Structure activity relationship studies on Amb639752: toward the identification of a common pharmacophoric structure for DGKα inhibitors

Article abstract of DOI:10.1080/14756366.2019.1684911

A series of analogues of Amb639752, a novel diacylglycerol kinase (DGK) inhibitor recently discovered by us via virtual screening, have been tested. The compounds were evaluated as DGK inhibitors on α, θ, and ζ isoforms, and as antagonists on serotonin re

Full text of DOI:10.1080/14756366.2019.1684911

Journal of Enzyme Inhibition and Medicinal Chemistry
ISSN: 1475-6366 (Print) 1475-6374 (Online) Journal homepage: https://www.tandfonline.com/loi/ienz20
Structure activity relationship studies on
Amb639752: toward the identification of a
common pharmacophoric structure for DGKα
inhibitors
Suresh Velnati, Alberto Massarotti, Annamaria Antona, Maria Talmon, Luigia
Grazia Fresu, Alessandra Silvia Galetto, Daniela Capello, Alessandra Bertoni,
Valentina Mercalli, Andrea Graziani, Gian Cesare Tron & Gianluca Baldanzi
To cite this article: Suresh Velnati, Alberto Massarotti, Annamaria Antona, Maria Talmon, Luigia
Grazia Fresu, Alessandra Silvia Galetto, Daniela Capello, Alessandra Bertoni, Valentina Mercalli,
Andrea Graziani, Gian Cesare Tron & Gianluca Baldanzi (2020) Structure activity relationship
studies on Amb639752: toward the identification of a common pharmacophoric structure for
DGKα inhibitors, Journal of Enzyme Inhibition and Medicinal Chemistry, 35:1, 96-108, DOI:
10.1080/14756366.2019.1684911
To link to this article: https://doi.org/10.1080/14756366.2019.1684911
© 2019 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group.
View supplementary material
Submit your article to this journal
View related articles
Published online: 05 Nov 2019.
Article views: 53
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=ienz20

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
2020, VOL. 35, NO. 1, 96–108
https://doi.org/10.1080/14756366.2019.1684911
RESEARCH PAPER
Structure activity relationship studies on Amb639752: toward the identification of
a common pharmacophoric structure for DGKa inhibitors
c
a
d
d
Suresh Velnatia,b , Alberto Massarotti , Annamaria Antona , Maria Talmon , Luigia Grazia Fresu ,
ꢀ
ꢀ
Alessandra Silvia Galettoa,e, Daniela Capelloa, Alessandra Bertonia, Valentina Mercallic, Andrea Grazianif,g,
Gian Cesare Tronc† and Gianluca Baldanzia,b
†
b
aDepartment of Translational Medicine, University of Piemonte Orientale, Novara, Italy; Institute for Research and Cure of Autoimmune
c
Diseases, CAAD, University of Piemonte Orientale, Novara, Italy; Department of Pharmaceutical Sciences, University of Piemonte Orientale,
d
e
Novara, Italy; Department of Health Sciences, School of Medicine, University of Piemonte Orientale, Novara, Italy; Palliative Care Division,
f
g
ꢀ
A.S.L., Vercelli, Italy; Universita Vita-Salute San Raffaele, Milan, Italy; Department of Molecular Biotechnology and Health Sciences, Molecular
Biotechnology Center, University of Torino, Turin, Italy
ABSTRACT
ARTICLE HISTORY
Received 8 August 2019
Revised 21 October 2019
Accepted 21 October 2019
A series of analogues of Amb639752, a novel diacylglycerol kinase (DGK) inhibitor recently discovered by
us via virtual screening, have been tested. The compounds were evaluated as DGK inhibitors on a, h, and
f isoforms, and as antagonists on serotonin receptors. From these assays emerged two novel compounds,
namely 11 and 20, which with an IC50 respectively of 1.6 and 1.8 mM are the most potent inhibitors of
DGKa discovered to date. Both compounds demonstrated the ability to restore apoptosis in a cellular
model of X-linked lymphoproliferative disease as well as the capacity to reduce the migration of cancer
cells, suggesting their potential utility in preventing metastasis. Finally, relying on experimental biological
data, molecular modelling studies allow us to set a three-point pharmacophore model for DGK inhibitors.
KEYWORDS
Diacylglycerol kinase; kinase
inhibitors; structure–activity
relationship; enzyme assays;
molecular modelling
1. Introduction
patient-derived lymphocytes resistant to reactivation-induced cell
death (RICD). Thus, antigen-activated lymphocytes accumulate in
lymphonodes and liver, resulting in severe immunopathology8.
Importantly, DGKa inhibitors restore RICD sensitivity in vitro and
in vivo, thus avoiding immunopathology and suggesting a puta-
tive therapeutic use of those molecules in XLP-19.
Diacylglycerol kinases (DGKs) are a large family of enzymes that
share a common catalytic activity: the phosphorylation of diacyl-
glycerol (DAG) to phosphatidic acid (PA). Remarkably, both the
substrate (DAG) and the product (PA) of the DGK-catalysed reac-
tion, are bioactive lipids that can act as second messengers1. DGK
activity consequently serves as a switch to simultaneously dampen
DAG-mediated signals and boost PA-mediated signals2. Ten mam-
malian DGK isoforms (a, b, c, d, e, f, g, h, i, and r) have been iden-
tified and divided into five groups (type I–V) according to their
structural features3,4. The expression of these isoforms varies
depending on the cell type. Among the 10 isoforms, the a isoform
is among the most studied and characterised. This kinase is highly
expressed in the brain, spleen, and thymus and, along with h iso-
form, in the bone marrow. This enzyme is also highly expressed in
T-lymphocytes, where it acts together with DGKf as negative
regulator of the T-cell receptor (TCR) response, and a mediator of
IL-2 mediated proliferation3,5. The biological relevance of DGKa is
best demonstrated in patients with X-linked lymphoproliferative
disease (XLP-1), who experience life-threatening, uncontrolled
accumulation of CD8þ T cells in response to the Epstein–Barr virus
(EBV) infection6. In those patients, germline mutations of the
adaptor protein SAP (SH2D1A) perturb TCR signalling and render
Apart from T-cell regulation, DGKa also plays a role in cancer,
mediating numerous aspects of cancer cell progression including
survival10,11, migration and invasion of cancer cells12–14. In particu-
lar, it has been reported that DGKa is over expressed in hepatocel-
lular carcinoma15, and melanoma cells11 while other reports
suggested that the growth of colon and breast cancer cell lines
was significantly inhibited by DGKa-siRNA16 and DGKa/atypical
PKC/b1 integrin signalling pathway was crucial for matrix invasion
of breast carcinoma cells17. In addition, expression is also higher
in lymphonodal metastasis than in breast and gastric original
tumour18,19. Finally, knock down of DGKa impairs glioblastoma
tumorigenesis20,21
.
For all these reasons, the identification of strong and selective
DGKa inhibitors, it is an important field of research. To date, only
a handful of two-digit micromolar inhibitors of DGKa have been
identified, but only three were the most characterised, namely,
R59949, R59022, and ritanserin (Figure 1).
In our assay system, R59949 and R59022 have an IC50 of 11
DGKa constitutively active7. Deregulated DGKa activity renders and 20 mM, respectively9. Their efficacy has been evaluated in vivo
ꢀ
CONTACT Andrea Graziani
andrea.graziani@hsr.it
Universita Vita-Salute San Raffaele, Milan 20132, Italy; Gian Cesare Tron
giancesare.tron@uniupo.it
Department of Pharmaceutical Science, University of Piemonte Orientale, Novara 28100, Italy
These authors contributed equally to this work.
†These authors shared senior authorship.
ꢀ
Supplemental data for this article can be accessed here.
ß 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
97
studies on mice, and is limited by their rapid clearance from Na/benzophenone under a slight positive atmosphere of dry
(t1/2¼ꢁ2 h)22. Furthermore, these two inhibitors are also able to nitrogen. Dichloromethane was dried by distillation from P2O5 and
target different isoforms of DGK, in particular R59022 acts on type stored on activated molecular sieves (4 Å). When needed the reac-
III and V (e e h), while R59949 on type I and II (c, d e j)23,24 and a tions were performed in flame- or oven-dried glassware under a
study conducted by Boroda et al. recently demonstrated their positive pressure of dry nitrogen. Melting points were determined
strong antagonistic activity on 5-HT2 receptors (R59022 IC50
in open glass capillary with a Stuart scientific SMP3 apparatus and
are uncorrected. All compounds were checked by IR (FT-IR
THERMO-NICOLET AVATAR), 1H and 13C APT (JEOL ECP 300 MHz
spectrometer), and mass spectrometry (Thermo Finningan LCQ-
deca XP-plus, San Jose, CA) equipped with an ESI source and an
ion trap detector. Chemical shifts are reported in parts per million
(ppm). Flash column chromatography was performed on silica gel
(Merck Kieselgel 60, 230–400 mesh ASTM, Kenilworth, NJ). Thin
layer chromatography (TLC) was carried out on 5 ꢂ 20 cm plates
5HT2A¼2.2 nM; R5994 IC50 5HT2A¼9.2 nM)25.
