Multistep continuous-flow synthesis in medicinal chemistry: Discovery and preliminary structure-activity relationships of CCR8 ligands

Article abstract of DOI:10.1002/chem.201204350

A three-step continuous-flow synthesis system and its application to the assembly of a new series of chemokine receptor ligands directly from commercial building blocks is reported. No scavenger columns or solvent switches are necessary to recover the desired test compounds, which were obtained in overall yields of 49-94 %. The system is modular and flexible, and the individual steps of the sequence can be interchanged with similar outcome, extending the scope of the chemistry. Biological evaluation confirmed activity on the chemokine CCR8 receptor and provided initial structure-activity-relationship (SAR) information for this new ligand series, with the most potent member displaying full agonist activity with single-digit nanomolar potency. To the best of our knowledge, this represents the first published example of efficient use of multistep flow synthesis combined with biological testing and SAR studies in medicinal chemistry. Copyright

Full text of DOI:10.1002/chem.201204350

FULL PAPER
DOI: 10.1002/chem.201204350
Multistep Continuous-Flow Synthesis in Medicinal Chemistry: Discovery and
Preliminary Structure–Activity Relationships of CCR8 Ligands
Trine P. Petersen,[a, b] Sahar Mirsharghi,[a] Pia C. Rummel,[c] Stefanie Thiele,[c]
Mette M. Rosenkilde,[c] Andreas Ritzꢀn,*[b, d] and Trond Ulven*[a]
Abstract: A three-step continuous-flow
synthesis system and its application to
the assembly of a new series of chemo-
kine receptor ligands directly from
commercial building blocks is reported.
No scavenger columns or solvent
switches are necessary to recover the
desired test compounds, which were
obtained in overall yields of 49–94%.
The system is modular and flexible,
and the individual steps of the se-
quence can be interchanged with simi-
lar outcome, extending the scope of the
chemistry. Biological evaluation con-
firmed activity on the chemokine
CCR8 receptor and provided initial
structure–activity-relationship (SAR)
information for this new ligand series,
with the most potent member display-
ing full agonist activity with single-digit
nanomolar potency. To the best of our
knowledge, this represents the first
published example of efficient use of
multistep flow synthesis combined with
biological testing and SAR studies in
medicinal chemistry.
Keywords: chemokine receptor li-
gands · flow chemistry · medicinal
chemistry · small molecule agonist ·
structure–activity relationships
Introduction
in principle the convenient synthesis of complex products
from simple starting materials in a single continuous flow in
a fraction of the time required for batch synthesis. Although
the initial development of an efficient multistep flow synthe-
sis protocol may require some effort, once developed, such
a system allows for multistep library synthesis of screening
compounds and follow-up by repeated or scaled-up synthe-
sis of interesting single library compounds without the need
for any further development. A number of multistep flow
synthesis systems have been reported,[1a,c,2e,3] but the poten-
tial of such systems in medicinal chemistry is still unrealized.
Although continuous-flow synthesis offers a number of
potential advantages, important challenges remain. Precipi-
tation causing clogging and blockage should be taken into
account in the design of flow synthesis systems. Multistep
systems present additional important issues that must be ad-
dressed. First, different reactions frequently require differ-
ent solvents, which demands disruption of the continuous
stream since there are as yet no satisfactory means of ach-
ieving continuous solvent switching. Second, byproducts
may interfere with downstream steps. A third significant
challenge is the “third-stream problem”, that is, timing the
introduction of additional reagents downstream, which is
complicated by dispersion in the first step(s).[4]
Organic synthesis by using continuous-flow methods has re-
ceived a great deal of attention in recent years due to the
potential advantages over traditional batch synthesis that
this technology has to offer.[1] Some of the more prominent
features include precise and rapid heating and cooling, safer
handling of exothermic reactions and explosive or toxic re-
agents, safe and convenient handling of reactions under high
pressure, easy implementation of immobilized catalysts, and
facile scale-up. The improved safety and more accurate tem-
perature control have made continuous-flow synthesis an at-
tractive option for the manufacture of active pharmaceutical
ingredients.[2] Telescoping a multistep synthesis into a single
continuous-flow process by coupling several flow reactions
in series is one of the most intriguing possibilities, enabling
[a] Dr. T. P. Petersen, S. Mirsharghi, Prof. Dr. T. Ulven
Department of Physics, Chemistry and Pharmacy
University of Southern Denmark
Campusvej 55, 5230 Odense M (Denmark)
E-mail: ulven@sdu.dk
[b] Dr. T. P. Petersen, Dr. A. Ritzꢀn
Discovery Chemistry & DMPK, H. Lundbeck A/S
Ottiliavej 9, 2500 Valby (Denmark)
Herein, we report an efficient three-step flow synthesis in
which all of these challenges have been overcome by careful
selection of the reactions. The system has been designed to
synthesize new ligands directed towards the chemokine re-
ceptor CCR8, a potential drug target expressed on mono-
cytes, splenocytes, and thymocytes, and implicated in allergic
disease.[5] Relatively few exogenous ligands have been de-
scribed for this receptor.[6] Inspired by known CCR8 ligands
(Figure 1), we have designed a three-step continuous-flow
[c] Dr. P. C. Rummel, Dr. S. Thiele, Prof. Dr. M. M. Rosenkilde
Laboratory for Molecular Pharmacology
The Panum Institute, University of Copenhagen
Blegdamsvej 3, 2200 Copenhagen N (Denmark)
[d] Dr. A. Ritzꢀn
Present address: Medicinal Chemistry II, LEO Pharma A/S
Industriparken 55, 2750 Ballerup (Denmark)
E-mail: andreas.ritzen@leo-pharma.com
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201204350.
