A novel diacylglycerol kinase α-selective inhibitor, CU-3, induces cancer cell apoptosis and enhances immune response[S]

Diacylglycerol kinase (DGK) consists of 10 isozymes. The α-isozyme enhances the proliferation of cancer cells. However, DGKα facilitates the nonresponsive state of immunity known as T-cell anergy; therefore, DGKα enhances malignant traits and suppresses immune surveillance. The aim of this study was to identify a novel small molecule that selectively and potently inhibits DGKα activity. We screened a library containing 9,600 chemical compounds using a newly established high-throughput DGK assay. As a result, we have obtained a promising compound, 5-[(2E)-3-(2-furyl)prop-2-enylidene]-3-[(phenylsulfonyl)amino]2-thioxo-1,3-thiazolidin-4-one) (CU-3), which selectively inhibited DGKα with an IC50 value of 0.6 μM. CU-3 targeted the catalytic region, but not the regulatory region, of DGKα. CU-3 competitively reduced the affinity of DGKα for ATP, but not diacylglycerol or phosphatidylserine. Moreover, this compound induced apoptosis in HepG2 hepatocellular carcinoma and HeLa cervical cancer cells while simultaneously enhancing the interleukin-2 production of Jurkat T cells. Taken together, these results indicate that CU-3 is a selective and potent inhibitor for DGKα and can be an ideal anticancer drug candidate that attenuates cancer cell proliferation and simultaneously enhances immune responses including anticancer immunity.

Chemical Industries), 100 U/ml penicillin, and 100 g/ml streptomycin. Jurkat T cells were maintained in 75 cm 2 fl asks in RPMI-1640 medium (Wako Pure Chemicals Industries) containing 10% positive regulator of the proliferative activity of hepatocellular carcinoma through the Ras/Raf/MEK (mitogen-activated protein kinase/ERK kinase)/ERK pathway ( 9 ). In melanoma cells, DGK ␣ positively regulates the tumor necrosis factor-␣dependent nuclear factor-B (p65) activation via the protein kinase C -mediated Ser311 phosphorylation of p65 ( 11 ). Therefore, the suppression of DGK ␣ activity is expected to inhibit the progression of these cancers. On the other hand, DGK ␣ is abundantly expressed in T lymphocytes where it facilitates the nonresponsive state known as anergy ( 12,13 ). Anergy induction in T cells represents the main mechanism for advanced tumors to avoid immune action.
To develop highly effective and DGK ␣ -selective inhibitors, a system for high-throughput screening (HTS) is required; however, the conventional DGK assay is quite laborious and requires technical skill. For example, the conventional assay requires the use of a radioisotope ([ ␥ -32 P]ATP) and the manipulation of thin-layer chromatography with multiple extraction steps. We recently established a simple DGK assay ( 18 ) that is useful for constructing an HTS system for detecting DGK inhibitors from chemical compound libraries.

HTS
For HTS, 384-well plates were predispensed with 60 nl (fi nal concentration: 30 or 50 M) of each compound. Glutathione FBS, 100 U/ml penicillin, and 100 g/ml streptomycin. The cells were maintained at 37°C in an atmosphere containing 5% CO 2 . COS-7 cells were seeded in 60 mm dishes at a density of 2.5 × 10 5 cells/dish. cDNA was transfected into COS-7 cells by electroporation with a Gene Pulser Xcell TM Electropolation System (Bio-Rad Laboratories) according to the manufacturer's instructions.