A
search on ChEMBL database26 (https://www.ebi.ac.uk/
chembl/) show how these two molecules have activity at the
same range of concentration with other biological targets, behav-
ing like a sort of promiscuous ligands. Ritanserin, a well-known
serotoninergic antagonist, is structurally similar to R59022, differ-
ing for an H-F isosteric substitution on a phenyl ring. Despite this
small modification, Boroda et al. showed that ritanserin was a
DGKa inhibitor (IC50¼15 mM) more potent than R59022 and
R59949 and with a better pharmacokinetic profile (t1/2¼40 h in
human)25. However, the comparison of ritanserin IC50 as serotonin
antagonist and as DGKa inhibitor, 0.9 nM and 15,000 nM, respect-
ively, reveal that ritanserin is much a powerful serotonin antagon-
ist than a DGK inhibitor. In addition, ritanserin is also a potent
inhibitor on dopaminergic receptors with an IC50 of 69 nM27.
Due to these drawbacks, at the beginning, in order to elimin-
ate the strong serotoninergic activity of R59949, we reasoned to
with a layer thickness of 0.25 mm (Merck Silica gel 60 F254
,
Kenilworth, NJ). When necessary they were developed with
KMnO4 reagent. Purity of tested compounds was established by
elemental analysis. Elemental analysis (C, H, N) of the target com-
pounds is within 0.4% of the calculated values, confirming
ꢃ95% purity.
2.1.1. Preparation of 2-chloro-1-(2,6-dimethyl-1H-indol-3-yl)ethan-
replace its protonable nitrogen atom, which at physiologically pH 1-one (5)
mimics the amino group of serotonin, with a carbon atom. We In a Schlenk tube, under nitrogen, 2,6-dimethyl-1H-indole (3)
decided therefore to synthesise compound 1 (Figure 2) (see sup- (0.20 g, 1.38 mmol, 1 eq) was dissolved in 4 mL of dichloroethane
porting information for its synthesis and a complete characterisa-
tion) and to test it as DGKa inhibitor.
dry and 0.25 mL of DBU (1.66 mmol, 1.2 eq) were added. The
resulting solution was heated at 90 ꢄC. When reached this tem-
perature, chloroacetyl chloride (4) (0.12 mL, 1.52 mmol, 1.1 eq) was
added. The reaction was stirred for 30 min, then solvent was
evaporated and the crude purified by column chromatography
using PE/EtOAc 7:3 and PE/EtOAc 5:5 as eluants to give 270 mg of
product as violet solid: yield 90%; m.p. 243.7–244.2 ꢄC; 1H NMR
(300 MHz, DMSO-d6) d 11.88 (br s, NH), 7.85 (d, J ¼ 7.9 Hz, 1H), 7.18
(s, 1H), 6.99 (d, J ¼ 8.0 Hz, 1H), 4.89 (s, 2H), 2.68 (s, 3H), 2.39 (s,
3H). MS (ESI) m/z: 222 [M þ H]þ.
Interestingly, the compound was totally devoid of inhibitory
activity on the enzyme, showing the importance of the basic
nitrogen atom not only for the anti-serotoninergic activity, but
also for the interaction with the kinase. With this in mind, we
recently used an in-silico approach based on chemical homology
with the two commercially available DGKa inhibitors R59022 and
R59949 using the programmes ROCS28 and EON29. From this
study, we identified a compound, Amb639752 (Figure 2), featuring
a lower IC50 for DGKa than ritanserin (IC50¼17 mM), a better select-
ivity for the a-isoform and devoid of anti-serotoninergic activity.
Along with CU-3, which features an IC50 of 0.6 mM on DGKa30 but
contains a reactive Michael acceptor31, Amb639752 is the most
effective pharmacological tool available to study DGKa9. In this
manuscript, we report the structure–activity studies on
Amb639752 and, in combination with data on ritanserin, the gen-
eration of a pharmacophore model for this class of compounds,
which could be useful for the identification of other potential
DGKa inhibitors.
2.1.2. Preparation of tert-butyl 4-(2-(2,6-dimethyl-1H-indol-3-yl)-2-
oxoethyl)piperazine-1-carboxylate (6)
Under nitrogen, 200 mg of 5 (0.90 mmol, 1 eq) was dissolved in
toluene dry, then N-Boc-piperazine (0.17 g, 0.90 mmol, 1 eq),
K2CO3 (0.32 g, 2.25 mmol, 2.5 eq), and KI (0.015 g, 0.09 mmol,
2. Methods
2.1. Chemistry procedures
Commercially available reagents and solvents were used without
further purification. Toluene were distilled immediately before use
Figure 2. Structures of the deaza analogue of R59949 and Amb639752.
Figure 1. Three of the most studied DGKa inhibitors.

98
S. VELNATI ET AL.
0.1 eq) were added. The reaction was heated at 80 ꢄC overnight. 131.5, 129.5, 129.1, 125.1, 123.5, 121.2, 112.7, 111.6, 53.14, 53.14,
Solvent was evaporated and the crude was purified by column
chromatography using PE/EtOAc 4:6 and PE/EtOAc 2:8 as eluants
to give 215 mg of product as yellow amorphous solid: yield 63%;
1H NMR (300 MHz, CDCl3) d 8.52 (br s, NH), 7.77 (d, J ¼ 8.2 Hz, 1H),
7.12 (s, 1H), 7.08 (d, J ¼ 8.0 Hz, 1H), 3.79 (s, 2H), 3.47–3.43 (m, 6H),
2.75 (s, 3H), 2.64 (br s, 2H), 2.44 (s, 3H), 1.47 (s, 9H). IR (KBr): 3190,
2964, 1698, 1413, 1417, 1364, 1126, 806 ꢀmax/cmꢅ1. MS (ESI) m/z:
372 [M þ H]þ.
52.9, 21.7, 15.7. IR (KBr): 3225, 2793, 2358, 1609, 1442, 1261, 864
ꢀmax/cmꢅ1. MS (ESI) m/z: 410 [M þ H]þ. Anal. Calcd. for C23H24
ClN3O2: C, 67.39; H, 5.90; N, 10.25; found C, 67.11; H, 6.12;
N, 10.54.
2.1.7. 1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-(4-methoxybenzoyl)piper-
azin-1-yl)ethan-1-one (12)
Yellow solid; yield 53%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 219.9–220.8 ꢄC.; 1H NMR (300 MHz, DMSO-d6) d 11.70 (br s,
NH), 7.85 (d, J ¼ 8.2 Hz, 1H), 7.38 (br d, AA’XX’, 2H), 7.14 (s, 1H),
6.99–6.94 (m, 3H), 3.79 (br d, 3H), 3.67 (br s, 2H), 3.51 (br s, 4H)
2.67 (s, 3H), 2.58 (br s, 4H), 2.38 (s, 3H); 13C NMR (75 MHz, DMSO-
d6) d 192.8, 169.5, 144.6, 136.5, 135.7, 131.5, 130.0, 129.0, 127.5,
125.1, 123.4, 121.2, 112.8, 111.6, 66.9, 55.8, 53.2, 53.1, 21.7, 15.7; IR
(KBr): 3235, 3003, 2807, 1613, 1463, 1253, 977 ꢀmax/cmꢅ1; MS (ESI)
m/z: 406 [M þ H]þ. Anal. Calcd. for C24H27N3O3: C, 71.09; H, 6.71;
N, 10.36; found C, 71.10; H, 6.75; N, 10.32.
2.1.3. Preparation of 1-(2,6-dimethyl-1H-indol-3-yl)-2-(piperazin-1-
yl)ethan-1-one (7)
Two hundred and fifteen milligrams of 6 (0.58 mmol, 1 eq) were
dissolved in dichloromethane dry. The resulting solution was
cooled at 0 ꢄC and 0.69 mL of trifluoroacetic acid (9.28 mmol,
16 eq) was added. After 3 h, the reaction was worked up adding
NaOH 2 M solution until pH ¼ 12. Then, NaCl solid was added
and the solution was extracted with THF (x2). The combined
organic phases were dried on sodium sulphate. After evaporation
of the solvent, the crude was purified by column chromatography
using EtOAc/MeOH 9:1 and MeCN/NH3 9:1 as eluants to give
103 mg of the product as yellowish solid: yield: 65%; m.p.:
232.9–233.6 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.79 (br s, NH),
7.85 (d, J ¼ 7.9 Hz, 1H), 7.15 (s, 1H), 6.95 (d, J ¼ 8.2 Hz, 1H), 3.83(br
s, 4H), 3.62 (s, 2H), 2.86 (br s, 4H), 2.67 (s, 3H), 2.38 (s, 3H); IR (KBr):
3446, 3181, 2813, 1634, 1455, 1330, 821 ꢀmax/cmꢅ1; MS (ESI) m/z:
272 [M þ H]þ.