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

target compounds directly from commercial starting materi-
als in three continuous steps with no handling of intermedi-
ates and a capacity of two compounds per day. However,
the productivity was hampered by the three columns con-
taining PS-supported scavenger resins. Since the scavenger
resins are consumed, they must be replaced or regenerated
after the synthesis of every few compounds, which is expen-
sive, requires manual intervention, and lowers the productiv-
ity of the system. For a multistep system, the dispersion of
reactants and intermediates caused by the scavenger col-
umns is especially problematic. The frequency of resin re-
placement can be reduced by using larger columns, but this
typically results in unacceptable dispersion. Furthermore,
the presence of the scavenger columns necessitates long
washing cycles between the synthesis of each compound,
adding a significant unproductive-time penalty to each syn-
thesis. Notably, the scavenger columns did not eliminate the
need for final chromatographic purification of the test
compounds.
In further optimization of the system, a critical question
was whether the benefit of the scavenger columns outweigh-
ed their disadvantages. We reasoned that if stoichiometric
amounts of reagents were to be used and complete conver-
sion could be achieved, there would be no need for scaveng-
ers. A simplified two-step model system was investigated by
using a preformed alkylated and Cbz-protected piperazine
as the substrate (Table 1). The choice of starting material
and reaction sequence was based on the assumption that
complete conversion by using stoichiometric amounts of re-
agents would most likely be easier to achieve for the isocya-
nate addition and hydrogenolysis steps than for the N-alkyl-
Figure 1. Inspirational CCR8 ligands.
system for the synthesis of related structures incorporating a
urea group.[7] In the flow system presented herein, reactions
that can be performed in DMF have been chosen to avoid
problems related to precipitation and solvent switching. Fur-
thermore, the reactions give either no or unproblematic stoi-
chiometric byproducts, thus solving the second of the afore-
mentioned problems. The “third-stream problem” has been
addressed by ensuring a moderate excess of the third build-
ing block over an appropriate time period. In the optimized
version of the system, all scavenger resins have been elimi-
nated. This results in a functional system capable of assem-
bling three variable building blocks into diverse test com-
pounds in a fraction of the time required for batch synthesis.
We also report that the compounds synthesized by this
system indeed act as potent CCR8 modulators, and we de-
scribe preliminary structure–activity-relationship studies of
the 4-benzylpiperazine-1-carboxamide compound series.
Results and Discussion
Table 1. Two-step flow system.
Continuous-flow synthesis: Our
first-generation system comprised
three consecutive steps in a contin-
uous flow: isocyanate addition to a
carbobenzyloxy (Cbz)-monoprotect-
ed diamine, removal of the protect-
ing group by catalytic hydrogena-
Product
Purity [%][a]
99
pEC50 (EC50 [nm]; Emax [%])[b]
6.61Æ0.34 (245; 68)
tion, and alkylation of the liberated
amine with benzyl bromide or a
similar reagent (Scheme 1).[7] To
15
ensure complete conversion, an
excess of isocyanate was used in the
urea formation, and a polystyrene
(PS)-trisamine scavenger column
was inserted directly after the reac-
tor to remove the excess. The final
N-alkylation step was carried out in
a PS-NMM (NMM=N-methylmor-
pholine) column and then the mix-
ture proceeded to a PS-trisamine
column to scavenge the excess al-
16
99
99
7.39Æ0.10 (41; 62)
7.46Æ0.18 (34; 74)
7.62Æ0.20 (24; 85)
17
18
94
ACHTUNGTRENNUNGkylating reagent. This system dem-
onstrated a significant advantage of
flow chemistry: the synthesis of
1
[a] Purity by ELSD (evaporative light scattering detector), H and 13C NMR spectroscopy. [b] Activation
of COS-7 cells transfected with CCR8 and promiscuous Gqi4myr protein in an inositol triphosphate-
turnover assay; Emax is given as a percentage of maximal induced activation by CCL1.
&
2
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Library Synthesis Ligands by a Continuous-Flow System
FULL PAPER
Scheme 1. First-generation three-step flow system. Method A: stock solutions of Cbz-diamine and alkylating agent in ethanol and the required isocyanate
in DMF. Method B: all stock solutions in DMF.[7] Method I: see Table 2. Yields are after purification by semi-automatic flash chromatography (Combi-
flash Companion) and are reproduced from ref. [7]. Activation of COS-7 cells transfected with CCR8 and promiscuous Gqi4myr protein in an inositol
triphosphate-turnover assay; Emax is given as a percentage of maximal induced activation by CCL1 at 10 mm concentration of test compound.
