia 2010, 58, 1017. (c) Morgan, N. H., Stanford, I. M., Woodhall, G. L.
Neuropharmacology 2009, 57, 356. (d) den Boon, F. S., Chameau, P., Schaafsma-Zhao, Q.,
van Aken, W., Bari, M., Oddi, S., Kruse, C. G.; Maccarrone, M., Wadman, W. J., Werkman,
T. R. Proc. Natl. Acad. Sci. USA 2012, 109, 3534.
5 (a) Wiskerke, J., Pattij, T., Schoffelmeer, A. N., De Vries, T. J. Addict. Biol., 2008, 13, 225-
238. (b) Sidhpura, N., Parsons, L. H. Neuropharmacology 2011, 61, 1070.
6 (a) Murineddu, G., Asproni, B., Pinna, G. A. Recent Pat. CNS Drug Discov. 2012, 7, 4. (b)
Riether, D. Exp. Opin. Ther. Pat. 2012, 22, 495. (c) Han, S., Thatte, J., Jones, R. M. Ann.
Rep. Med. Chem. 2009, 44, 227.
7 Huang, Y.; Doemling, A. Molecular Diversity 2011, 15, 3.
8 Recently 2-amido-3-carboxamide thiophene-based CB2 receptor agonists that show efficacy
in two different neuropathic models in rats have been reported: Nelson, D. W., Frost, J. M.,
Tietje, K. R., Florjancic, A. S., Ryther, K., Carroll, W. A., Dart, M. J., Daza, A. V., Hooker,
B. A., Grayson, G. K., Fan, Y., Garrison, T. R., El-Kouhen, O. F., Yao, B., Pai, M.,
Chandran, P., Zhu, C., Hsieh, G. C., Meyer, M. D. Bioorg. Med. Chem. Lett. 2012, 22, 2604.
9
Functional Assay (Luciferase Reporter-Gene Assay): Stimulation of the cells was done
in all tests with Forskolin (SIGMA #F6886). The Luciferase buffer contained 30 mM
HEPES, 25 mM Glycine, 0.5 mM EDTA, 2.5 mM MgCl2, 0.15% TritonX100, 0.5 mM AMP,
1.25 mM ATP, 1 mM DTT, 0.125 mM Coenzyme A (PJK), 0.25 mM firefly D-Luciferin
(PJK), pH of 7.8. In each experiment a concentration-response curve with forskolin was
performed in order to cross-check that agonist activity was determined at an EC80 of
forskolin. Briefly, cells were seeded out 24h before testing in 384-well micro-plates (Greiner)
with DMEM/F12 + Glutamax with 2% FCS with a concentration of 3,000 cells/well for CB2
receptors and 1,000 cells/well for CB1 receptors. Four replicates of one respective compound
in each dilution step were generated. Micro-plates were incubated at 37°C for 4 h. Before
luminescence reading cell lysis and Luciferase reaction were triggered by the addition of
Luciferase buffer. Plates were incubated for 4 min at room temperature before reading. The
readout was done with an EnVision2102 multi-label reader (Perkin Elmer) by reading each
well for 0.5 sec with the integrated ultrasensitive luminescence detector. Data analysis was
done with Excel and GraphPad Prism.
16