2.1.8. 4-(4-(2-(2,6-Dimethyl-1H-indol-3-yl)-2-oxoethyl)piperazine-1-
carbonyl)benzonitrile (13)
Yellow solid; yield 23%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 243.9–244.8 ꢄC. 1H NMR (300 MHz, DMSO-d6) d 11.70 (br s,
NH), 7.92–7.84 (m, 3H), 7.58 (br d, AA’XX’, 2H), 7.15 (s, 1H), 6.96 (br
d, 1H), 3.68 (br s, 4H), 3.29 (br s, 2H), 2.67 (s, 3H), 2.51 (br s, 4H),
2.38 (s, 3H); 13C NMR (75 MHz, DMSO-d6) d 192.5, 166.3, 143.2,
139.4, 134.1, 131.6, 129.9, 126.8, 123.5, 121.9, 119.6, 117.4, 111.2,
111.1, 110.1, 65.1, 51.7, 51.2, 20.1, 14.1; IR (KBr): 3410, 3254, 2816,
2790, 2233, 1609, 1454, 1291, 979 ꢀmax/cmꢅ1; MS (ESI) m/z:
401[M þ H]þ. Anal. Calcd. for C24H24N4O2: C, 71.98; H, 6.04; N,
13.99; found C, 72.13; H, 6.23; N, 14.08.
2.1.4. General procedure for the synthesis of final compounds
8, 11–22
1-(2,6-Dimethyl-1H-indol-3-yl)-2-(piperazin-1-yl)ethan-1-one
(7)
(1 eq) was dissolved in dichloromethane dry. To the resulting solu-
tion EDCI (1 eq), TEA (2 eq), DMAP (0.1 eq) and the appropriate
carboxylic acid (1 eq) were sequentially added. The reaction was
stirred under nitrogen at room temperature overnight.
Evaporation of the solvent gave a crude which was directly puri-
fied by column chromatography.
2.1.9. 1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-(thiophene-2-carbonyl)pi-
perazin-1-yl)ethan-1-one (14)
Yellow solid; yield 29%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 200.3–201.2 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.74 (br s,
NH), 7.86 (d, J ¼ 8.2 Hz, 1H), 7.76 (br d, 1H), 7.41–7.40 (m, 1H),
7.15–7.10 (m, 2H), 6.96 (d, J ¼ 8.2 Hz, 1H), 3.75–3.67 (m, 4H), 3.37
(br s, 2H) 2.68 (s, 3H), 2.62 (s, 4H), 2.38 (s, 3H); 13C NMR (75 MHz,
DMSO-d6) d 192.8, 167.8, 144.7, 137.8, 135.7, 131.5, 130.0, 129.6,
127.7, 125.1, 124.1, 123.4, 121.2, 112.8, 111.6, 66.7, 53.3, 21.7, 15.7;
IR (KBr): 3270, 2927, 2793, 1642, 1454, 1261, 809 ꢀmax/cmꢅ1; MS
(ESI) m/z: 382[M þ H]þ; Anal. Calcd. for C21H23N3O2S: C, 66.12; H,
6.08; N, 11.01; found C, 66.23; H, 6.26; N, 10.93.
2.1.5. 2-(4-Benzoylpiperazin-1-yl)-1-(2,6-dimethyl-1H-indol-3-yl)e-
than-1-one (8)
Yellow solid; yield 29%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 213.8–214.3 ꢄC. 1H NMR (300 MHz, DMSO-d6) d 11.72 (br s,
NH), 7.85 (d, J ¼ 8.2 Hz, 1H), 7.46–7.40 (m, 5H), 7.15 (s, 1H), 6.96 (d,
J ¼ 7.9 Hz, 1H), 3.67 (br s, 4H), 3.18 (br s, 2H) 2.67 (s, 3H), 2.61–2.56
(m, 4H), 2.38 (s, 3H); 13C NMR (75 MHz, DMSO-d6) d 192.8, 169.5,
144.7, 136.5, 135.7, 131.5, 130.0, 129.0, 127.5, 125.1, 123.4, 121.2,
112.9, 111.6, 66.9, 53.2, 53.1, 21.7, 15.7; IR (KBr) 3189, 2990, 2828,
1609, 1446, 1282, 807 ꢀmax/cmꢅ1. MS (ESI) m/z: 374 [M–H]þ. Anal.
Calcd. for C23H25N3O2: C, 73.57; H, 6.71; N, 11.19; found C, 73.76;
H, 6.94; N, 10.85.
2.1.10.
1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-nicotinoylpiperazin-1-
yl)ethan-1-one (15)
Yellow solid; yield 39%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 216.2–216.8 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.70 (br s,
NH), 8.65–8.61 (m, 2H), 7.87–7.82 (m, 2H), 7.50–7.45 (m, 1H), 7.15
(s, 1H), 6.95 (br d, 1H), 3.68 (br s, 4H), 3.37 (br s, 2H), 2.67 (s, 3H),
2.56 (br s, 4H), 2.37 (s, 3H); 13C NMR (75 MHz, DMSO-d6) d 192.6,
167.2, 151.0, 148.1, 144.7, 135.7, 135.4, 132.3, 131.5, 125.1, 124.1,
123.5, 121.2, 116.2, 111.6, 66.7, 53.4, 52.9, 21.7, 15.7; IR (KBr): 3414,
3213, 2828, 1621, 1454, 1267, 1301, 817 ꢀmax/cmꢅ1; MS (ESI) m/z:
377 [M þ H]þ; Anal. Calcd. for C22H24N4O2: C, 70.19; H, 6.43; N,
2.1.6.
2-(4-(4-Chlorobenzoyl)piperazin-1-yl)-1-(2,6-dimethyl-1H-
indol-3-yl)ethan-1-one (11)
Yellow solid; yield 33%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 240.7–241.3 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.70 (br s,
NH), 7.85 (d, J ¼ 7.9 Hz, 1H), 7.50 (br d, AA’XX’, 2H), 7.42 (br d,
AA’XX’, 2H), 7.14 (s, 1H), 6.95 (d, J ¼ 8.2 Hz, 1H), 3.67 (br s, 4H),
3.33 (br s), 2.67 (s, 3H), 2.61–2.56 (m, 4H), 2.38 (s, 3H); 13C NMR
(75 MHz, DMSO-d6) d 193.17, 168.4, 144.8, 135.7, 135.2, 134.8, 14.88; found C, 70.21; H, 6.44; N, 14.73.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
99
2.1.11. 1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-(2-methylbenzoyl)pipera- m/z: 390 [M þ H]þ; Anal. Calcd. for C24H27N3O2: C, 74.01; H, 6.99;
N, 10.79; found C, 74.12; H, 7.02; N, 10.79.
zin-1-yl)ethan-1-one (16)
Yellow solid; yield 41%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 200.5–201.6 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.70 (br s,
NH), 7.85 (d, J ¼ 8.2 Hz, 1H), 7.31–7.27 (m, 3H), 7.26–7.24 (m, 2H),
7.15 (d, 1H), 3.69 (br s, 2H), 3.34 (br s, 2H), 3.15 (br s, 2H), 2.67 (s,
3H), 2.51 (br s, 4H), 2.44 (s, 3H), 2.38 (s, 3H); 13C NMR (75 MHz,
DMSO-d6) d 192.6, 169.0, 144.7, 135.7, 134.2, 131.8, 131.5, 130.7,
129.2, 126.4, 126.2, 123.4, 121.2, 112.8, 66.8, 53.5, 53.0, 46.8, 41.4,
21.7, 19.2, 15.7; IR (KBr): 3431, 3221, 2919, 2797, 1615, 1454, 1257,
748 ꢀmax/cmꢅ1; MS (ESI) m/z: 390 [M þ H]þ. Anal. Calcd. for
C24H27N3O2: C, 74.01; H, 6.99; N, 10.79; found C, 74.01; H, 7.01;
N, 10.63.