A
mediate stream and the alkylating agent stream. Such mis-
matching could occur as a result of excessive dispersion in
the first two steps. The problem of matching the concentra-
tion and timing of a third stream to that of a reaction mix-
ture exiting a flow reactor, the concentration–time profile of
which may have been changed by dispersion, is a recognized
challenge in multistep flow synthesis (the “third-stream
problem”).[4] One solution to this problem has been the use
of rather sophisticated equipment to measure reactant con-
centrations in-line, for example, by IR or UV detection, to
automatically control the flow rate of the third stream.[4a–d]
We opted for the more low-tech solution of simply adjusting
the concentration–time profile of the third reagent stream
to overlap with that of the intermediate stream so that at
least a stoichiometric amount of the third reagent would be
present at all times. To reduce costs and increase capacity,
the PS-NMM base in the column reactor was replaced by a
mixture of sodium carbonate and sand. We found that feed-
ing a threefold excess of the alkylating agent relative to the
amine/isocyanate solution of the first step with a 40 min
with its integral pump feeding Cbz-phenoxybenzylpiperazine
through the hydrogenation module and an additional pump
feeding the isocyanate solution into a T-piece, where it was
mixed with the deprotected piperazine. A steady-state 1:1
stoichiometry between the piperazine and the isocyanate
was maintained by using 0.1m reagent solutions and both
pumps running at 0.50 mLminÀ1, giving a reaction time of
2 min at RT. We were pleased to observe that this simple
system facilitated quantitative conversion to the products,
and that the compounds produced were of a suitable quality
for direct testing without the need for a final purification
(Table 1).
We next turned our attention to the N-alkylation step.
Based on the results obtained with the two-step system, it
seemed likely that the sometimes low to moderate yields
and formation of side products seen with the first-generation
three-step system were related to this step and possibly to
the scavenger columns. Poor conversion might result from a
mismatching of concentration–time profiles of the inter-
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

A. Ritzꢀn, T. Ulven et al.
delay ensured that the alkylating agent was in excess at all
times and resulted in complete conversion (at the same con-
centration, 0.050m, and flow rate, 0.20 mLminÀ1, at 758C
with a residence time of 12.5 min). Since a final purification
was necessary under any circumstances and excess alkylating
agent could thus be easily removed, the second PS-trisamine
scavenger column was also removed.
eration, whereas the yields from the second-generation
system (Table 2) are based on the total amounts injected.
The yields reported in Tables 1 and 2 are therefore not di-
rectly comparable.
One remaining limitation that we sought to address is the
need for the urea substituent to withstand catalytic hydroge-
nation with a Pd/C catalyst. Under these conditions, aromat-
ic halides and other readily reduced functional groups will
not survive. Reversing the sequence of the steps, that is,
placing the isocyanate addition after the H-Cube, would
provide a solution to this problem. This should permit the
presence of reduction-sensitive functionalities in either R1
or R2 (but not in both at the same time) depending on the
order of the steps. Performing the alkylation step first im-
plies formation of HBr early in the sequence, with a risk of
poor conversion and precipitation downstream in the case
of incomplete scavenging of this acid. To investigate the fea-
sibility of a reversed system, we swapped the tube reactor
and sodium carbonate/sand column along with the corre-
sponding reagents (Table 2, configuration II). As before, the
first step was performed by using equimolar amounts of re-
agents whereas for the third step a threefold excess of iso-
cyanate was used. Employing the same flow rates as in the
original system, the residence time in the alkylation reactor
was doubled while the residence time in the isocyanate addi-
tion reactor was halved due to the third stream. We were
pleased to find that the reversed version of the system
worked just as well as the original sequence (compare, for
example, 21 with 33 in Table 2). The unsubstituted 3-phenyl-
benzyl and piperazine building blocks performed well with
4-chlorobenzyl isocyanate, but the lack of a final PS-tris-
With these improvements implemented, a modified three-
step system was configured by using stoichiometric amounts
of the two first reagents and with the scavenger columns re-
moved; see Table 2, configuration I. A 0.05m solution of the
starting amine and isocyanate in DMF was pumped at a
flow rate of 0.20 mLminÀ1 through a 5 mL tube reactor at
1008C, from which it proceeded to the H-Cube (full H2
mode, 808C, catalyst cartridge: 10% Pd/C). The output
stream from the H-Cube was fed into a gas–liquid separator
to vent out excess hydrogen gas and carbon dioxide. The
stream containing the deprotected piperazine was combined
with a solution of the alkylating agent (both pumped at
0.20 mLminÀ1) by using a T-piece and the mixture was fed
through the sodium carbonate/sand column at 758C with a
residence time of 12.5 min. We were delighted to observe
that when the synthesis of 6 was performed with the im-
proved system (Table 2, configuration I), a 90% yield was
obtained, as compared with 6 or 28% yields with the initial
system (Scheme 1). The use of unsubstituted 3-phenoxyben-
zyl bromide, Cbz-piperazine, and benzyl isocyanate gave a
94% overall yield (19, Table 2). The yield dropped to 49%
with homopiperazine (20, Table 2), but increased to 85%
and above when reverting to Cbz-piperazine and introduc-
ing small substituents on the isocyanate or the alkylating
agent (21, 22, Table 2). Substituents on the piperazine ring
introduce steric hindrance and can potentially interfere with
both the urea formation and the alkylation step. However,
methyl groups in both positions were well tolerated, result-
ing in 62–89% yields of the respective products (23, 24, 27,
29, 30, Table 2). Gratifyingly, piperazines with the larger
phenyl substituent also performed well, giving 57–66%
yields (25, 26, Table 2). The use of 4-(Cbz-amino)piperidine
instead of Cbz-piperazines as the central building block re-
sulted in 50–82% yields (31, 32, Table 2). Compounds 29–33
were produced with configuration II, see below.
It is noteworthy that the above results were obtained with
a 1:1 ratio of the Cbz-diamine and isocyanate building
blocks, with only the alkylating agent in the third step being
used in excess (3 equiv). Removal of the scavenger resins
drastically reduced the cost and down time associated with
their periodic replacement, and did not complicate the final
purification (except in the case of compound 33; see below).