Determination of the DGK activity in cells by LC/MS
Determination of the DGK activity in cells by LC/MS was carried out as described previously ( 30,31 ). COS-7 cells expressing DGK ␣ -⌬ 1-196 (a constitutively active mutant) were harvested in phosphate-buffered saline. Total lipids were extracted from the cells according to the method of Bligh and Dyer ( 32 ). The extracted cellular lipids (10 l) containing 65 pmol of the 28:0-PA internal standard (Sigma-Aldrich) were separated on the LC system (Accela LC Systems, Thermo Fisher Scientifi c, Tokyo, Japan) using a UK-Silica column (3 m, 150 × 2.0 mm inner diameter; Imtakt, Kyoto, Japan) ( 30,31 ). Mobile phase A consisted of chloroform-methanol-ammonia (89:10:1), and mobile phase B S -transferase-fused pig DGK ␣ was expressed in Sf9 insect cells and purifi ed with a glutathione-Sepharose column (GE Healthcare). The DGK assay was performed using the ADP-Glo™ Kinase Assay Kit (Promega, Tokyo, Japan) at 30°C for 2 h as described previously ( 18 ). The chemiluminescence generated in this assay correlates to the amount of ADP generated, which is equivalent to the phosphatidic acid (PA) produced, in the kinase assay, indicating kinase activity. The chemiluminescence was measured using a PHERAstar microplate reader (BMG LABTECH, Offenburg, Germany). The assay performance was consistent across all plates, which was evident due to the robust Z' factor.

Chemical compounds
CU-1, -2, -3, and -4 were obtained from Drug Discovery Initiative, University of Tokyo. For further characterization of CU-3 [inhibition mechanisms of CU-3, inhibition effects of CU-3 in cells, induction of apoptosis in cancer cells by CU-3, and enhancement of the interleukin-2 (IL-2) production by CU-3 (see Results)], highly pure CU-3 was resynthesized and supplied by Namiki Shoji (Tokyo, Japan).

RNA interference of DGK ␣
DGK ␣ Stealth select RNAi (catalog number HSS102626; Invitrogen, Tokyo, Japan) was used. Jurkat T cells suspended in unsupplemented, prewarmed RPMI, with 500 nM of human si-DGK ␣ or nontargeting control siRNA (Invitrogen) in 4 mm cuvettes (Nepa Gene) were electroporated at 210 V and 950 microfarads using the GenePulsar Xcell Electroporation System (Bio-Rad Laboratories). After electroporation, cells were incubated for 48 h in culture medium.

Western blot analysis
Lysates of HepG2, HeLa, Jurkat, or COS-7 cells were separated on SDS-polyacrylamide gel electrophoresis. The separated proteins were transferred to a PVDF membrane (Bio-Rad Laboratories) and blocked with Block Ace (Dainippon Pharmaceutical). The membrane was incubated with anti-DGK ␣ ( 34 ) antibody in 10% Block Ace for 1 h. The immunoreactive bands were then visualized using a peroxidase-conjugated anti-rabbit IgG antibody (Jackson ImmunoResearch Laboratories) and the Enhanced Chemiluminescence Western Blotting Detection System (GE Healthcare).

Core library screen to identify compounds that inhibit DGK ␣ activity
To identify specifi c inhibitors for DGK ␣ activity, we screened the Core Library (Drug Discovery Initiative, consisted of chloroform-methanol-ammonia-water (55:39:1:5). The gradient elution program was as follows: 30% B for 5 min, 30-60% B over 25 min, 60-70% over 10 min, followed by 70% B for 5 min. The fl ow rate was 0.3 ml/min, and the chromatography was performed at 25°C.
The LC system described above was coupled online to an Exactive Orbitrap MS (Thermo Fisher Scientifi c) equipped with an ESI source. The ion spray voltage was set to -5 kV and 5 kV in the negative and positive ion mode, respectively. The capillary temperature was set to 300°C. The other parameters were set according to the manufacturer's recommendations. Individual phospholipids were measured by scanning from m/z 450 to 1,100 in the negative or positive ion modes using an Orbitrap Fourier Transform MS with a resolution of 50,000. The MS peaks were identifi ed based on their m/z value and were presented in the form of X : Y , where X is the total number of carbon atoms and Y is the total number of double bonds in both acyl chains of the phospholipid.

Apoptosis analysis
HepG2, HeLa, and COS-7 cells were incubated in a 96-well plate in the presence or absence of CU-3 (5 M) for 24 h. The caspase-3/7 assay (Caspase-Glo ® 3/7; Promega) was conducted according to the manufacturer's description. After a 1 h incubation at 25°C, each sample was measured in a microplate reader (GloMax ® -Multi+ Detection System; Promega).