2.1.16. 2-(4-(Cyclopentanecarbonyl)piperazin-1-yl)-1-(2,6-dimethyl-
1H-indol-3-yl)ethan-1-one (21)
Brown oil; yield 61%; column eluants: EtOAc, EtOAc/MeOH 9:1; H
NMR (300 MHz, CDCl3) d 9.00 (br s, NH), 7.73 (d, J ¼ 8.2 Hz, 1H),
7.26 (br d, 1H), 7.12 (s, 1H), 7.05 (d, J ¼ 8.2 Hz, 1H), 3.83 (s, 2H),
3.75 (br s, 2H), 3.63 (br s, 2H), 2.80–2.71 (m, 7H), 2.45 (s, 3H),
1.92–1.46 (m, 7H); IR (KBr): 3253, 2944, 2862, 1650, 1620, 1455,
1234, 957 ꢀmax/cmꢅ1; MS (ESI) m/z: 368 [M þ H]þ; Anal. Calcd. for
C22H29N3O2: C, 71.90; H, 7.95; N, 11.43; found C, 72.23; H, 8.31;
N, 11.32.
1
2.1.12. 1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-(3-methoxybenzoyl)pi-
perazin-1-yl)ethan-1-one (17)
2.1.17. 1-(4-(2-(2,6-Dimethyl-1H-indol-3-yl)-2-oxoethyl)piperazin-1-
yl)heptan-1-one (22)
Brown oil; yield 76%; column eluants: EtOAc, EtOAc/MeOH 9:1; H
Yellow solid; yield 32%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 209.8–210.4 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.71 (br s,
NH), 7.85 (d, J ¼ 7.9 Hz, 1H), 7.34 (t, 1H), 7.15 (br s, 1H), 7.02–6.92
(m, 4H), 3.78 (s, 3H), 3.68 (br s, 3H), 3.35 (br s, 3H), 2.67 (br s, 7H),
2.38 (s, 3H); 13C NMR (75 MHz, DMSO-d6) d 192.7, 169.1, 159.7,
144.7, 137.9, 135.7, 131.5, 130.2, 125.1, 123.4, 121.2, 119.4, 115.7,
112.7, 111.6, 66.8, 55.8, 53.3, 53.11, 47.4, 21.7, 15.7; IR (KBr): 3131,
3049, 2944, 1651, 1455, 1292, 1130, 968 ꢀmax/cmꢅ1; MS (ESI) m/z:
406 [M þ H]þ. Anal. Calcd. for C24H27N3O3: C, 71.09; H, 6.71; N,
10.36; found C, 70.85; H, 6.45; N, 10.76.
1
NMR (300 MHz, CDCl3) d 10.25 (br s, NH), 7.68 (d, J ¼ 8.2 Hz, 1H),
7.08 (s, 1H), 7.00 (d, J ¼ 7.9 Hz, 1H), 3.83 (s, 2H), 3.73 (br s, 2H),
3.56 (br s, 2H), 2.77–2.71 (m, 3H), 2.62 (s, 3H), 2.38 (s, 3H),
2.32–2.27 (m, 5H), 1.59–1.54 (m, 2H), 1.26 (br s, 4H), 0.85 (br t, 3H);
13C NMR (75 MHz, CDCl3) d 191.7, 172.4, 145.4, 135.5, 132.1, 124.0,
123.6, 120.4, 112.4, 111.6, 66.2, 53.3, 53.1, 45.3, 41.2, 34.5, 31.6,
28.9, 25.0, 22.5, 21.5, 15.6, 14.1; IR (KBr): 2927, 2857, 1731, 1645,
1455, 1434, 1234, 668 ꢀmax/cmꢅ1; MS (ESI) m/z: 384 [M þ H]þ;
Anal. Calcd. for C23H33N3O2: C, 72.03; H, 8.67; N, 10.96; found C,
72.03; H, 8.73; N, 11.21.
2.1.13. 2-(4-(3,4-Difluorobenzoyl)piperazin-1-yl)-1-(2,6-dimethyl-1H-
indol-3-yl)ethan-1-one (18)
2.2. Cell lines
Yellow solid; yield 27%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 233.8–235.0 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.71 (br s,
NH), 7.85 (d, J ¼ 7.9 Hz, 1H), 7.56–7.47 (m, 2H), 7.28 (br s, 1H), 7.15
(s, 1H), 6.96 (d, J ¼ 7.9 Hz, 1H), 3.68 (br s, 2H), 3.33 (br s, 2H), 2.67
(s, 3H), 2.51 (br s, 6H), 2.38 (s, 3H); MS (ESI) m/z: 412 [M þ H]þ; IR
Madin-Darby canine kidney (MDCK) cells stably expressing One
Strep Tag DGKa (OST-DGKa) were prepared by infecting MDCK
cells with a vector expressing an inducible OST tagged DGKa con-
structs17. MDCK cells infected with empty vector were used as
controls. MDCK cells were cultured in MEM (minimal essential
medium) with 5% FBS (foetal bovine serum) and 1% antibiotic–an-
timycotic solution. Routinely, cells were splitted every 3–4 days
with trypsin–EDTA 0.25% in standard 100 mm dishes.
(KBr): 3252, 2919, 2795, 1618, 1469, 1286, 1046, 980 ꢀmax/cmꢅ1
;
MS (ESI) 412 [M þ H]þ. Anal. Calcd. for C23H23F2N3O2: C, 67.14; H,
5.63; N, 10.21; found C, 67.43; H, 5.79; N, 10.59.
Human embryonic kidney 293T cells (10 cm2 plates) were cul-
tured in RPMI with 10% FBS and 1% penicillin/streptomycin and
cultures were maintained by splitting them for every 2–3 days
using trypsin–EDTA 0.25%.
2.1.14.
2-(4-(3-Chlorobenzoyl)piperazin-1-yl)-1-(2,6-dimethyl-1H-
indol-3-yl)ethan-1-one (19)
Yellow solid; yield 32%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 222.3–223.5 ꢄC; 1H NMR (300 MHz, CDCl3) d 9.42 (br s, NH),
7.71 (d, J ¼ 7.9 Hz, 1H), 7.40–7.25 (m, 4H), 7.04 (d, J ¼ 8.2 Hz, 2H),
3.85 (br s, 4H), 3.47 (br s, 2H), 2.80 (br s, 1H), 2.68 (br s, 6H), 2.40
(s, 3H); 13C NMR (75 MHz CDCl3,) d 192.5, 169.0, 144.7, 137.5,
135.3, 134.7, 132.3, 130.0, 127.3, 125.2, 124.0, 123.7, 120.6, 112.8,
111.4, 67.0, 53.7, 53.7, 29.6, 21.5, 15.7; IR (KBr): 3264, 2916, 2795,
1646, 1454, 1256, 978, 809 ꢀmax/cm–1; MS (ESI) m/z: 410 [M þ H]þ.
Anal. Calcd. for C23H24ClN3O2: C, 67.39; H, 5.90; N, 10.25; found C,
67.38; H, 5.90; N, 10.24.
Michigan Cancer Foundation 7 (MCF7) cells were cultured in
DMEM with 10% FBS þ 1% penicillin/streptomycin and cultures
were maintained by splitting them for every 2–3 days using
trypsin–EDTA 0.25%.
2.3. Primary cells
PBL were isolated from healthy anonymous human donors by
Ficoll-Paque PLUS (GE Healthcare, Chicago, IL) density gradient
centrifugation, washed, and resuspended at 2 ꢂ 106 cell/mL in
RPMI-GlutaMAX containing 10% heat inactivated FCS, 2 mM glu-
tamine, and 100 U/mL of penicillin and streptomycin. T cells were
activated with 1 mg/mL anti-CD3 (UCHT1) and anti-CD28 (clone
CD28.2) antibodies. After three days, activated T cells were
washed and cultured in medium additionated of 100 IU/mL rhIL-2
(Peprotech, Rocky Hill, NJ) at 1.2 ꢂ 106 cells/mL for ꢃ7 days by
changing media for every 2–3 days.