Another improvement was achieved by taking dispersion
into account and compensating for it in the third step as de-
scribed above. No effort to control dispersion was made in
the first-generation system, which resulted in significant por-
tions of starting materials being wasted in the leading and
trailing edges of the main cut. The yields reported for our
first-generation system (Scheme 1) are based on the product
collected over a typical period of 10 min at steady-state op-
AHCTUNGTREGaNNNU mine scavenger column prevented direct chromatographic
separation of excess isocyanate from the final product in
one case, that of compound 33. However, this issue was
solved by collecting the final product in a flask containing
diethylenetriamine (DTA), thus converting excess isocya-
nate into a polar urea that was readily removed in the final
chromatography.
Biological testing: The compounds were tested in an inositol
phosphate turnover assay (IP3 assay) in COS-7 cells transi-
ently transfected with human CCR8. Screening of the 4-ben-
zylpiperazine-1-carboxamide compounds from the first-gen-
eration system resulted in the identification of several active
compounds serving as partial or full agonists (Scheme 1). In
agreement with the inspirational structures (Figure 1),[6a,b]
a
western 3-phenoxyphenyl moiety (2 and 5) afforded the
most potent compounds with EC50 around 100 nm, whereas
3-biphenyl (1, 3, 7, and 11) and 2-naphthyl (compounds 4, 8,
and 12) compounds were also active with EC50 values of the
most potent compounds (8 and 11) around 1 mm. The 3-bi-
phenyl/2-naphthyl pairs 7/8 and 11/12 demonstrate that pref-
erence for either group depends on the rest of the molecule.
Simple substituted benzyl (6), 2-biphenylmethyl (10), non-
aromatic (13), or heteroaromatic (9, 14) groups that deviat-
ed much from this pattern were mostly inactive, although it
cannot be excluded that the eastern 3,5-dimethylphenyl or
&
4
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Library Synthesis Ligands by a Continuous-Flow System
Table 2. Improved three-step flow system.[a]
FULL PAPER
Product
Method
I
Yield [%]
94
pEC50 (EC50 [nm]; Emax [%])[b]
7.10Æ0.21 (80; 92)
19
20
21
22
I
I
I
49
85
92
6.37Æ0.06 (422; 88)
6.13Æ0.02 (743; 81)
7.54Æ0.09 (29; 38)
23
I
62
6.94Æ0.03 (114; 46)
24
25
I
I
87
66
5.86Æ0.16 (1370; 48)
5.41Æ0.32 (3870; 10)
26
27
I
I
57
62
4%[b]
6.85Æ0.63 (142; 52)
28
29
I
84
79
8.46Æ0.25 (3; 60)
7.44Æ0.42 (37; 32)
II
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

A. Ritzꢀn, T. Ulven et al.
Table 2. (Continued)
Product
Method
II
Yield [%]
89
pEC50 (EC50 [nm]; Emax [%])[b]
30
31
1%[b]
II
50
6.69Æ0.12 (203; 78)
32
33
II
82
74
7.52Æ0.07 (30; 81)
6.34Æ0.13 (462; 83)
II/DTA[c]
[a] Method I refers to synthesis by the general procedure for configuration I. Method II refers to synthesis by the general procedure for configuration II.
BPR=back-pressure regulator; GLS=gas–liquid separator. [b] Activation of 10 mm concentration. [c] DTA=diethylenetriamine.
2-ethoxyphenyl substituents also contributed to the low ac-
tivity. The eastern part of the molecule could in general ac-
commodate a broader variety of structures, with the N’-urea
substituents of the most active compounds comprising sub-
stituted phenyl (2, 8), phenylethyl (11), and aliphatic (5)
groups.
creased the potency (EC50) to 37 nm. A tert-butyl substituent
at the para-position of the western phenoxy moiety (30) re-
sulted in a complete loss of activity. The larger phenyl sub-
stituent on the piperazine ring reduced the potency tenfold
when flanking the urea moiety (25), and resulted in almost
complete loss of activity even at a concentration of 10 mm
when flanking the phenoxybenzyl moiety (26). Unsubstitut-
ed 3-phenoxybenzylpiperazines with western N-benzyl (19)
and N-cyclohexyl (18) substituents exhibited potent activity.
Gratifyingly, combining the two in 28 resulted in a highly
potent CCR8 agonist with an EC50 of 3 nm, the highest po-
tency observed for any of the tested compounds.
The central 4-aminopiperidine moiety occurs in related
CCR8 ligands (e.g., 11c in Figure 1), and the mono-Cbz-pro-
tected derivative was confirmed to be a successful substrate
in the flow system. Compound 31, and more particularly 32,
were among the most potent, but did not rival 28.
The favored 3-phenoxybenzylpiperazine was chosen as
the substrate for the exploratory two-step system (Table 1),
as this scaffold was expected to be active by comparison
with the inspirational compounds. Appending the 2-chloro-
benzylurea (15) produced a CCR8 agonist with an EC50 of
245 nm (Table 1). Ureas substituted with phenylethyl (16),
isopropyl (17), or cyclohexyl (18) group exhibited potencies
that were further increased by an order of magnitude.