Assay for IL-2 mRNA expression in Jurkat T cells
The assay for IL-2 mRNA expression in Jurkat T cells was carried out as previously reported ( 33 ). Jurkat cells were preincubated in 35 mm culture dishes fi lled with 2 ml of RPMI in the presence or absence of CU-3 (5 M) for 5 min. Concanavalin A (Con A) was then added to the media, and the cells were further incubated for 3 h, collected by centrifugation (400 g , 5 min), and lysed with 1 ml of ISOGEN (Wako Pure Chemical Industries). Total RNA was prepared, and 1 g of total RNA was reverse transcribed into cDNA according to the manufacturer's instructions (DGK ␣ -⌬ 1-196) and the recoverin homology domain-the C1 domains (DGK ␣ -⌬ 1-332) ( Fig. 4A ). CU-3 inhibited the DGK activities of the wild-type enzyme and these mutants to a similar extent ( Fig. 4B ). These results indicate that CU-3 targets the catalytic domain, not the regulatory region, of DGK ␣ . Although DGK ␣ is activated by Ca 2+ ( 7,35 ), these mutants commonly lack the Ca 2+ binding EF-hand motifs and show strong Ca 2+ -independent activity ( 28,36 ). Therefore, it is likely that Ca 2+ is not involved in the inhibition mechanism of CU-3.
It was reported that in addition to wild-type DGK ␣ ( 35 ), DGK ␣ -⌬ 1-332 (catalytic domain alone) was also activated by PS ( 28 ). Therefore, we next determined the EC 50 values of PS for DGK ␣ activity in the presence and absence of CU-3. PS strongly activated DGK ␣ , with activation reaching a maximum at ‫ف‬ 23.0 mol% in the absence of CU-3 ( Fig. 5 ). The apparent EC 50 of the isozyme for PS was 12.0 mol% ± 0.0 (n = 3) in the absence of CU-3 ( Fig. 5 and University of Tokyo) that consists of 9,600 compounds using a recently established, simple DGK assay ( 18 ). Purifi ed DGK ␣ was incubated with these compounds at a concentration of 30 M. We set the threshold for hit compounds at >20% inhibition for all compounds and identifi ed 103 hit compounds. These hit compounds were then subjected to the second screen at a concentration of 50 M to assess the reproducibility.

Inhibition mechanisms of CU-3
We next attempted to reveal the inhibition mechanisms of CU-3. We fi rst examined which region of DGK ␣ was targeted by CU-3. We prepared truncation mutants lacking the recoverin homology domain-the EF-hand motifs  CU-3 on DGK ␣ in cell were determined using our newly established LC/MS method ( 30,31 ). The overexpression of DGK ␣ -⌬ 1-196 ( Fig. 8A ), a constitutively active mutant, clearly increased the total amount of PA (approximately a 15% increase) in COS-7 cells ( Fig. 8B ). Moreover, the addition of CU-3 signifi cantly reduced the total PA amount that was increased by the overexpression of DGK ␣ -⌬ 1-196 ( Fig. 8B ).

Induction of apoptosis in cancer cells by CU-3
We next examined whether CU-3 induces apoptosis in a human hepatocellular carcinoma cell line HepG2 Table 3 ). The addition of CU-3 did not markedly affect the EC 50 value of PS (13.0 ± 0.6 mol%, n = 3) ( Fig. 5 and Table 3 ). We next measured the kinetic parameter for DG. The activity of DGK ␣ increased in a DG dose-dependent manner ( Fig. 6A ). A double reciprocal plot provided the K m value for DG in the absence of CU-3 (3.4 ± 1.0 mol%, n = 3; Fig. 6B and Table 4 ). As shown in Fig. 6B and Table 4 , CU-3 did not signifi cantly affect the apparent K m value for DG (2.9 ± 0.5 mol%, n = 3).
We next determined the kinetic parameter for ATP in the presence and absence of CU-3. The activity of DGK ␣ was increased in an ATP dose-dependent manner ( Fig. 7A ). A double reciprocal plot provided the apparent K m value for ATP in the absence of CU-3 (0.25 ± 0.07 mM, n = 3; Fig. 7B and Table 5 ). However, the K m value for ATP in the presence of CU-3 was signifi cantly increased to 0.48 ± 0.07 mM, n = 3 ( Fig. 6B and Table 5 ). The V max value was not markedly changed in the presence of CU-3 ( Fig. 7B ). These results strongly suggest that CU-3 competitively inhibited the affi nity of DGK ␣ for ATP.