2.1.15. 1-(2,6-Dimethyl-1H-indol-3-yl)-2-(4-(4-methylbenzoyl)pipera-
zin-1-yl)ethan-1-one (20)
Yellow solid; yield 48%; column eluants: EtOAc, EtOAc/MeOH 9:1;
m.p.: 230.9–231.2 ꢄC; 1H NMR (300 MHz, DMSO-d6) d 11.71 (br s,
NH), 7.86–7.82 (m, 1H), 7.31–7.23 (m, 4H), 7.15 (s, 1H), 6.96 (d,
J ¼ 7.9 Hz, 1H), 3.69 (br s, 2H), 3.35 (br s, 4H), 2.67 (s, 3H), 2.51. IR
Human monocytes were isolated from healthy anonymous
(KBr): 3228, 2915, 2792, 1607, 1454, 1260, 979 ꢀmax/cmꢅ1; MS (ESI) human buffy coats (provided by the Transfusion Service of

100
S. VELNATI ET AL.
ꢀ
Ospedale Maggiore della Carita, Novara, Italy) by the standard
technique of dextran sedimentation and Histopaque (density ¼
1.077 g cm3, Sigma-Aldrich, Milano, Italy) gradient centrifugation
(400ꢂg, 30 min, room temperature) and recovered by fine suction
at the interface, as described previously32. Purified monocytes
populations were obtained by adhesion (90 min, 37 ꢄC, 5% CO2) in
serum-free RPMI 1640 medium (Sigma-Aldrich, Milano, Italy) sup-
Percentage residual activity was calculated as follows: (OST-
DGKa homogenate with inhibitor – vector homogenate)/(OST-
DGKa homogenate with DMSO – vector homogenate)ꢂ100.
2.7. Superoxide anion (O22) production
All the experiments were performed in triplicate using cells iso-
plemented with 2 mM glutamine and antibiotics. Cell viability (try- lated from each single donor.
Monocytes (250,000 cells/well) were treated for 1 h with the
pan blue dye exclusion) was usually >98%.
indicated drugs (10 mM) with or without serotonin (1 mM). Then,
cells were stimulated with phorbol 12-myristate 13-acetate (PMA;
Sigma-Aldrich, Milano, Italy) 1 mM for 30 min. PMA is a well-known
stimulus that induces a strong and significant respiratory burst via
PKC activation33. Superoxide anion production was then evaluated
by the superoxide dismutase (SOD)-sensitive cytochrome C (CytC)
reduction assay and expressed as nmoles CytC reduced/106 cells/
30 min, using an extinction coefficient of 21.1 mM. To avoid inter-
ference with spectrophotometrical recordings, cells were incu-
bated with RPMI 1640 without phenol red, antibiotics, and FBS.
2.4. Preparation of DGKa enriched homogenates
Large cultures of MDCK cells for enzyme preparation were done
by plating 5 ꢂ 106 cells in 245 mm2 dishes. Once they reached
nearly 70% confluence, cells were treated with doxycycline (1 mg/
mL, two days). After two days of treatment, each plate was
washed in cold PBS and cells homogenised in 5 mL of homogen-
ate buffer (25 mM Hepes (pH 8), 20% glycerol, 135 mM NaCl, 5 mM
ethylenediaminetetraacetic acid (EDTA), 1 mM ethylene glycol-bis(-
beta-aminoethyl ether)-N,N,N0,N0-tetraacetic acid (EGTA), 1 mM
sodium orthovanadate, and protease inhibitor cocktail) for each
dish. Cells were collected with a rubber scraper, homogenised by
passing them through a 29 G-needle syringe 20 times and stored
in aliquots at –80 ꢄC. Presence of OST-DGKa was confirmed by
western blotting and enzyme assay, transduced DGKa has an
activity >100 folds the endogenous DGK.
2.8. RICD assay in SAP silenced T cells
Activated human PBLs were transfected with 200pmol of siRNA oli-
gonucleotides specific for the target protein (Stealth Select siRNA;
Life Technologies, Carlsbad, CA) or a non-specific control oligo (Life
Technologies, Carlsbad, CA). Transient transfections were performed
using Amaxa nucleofector kits for human T cells (Lonza, Basel,
Switzerland) and the Amaxa Nucleofector II or 4D systems (pro-
grammes T-20 or EI-115). Cells were cultured in IL-2 (100 IU/mL) for
four days to allow target gene knockdown. Knockdown efficiency
was periodically evaluated by Western blotting.
Non-specific Stealth RNAi Negative Control Duplexes (12935-300,
Life Technologies, Carlsbad, CA) were used as a negative control.
siRNA SAP: sense strand UGUACUGCCUAUGUGUGCUGUAUCA,
antisense strand UGAUACAGCAGACAUAGGCAGUACA.
2.5. Preparation of DGKf and DGKh enriched homogenates
293T cells were transiently transfected with indicated DGK isoform
plasmid DNA using Lipofectamine 3000, Invitrogen (Carlsbad, CA).
Forty eight hours after transfection, cells were harvested and
homogenised with a 29 G-needle using 500 mL of homogenate
buffer for each dish and immediately stored in aliquots at –80 ꢄC.
Cells transfected with empty vector were used as controls, overex-
pressed DGK has an activity >50 folds the endogenous.
To test restimulation induced cell death, activated T cells (105
cells/well) were plated in triplicate in 96-well round-bottom plate and
treated with anti-CD3 (clone OKT3) (10 ng/mL) in RPMI-GlutaMAX
supplemented with 100 IU/mL rhIL-2 for 24 h. In these assays, inhibi-
tors (10 mM) were added 30 min before the restimulation with OKT3.
24 h after treatment, cells were stained with 20 ng/mL propidium iod-
ide and collected for a constant time of 30 s per sample on Attune
Nxt Flow Cytometer (Thermo Fisher Scientific, Waltham, MA). Cell
death is expressed as % cell loss and calculated as:
2.6. DGK assay
Essentionally, the same procedure was followed as reported previ-
ously in Velnati et al.9 In brief, DGK activity was assayed by meas-
uring initial velocities (5 min at 27 ꢄC) in presence of saturating
substrate concentrations. Reaction conditions: 0.9 mg/mL 1,2-dio-
leoyl-sn-glycerol, 5 mM ATP, 0.01 mCi/mL [c32P]-ATP, 1 mM sodium
orthovanadate, 10 mM MgCl2, 1.2 mM EGTA in 7.5 mM Hepes pH
812. Reaction mixture is assembled mixing enzyme (24.5 mL of
homogenate), 100ꢂ inhibitor or DMSO (0.5 mL), 5ꢂ ATP solution
(10 mL of 25 mM ATP, 0.05 mCi/mL [c32P]-ATP (Perkin-Elmer, Milan,
Italy), 5 mM sodium orthovanadate, 50 mM MgCl2), and 3.3ꢂ DAG
solution (15 mL of 3 mg/mL 1,2-dioleoyl-sn-glycerol resuspended by
sonication in 4 mM EGTA in 25 mM Hepes pH 8). The reaction was
stopped after 5 min by adding 200 mL of freshly prepared 1 M HCl
and lipid was extracted by adding 200 mL of CH3OH:CHCl3 1:1
solution and vortexing for 1 min. The two phases were separated
by centrifugation (12,000 RCF for 2 min). Twenty-five microlitres of
the lower organic phase was spotted in small drops on silica TLC
plates. TLC was run 10 cm and dried before radioactive signals
were detected by GS-250 molecular imager and was quantified by
quantity one (Bio-Rad, Hercules, CA) software assuring the absence
of saturated spots.
ꢀ
ꢀ
ꢁ
ꢁ
number of viable cells in sample
number of viable cells in control
% cell loss ¼ 1 ꢅ
ꢂ 100
Results were expressed as mean standard error of the mean
(SEM). We always compared controls and SAP silenced lympho-
cytes from the same donors as there is a large individual variabil-
ity in RICD sensitivity.
2.9. Migration assays
Cell migration assays were performed using the Culture-Insert 2
well in m-Dish (ibidi GmbH, Martinsried, Germany).
Briefly, 25,000 MCF7 cells were plated in each well and cultured
for 24 h. The day after, the culture insert was removed and the cells
were washed with PBS before treating them with respective DGKa
inhibitors (10mM) or DMSO for 15 h in complete medium (DMEM
10% FBS þ 1% penicillin/streptomycin), while medium without FBS
was used as a negative control for migration.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
101
Phase-contrast pictures were taken immediately after treatment analysed by using one-way ANOVA with multiple comparisons cor-
(0 h) and after 15 h under 5ꢂ magnification.
rection using GraphPad PRISM 8.0 software (GraphPad Software,
La Jolla, CA). Error bars are described in figure legends as SEM
or SD where appropriate. A single, double, triple and four aster-
isk denotes significance of a p value ꢆ0.05, ꢆ0.01, ꢆ0.001, and
ꢆ0.0001 respectively in all experiments.
Finally, wound areas were determined using ImageJ software
(NIH, , Bethesda, MD). Wound reduction was calculated by using
the following formula: (wound area at 15 h/wound at 0 h)ꢂ100,
the values obtained were expressed as the percentage of wound
area compared to the initial area.