The compounds synthesized by the optimized three-step
system were also mostly based on the 3-phenoxybenzylpi-
perazine scaffold. Compared to the unsubstituted N-benzyl-
piperazine 19, the corresponding homopiperazine 20 proved
to be less potent. The same was true for the products with a
para-methyl- (21) or para-chloro-substituted (33) eastern N-
benzyl moiety. However, this information was not available
prior to the synthesis of the remaining analogues, most of
which bear an eastern N-benzyl group with a para-methyl or
para-chloro substituent. An ortho-fluoro substituent on the
western terminal phenoxy analogue (22) resulted in in-
creased potency and a reduction in efficacy to 38% relative
to the endogenous agonist CCL1. An (S)-methyl substituent
on the piperazine ring of 21 flanking the phenoxybenzyl
moiety (23) also increased the potency to the level seen for
the unsubstituted analogue 19. A corresponding methyl sub-
stituent in combination with western 3,5-dichlorophenoxy
and eastern N-cyclopentyl moieties (27) resulted in similar
activity. Adding a para-fluoro substituent to the western
benzyl group of 21 in combination with an (R)-methyl group
flanking the phenoxybenzyl moiety of the piperazine (24)
resulted in substantial reductions in both potency and effica-
cy, whereas an (R)-methyl group flanking the urea moiety in
combination with an eastern para-chloro substituent (29) in-
Conclusion
Multistep flow-synthesis systems show great promise in me-
dicinal chemistry, but their full potential has not yet been re-
alized. This is likely due to the challenges related to precipi-
tation of intermediates or products, the need for different
solvents for different reactions, interference of byproducts
or excess reagents with downstream steps, and timing of re-
agent addition in downstream steps. In this report, we have
shown how these limitations can be overcome in an efficient
and inexpensive manner. By combining reactions that do
not produce byproducts likely to be problematic in down-
stream steps and that can all be performed in DMF, we
have devised an efficient three-step continuous-flow system
for the synthesis of diverse test compounds from commercial
building blocks. In the initial version of the system, three
polystyrene-based scavenger columns were used to remove
excess reagents and to scavenge acid produced in the alkyl-
AHCTUNGTREGaNNNU tion step. This, however, led to significant problems with
&
6
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Library Synthesis Ligands by a Continuous-Flow System
FULL PAPER
dissolved in DMF (30 mL) and pumped by a third pump (0.20 mLminÀ1
,
dispersion and feeding of the third building block, in addi-
tion to the need to replace the resin between every few re-
actions. The optimized system does not contain any scaveng-
er resins; instead, an Na2CO3/sand column is used, with ad-
justed stoichiometry of the reagents. This resulted in a
system for the practical and efficient synthesis of ligand ana-
logues in generally excellent yields over the three steps. The
substrate scope of the system has been demonstrated to be
broad for all three building blocks. For example, the applica-
tion of 4-(Cbz-amino)piperidine in place of Cbz-piperazine
allowed access to structures closely related to those of previ-
ously reported CCR8 ligands. Testing on CCR8 established
that the 3-phenoxybenzylpiperazine scaffold generally dis-
played the highest potency. Various small substituents were
generally well tolerated, but only in a few cases produced
substantially increased potency. A larger degree of variation
was accommodated in the eastern part of the compounds.
Fusing the eastern parts of two potent compounds led to the
identification of the most potent N-tetraline 28, a single-
digit nanomolar CCR8 agonist. The multistep flow system
has thus been demonstrated as an efficient tool in the opti-
mization of receptor ligands and implies drastic reductions
in both time and labor compared to batch synthesis after im-
plementation. To the best of our knowledge, this represents
the first example of efficient use of a multistep flow system
for the synthesis of test compounds in the optimization pro-
started 40 min after the first pump) to mix with the intermediate from
the second pump. The mixture continued through a glass column filled
with Na2CO3/sand (9+9 g, 5 mL, 758C). The solution was collected, the
solvent was evaporated under reduced pressure, and the product was pu-
rified by column chromatography using a semi-automatic Combiflash
system (24 g silica; eluent gradient: heptane/EtOAc).
4-(3-Phenoxybenzyl)-N-(1,2,3,4-tetrahydronaphthalen-1-yl)piperazine-1-
carboxamide (28): The title compound was prepared according to the
general procedure (configuration I) from benzyl piperazine-1-carboxylate
(109.6 mg,
(87.5 mg,
0.497 mmol),
0.505 mmol),
1-isocyanato-1,2,3,4-tetrahydronaphthalene
and 1-(bromomethyl)-3-phenoxybenzene
(393.9 mg, 1.50 mmol). It was obtained in a yield of 185.6 mg (84%).
1H NMR (500 MHz, CDCl3): d=7.36 (t, J=7.2 Hz, 3H), 7.30 (dd, J=
13.1, 5.1 Hz, 1H), 7.21–7.16 (m, 2H), 7.12 (dt, J=11.2, 7.3 Hz, 3H), 7.08–
7.01 (m, 3H), 6.93 (d, J=8.1 Hz, 1H), 5.15–5.08 (m, 1H), 4.81 (d, J=
8.0 Hz, 1H), 3.53 (s, 2H), 3.43–3.32 (m, 4H), 2.89–2.72 (m, 2H), 2.46 (t,
J=4.7 Hz, 4H), 2.08 (dd, J=12.8, 8.3 Hz, 1H), 1.85 ppm (d, J=4.7 Hz,
3H); 13C NMR (126 MHz, CDCl3): d=157.23, 157.20, 157.01, 140.05,
129.71, 129.52, 123.86, 123.18, 119.39, 118.76, 117.54, 62.50, 52.68, 50.68,
43.76, 29.46 ppm; HRMS: m/z calcd for C28H32N3O2 [M+H]: 442.2489;
found: 442.2488.