Inhibition effects of CU-3 in cells
We showed that CU-3 intensely inhibited DGK ␣ activity in vitro ( Figs. 1-7 ). Next, the inhibition effects of The values are presented as the mean ± SD (n = 3).   ( Fig. 9D ), suggesting that CU-3 is not nonspecifi cally toxic to cells.

Enhancement of the IL-2 production by CU-3
We next tested whether CU-3 enhanced the function of T cells. To assess this possibility, we determined the Con A-induced IL-2 mRNA production activity of Jurkat cells (a human T-cell line). As shown in Fig. 10 , IL-2 mRNA in Jurkat T cells was not detectable in the absence of Con A, and Con A markedly induced the IL-2 mRNA production of the cells. CU-3 further enhanced IL-2 mRNA production activity (an ‫ف‬ 50% increase) of Jurkat T cells ( Fig. 10 ).
These results indicate that CU-3 enhanced the function of T cells.
To demonstrate that CU-3 causes IL-2 production through the inhibition of DGK ␣ , we compared the effects of CU-3 and DGK ␣ -siRNA side by side and examined whether the RNA interference further enhances the effects of the inhibitor. CU-3 and DGK ␣ -siRNA increased IL-2 production to almost the same extents ( Fig. 11 ). Moreover, DGK ␣ -siRNA did not further enhance the because DGK ␣ was highly expressed in the cell lines ( Fig. 9A ) and signifi cantly enhanced their cell growth ( 9 ). To assess this possibility, caspase-3/7 activity was determined. As shown in Fig. 9B , CU-3 markedly enhanced the caspase-3/7 activity (an ‫ف‬ 50% increase) of HepG2 cells. Moreover, we utilized HeLa cells (a human cervical carcinoma cell line) because DGK ␣ was highly expressed in the cells ( Fig. 9A ) and a DGK inhibitor R59022 induced their cell death ( 37 ). As shown in Fig. 9C , essentially the same results (an ‫ف‬ 24% increase of caspase-3/7 activity) were obtained with HeLa cells. However, the caspase-3/7 activity of COS-7 epithelial cells (monkey kidney-derived, noncancerous cell line) was not augmented by CU-3 The values are presented as the mean ± SD (n = 3). CU-3 did not affect the dependence on an activator PS ( Fig. 5 and Table 3 ). Moreover, CU-3 did not change the K m value of DGK ␣ for DG ( Fig. 6 and Table 4 ). These results indicate that the affi nities for PS and DG failed to be affected by the inhibitor. However, CU-3 signifi cantly decreased the K m value of DGK ␣ for ATP ( Fig. 7 and Table  5 ). Therefore, it is likely that the inhibitor competitively decreased the affi nity for ATP.
CU-3 is not strikingly similar to R59022 ( 14 ), R59949 ( 16 ), or stemphone ( 39 ). CU-3 contains 2-thioxo-1,3-thiazolidin-4-one ( Table 1 ), which is partly similar to ATP. Therefore, it is reasonable to speculate that the inhibitor may partly mimic ATP and eventually decrease the affi nity for the substrate of DGK, ATP, as observed in Fig. 7 . The primary structures of the catalytic domains of DGK isozymes are highly similar to each other (1)(2)(3)(4)(5)(6). Therefore, it is unknown how the isozyme selectivity occurs. To evaluate this issue, a determination of the 3D structure is needed.
In addition to in vitro, CU-3 inhibited cellular DGK ␣ activity in COS-7 cells ( Fig. 8 ). It was noted that a concentration of 2 M, which is only ‫ف‬ 3-fold higher than the IC 50 value (0.6 M) ( Fig. 1 and Table 2 ), strongly inhibited PA production by overexpressed DGK ␣ ( Fig. 8 ). This result suggests that CU-3 has high cell permeability. Therefore, CU-3 is expected to effectively inhibit DGK ␣ in vivo as well.
We demonstrated that CU-3 induced apoptosis in HepG2 hepatocellular carcinoma and HeLa cervical cancer cells ( Fig. 9B, C ). We also reported that DGK ␣ -specifi c siRNA attenuated the proliferation of hepatocellular carcinoma ( 9 ) and induced apoptosis in melanoma cells ( 10 ). Supporting our results, Torres-Ayuso et al. ( 40 ) also demonstrated that the growth of colon and breast cancer cell lines was effect of CU-3 ( Fig. 11 ), indicating that CU-3 and DGK ␣ -siRNA inhibit the same target, DGK ␣ .