2.11. Pharmacophoric model
2.10. Quantification and statistical analysis
A representative 3D structure of each compound was generated
using OMEGA2 software34–36. The generated file was used to gen-
erate a pharmacophore model with the Pharmagist web server
(bioinfo3d.cs.tau.ac.il/PharmaGist/)37.
Data for the screen on OST-DGKa homogenates are the mean of
duplicates. The compounds showing inhibitory activity in this
assay were tested >4 times and the mean SEM is reported.
To calculate IC50 values of active inhibitors, the inhibitor activ-
ity was measured at least three times at 0.1, 1.0, 10.0, and
100.0 mM concentration. Data were analysed using [inhibitor] vs.
normalised response parameters with least square [ordinary] curve
fitting method in GraphPad PRISM 8.0 software (GraphPad
Software, La Jolla, CA) mentioning 95% confidence interval and
IC50 values always greater than 0.0. Graph shows the mean SEM
of inhibitor activity at the indicated concentration. In all the
experiments, the data were normalised with the controls.
3. Results
3.1. Chemistry
At the beginning, we purchased 14 analogues of Amb639752 by
vendors (Figure 3), while one analogue (2), being not commer-
cially available, was synthesised (see Supplementary material). All
Evaluation of in vitro assays across multiple treatments (RICD), the compounds were evaluated for their inhibitory activity on
SOD-sensitive CytC reduction assay, migration assays were DGKa at a concentration of 100 mM (Table 1).
Figure 3. First set of compounds tested for their inhibitory activity on DGKa.

102
S. VELNATI ET AL.
Each inhibitor was tested in duplicate at least once, and DGKa
In the first one, the commercially available 2,6-dimethyl-1H-
indole 3 was acylated with 2-chloroacetyl chloride 4, in the pres-
ence of DBU in dichloroethane39 to give the derivative 5 90%
yield. Then, the acylated compound 5 was reacted with N-Boc
piperazine in the presence of potassium carbonate and potassium
iodide to afford the piperazinic derivative 6 in 63% yield. Boc
deprotection with trifluoroacetic acid, followed by coupling with
benzoic acid using the condensing agent EDCI afforded the final
compound 8 (Scheme 1).
activity was expressed as percentage of residual DGKa activity
compared to DMSO control in the same assay. Assay uses OST-
DGKa overexpressing cell lysates in presence of saturating exogen-
ous DAG and ATP. We considered R59022 (commercially available)
and the lead compound Amb63975230 as our reference mole-
cules. As expected, our reference inhibitors R59022 and
Amb639752 featured 73% and 96% inhibition respectively, con-
firming the quality of data obtained.
In the second synthetic strategy, we initially coupled the ben-
zoic acid with N-Boc piperazine in the presence of EDCI to give
piperazinic derivative 9 in 49% yield. Boc deprotection gave in
quantitative yield the compound 10. Due to its high aqueous
solubility, solvent was evaporated and the crude as trifluoroace-
tate salt was directly used for the next step, where it was reacted
with the acylated indole 5 to give the final compound 8 in 25%
yield (Scheme 2) (see Supplementary material for full syn-
thetic details).
This first screening showed us how Amb639752 exhibits a rigid
structure activity relationship. Indeed, both the methyl groups on
the 2,6 position of indole are mandatory, the NH indole cannot be
alkylated as well as ramifications on the alkyl chain are detrimen-
tal. We then focussed our attention on furan ring knowing its
intrinsic toxicity via metabolic activation38. Unfortunately, there
were no analogues available by vendors. Our first goal was to
replace the furan moiety with the phenyl ring, investigating two
different synthetic pathways.
Overall yield calculation was 11% for both strategies, but with
the first route it was possible to use a common synthetic inter-
mediate 7 which can be coupled with different carboxylic acids.
Furthermore, the second route requires more purification steps.
For this reason, we applied the first route and coupled the
advanced intermediate 7 with 12 different carboxylic acids (Figure
4) to afford compounds 11–22 (Figure 5).
All those molecules were dissolved in DMSO and tested at a
concentration of 100 mM for the ability to inhibit DGKa using
equal amounts of DMSO as control. We identified eight com-
pounds capable of reducing OST-DGKa activity similar or superior
to R59022 (Table 2).
Table 1. Inhibitory activity on DGKa (I).
Compound
Residue activity at 100 mM
R59022
R59949
27
28
Amb639752
Amb758897
Amb37496
Amb626577
Amb1926062
Amb94966
Amb730692
Amb98433
Amb18718
Amb726054
Amb731111
Amb22018852
Amb687392
Amb758592
Amb629700
2
4
67
114
127
88
196
45
81
135
126
130
160
115
150
101
135
3.2. Potency and isoform specificity of active molecules
To measure the inhibitor potency, we determined the IC50 values
for the compounds that resulted active when tested at 100 mM by
measuring the residual OST-DGKa activity over a dose range of
inhibitor concentrations (0.1 mM, 1.0 mM, 10.0 mM, and 100.0 mM).
Scheme 1. The first synthetic route for the compound 8.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
103
Scheme 2. The second synthetic route for the compound 8.
Figure 4. Carboxylic acids used.
For R59022 and Amb639752, we measured IC50 values of
To this purpose, we measured the effect of serotonin on
15.2 5.8 mM and 6.9 3.0 mM respectively which were comparable PMA-induced oxidative burst in human monocytes. As previously
to previous reports using similar assay conditions9. Considering shown, 1 mM serotonin reverses the oxidative burst to control
those two as reference/template compounds, we measured the values33. Known serotonin receptor antagonists ritanserin and
IC50 values of 8, 11, 12, 13, 14, 16, 19, and 20 as 3.2 1.0, ketanserin (10 mM) impaired serotonin action, while pure DGKa
1.6 0.4, 3.6 1.2, 6.9 2.3, 3.0 1.0, 32.8 11.5, 49.7 31.7, and inhibitors such as Amb639752 had no effect (Figure 8). These
1.8 0.4 mM, respectively, signifying that their activity is equal or data indicate that this assay is sensitive to perturbations in sero-
superior to the template compounds (Figure 6).
tonin signalling independently of activity against DGKa.
Interestingly, as like Amb639752, all the newly synthesised
active molecules did not affect serotonin action (Figure 8), indi-
cating that all those molecules are not serotonin receptor
antagonists.
Due to their higher IC50 values, we thus decided to exclude 16
and 19 for further experiments. In summary, we recognised six
compounds with equal or superior inhibitory activity compared to
commercially available DGKa inhibitors.
To check the isoform specificity of those active molecules, we
tested them, along with Amb639752, for their ability to inhibit
DGKa, DGKf (the other major DGK isoform expressed in lympho-
cytes), and the more distantly related and widely expressed DGKh.
At the highest concentration of 100 mM, all those molecules
resulted in highly specific against DGKa as like their parent mol-
ecule, Amb639752 by completely inhibit DGKa whereas they do
not have significant effects on DGKf and DGKh apart from 20
which, at the contrary, acts as an activator of DGKh (Figure 7).
3.4. DGKa inhibitors restore RICD in SAP deficient T cells
Ruffo et al. demonstrated that the defective RICD observed in T
cells from XLP-1 patients was rescued by silencing DGKa expres-
sion or by pre-treatment with DGKa inhibitors R59949 or R590228.
Interestingly, R59022 also showed beneficial effects in an in vivo
model of XLP-1, but due to its poor pharmacological proprieties,
its use in human patients results unlikely. We therefore tested the
effect all those active molecules along with Amb639752 on RICD
sensitivity of SAP-deficient T cells. As additional controls, we also
included ritanserin and ketanserin to evaluate the contribution of
serotonin antagonism to the effects observed.
3.3. Activity of compounds on serotonin receptors
R59022, R59949, and ritanserin feature a dual activity as DGKa
inhibitors and serotonin receptor antagonists25. Conversely,
To evaluate inhibitor efficacy in physiological context, we mod-
Amb639752 was reported as a selective DGKa inhibitor which has elled XLP-1 by silencing SAP in primary peripheral blood T lym-
no effects on serotonin activity9. Thus, we investigated whether phocytes (PBLs) and restimulating them with anti-CD3 antibody
the active molecules identified affect serotonin signalling.
(OKT3 10 ng/mL, 24 h). We pre-treated the cells with the indicated

104
S. VELNATI ET AL.
Figure 5. Putative DGKa inhibitors synthesised. In brackets the yield of the coupling reaction with the common intermediate 7.
inhibitors for 30 min at a concentration of 10 mM8. In control
Table 2. Inhibitory activity on DGKa (II).