General procedure for the three-step flow synthesis (configuration II):
The Cbz-protected diamine (0.50 mmol) and alkylating agent
(0.50 mmol) were dissolved in DMF (10.0 mL). The solution was pumped
at a flow rate of 0.20 mLminÀ1 through a glass column filled with
Na2CO3/sand (9+9 g, 5 mL, 758C) followed by an H-Cube (catalyst:
10% Pd/C, 30ꢂ4 mm cartridge, full H2 mode, 808C). The product was
collected in a flask (5 mL) to release excess H2 and from there passed on
by a second pump (0.20 mLminÀ1). The isocyanate (1.5 mmol) was dis-
solved in DMF (30 mL) and pumped by a third pump (0.20 mLminÀ1
,
started 40 min after the first pump) to mix with the intermediate from
the second pump. The mixture continued through a coiled reactor (PFA,
5 mL, 1008C). The solution was collected, the solvent was evaporated
under reduced pressure, and the product was purified by column chroma-
tography using a semi-automatic Combiflash system (24 g silica; eluent
gradient: heptane/EtOAc).
ACHTUNGTRENNUNGcess of an active medicinal chemistry project.
Experimental Section
(R)-4-{3-[4-(tert-Butyl)phenoxy]benzyl}-N-(4-chlorobenzyl)-2-methylpi-
perazine-1-carboxamide (30): The title compound was prepared accord-
ing to the general procedure (configuration II) from (R)-benzyl 2-methyl-
piperazine-1-carboxylate (116.6 mg, 0.498 mmol), 1-bromomethyl-3-[4-
(tert-butyl)phenoxy]benzene (162.3 mg, 0.508 mmol), and 4-chlorobenzyl
isocyanate (259.8 mg, 1.55 mmol). It was obtained in a yield of 224.5 mg
(89%). 1H NMR (500 MHz, DMSO): d=7.39 (d, J=8.6 Hz, 2H), 7.35
(d, J=8.3 Hz, 2H), 7.26 (d, J=8.1 Hz, 2H), 7.03 (d, J=6.7 Hz, 2H), 6.95
(d, J=8.3 Hz, 3H), 6.89 (d, J=7.9 Hz, 1H), 4.22 (qd, J=15.6, 5.7 Hz,
2H), 4.10 (s, 1H), 3.70 (d, J=12.6 Hz, 1H), 3.53 (d, J=13.9 Hz, 1H),
3.34 (d, J=13.7 Hz, 1H), 2.92 (t, J=11.3 Hz, 1H), 2.74 (d, J=10.6 Hz,
1H), 2.55 (d, J=11.0 Hz, 1H), 2.01 (dd, J=10.8, 2.8 Hz, 1H), 1.92 (t, J=
10.2 Hz, 1H), 1.28 (s, 9H), 1.05 ppm (d, J=6.5 Hz, 3H); 13C NMR
(126 MHz, DMSO): d=157.42, 157.04, 153.96, 145.89, 140.64, 140.24,
130.87, 129.66, 128.80, 127.97, 126.63, 122.86, 118.63, 117.28, 116.69, 61.22,
56.91, 52.91, 46.12, 42.80, 38.54, 34.00, 31.21, 15.23 ppm; HRMS: m/z
calcd for C30H37ClN3O2 [M+H]: 506.2569; found: 506.2562.
General procedure for the two-step flow synthesis: Benzyl 4-(3-phenoxy-
benzyl)piperazine-1-carboxylate (2.5 mmol) was dissolved in EtOH
(25 mL). The solution was pumped at a flow rate of 0.50 mLminÀ1
through an H-Cube (catalyst: 10% Pd/C, 30ꢂ4 mm cartridge, full H2
mode, 258C). The required isocyanate (1.0 mmol) was dissolved in DMF
(10 mL) and pumped by a second pump (0.50 mLminÀ1) to mix with the
intermediate from the H-Cube. The mixture continued through a stain-
less steel coiled reactor (2 mL, RT). The product was collected, diluted
with water (20 mL), and extracted with EtOAc (3ꢂ15 mL). The com-
bined extracts were dried over MgSO4, filtered, and the solvent was
evaporated under reduced pressure.
4-(3-Phenoxybenzyl)piperazine-1-carboxylic acid cyclohexylamide (18):
The title compound was prepared according to the general procedure for
the two-step flow synthesis by using cyclohexane isocyanate (126.1 mg,
1.01 mmol). 1H NMR (600 MHz, CDCl3): d=7.29–7.23 (m, 2H), 7.19 (t,
J=7.8 Hz, 1H), 7.05–7.00 (m, 1H), 6.98 (d, J=7.5 Hz, 1H), 6.96–6.89 (m,
3H), 6.83–6.78 (m, 1H), 4.54 (d, J=7.3 Hz, 1H), 3.59–3.50 (m, 1H), 3.41
(s, 2H), 3.30–3.24 (m, 4H), 2.37–2.31 (m, 4H), 1.89–1.81 (m, 2H), 1.67–
1.58 (m, 2H), 1.57–1.49 (m, 1H), 1.34–1.21 (m, 2H), 1.10–0.97 ppm (m,
3H); 13C NMR (151 MHz, CDCl3): d=157.09, 157.06, 140.00, 129.62,
129.41, 123.78, 123.08, 119.27, 118.63, 117.40, 77.37, 77.16, 76.95, 62.43,
52.59, 49.34, 43.60, 33.79, 25.57, 25.05 ppm; HRMS: m/z calcd for
C24H32N3O2 [M+H]: 394.2489; found: 394.2489.
Acknowledgements
We thank the Danish Agency for Science, Technology, and Innovation
for financial support.