10 The 2-iPr analogues (derived from 2g) retained high potency at hCB2. However, even with
the most potent analogues selectivity no more than 1log was achieved. Microsomal stability
of these compounds was found to very low as compared to compounds reported in Table 5.
11 (a) A gem-Liu, P.; Sharon, A.; Chu, C. K. J Fluor Chem. 2008, 129, 743. (b) Xu, Y., Qian,
L., Pontsler, A. V., McIntyre, T. M., Prestwich, G. D. Tetrahedron 2004, 60, 43. (c) Hunault,
J., Diswall, M., Frison, J-C., Blot, V., Rocher, J., Marionneau-Lambot, S., Oullier, T.,
Douillard, J-Y., Guillarme, S., Saluzzo, C., Dujardin, G., Jacquemin, D., Graton, J., Le
Questel, J-Y., Evain, M., Lebreton, J., Dubreuil, D., Le Pendu, J., Pipelier, M. J. Med. Chem.
2012, 55, 1227.
12
Membrane preparations and radioligand binding assay: Membranes from cells stably
expressing human CB1 and human CB2 receptors were purchased from Millipore USA
(HTS019M, ChemiScreen). The human CB2 and human CB1 receptors were tested at
membrane concentrations of 1 µg/well (CB2) and 10 µg/well (CB1). Reactions were
performed in 200 µL binding buffer (50 mM HEPES, 5 mM MgCl2, 1 mM CaCl2, 0.2% BSA
fraction V, pH 7.4) containing 1 nM ³H-CP55,940 (CB2 binding assay) or 2 nM ³H-CP55,940
(CB1 binding assay) in a In a 96-well micro-plate (Matrix). After 60 min at 22 °C the
reaction mix was filtered through GF-B filter-plates (Perkin Elmer) followed by 8 wash steps
with 200 µL ice cold washing buffer (50 mM HEPES pH 7.4, 500 mM NaCl, 0.1% BSA
fraction V). At the end, 40 µL Microscint (PerkinElmer) were added on top of the filters, the
filter-plates sealed on top and incubated for at least 20 min at room temperature. The filter-
plates were read in a Wallac TopCount (Perkin Elmer). Three replicates of one respective
compound in each dilution step were used.
13 t1/2 for 3t in HLM was found to be 25.4 min.
14
Inhibition of acetic acid induced writhing: Female mice, NMRI, were obtained from
breeder Charles River, Sulzfeld. Animals were included in the study following an
acclimatization period of at least 5 days after arrival. Body weight of test animals on Day 0 of
the study was between 25 – 30g. Number of writhings induced by the i.p. injection of 0,75%
AcOH (10 ml/kg, pH 4.5) were observed 30 min after a single po administration of vehicle or
suspension of compound. Animals (n=5 per group) were directly placed in small boxes after
i.p. injection of acetic acid and writhing behavior was observed for the following 20 minutes.
Number of writhing was counted by the observer using the manual counter for 20 minutes in
5 minutes blocks. 30% decrease in the writhing behaviour compared to the control group was
considered to be as a relevant compound effect. The animals were observed for possible side
effect and those noted. Number of animal’s death was noted if occurring, moribund animals
were euthanized.
15 As a cut-off, 30% inhibition of acetic-acid induced writhing response was considered
meaningful.
16 Compound 3t and other compounds of interest were found to have activity and selectivity
in rCB2/1 in the similar range to hCB2/1.
17
Carrageenan induced mechanical allodynia: Male rats, Sprague Dawley, were obtained
from breeder Charles River, Sulzfeld. Animals were included in the study following an
acclimatisation period of at least 5 days after arrival. Body weight of test animals was
between 160 - 200 g. Basal values for mechanical allodynia and paw volumes were measured
17

for every animals from each group (n=8). The rats were fasten in the rat holder, their left foot
swabbed with ethanol, and 0.05 ml Carrageenan (1.5%, Cargill) was injected subcutaneously
into the plantar surface of the left hind paw of the rat. The Carrageenan injection induces
local inflammation, paw swelling and pain. 1.5 hours post-Carrageenan, the rats were
administered with the vehicle or testing compound. For behaviour readout, rats were placed
in a plexiglass chamber on an elevated wire mesh platform and left to acclimate for around 15
minutes. Mechanical paw withdrawal thresholds (PWT) were measured by using electronic
von Frey filament system (Bioseb, France). Each, contralateral and ipsilateral paw was tested
3 times to obtain a mean value.
18

Table 1. Screening of substitutions on the aryl ring
Compound # R1
R2
EC50 (hCB2) (nM/%) EC50 (hCB1) (nM/%) Selectivity
2-Cl
H
227 (112)
3270 (85)
4639 (78)
>10000
14
2a
2b
2c
2d
2e
2f
695
(76)
2-Cl
6-Cl
7
2-Cl
4-Cl 9354 (108)
NA
NA
NA
12
3-Cl
H
H
H
H
H
H
H
H
H
6-F
1421 (98)
>10000
>10000
4-Cl
>10000
2-Me
2-iPr
646 (105)
7452 (97)
1500 (105)
>10000
35
(101)
42
2g
2h
2i
2-OMe
2-OCF2
2-OCF3
3-OCF3
2-CF3
2-CF3
>10000
NA
>20
>90
>10
50
488
(96)
>10000
103 (106)
>10000
2j
749
(77)
>10000
2k
2l
35
(102)
1753 (84)
4538 (56)
95
(100)
48
2m