DISCUSSION
We screened a library containing 9,600 compounds using a high-throughput chemiluminescence-based assay that was recently established ( 18 ). Among the compounds, CU-3 was identifi ed as a potent and selective inhibitor against the ␣ -isozyme of DGK ( Figs. 1 and 2 ; Table 2 ). Compared with commercially available DGK inhibitors, R59022 and R59949 ( 18 ), CU-3 exhibited higher effi ciency and selectivity against DGK ␣ . The IC 50 value of CU-3 (0.6 M) ( Fig. 1 ) was markedly lower than that of R59022 and R59949 ( ‫ف‬ 25 M and 18 M, respectively) ( 18 ). R59022 and R59949 only semiselectively inhibited type I, III, and V DGKs ␣ , , and , and type I and II DGKs ␣ , ␥ , ␦ , and , respectively ( 18 ). Moreover, the K m value of CU-3 for DGK ␣ was at least ‫ف‬ 12 times lower than those for other DGK isozymes ( Fig. 2 and Table 2 ). Stemphone, a fungal metabolite, has also been reported to be a DGK inhibitor ( 38 ). However, its DGK-isozyme selectivity is not known. Therefore, this study is the fi rst report of a highly ␣ -isozyme selective inhibitor.
DGK ␣ -⌬ 1-332 (catalytic domain alone) ( Fig. 4A ) lacks the regulatory region including the Ca 2+ binding EF-hand motifs. CU-3 inhibited DGK activities of the full-length enzyme and the mutant to a similar extent ( Fig. 4B ). The results indicate that the target of CU-3 is the catalytic domain of DGK ␣ . Therefore, it is likely that Ca 2+ is not involved in the inhibition mechanism of CU-3. angiogenesis of cancer cells in addition to their cell death. It is notable that DGK ␣ -knockout mice are generally healthy, although the mice have a defect in T-cell anergy ( 12,13,42 ). These results allow us to speculate that DGK ␣specifi c inhibitors would not have severe side effects.
In addition to the induction on cancer cell apoptosis ( Fig. 9 ), the inhibitor promotes IL-2 production ( Fig. 10 ), which is one of the indicators of T-cell activation. Olenchock et al. ( 12 ) and Zha et al. ( 13 ) previously reported that DGK ␣ is abundantly expressed in T cells, where it facilitates the nonresponsive state, anergy. Thus, the inhibitor of DGK ␣ was expected to enhance T-cell activity. In this study, we found that CU-3 indeed enhanced the IL-2 production of T cells ( Fig. 10 ). Although it has already been reported that the DGK ␣ inhibitor inhibited cancer cell proliferation ( 37 ), this study provides a new aspect of the DGK ␣ inhibitor, an enhancer of the immune system. Inactivation (anergy induction) of T cells is the main mechanism for advanced tumors to avoid immune action. Therefore, it is expected that CU-3 is able to activate cancer immunity. signifi cantly inhibited by DGK ␣ -siRNA and R59949, which attenuates DGK ␣ activity ( 18 ). In addition, Dominguez et al. ( 37 ) reported that R59022, which most strongly inhibits DGK ␣ ( 18 ), negatively affected the proliferation of glioblastoma, melanoma, breast cancer, and cervical cancer cells. It is interesting to investigate the effect of CU-3 on the cell growth of a variety of cancer cells. They also observed that in marked contrast to cancer cells, R59022 did not weaken the growth of noncancerous astrocytes and fibroblasts ( 37 ). In this study, we also observed that although CU-3 enhanced the caspase-3/7 activities of HepG2 hepatocellular carcinoma and HeLa cervical