General procedure for the three-step flow synthesis (configuration I):
The required Cbz-protected diamine (0.50 mmol) and isocyanate
(0.50 mmol) were dissolved in DMF (10.0 mL) in a pre-dried flask under
argon. The solution was pumped at a flow rate of 0.20 mLminÀ1 through
a coiled reactor (PFA, 5 mL, 1008C) followed by an H-Cube (catalyst:
10% Pd/C, 30ꢂ4 mm cartridge, full H2 mode, 808C). The product was
collected in a flask (5 mL) to release excess H2 and from there passed on
by a second pump (0.20 mLminÀ1). The alkylating agent (1.5 mmol) was
[1] a) J. Wegner, S. Ceylan, A. Kirschning, Adv. Synth. Catal. 2012, 354,
17–57; b) J. Wegner, S. Ceylan, A. Kirschning, Chem. Commun.
2011, 47, 4583–4592; c) D. Webb, T. F. Jamison, Chem. Sci. 2010, 1,
675–680; d) C. Wiles, P. Watts, Chem. Commun. 2011, 47, 6512–
6535; e) R. L. Hartman, J. P. McMullen, K. F. Jensen, Angew. Chem.
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

A. Ritzꢀn, T. Ulven et al.
2011, 123, 7642–7661; Angew. Chem. Int. Ed. 2011, 50, 7502–7519;
f) B. P. Mason, K. E. Price, J. L. Steinbacher, A. R. Bogdan, D. T.
McQuade, Chem. Rev. 2007, 107, 2300–2318; g) L. Lengyel, V. Gyol-
lai, T. Nagy, G. Dorman, P. Terleczky, V. Hada, K. Nogradi, F. Sebok,
L. Urge, D. Ferenc, Mol. Diversity 2011, 15, 631–638; h) M. York,
Tetrahedron Lett. 2011, 52, 6267–6270; i) A. R. Bogdan, N. W. Sach,
Adv. Synth. Catal. 2009, 351, 849–854; j) P. Watts, C. Wiles, Org.
Biomol. Chem. 2007, 5, 727–732.
S. V. Ley, I. R. Baxendale, B. Wittkamp, J. G. Goode, N. L. Gaunt,
Org. Process Res. Dev. 2010, 14, 393–404; e) C. J. Smith, C. D. Smith,
N. Nikbin, S. V. Ley, I. R. Baxendale, Org. Biomol. Chem. 2011, 9,
1927–1937.
[5] a) J. E. Pease, R. Horuk, Expert Opin. Ther. Pat. 2009, 19, 199–221;
b) E. Mitsuyama, Y. Kunori, T. Kamimura, O. Kaminuma, Mini-Rev.
Med. Chem. 2006, 6, 463–466; c) S. A. Islam, D. S. Chang, R. A.
Colvin, M. H. Byrne, M. L. McCully, B. Moser, S. A. Lira, I. F. Charo,
A. D. Luster, Nat. Immunol. 2011, 12, 167–186.
[2] a) M. D. Hopkin, I. R. Baxendale, S. V. Ley, Chem. Commun. 2010,
46, 2450–2452; b) T. L. LaPorte, C. Wang, Curr. Opin. Drug Discov-
ery Dev. 2007, 10, 738–745; c) M. Viviano, T. N. Glasnov, B. Reichart,
G. Tekautz, C. O. Kappe, Org. Process Res. Dev. 2011, 15, 858–870;
d) F. Lꢀvesque, P. H. Seeberger, Angew. Chem. 2012, 124, 1738–1741;
Angew. Chem. Int. Ed. 2012, 51, 1706–1709; e) A. R. Bogdan, S. L.
Poe, D. C. Kubis, S. J. Broadwater, D. T. McQuade, Angew. Chem.
2009, 121, 8699–8702; Angew. Chem. Int. Ed. 2009, 48, 8547–8550.
[3] a) I. R. Baxendale, J. Deeley, C. M. Griffiths-Jones, S. V. Ley, S.
Saaby, G. K. Tranmer, Chem. Commun. 2006, 2566–2568; b) N.
Pagano, A. Herath, N. Cosford, J. Flow Chem. 2012, 1, 28–31; c) W.
Shu, L. Pellegatti, M. A. Oberli, S. L. Buchwald, Angew. Chem. 2011,
123, 10853–10857; Angew. Chem. Int. Ed. 2011, 50, 10665–10669;
d) I. R. Baxendale, S. V. Ley, A. C. Mansfield, C. D. Smith, Angew.
Chem. 2009, 121, 4077–4081; Angew. Chem. Int. Ed. 2009, 48, 4017–
4021; e) H. R. Sahoo, J. G. Kralj, K. F. Jensen, Angew. Chem. 2007,
119, 5806–5810; Angew. Chem. Int. Ed. 2007, 46, 5704–5708; f) T.
Noꢃl, S. Kuhn, A. J. Musacchio, K. F. Jensen, S. L. Buchwald, Angew.
Chem. 2011, 123, 6065–6068; Angew. Chem. Int. Ed. 2011, 50, 5943–
5946; g) A. Herath, R. Dahl, N. D. P. Cosford, Org. Lett. 2010, 12,
412–415; h) P. F. Li, S. L. Buchwald, Angew. Chem. 2011, 123, 6520–
6524; Angew. Chem. Int. Ed. 2011, 50, 6396–6400; i) L. Kupracz, J.
Hartwig, J. Wegner, S. Ceylan, A. Kirschning, Beilstein J. Org. Chem.