2-Ph
H
7892 (63)
>10000
NA
2n

Table 2. Screening of modifications on the amide part of the carboxylic acid end
Compound R1
#
R2
R3
EC50
(hCB2)
EC50
(hCB1)
Selectivity
(nM/%) (nM/%)
2-CF3
H
H
H
H
-tBu
-tBu
-OH
897
(88)
>10000
>10000
>10000
11
3a
3b
3c
3d
2-OCF3
2-Cl
939
(102)
11
>10000
NA
204
2-CF3
3.0
607
(99)
(93)
2-OCF3
2-CF3
H
9.5
(101)
925
(85)
98
71
3e
3f
6-F
14
(103)
1000
(94)
2-OCF3
2-CF3
H
35
(94)
519
(88)
15
3g
3h
3i
6-F
H
11
(112)
812
(90)
74
2-CF3
17
(93)
8710
(45)
525
2-OCF3
2-CF3
H
96
(100)
>10000
>104
385
3j
3k
3l
6-F
H
24
(96)
9360
(58)
2-CF3
41
>10000
>243
(93)

2-OCF3
2-CF3
H
101
(81)
>10000
>10000
>100
>135
3m
3n
6-F
74
(94)
2-CF3
2-OCF3
2-CF3
2-CF3
H
15
(98)
5960
(53)
387
37
3o
3p
3q
3r
H
38
(100)
1416
(45)
6-F
H
24
(100)
6611
(55)
275
118
2.5
296
(106)
(84)
2-OCF3
2-CF3
H
4.7
(104)
1974
(104)
420
122
57
3s
3t
6-F
H
6.5
(96)
798
(83)
2-CF3
1.4
80
3u
(99)
(90)
2-OCF3
2-CF3
H
3.3
(96)
125
(79)
38
71
3v
6-F
3.6
251
3w
(97)
(94)

Table 3. Screening of non-aryl substituents
Compound X
#
R1
R2
EC50
(hCB2) (hCB1)
(nM/%) (nM/%)
EC50
Selectivity
CH2
23
(94)
1905
(90)
84
4a
4b
4c
4d
4e
4f
O
489
(101)
>10000 >20
CH2
O
71
(81)
8030
(61)
113
313
(84)
>10000 >32
>10000 >49
>10000 >65
CH2
CH2
201
(101)
152
(98)
CH2
344
>10000 >29
4g
(88)
CH2
CH2
47
(110)
3972
(63)
85
4h
4i
219
>10000 >45
(92)

CH2
CH2
5.4
(99)
336
(91)
63
4j
56
(103)
>10000 >178
>10000 >25
4k
CH2
394
4l
(89)
CH2
CH2
4.7
(82)
113
(86)
24
4m
4n
26
(104)
3217
(100)
126
CH2
209
>10000 >48
4o
(87)

Table 4. Modifications on the ring fused to thiophene
Compoun A
d #
R1
R2
EC50
(hCB2)
EC50
(hCB1)
Selectivity
(nM/%) (nM/%)
137
(52)
>10000 >72
5a
5b
5c
106
(76)
>10000 >94
0.55
16
30
(116)
(86)
129
(96)
>10000 >77
5d
5e
5f
6.4
(101)
243
38
62
42
36
(96)
5.9
(100)
368
(98)
40
(91)
1656
(88)
5g
5h
6.0
217
(105)
(94)

49
81
2
5i
(98)
(95)
46
(105)
958
(97)
21
8
5j
5k
5l
18
(99)
139
(93)
63
>10000 >145
(96)

Table 5. Binding affinity, h/rCB2/1 and microsomal stability data of key compounds
Compound Ki (hCB2) Ki (hCB1) Ki (rCB2) Ki (rCB1) t1/2
t1/2
#
(nM)
241
347
999
904
899
81.5
240
444
(nM)
(nM)
1610
2130
2540
1,700
3,110
594
(nM)
(MLM)
(min.)
(RLM)
(min.)
11,700
19,400
>30,000
>30,000
>30,000
7,190
15,100
22,600
>30,000
>30,000
>30,000
>30,000
>30,000
13,500
15.5
10.0
19.8
21.0
37.3
37.4
25.5
51.3
5.28
3.52
17.8
13.6
50.4
8.0
3d
3e
3i
3k
3o
3r
3s
3t
>30,000
11,100
4,620
630
6.4
14.2

Figure 1. HTS hit

Scheme 1. Synthesis route for synthesis of compounds studied (D)