cancer cells ( Fig.  9B, C ), the compound did not increase the caspase-3/7 activity of the noncancer-derived COS-7 cells ( Fig. 9D ). Therefore, we reproduced their results ( 37 ). These fi ndings suggest that cancer-derived cells and noncancerderived cells may utilize different pathways to induce apoptosis. DGK ␣ seems to be particularly relevant for cancer cells. For example, we have revealed that the expression of DGK ␣ was not detectable in normal hepatocytes, whereas this isozyme was expressed in several hepatocellular carcinoma cell lines ( 9 ). Moreover, although normal melanocytes did not express DGK ␣ , several melanoma cell lines did express this isozyme ( 10 ). It has been shown that DGK ␣ positively regulated angiogenesis signaling ( 41 ). Therefore, it is possible that CU-3 also attenuates the Fig. 10. Effect of CU-3 on the IL-2 production of Jurkat T cells. A: Jurkat T cells were preincubated in 35 mm culture dishes fi lled with 2 ml RPMI in the presence or absence of CU-3 (5 M) for 5 min. Con A was then added to the media, and the cells were further incubated for 3 h. Total RNA was reverse transcribed into cDNA, and PCR amplifi cation (34 cycles) was performed using primers for IL-2 or GAPDH. The PCR products were then separated by agarose gel electrophoresis and visualized with ethidium bromide. The visualized bands were digitized and quantifi ed using Adobe Photoshop and NIH Image software. B: The value in the presence of Con A and in the absence of CU-3 was set to 100%. The values are presented as the mean ± SD (n = 5). ** P < 0.01. Con A was then added to the media, and the cells were further incubated for 3 h. Total RNA was reverse transcribed into cDNA, and PCR amplifi cation (34 cycles) was performed using primers for IL-2 or GAPDH. The PCR products were then separated by agarose gel electrophoresis and visualized with ethidium bromide. The visualized bands were digitized and quantifi ed using Adobe Photoshop and NIH Image software. The value of control siRNA-treated samples without CU-3 was set to 100%. The values are presented as the mean ±SD (n = 3). *** P < 0.005; n.s., not signifi cant.
General anticancer drugs attenuate the proliferation and function of both cancer and bone marrow cells ( 43,44 ). Therefore, general anticancer drugs induce not only the attenuation of cancer cell proliferation but also bone marrow suppression/myelosuppression, which is one of the most commonly observed side effects of chemotherapy. Bone marrow suppression results in a decreased number of T cells and, consequently, causes severe side effects, such as recurrent infectious diseases. However, there is no drug that has both protumoral and anti-immunogenic effects. Interestingly, because DGK ␣ has both protumoral and anti-immunogenic properties ( 4 ), the DGK ␣ -selective inhibitor CU-3 would simultaneously have antitumoral and proimmunogenic effects. Therefore, CU-3 can be a lead compound to develop an ideal anticancer drug without infectious side effects. Moreover, in addition to the direct effects on apoptosis induction in cancer cells, CU-3 can indirectly induce the death of cancer cells through activation of the immune system. In addition to the development of an anticancer drug, CU-3 will be a great tool for biological science concerning cancer and immunity.