2011, 7, 1441–1448; j) T. N. Glasnov, C. O. Kappe, Adv. Synth. Catal.
2010, 352, 3089–3097; k) M. Brasholz, J. M. Macdonald, S. Saubern,
J. H. Ryan, A. B. Holmes, Chem. Eur. J. 2010, 16, 11471–11480; l) E.
Riva, A. Rencurosi, S. Gagliardi, D. Passarella, M. Martinelli, Chem.
Eur. J. 2011, 17, 6221–6226; m) A. Sniady, M. W. Bedore, T. F. Jami-
son, Angew. Chem. 2011, 123, 2203–2206; Angew. Chem. Int. Ed.
2011, 50, 2155–2158.
[6] a) C. A. Haskell, R. Horuk, M. Liang, M. Rosser, L. Dunning, I.
Islam, L. Kremer, J. Gutierrez, G. Marquez, C. Martinez, M. J. Bis-
cone, R. W. Doms, S. Ribeiro, Mol. Pharmacol. 2006, 69, 309–316;
b) S. Ghosh, A. Elder, J. Guo, U. Mani, M. Patane, K. Carson, Q. Ye,
R. Bennett, S. Chi, T. Jenkins, B. Guan, R. Kolbeck, S. Smith, C.
Zhang, G. LaRosa, B. Jaffee, H. Yang, P. Eddy, C. Lu, V. Uttamsingh,
R. Horlick, G. Harriman, D. Flynn, J. Med. Chem. 2006, 49, 2669–
2672; c) P. C. Jensen, R. Nygaard, S. Thiele, A. Elder, G. M. Zhu, R.
Kolbeck, S. Ghosh, T. W. Schwartz, M. M. Rosenkilde, Mol. Pharma-
col. 2007, 72, 327–340; d) J. Jin, Y. Wang, F. Wang, J. K. Kerns, V. M.
Vinader, A. P. Hancock, M. J. Lindon, G. I. Stevenson, D. M.
Morrow, P. Rao, C. Nguyen, V. J. Barrett, C. Browning, G. Hartmann,
D. P. Andrew, H. M. Sarau, J. J. Foley, A. J. Jurewicz, J. A. Fornwald,
A. J. Harker, M. L. Moore, R. A. Rivero, K. E. Belmonte, H. E.
Connor, Bioorg. Med. Chem. Lett. 2007, 17, 1722–1725; e) T. J. Jen-
kins, B. Guan, M. Dai, G. Li, T. E. Lightburn, S. Huang, B. S. Freeze,
D. F. Burdi, S. Jacutin-Porte, R. Bennett, W. Chen, C. Minor, S.
Ghosh, C. Blackburn, K. M. Gigstad, M. Jones, R. Kolbeck, W. Yin,
S. Smith, D. Cardillo, T. D. Ocain, G. C. Harriman, J. Med. Chem.
2007, 50, 566–584; f) I. Shamovsky, C. de Graaf, L. Alderin, M.
Bengtsson, H. k. Bladh, L. Bçrjesson, S. Connolly, H. J. Dyke, M. van
den Heuvel, H. Johansson, B.-G. Josefsson, A. Kristoffersson, T. Lin-
nanen, A. Lisius, R. Mꢅnnikkç, B. Nordꢀn, S. Price, L. Ripa, D.
Rognan, A. Rosendahl, M. Skrinjar, K. Urbahns, J. Med. Chem. 2009,
52, 7706–7723; g) A. K. Karlsson, K. Walles, H. Bladh, S. Connolly,
M. Skrinjar, A. Rosendahl, Biochem. Biophys. Res. Commun. 2011,
407, 764–771; h) P. Jensen, S. Thiele, A. Steen, A. Elder, R. Kolbeck,
S. Ghosh, T. Frimurer, M. Rosenkilde, Br. J. Pharmacol. 2012, 166,
258–275.
[4] a) H. Lange, C. F. Carter, M. D. Hopkin, A. Burke, J. G. Goode, I. R.
Baxendale, S. V. Ley, Chem. Sci. 2011, 2, 765–769; b) Z. Z. Qian,
I. R. Baxendale, S. V. Ley, Chem. Eur. J. 2010, 16, 12342–12348;
c) C. F. Carter, I. R. Baxendale, M. OꢄBrien, J. B. J. Pavey, S. V. Ley,
Org. Biomol. Chem. 2009, 7, 4594–4597; d) C. F. Carter, H. Lange,
[7] T. P. Petersen, A. Ritzen, T. Ulven, Org. Lett. 2009, 11, 5134–5137.
Received: December 6, 2012
Revised: April 2, 2013
Published online: && &&, 0000
&
8
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Library Synthesis Ligands by a Continuous-Flow System
FULL PAPER
Flow Synthesis
T. P. Petersen, S. Mirsharghi,
P. C. Rummel, S. Thiele,
M. M. Rosenkilde, A. Ritzꢀn,*
T. Ulven* ...................... &&&& — &&&&
Multistep flow synthesis: Efficient
assembly of chemokine receptor
(see scheme; BPR=back-pressure reg-
Multistep Continuous-Flow Synthesis
in Medicinal Chemistry: Discovery and
Preliminary Structure–Activity Rela-
tionships of CCR8 Ligands
ulator). The compounds represent a
new series of CCR8 ligands, with the
most potent member displaying full
agonist activity and single-digit nano-
molar potency.
ligands in a three-step continuous-flow
synthesis system is reported. The
system combines three diverse building
blocks in high yields and allows modi-
fications in any part of the scaffold
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!