Compound
Residue activity at 100 mM
IC50 (mM)
siRNA-transfected cells, DGKa inhibitors poorly affect RICD, with
Amb639752, 11 and 14 slightly reducing it (Figure 9). Conversely,
DGKa inhibitors significantly rescued the apoptotic defect of SAP-
deficient T cells although not reaching control levels. At 10 mM, all
the new molecules showed an efficacy comparable to Amb637952
and ritanserin used as positive reference molecules. Conversely,
the serotonin antagonist ketanserin is inactive, excluding the
involvement of serotonin receptors in rescuing the RICD in SAP-
deficient T cells (Figure 9).
8
5
6
6
14
6
89
27
41
47
26
3
3.2
1.6
3.6
6.9
3.0
–
32.8
–
–
11
12
13
14
15
16
17
18
19
20
21
22
49.7
1.8
–
In summary, these data confirm that the newly identified DGKa
inhibitors can rescue RICD susceptibility in T cell models of XLP-1
suggesting a putative use for XLP-1 therapy.
39
48
–

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
105
Figure 6. Dose–response curves for novel DGKa inhibitors. Dose–response of the most active compounds along with their IC50 values. Data from at least three inde-
pendent experiments performed in triplicate.
3.5. DGKa inhibitors reduce migration of the cancer cells (MCF7)
impairing cancer cell migration, we measured serum induced
Previous studies conducted in our laboratory demonstrated that
the inhibition of DGK activity decreases chemotaxis, proliferation,
wound healing in MCF7 breast cancer cells in presence of 10 mM
inhibitor. In presence of serum, none of the inhibitors is toxic for
migration, and invasion of many cancer cell lines12–14. To evaluate MCF7 cells even after prolonged treatment (data not shown).
if our newly synthetised DGKa inhibitors were effective in After 15 h of treatment, all the newly synthesised active molecules

106
S. VELNATI ET AL.
Figure 7. Isoform specificity of novel DGKa inhibitors. 293T cells were transfected with different DGK isoforms (A – DGKa, B – DGKf, C – DGKh, respectively) or empty
vectors and homogenised. All the molecules were tested at 100 lM for their capacity to inhibit the DGK activity of the different isoform homogenates. Data are
means SEM of at least three independent experiments performed in triplicate.
15
10
5
unstimulated
PMA stimulation
serotonin + PMA
10% FBS DMEM
0% FBS DMEM
****
****
80
60
40
20
0
***
**
****
****
****
**
****
****
***
***
**
ns
ns
****
0
2
8
8
n
n
n
i
O
O
O
52
S
i
i
in
5
11 12 13 14 20
11 12 13 14 20
9
2
8
1
1
2
1
3
1
4
1
0
2
n
i
S
S
7
er er
5
9
97
r
ser ser
s
s
SO
R5
SO
DM
3
3
94
DM
n
DM DM
n
e
6
97
a
an
an
9
s
t
ta
i
b6
b
i
3
et
r
r
DM
n
ket
k
a
Am
Am
t
b6
ri
Am
Figure 8. Novel DGKa inhibitors do not affect serotonin signalling. Human mono-
cytes were pre-incubated for 1 h with the indicated drugs in absence or presence
of serotonin and then stimulated with PMA 1 mM for 30 min (᭿ control unstimu-
Figure 10. Novel DGKa inhibitors slow tumour cell migration. MCF7 monolayer
was wounded and treated for 15 h with serum in presence of our new DGKa
inhibitors (10lM) or vehicle (DMSO). Results are expressed as the percentage of
wound area compared to the initial area. Data are the mean SEM of nine inde-
pendent experiments.
lated cells,
PMA stimulated cells, ᭿ PMA and serotonin stimulated cells).
Results are expressed as n moles of reduced cytochrome C/106 cells. Data are
means SEM of 10 independent experiments performed in triplicate.
equally reduced cell migration when compared to the vehicle
(DMSO) delaying wound closure (Figure 10).
Besides being in good agreement with the notion that DGKa is
required for cancer cells migration, this observation indicates that
our new DGKa inhibitors reduce cancer cell motility, suggesting a
potential utility in a metastasis setting.
****
60
40
20
0
ns
control siRNA
SAP siRNA
**** **** ****
****
****
***
ns
** **
3.6. Generation of a pharmacophore hypothesis
From the data obtained, it is possible to identify some key phar-
macophoric points crucial for the biological activity of the Amb
compounds on DGKs namely: (i) a basic nitrogen; (ii) the methyl
groups at the 2 and 6 position of the indole nucleus, and (iii) a
(hetero)aromatic ring. This information allows us to build a four-
point pharmacophoric model represented in Figure 11 superim-
posed with the minimised structure of compound 11. Although,
we are not able to evaluate the importance of the two carbonyl
groups, it represents the first attempts in order to identify the
minimum structural request to interact with DGK catalytic site
considering the molecular structure of the four most active inhibi-
tors discovered to date (Amb639752, ritanserin, R59022, and
R59949). We feel that this model might be useful to identify novel
2
8
2
75
8
1
0
2
n
n
i
n
i
n
i
O
O
9
i
11 12 13 14 20
1
12 13 14
S
S
75
39
er
ser ser
s
ser
n
a
an
3
DM
n
DM
6
6
a
an
t
t
t
b
b
ri
ri
m
ket
ke
A
Am
Figure 9. DGKa inhibitors partially restores apoptosis in SAP deficient lympho-
cytes. Lymphocytes from normal subjects were transfected with control or SAP
specific siRNA (᭿ control siRNA, ᭿ SAP siRNA). After 4 days, the cells were
restimulated with CD3 agonist OKT3 (10ng/mL) in presence of respective inhibi-
tor. Vehicle (DMSO) was used as control. Twenty-four hours later, the % of cell
loss was evaluated by PI staining. Data are the mean SEM of nine independent
experiments performed in triplicate.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
107
(CIB) bando “Network-CIB: Catalisi dell’Innovazione nelle bio-
tecnologie” to GB.
References
1. Cai J, Abramovici H, Gee SH, Topham MK. Diacylglycerol kin-
ases as sources of phosphatidic acid. Biochim Biophys Acta
2009;1791:942–8.
2. Krishna S, Zhong X. Role of diacylglycerol kinases in T cell
development and function. Crit Rev Immunol 2013;33:
97–118.
Figure 11. Proposed pharmacophoric model.
3. Merida I, Avila-Flores A, Merino E. Diacylglycerol kinases: at
the hub of cell signalling. Biochem J 2008;409:1–18.
4. Sakane F, Imai S, Kai M, et al. Diacylglycerol kinases: why so
many of them? Biochim Biophys Acta 2007;1771:793–806.
5. Baldanzi G, Bettio V, Malacarne V, Graziani A. Diacylglycerol
kinases: shaping diacylglycerol and phosphatidic acid gra-
dients to control cell polarity. Front Cell Dev Biol 2016;4:140.
6. Tangye SG. XLP: clinical features and molecular etiology due
to mutations in sh2d1a encoding sap. J Clin Immunol 2014;
34:772–9.
7. Baldanzi G, Pighini A, Bettio V, et al. Sap-mediated inhibition
of diacylglycerol kinase alpha regulates TCR-induced diacyl-
glycerol signaling. J Immunol 2011;187:5941–51.
8. Ruffo E, Malacarne V, Larsen SE, et al. Inhibition of diacylgly-
cerol kinase a restores restimulation-induced cell death and
reduces immunopathology in XLP-1. Sci Transl Med 2016;8:
321–327.
9. Velnati S, Ruffo E, Massarotti A, et al. Identification of a
novel DGKalpha inhibitor for XLP-1 therapy by virtual
screening. Eur J Med Chem 2019;164:378–90.
10. Bacchiocchi R, Baldanzi G, Carbonari D, et al. Activation of
alpha-diacylglycerol kinase is critical for the mitogenic prop-
erties of anaplastic lymphoma kinase. Blood 2005;106:
2175–82.
11. Yanagisawa K, Yasuda S, Kai M, et al. Diacylglycerol kinase
alpha suppresses tumor necrosis factor-alpha-induced apop-
tosis of human melanoma cells through NF-kappaB activa-
tion. Biochim Biophys Acta 2007;1771:462–74.
12. Baldanzi G, Cutrupi S, Chianale F, et al. Diacylglycerol kin-
ase-alpha phosphorylation by Src on Y335 is required for
activation, membrane recruitment and Hgf-induced cell
motility. Oncogene 2008;27:942–56.
13. Filigheddu N, Cutrupi S, Porporato PE, et al. Diacylglycerol
kinase is required for Hgf-induced invasiveness and anchor-
age-independent growth of mda-mb-231 breast cancer cells.
Anticancer Res 2007;27:1489–92.
14. Rainero E, Caswell PT, Muller PA, et al. Diacylglycerol kinase
alpha controls RCP-dependent integrin trafficking to pro-
mote invasive migration. J Cell Biol 2012;196:277–95.
15. Takeishi K, Taketomi A, Shirabe K, et al. Diacylglycerol kinase
alpha enhances hepatocellular carcinoma progression by
activation of Ras–Raf–MEK–ERK pathway. J Hepatol 2012;57:
77–83.
16. Torres-Ayuso P, Daza-Martin M, Martin-Perez J, et al.
Diacylglycerol kinase alpha promotes 3D cancer cell growth
and limits drug sensitivity through functional interaction
with Src. Oncotarget 2014;5:9710–26.
17. Rainero E, Cianflone C, Porporato PE, et al. The diacylgly-
cerol kinase alpha/atypical PKC/beta1 integrin pathway in
SDF-1alpha mammary carcinoma invasiveness. PLoS One
2014;9:e97144.
compounds active on DGKa through more targeted virtual screen-
ing campaigns, overcoming the current scaffolds.
4. Discussion
As a key component of several signal transduction pathways,
DGKa represent an emerging pharmacological target. We have
demonstrated the efficacy of DGKa inhibitors for XLP-1 treatment8,
while others have proposed them for cancer treatment22 and to
remove immune-checkpoints promoting immune vigilance against
cancer40. Commercially available DGKa inhibitors are limited by
poor specificity25,41 and pharmacokinetic20. The CU-3 molecule
described by other features a noteworthy activity and specificity
but its reactive chemical structure make unlikely an in vivo use30.
With intent of developing molecules suitable for therapeutic use
we selected Amb639752 as a novel inhibitor with remarkable
DGKa activity. Amb639752 also features improved selectivity for
DGKa as it does not affect serotonin signalling9. Despite numerous
efforts a structure of mammalian DGKs is still missing, thus we
decided to explore the structure–activity relationship of this mol-
ecule to improve its activity and pave the way for further develop-
ments. Our efforts allowed us to build a pharmacophoric model
for DGKa inhibitors characterised by three required features. We
also characterised a set of novel compounds with improved IC50
in the low mM range and identified the most profitable synthetic
route for them. The mode of DGKa inhibition by those molecules
is still unknown apart for ritanserin, which binds at the same time
the DGKa catalytic accessory domain and the C1 domain puta-
tively promoting a close inactive conformation41.
The second-generation inhibitors we described in this work
maintain the specificity of Amb639752 as they not affect DGKf,
the predominant isoform of lymphocytes42 and the broadly
expressed DGKh43. Those DGKa inhibitors are active in a lympho-
cyte based XLP-1 assay and in a cancer cell migration assay, hold-
ing the promise for a potential therapeutic application. However,
their efficacy is still to be determined in in vivo models of disease
where some of the parental compounds showed efficacy but poor
pharmacokinetic8,20
.
Disclosure statement
The authors declare no competing financial interest.
Funding
ꢀ
This work was supported by Universita del Piemonte Orientale,
Telethon Foundation [Grant GGP16252 to AG and GB], National
Ministry of University and Research PRIN 2017 [Grant
201799WCRH to GB], Consorzio Interuniversitario di Biotecnologie

108
S. VELNATI ET AL.
18. Hao X, Sun B, Hu L, et al. Differential gene and protein 31. Baell JB, Holloway GA. New substructure filters for removal
expression in primary breast malignancies and their lymph
node metastases as revealed by combined cDNA microarray
and tissue microarray analysis. Cancer 2004;100:1110–22.
19. Marchet A, Mocellin S, Belluco C, et al. Gene expression pro-
file of primary gastric cancer: towards the prediction of
lymph node status. Ann Surg Oncol 2007;14:1058–64.
20. Dominguez CL, Floyd DH, Xiao A, et al. Diacylglycerol kinase
a is a critical signaling node and novel therapeutic target in
glioblastoma and other cancers. Cancer Discov 2013;3:
782–97.
21. Kefas B, Floyd DH, Comeau L, et al. A miR-297/hypoxia/DGK-
a axis regulating glioblastoma survival. Neuro Oncol 2013;
15:1652–63.
22. Purow B. Molecular pathways: targeting diacylglycerol kinase
alpha in cancer. Clin Cancer Res 2015;21:5008–12.
23. Jiang Y, Sakane F, Kanoh H, Walsh JP. Selectivity of the diac-
ylglycerol kinase inhibitor 3-[2-(4-[bis-(4-fluorophenyl)methy-
lene]-1-piperidinyl)ethyl]-2, 3-dihydro-2-thioxo-4(1h)quinazo
linone (r59949) among diacylglycerol kinase subtypes.
Biochem Pharmacol 2000;59:763–72.
24. Sato M, Liu K, Sasaki S, et al. Evaluations of the selectivities
of the diacylglycerol kinase inhibitors r59022 and r59949
among diacylglycerol kinase isozymes using a new non-
radioactive assay method. Pharmacology 2013;92:99–107.
25. Boroda S, Niccum M, Raje V, et al. Dual activities of ritan-
serin and r59022 as DGKa inhibitors and serotonin receptor
antagonists. Biochem Pharmacol 2017;123:29–39.
26. Gaulton A, Hersey A, Nowotka M, et al. The ChEMBL data-
base in 2017. Nucleic Acids Res 2017;45:D945–D954.
27. Leysen JE, Gommeren W, Van Gompel P, et al. Receptor-
binding properties in vitro and in vivo of ritanserin: a very
potent and long acting serotonin-S2 antagonist. Mol
Pharmacol 1985;27:600–11.
28. Hawkins PC, Skillman AG, Nicholls A. Comparison of shape-
matching and docking as virtual screening tools. J Med
Chem 2007;50:74–82.
of pan assay interference compounds (pains) from screening
libraries and for their exclusion in bioassays. J Med Chem
2010;53:2719–40.
32. Lavagno L, Gunella G, Bardelli C, et al. Anti-inflammatory
drugs and tumor necrosis factor-alpha production from
monocytes: role of transcription factor NF-kappa B and
implication for rheumatoid arthritis therapy. Eur J Pharmacol
2004;501:199–208.
33. Talmon M, Rossi S, Pastore A, et al. Vortioxetine exerts anti-
inflammatory and immunomodulatory effects on human
monocytes/macrophages. Br J Pharmacol 2018;175:113–24.
34. Omega, version 2.4.6; OpenEye Scientific Software: Santa Fe,
NM. Available from: http://www.eyesopen.com
35. Hawkins PCD, Skillman AG, Warren GL, et al. Conformer gen-
eration with omega: algorithm and validation using high
quality structures from the Protein Databank and Cambridge
Structural Database. J Chem Inf Model 2010;50:572–84.
36. Hawkins PCD, Nicholls A. Conformer generation with omega:
learning from the data set and the analysis of failures. J
Chem Inf Model 2012;52:2919–36.
37. Schneidman-Duhovny D, Dror O, Inbar Y, et al. Pharmagist:
a webserver for ligand-based pharmacophore detection.
Nucleic Acids Res 2008;36:W223–228.
38. Wermuth C, Aldous D, Raboisson P, Rognan D. The practice
of medicinal chemistry. 4th ed. London, UK: Academic Press;
2015.
39. Johansson H, Urruticoechea A, Larsen I, Sejer Pedersen D. A
scalable method for regioselective 3-acylation of 2-substi-
tuted indoles under basic conditions. J Org Chem 2015;80:
471–81.
40. Sakane F, Mizuno S, Komenoi S. Diacylglycerol kinases as
emerging potential drug targets for a variety of diseases: an
update. Front Cell Dev Biol 2016;4:82.
41. Franks CE, Campbell ST, Purow BW, et al. The ligand binding
landscape of diacylglycerol kinases. Cell Chem Biol 2017;24:
870–80.e5.
29. Muchmore SW, Souers AJ, Akritopoulou-Zanze I. The use of
three-dimensional shape and electrostatic similarity search-
ing in the identification of a melanin-concentrating hor-
mone receptor 1 antagonist. Chem Biol Drug Des 2006;67:
174–6.
42. Joshi RP, Schmidt AM, Das J, et al. The f isoform of diacyl-
glycerol kinase plays a predominant role in regulatory T cell
development and TCR-mediated ras signaling. Sci Signal
2013;6:ra102.
43. Tu-Sekine B, Goldschmidt HL, Raben DM. DGK-h: structure,
enzymology, and physiological roles. Front Cell Dev Biol
2016;4:101.
30. Liu K, Kunii N, Sakuma M, et al. A novel diacylglycerol kinase
a-selective inhibitor, CU-3, induces cancer cell apoptosis and
enhances immune response. J Lipid Res 2016;57:368–79.